collected articles on type 1 diabetes.pdf

Collected articles on Type 1 Diabetes

Index

- Mechanisms of Pancreatic β-Cell Death in Type 1 and Type 2 Diabetes - Many Differences -Few Similarities

- Immunology and Genetics of Type 1 Diabetes- Type 1 Diabetes Aetiology Immunology and Therapeutic Strategies- Autoimmune Destruction of Pancreatic β-Cells- Viral infections as potential triggers of type 1- Virus Infections in Type 1 Diabetes- Interleukine & β-cell regeneration- Remission of Diabetes by β-Cell Regeneration in Diabetic Mice Treated With a Recombinant

Adenovirus Expressing βetacellulin- Vitamin D and Type 1 diabetes Mellitus State of the Art

Mechanisms of Pancreatic �-Cell Death in Type 1 andType 2 DiabetesMany Differences, Few SimilaritiesMiriam Cnop,

1,2Nils Welsh,

3Jean-Christophe Jonas,

4Anne Jorns,

5,6Sigurd Lenzen,

6

and Decio L. Eizirik1

Type 1 and type 2 diabetes are characterized by progres-sive �-cell failure. Apoptosis is probably the main form of�-cell death in both forms of the disease. It has beensuggested that the mechanisms leading to nutrient- andcytokine-induced �-cell death in type 2 and type 1 diabetes,respectively, share the activation of a final common path-way involving interleukin (IL)-1�, nuclear factor (NF)-�B,and Fas. We review herein the similarities and differencesbetween the mechanisms of �-cell death in type 1 and type2 diabetes. In the insulitis lesion in type 1 diabetes,invading immune cells produce cytokines, such as IL-1�,tumor necrosis factor (TNF)-�, and interferon (IFN)-�.IL-1� and/or TNF-� plus IFN-� induce �-cell apoptosis viathe activation of �-cell gene networks under the control ofthe transcription factors NF-�B and STAT-1. NF-�B activa-tion leads to production of nitric oxide (NO) and chemo-kines and depletion of endoplasmic reticulum (ER)calcium. The execution of �-cell death occurs throughactivation of mitogen-activated protein kinases, via trig-gering of ER stress and by the release of mitochondrialdeath signals. Chronic exposure to elevated levels of glu-cose and free fatty acids (FFAs) causes �-cell dysfunctionand may induce �-cell apoptosis in type 2 diabetes. Expo-sure to high glucose has dual effects, triggering initially“glucose hypersensitization” and later apoptosis, via dif-ferent mechanisms. High glucose, however, does not induceor activate IL-1�, NF-�B, or inducible nitric oxide synthasein rat or human �-cells in vitro or in vivo in Psammomysobesus. FFAs may cause �-cell apoptosis via ER stress,which is NF-�B and NO independent. Thus, cytokines andnutrients trigger �-cell death by fundamentally differentmechanisms, namely an NF-�B–dependent mechanism that

culminates in caspase-3 activation for cytokines and anNF-�B–independent mechanism for nutrients. This arguesagainst a unifying hypothesis for the mechanisms of �-celldeath in type 1 and type 2 diabetes and suggests thatdifferent approaches will be required to prevent �-celldeath in type 1 and type 2 diabetes. Diabetes 54 (Suppl. 2):S97–S107, 2005

Clinical definitions of disease often obscure dif-ferent mechanistic subtypes. This is particularlyrelevant for complex diseases such as diabetes,where combinations of multiple genes and en-

vironmental factors eventually lead to loss of functional�-cell mass and hyperglycemia. The mechanisms leadingto �-cell loss may be quite diverse in the various subtypesof the disease. As our knowledge of disease pathogenesesincreases, better classifications of diabetes may beproposed.

The two main forms of diabetes are type 1 and type 2diabetes (1). Both types are characterized by progressive�-cell failure. In type 1 diabetes, this is typically caused byan autoimmune assault against the �-cells, inducing pro-gressive �-cell death. The pathogenesis of type 2 diabetesis more variable, comprising different degrees of �-cellfailure relative to varying degrees of insulin resistance.

The genetics (i.e., HLA-related in type 1 diabetes vs.non–HLA-related in type 2 diabetes), putative environmen-tal triggers (for instance viral infection in type 1 diabetes,obesity in type 2 diabetes), and natural history of thedisease are different between type 1 and type 2 diabetes.Because these topics are covered in detail in other articlesin this supplement issue, they will not be discussed furtherhere.

In type 1 diabetes, �-cell mass is reduced by 70–80% atthe time of diagnosis. Because of the variable degrees ofinsulitis and absence of detectable �-cell necrosis, it wassuggested that �-cell loss occurs slowly over years (2).These pathology findings are in line with the progressivedecline in first-phase insulin secretion in antibody-positiveindividuals, long before the development of overt diabetes(3). It was later shown that �-cell apoptosis causes agradual �-cell depletion in rodent models of type 1 diabe-tes (rev. in 4). In type 2 diabetic subjects, initial patholog-ical studies suggested a �-cell loss of 25–50% (2,5), but thiswas debated by others (6). Recent studies, which matcheddiabetic patients and control subjects for BMI, showed asignificant reduction in �-cell mass (7,8) and a threefoldincrease in �-cell apoptosis (8). These observations sug-gest that �-cell mass is decreased in type 2 diabetes,secondary to increased rates of �-cell apoptosis, but it

From the 1Laboratory of Experimental Medicine, Faculty of Medicine, Eras-mus Hospital, Universite Libre de Bruxelles, Brussels, Belgium; the 2Divisionof Endocrinology, Erasmus Hospital, Universite Libre de Bruxelles, Brussels,Belgium; the 3Department of Medical Cell Biology, Uppsala University,Biomedicum, Uppsala, Sweden; the 4Unit of Endocrinology and Metabolism,Faculty of Medicine, University of Louvain (UCL), Brussels, Belgium; the5Centre of Anatomy, Hannover Medical School, Hannover, Germany; and the6Institute of Clinical Biochemistry, Hannover Medical School, Hannover,Germany.

Address correspondence and reprint requests to Dr. Miriam Cnop, Labora-tory of Experimental Medicine, Universite Libre de Bruxelles (ULB), Route deLennik 808, CP-618, 1070 Brussels, Belgium. E-mail: [email protected].

Received for publication 23 February 2005 and accepted in revised form 30March 2005.

This article is based on a presentation at a symposium. The symposium andthe publication of this article were made possible by an unrestricted educa-tional grant from Servier.

ATF, activating transcription factor; CHOP, C/EBP (CCAAT/enhancer bind-ing protein) homologous protein; ER, endoplasmic reticulum; ERK, extracel-lular signal–regulated kinase; FACS, fluorescence-activated cell sorting; FFA,free fatty acid; GIIS, glucose-induced insulin secretion; I�B, inhibitory �B;IFN, interferon; IL, interleukin; iNOS, inducible nitric oxide synthase; JNK,c-Jun NH2-terminal kinase; MAPK, mitogen-activated protein kinase; NF,nuclear factor; SOCS, suppressor of cytokine signaling; TNF, tumor necrosisfactor.

© 2005 by the American Diabetes Association.

DIABETES, VOL. 54, SUPPLEMENT 2, DECEMBER 2005 S97

remains unclear whether this explains the observed func-tional loss (9).

�-Cell apoptosis may thus be a common feature of type1 and type 2 diabetes. Recent studies (rev. in 10 and 11)suggest that both forms of diabetes are characterized byintra-islet expression of inflammatory mediators (espe-cially the cytokine interleukin [IL]-1�), triggering a finalcommon pathway of �-cell apoptosis, progressive �-cellloss, and diabetes. Based on this hypothesis, a unifyingclassification of diabetes has been proposed (10). Againstthis background, we review herein the experimental evi-dence on the similarities and differences between themechanisms of �-cell death in type 1 and type 2 diabetes.

MECHANISMS OF �-CELL DEATH IN TYPE 1 DIABETES

Pancreatic �-cells are the target of an autoimmune assaultin type 1 diabetes, with invasion of the islets by mononu-clear cells in an inflammatory reaction termed “insulitis,”leading to loss of most �-cells after prolonged periods ofdisease (2). �-Cell death in the course of insulitis isprobably caused by direct contact with activated macro-phages and T-cells, and/or exposure to soluble mediatorssecreted by these cells, including cytokines, nitric oxide(NO), and oxygen free radicals (4). In vitro exposure of�-cells to IL-1�, or to IL-1� � interferon (IFN)-�, causesfunctional changes similar to those observed in pre-dia-betic patients, namely elevated proinsulin/insulin levels(12) and a preferential loss of first-phase insulin secretionin response to glucose, caused by an IL-1�–mediateddecrease in the docking and fusion of insulin granules tothe �-cell membrane (13). After prolonged exposure to

IL-1� � IFN-� and/or tumor necrosis factor (TNF)-�, butnot to either cytokine alone, this functional impairmentevolves to �-cell death (4).

Apoptosis, the main cause of �-cell death at the onset oftype 1 diabetes, is a highly regulated process, activatedand/or modified by extracellular signals, intracellular ATPlevels, phosphorylation cascades, and expression of pro-and anti-apoptotic genes (4). Cytokines induce stressresponse genes that are either protective or deleterious for�-cell survival. In extensive microarray experiments (14–17), we have identified �700 genes and expressed se-quence tags that are up- or downregulated in purified rat�-cells or insulin-producing cells after 1–24 h of exposureto IL-1� and/or IFN-�. The main findings of these studiesare summarized in Fig. 1. A detailed review of the genenetworks activated by cytokines in �-cells, and on the roleof chemokines produced by �-cells in the build up ofinsulitis, is provided by Eizirik et al. (18), while thecomplete list of �-cell–expressed genes, as detected byour microarray analyses, is accessible at the Beta CellGene Expression Bank (http://t1dbase.org/cgi-bin/enter_bcgb.cgi) (19). IL-1� activates the transcription factornuclear factor (NF)-�B (Fig. 1) in rodent and human isletcells (4), and prevention of NF-�B activation by an inhib-itory �B (I�B) “super-repressor” (20,21) protects pancre-atic �-cells against cytokine-induced apoptosis. In anadditional microarray analysis, we studied IL-1�–treated�-cells whose NF-�B activation was blocked by adenovi-rus-mediated expression of the I�B(SA)2 super-repressor(16). A total of 66 cytokine-responsive NF-�B–dependentgenes were identified, including genes coding for cyto-

FIG. 1. The transcription factor and gene networks putatively involved in the cytokine-promoted �-cell “decision” to undergo apoptosis. Thetranscription factors NF-�B and STAT-1 are the main regulators of the pathways triggered by IL-1� and IFN-�, respectively. The figure is basedon Refs. 14–18. MHC-1, major histocompatibility complex 1.

MECHANISMS OF �-CELL DEATH IN DIABETES

S98 DIABETES, VOL. 54, SUPPLEMENT 2, DECEMBER 2005

kines and chemokines and stress-related genes such asGADD153/CHOP [C/EBP (CCAAT/enhancer binding pro-tein) homologous protein] (a mediator of endoplasmicreticulum [ER] stress-induced cell death; see below) andc-myc. NF-�B was also found to downregulate (probablyindirectly, via NO production) the expression of othertranscription factors responsible for �-cell differentiationand function (e.g., PDX-1 and Isl-1). NF-�B regulatesexpression of inducible nitric oxide synthase (iNOS) in�-cells (22), and �50% of the �-cell genes modified after12 h of cytokine exposure are secondary to iNOS-mediatedNO formation (15). Of note, it has been recently describedthat transgenic expression of an NF-�B inhibitor under thecontrol of the pdx-1 promoter (blocking NF-�B during�-cell development) causes defective GLUT2 expressionand glucose-induced insulin secretion (GIIS) later inmouse life, suggesting that basal NF-�B is required fornormal insulin release (23). In our hands, however, block-ing basal NF-�B activity for 48–72 h in adult rat �-cellsaffected neither GLUT2 expression nor GIIS (16,21; A.K.Cardozo, D.L.E., unpublished data), suggesting that theputative physiological role of basal NF-�B activity is morerelevant during fetal �-cell development than during adultlife or is only detectable after prolonged NF-�B inhibition.In summary, IL-1�–induced NF-�B activation plays a cru-cial role in controlling multiple and distinct gene regula-tory networks, which affect the �-cell–differentiated stateand ER Ca2� homeostasis, attract and activate immunecells, and directly contribute to �-cell apoptosis.

Exposure of purified human or rodent �-cells to IL-1�alone is not sufficient to induce apoptosis, but when IL-1�is combined with IFN-�, �50% of these cells undergoapoptosis after 6–9 days (4). This suggests that IFN-�signal transduction must synergize with IL-1� signaling totrigger �-cell apoptosis (Fig. 1). IFN-� binds to cell surfacereceptors and activates the tyrosine kinases JAK1 andJAK2. These kinases phosphorylate the transcription fac-tor STAT-1, which dimerizes and translocates to the nu-cleus to bind to �-activated sites of diverse genes (4).STAT-1 mediates the potentiating effect of IFN-� on IL-1�–induced iNOS expression (22), and our recent observa-tions show that fluorescence-activated cell sorting(FACS)-purified �-cells from STAT-1–deficient mice(STAT-1�/�) are protected against IL-1� � IFN-�–inducedapoptosis (C. Gysemans, L. Ladriere, H. Callewaert, J.Rasschaert, D. Flamez, D.E. Levy, D.L.E., C. Mathieu,unpublished data). Because excessive activation of JAK/STAT signaling may lead to cell death, STAT transcrip-tional activity is regulated by multiple negative feedbackmechanisms. These include dephosphorylation of JAK andcytokine receptors by SHP, and inhibition of JAK enzy-matic activities by the suppressor of cytokine signaling(SOCS) family. Upregulation of SOCS-1 or SOCS-3 pro-tects �-cells in vitro and in vivo against cytokine-induceddeath (24,25). SOCS-3 also protects insulin-producing cellsagainst IL-1�–mediated apoptosis via NF-�B inhibition(26). The results summarized in Fig. 1 indicate that �-cellfate after cytokine exposure depends on the duration andseverity of perturbation of key �-cell gene networks. Theprecise identity and regulation of these gene networksremain to be elucidated, but the available data suggest animportant role for NF-�B and STAT-1.

It is of interest to understand how the cytokine-acti-vated gene expression patterns described in Fig. 1 actuallyresult in �-cell death. Some of the probable mechanisms

are outlined in Fig. 2. They include the following: 1)activation of the stress-activated protein kinases c-JunNH2-terminal kinase (JNK), p38 mitogen-activated proteinkinase (MAPK), and extracellular signal-regulated kinase(ERK); 2) triggering of ER stress; and 3) the release ofdeath signals from the mitochondria.

JNK is a member of the MAPK family. Pancreatic �-cellsexposed to IL-1� have an early and sustained increase inJNK activity, a phenomenon potentiated by IFN-� orTNF-� (4,11). Cell-permeable peptide inhibitors of JNKprevent cytokine-induced apoptosis in insulin-producingcells (27), but this remains to be confirmed in primary�-cells. p38 MAPK and ERK are also activated by cyto-kines, and pharmacological inhibition of these MAPKsdiminished cytokine-induced rat islet cell death (28,29),possibly by attenuating transcriptional activation of iNOS(28). However, when purified �-cells were exposed toIL-1� � IFN-� for 6–9 days, ERK, but not p38, inhibitorsprovided partial protection against apoptosis (30), sug-gesting that some of the protection by MAPK inhibitors inwhole islets is mediated via effects on other islet cells(such as resident macrophages). p38 may also increase theapoptotic propensity of the �-cell, since genetic downregu-lation of p38� results in a lowered sensitivity to cell deathinduced by the NO donor DETA/NONOate (N. Makeeva, J.Myers, N.W., unpublished data). In addition, the tumorsuppressor p53 is activated in response to cytokine-in-duced NO production (31). It is conceivable that stabiliza-tion of the pro-apoptotic protein p53 lies downstream ofthe NO-induced activation of MAPKs.

Disruption of ER homeostasis, as induced by changes inER Ca2� concentrations, triggers accumulation of un-folded proteins and activation of a specific stress re-sponse, known as the ER stress response (32). Thiscellular response is a coordinated attempt to restore ERhomeostasis and function, and it includes translationalattenuation, upregulation of ER chaperones, and degrada-tion of misfolded proteins. In case of prolonged and severeER stress, the apoptosis program is activated and exe-cuted by the transcription factor CHOP, the MAPK JNK,and caspase-12 (although it remains unclear whethercaspase-12 has a role in human cells) (32). Because of theirhigh rate of protein synthesis, �-cells are particularlysusceptible to ER stress (33), and NO donors trigger an ERstress response in �-cells leading to CHOP expression andapoptosis (34). We have recently shown that IL-1� � IFN-�inhibit SERCA2b expression, via NF-�B activation and NOproduction, and deplete ER Ca2� stores. This is followedby activation of diverse components of the ER stressresponse, including activation of IRE-1� and PERK/acti-vating transcription factor (ATF)-4, xbp1 mRNA process-ing, and induction of CHOP (35). Different from the �-cellresponse to chemical SERCA2b inhibitors or free fattyacids (FFAs) (36), cytokines neither activate ATF-6 norinduce the ER chaperone BiP (35). This defective ATF-6activation may deprive �-cells from a crucial defenseagainst ER stress, contributing to their exquisite vulnera-bility to cytokines.

Mitochondria are key organelles for �-cell function andsurvival (37). Paradoxically, mitochondria also play animportant role in triggering apoptosis (38). Members of theBcl-2 protein family regulate the mitochondrial responseto pro-apoptotic signals (38), preventing release of mito-chondrial proteins such as cytochrome c, which, whenliberated to the cytosol, sequentially activate caspase-9and -3 and execute cell death (39). Cytokines disrupt the

M. CNOP AND ASSOCIATES


mitochondrial membrane potential in RINm5F cells, whichis prevented by overexpression of the anti-apoptotic pro-tein Bcl-2 (40). Overexpression of Bcl-2 partially protectsmouse (41) and human (42) islets against cytokine-in-duced cell death, but does not prevent adenovirus-inducedislet cell death (43) or spontaneous diabetes in nonobesediabetic (NOD) mice (44). This suggests that other mech-anisms, bypassing Bcl-2, induce �-cell death in vivo and/orthat Bcl-2–regulated mitochondrial events and caspaseactivation are late steps in the apoptosis process, occur-ring when the cell fate has already been decided. In linewith the second possibility, blocking caspase-1 (induced in�-cells by cytokines [45]) decreases �-cell apoptosis after4 days of exposure to IL-1� � IFN-�, but it does notprevent their subsequent death by necrosis after 9 days (D.Liu, D.L.E., unpublished data). Other pro-apoptotic genesthat are induced by cytokines, as detected by microarrayanalysis (15), are indicated in Fig. 2 and include Bid, Bak,and caspase-3. An intriguing possibility is that an earlycytokine-induced “dialogue” between the nucleus, mito-chondria, and ER influences the decision of the �-cell toundergo apoptosis or not. In favor of this hypothesis,overexpression of free radical scavenging enzymes inmitochondria, but not in the cytosol, prevents IL-1�–induced NF-�B activation (46).


Insulin resistance, often associated with obesity, andinsulin secretion defects are major risk factors for type 2diabetes (9). A progressive decrease of �-cell functionleads to glucose intolerance, which is followed by type 2diabetes that inexorably aggravates with time (47). Thealterations of GIIS in human type 2 diabetes may theoret-ically result from changes in �-cell function, �-cell mass,or both. A decrease in �-cell mass is likely to play a role inthe pathogenesis of human type 2 diabetes (8,9) as it doesin rodent models of the disease (48,49). However, incontrast with type 1 diabetes, the 25–50% reduction in�-cell mass measured postmortem in type 2 diabeticpatients may not be important enough to account for theobserved loss of GIIS. Because �-cell mass cannot bemeasured in vivo, it remains unclear whether type 2diabetic patients had a lower �-cell mass early in life,failed to increase their �-cell mass in the face of insulinresistance, or had a progressive �-cell loss. The questionwhether the reduction in human �-cell mass results fromincreased �-cell apoptosis, reduced cell neogenesis/repli-cation, or both also remains unsettled (49). Based onresults obtained in rodent models of the disease and incultured rodent and human islet cells, it seems reasonableto assume that dyslipidemia and hyperglycemia negativelyaffect �-cell mass by increasing �-cell apoptosis in human

FIG. 2. Proposed model for the different pathways contributing to the execution of cytokine-induced �-cell apoptosis. Arrows indicate genes forwhich expression was modified by cytokines in a time course microarray analysis (15). �-Cell apoptosis is probably mediated by three mainpathways—namely JNK, ER stress, and liberation of pro-apoptotic proteins from the mitochondria.



type 2 diabetes (10,48). In the following paragraphs, wediscuss recent hypotheses on the mechanisms of glucotox-icity and lipotoxicity. Readers are directed to anotherrecent review (49) for information on other potentialagents causing �-cell dysfunction/death in type 2 diabetes.Glucotoxicity. Moderate or severe hyperglycemia cannotbe the primum movens in the pathophysiology of type 2diabetes, but it contributes to the reduction of GIIS (50).As such, it could contribute to the progression fromglucose intolerance to overt type 2 diabetes (47). Themechanisms by which hyperglycemia negatively affectsfunctional �-cell mass are still debated. Rodent �-cellschronically exposed to high glucose display several alter-ations of their phenotype, including changes in glucosestimulus-secretion coupling, gene expression, cell sur-vival, and cell growth (48,49). These alterations couldresult from cytokine-, oxidative stress–, or ER stress–induced changes in gene expression and cell survival(10,32,51) or from functional changes that are not directlyrelated to �-cell apoptosis, such as accumulation of glyco-gen (52).

Rodent �-cells display reduced expression of genesinvolved in GIIS in in vivo and in vitro models of prolongedexposure to high glucose. These include insulin, GLUT2,glucokinase and voltage-dependent Ca2� channels, andthe transcription factors that regulate their expression(53). These changes, which may play a role in the alter-ations of GIIS in rodent type 2 diabetes, have somesimilarities with those induced by cytokines (17,18). Onthe other hand, several genes expressed at low levels innormal �-cells are induced by hyperglycemia, includinghexokinase 1, lactate dehydrogenase and glucose-6 phos-phatase. In addition, pro- and anti-apoptotic factors, anti-oxidant enzymes, and some transcription factors areupregulated (53). Some of these genes, such as c-Myc, A20,and heme-oxygenase 1, are induced by hyperglycemia andcytokines, suggesting that both conditions share somecommon mechanisms to alter the �-cell phenotype. Thesuggestion that hyperglycemia increases �-cell productionof IL-1� in human islets provides such a unified hypothesisfor �-cell pathophysiology in type 1 and type 2 diabetes(10). However, the pattern of hyperglycemia-induced�-cell genes is only partly similar to that induced bycytokines (17,18,53). For instance, iNOS and I�B�, two

NF-�B–dependent genes markedly induced by IL-1�, arenot induced in rodent �-cells exposed to high glucose.Other genes, such as lactate dehydrogenase A, the mito-chondrial uncoupling protein UCP-2, and the transcriptionfactor CREM, are induced by hyperglycemia (53,54) anddownregulated by cytokines (17,55). Furthermore, hyper-glycemia induces �-cell hypertrophy, whereas cytokinetreatment does not, and the induction of �-cell apoptosisby high glucose is much lower than that produced bycytokines. These differences raise questions about the roleof IL-1�–induced NF-�B activation in �-cell glucotoxicity.These doubts are strengthened by our observations that,under various culture conditions, exposure of rat or hu-man islets or FACS-purified rat �-cells to high glucosedoes not increase IL-1� mRNA expression or NF-�BDNA-binding activity (56; Fig. 3; see also below).

It is generally assumed that oxidative stress activatesNF-�B activity in �-cells as in other cell types (53). Thisdoes not seem to be the case, since acute (57) or overnightexposure to low concentrations of hydrogen peroxidedoes not increase rat islet NF-�B activity and iNOS expres-sion (56). Islet c-Myc and heme-oxygenase 1 expressionare similarly induced by hydrogen peroxide and highglucose, and these effects are abrogated by the free radicalscavenger N-acetyl-L-cysteine. This suggests that �-cellglucotoxicity may, at least in part, result from an increasein �-cell oxidative stress and subsequent JNK activationthat is NF-�B independent (51,56). The main source ofreactive oxygen species in the �-cell is probably themitochondrial electron transport chain (58,59). It is there-fore possible that chronic stimulation of insulin secretionin states of insulin resistance induces oxidative stress.Other possible explanations for changes in �-cell functionand viability before overt hyperglycemia include activationof the ER stress pathway (also in the context of lipotox-icity; see below) and sustained elevation of cytosolic Ca2�

concentration (60).It is well established that chronic hyperglycemia leads

to �-cell degranulation and reduction in GIIS (48), but theeffect of hyperglycemia on the �-cell sensitivity to glucoseis controversial. A first group of studies indicates that theabsence of a glucose-induced rise in ATP production,perhaps due to hyperglycemia-induced expression of un-coupling protein 2, is responsible for defective GIIS (61).

FIG. 3. High glucose (28 mmol/l) does not induce NF-�B activationand DNA binding in �-cells, as assessed by immunofluorescenceusing an antibody directed against the p65 NF-�B subunit. FACS-purified rat �-cells were exposed to 10 mmol/l (control, A) or 28mmol/l glucose (B) for 12–24 h. The data at 12 h are shown here;similar observations were made at 24 h (not shown). As a positivecontrol, cells were exposed to IL-1� (30 units/ml, in mediumcontaining 10 mmol/l glucose [C]) during the last 30 min of culture.NF-�B is located in the cytosol at 10 mmol/l (A) or 28 mmol/l (B)glucose, whereas it translocates to the nucleus after exposure toIL-1�, indicating activation (C). Subcellular NF-�B localizationwas counted in 200–400 cells using the same experimental condi-tions as above (D). �, Cytoplasmic NF-�B localization; f, nuclearlocalization. The results are means � SE of three independentexperiments. *P < 0.001 vs. percent nuclear staining in the controlby two-sided paired t test (D). Original magnification �200.



These observations, conceptualized as �-cell “glucose de-sensitization,” seem in contradiction with other studiesshowing that �-cells exposed to hyperglycemia becomemore sensitive to glucose for the stimulation of mitochon-drial metabolism, proinsulin biosynthesis, and insulin se-cretion. This leads to maximal stimulation of triggeringand amplifying pathways of GIIS at low glucose (a conceptwe refer to as “glucose hypersensitization”) (62–64). Thisglucose hypersensitization, which results from higher ATPproduction at low glucose concentrations, may be due tothe accumulation of glycogen in �-cells (52,64). Glucosehypersensitization was observed together with a paradox-ical dissociation between glucose-induced Ca2� influx andinsulin secretion on one hand and a sustained elevation ofcytosolic Ca2� unaffected by glucose on the other, and it isassociated with a strong reduction of GIIS between 5 and10 mmol/l glucose. This concept of �-cell glucose hyper-sensitivity fits with the presence of fasting hyperinsulin-emia in type 2 diabetic patients and the observation (basedon autopsy material) that their �-cells are actively engagedin proinsulin synthesis (65).

Although both �-cell “glucose desensitization” and “glu-cose hypersensitization” may explain loss of GIIS at phys-iological glucose concentrations, these two hypotheseshave different implications for the role of apoptosis in�-cell glucotoxicity. Thus, �-cell glucose desensitization iscompatible with the concept that �-cell dysfunction(partly) results from �-cell apoptosis (10,51,61). In con-trast, �-cell glucose hypersensitization may result fromchanges in the expression of glycolytic enzymes (de-creased glucokinase and increased hexokinase 1 andlactate dehydrogenase expression), from the accumula-tion of glycogen at high glucose and its subsequent degra-dation at low glucose, or from other functional alterationsof �-cells (48,62,64), but not from apoptosis. We haverecently observed that overnight exposure of rat islets tolow concentrations of hydrogen peroxide induces glucosedesensitization and �-cell apoptosis that are both pre-vented by N-acetyl-L-cysteine. In contrast, a 1-week cultureat 30 mmol/l glucose, compared with 10 mmol/l, induces astate of glucose hypersensitization and a modest increasein �-cell apoptosis that are both unaffected by N-acetyl-L-cysteine (66). These results suggest that the various facetsof �-cell glucotoxicity may result from different patho-physiological mechanisms. Thus, after prolonged expo-sure to hyperglycemia, part of the surviving �-cells maystill be “glucose hypersensitive” while apoptosis is alreadyaffecting a small proportion of these cells. These differentpathophysiological mechanisms are compatible with theobservations that 1) after 3 weeks of diet-induced diabetesin the gerbil Psammomys obesus, a stage at which the�-cell mass is decreased (67), isolated islets were stillglucose hypersensitive (62); and 2) human islets trans-planted under the kidney capsule of hyperglycemic nudemice and maintained in vivo for 4 weeks have severelyimpaired GIIS, which can be dissociated from impairedglucose oxidation or protein synthesis and, under someconditions, from depleted insulin content or cell death(68,69).Lipotoxicity. Physical inactivity, energy-dense diets richin saturated fat, and central obesity predispose individualsto type 2 diabetes. Prospective studies in subjects at riskfor diabetes have shown that the development of abdom-inal obesity is correlated with loss of �-cell function andhence glucose intolerance (70; M.C., J. Vidal, R.L. Hull,K.M. Utzschneider, D.B. Carr, E.J. Boyko, W. Fujimoto,

S.E. Kahn, unpublished data). Autopsy data suggest thatthe progressive decline in insulin secretion in type 2diabetes is accompanied by a decrease in �-cell mass andthat this is secondary to increased �-cell apoptosis. Thus,it is conceivable that circulating adipose tissue–derivedproducts, such as FFAs and adipokines, play a direct rolein pancreatic �-cell dysfunction and death. A high plasmaconcentration of FFAs is indeed a risk factor for thedevelopment of type 2 diabetes, independently of itseffects on insulin sensitivity (71).

Circulating FFAs are solubilized and transported inmillimolar concentrations, by virtue of their tight bindingto albumin. Unbound FFA levels measure in the nanomo-lar range (5–20 nmol/l), a concentration at which they arerapidly taken up via a protein-mediated transport. FFAsacutely stimulate insulin secretion, but prolonged �-cellexposure to high FFA levels reduces GIIS in vitro (72) andin vivo, especially in individuals genetically predisposed totype 2 diabetes (73). Studies in the ZDF rat indicate thathigh circulating FFAs and triglyceride levels induce tri-glyceride accumulation in pancreatic islets (74). The asso-ciated rise in cytoplasmic FFA levels would increaseceramide formation and induce iNOS, resulting in NO-mediated �-cell apoptosis (75).

In our in vitro experiments (36,76,77), we used physio-logical concentrations of palmitate and oleate. FFAs aretoxic to FACS-purified rat �-cells (36,76) and insulin-producing INS-1E cells (36). Cytotoxicity depends on theunbound FFA concentration and is greater for palmitatethan oleate. FFA-induced cell damage results in apoptosisand, to a lesser extent, necrosis in �-cells (76) and mostlyin apoptosis in INS-1E cells (36). The toxic effects of FFAsare potentiated when �-cells are concomitantly exposed tohigh glucose levels (78,79). FFA cytotoxicity does notdepend on mitochondrial FFA oxidation, because eto-moxir, an inhibitor of carnitine palmitoyltransferase I, didnot alter FFA-induced �-cell death, and bromopalmitate, anonmetabolizable analog, was as toxic as palmitate (76).FFA-induced cell death occurred in the absence of iNOSexpression or NO production (36,76), and it was notcounteracted by antioxidant or free radical scavengingcompounds (76), suggesting that oxidative stress is not itsmain mediator. Moreover, oleate or palmitate did notactivate NF-�B in INS-1E or �-cells (36), at low (6.1mmol/l), medium (10 mmol/l), or high (28 mmol/l) glucoselevels (36; I. Kharroubi, D.L.E., M.C., unpublished data). Itwas suggested that FFA cytotoxicity could be counter-acted by the peroxisome proliferator–activated receptor-�agonist troglitazone through lowering islet triglyceridecontent (80). In our hands, however, troglitazone did notimprove survival of FFA-exposed �-cells, but rather sensi-tized them to necrosis and apoptosis at low FFA concen-trations (77). Furthermore, FFA-induced cytoplasmictriglyceride accumulation was inversely correlated to�-cell death (76). A mixture of oleate and palmitate causedthe lowest cell death and the highest triglyceride accumu-lation, whereas bromopalmitate, which did not increasecellular triglycerides, exerted the highest toxicity (76).These findings suggest that storage of excess FFAs astriglycerides protects the cell against accumulation ofpotentially deleterious fatty acyl-CoA.

FFA-induced �-cell toxicity might also occur at the ERlevel, where FFA esterification takes place. Using electronmicroscopy, we observed that the ER of FFA-exposed�-cells is dilated (M.C., unpublished data), and we there-fore examined whether FFAs induce ER stress. Both



oleate and palmitate caused alternative splicing of XBP-1,activation of ATF-6, and induction of the ER chaperoneBiP in INS-1E cells (36). In addition to these specific ERstress markers, there was induction of ATF-4 and CHOP(36). It is thus conceivable that a high FFA load, thatexceeds the �-cell’s esterification capacity, impairs ERfunctions and triggers an ER stress response, thus contrib-uting to �-cell toxicity. The mechanism by which FFAscause ER stress remains to be elucidated, but (over)stimu-lation of FFA esterification in the ER might result indelayed processing and export of newly synthesized pro-teins, whereas saturated triglycerides may precipitate attheir site of synthesis in the ER because of their highmelting point. FFAs might also impair ER Ca2� handling(81), whereas conditions that increase �-cell secretorydemand, such as insulin resistance or high glucose, mightamplify ER stress and �-cell death. ER stress has beenrecently proposed as the cellular/molecular mechanismlinking obesity with insulin resistance (82,83). FFAs mightthus be responsible for the ER stress response observed inthe hepatocytes and adipocytes of obese mice (82,83),while hampering in parallel pancreatic �-cell function/viability (36). If that is the case, these intriguing novelobservations (36,82,83) place ER stress as a commonmolecular pathway for the two main causes of type 2diabetes—namely insulin resistance and loss of �-cellmass.

SIMILARITIES AND DIFFERENCES BETWEEN THE


The development of novel approaches to prevent �-celldeath in diabetes depends on our knowledge of themechanisms leading to �-cell demise. Thus, if the mecha-nisms of �-cell apoptosis were similar in type 1 and type 2diabetes, it would be logical to search for common inter-

ventions in both forms of diabetes. Let us thereforeexamine the evidence pointing to the differences andsimilarities between the mechanisms of �-cell death andanalyze whether sufficient information is available to sup-port a similar “etiological” intervention in type 1 and type2 diabetes.Novel in vivo evidence for the in situ expression of

mediators of �-cell death in animal models of diabe-

tes. The NOD mouse and the BB rat are the most usedanimal models of type 1 diabetes (84), but a new model fortype 1 diabetes has been recently described—the IDDM(LEW.1AR1/Ztm-iddm) rat (85,86). This latter model is ofparticular interest, since IDDM rats have a well-preservedcellular immune system, there is no sex bias in theincidence of diabetes, and detailed studies of the eventsleading to �-cell death are possible (see below).

ED1� macrophages are the predominant infiltratingimmune cell species during the early stages of insulitis forall three animal models of type 1 diabetes. This is followedby an increasing infiltration by cytotoxic CD8� lympho-cytes, predominating at the onset of diabetes. Other im-mune cells participating in the insulitis are CD4�

lymphocytes, NK cells, and B-cells (rev. in 4,84). Theseimmune cells are activated (87) and express proinflamma-tory cytokines such as IL-1�, TNF-�, and IFN-� (88,89).IL-1� and TNF-�, but not IFN-�, are detected in theinfiltrating immune cells in the IDDM rat, but the pancre-atic �-cells do not express these cytokines at any of thestages leading to overt diabetes (90; Fig. 4A and B; A.J., A.Gunther, H.-J. Hedrich, D. Wedekind, M. Tiedge, S.L.,unpublished data). Data from other models suggest that�-cells express chemokines such as MCP-1 and IP-10,which may contribute to the buildup of insulitis (91).Detailed morphological studies in the IDDM rat, usingboth in situ PCR and immunohistochemistry (A.J., A.

FIG. 4. Morphology of an islet from a diabeticIDDM (LEW.1AR1/Ztm-iddm) rat (A–C) exhibit-ing hyperglycemia (21.4 mmol/l) and hypoinsu-linemia (0.5 ng/ml) 1 day after diabetesmanifestation and of an islet from a type 2diabetes Psammomys obesus (sand rat) (D–F)exhibiting hyperglycemia (17.5 mmol/l) and hy-perinsulinemia (1.8 ng/ml) after 3 weeks on ahigh-energy diet. The sections were immuno-stained for IL-1� (A and D), iNOS (B and E), andactivated caspase-3 (C and F) and show cyto-plasmic immunoreactivities only in the infil-trated islets of the type 1 diabetic animal (A–C).Infiltrating immune cells in the diabetic IDDMrat (A–C) are mostly ED1� macrophages (ar-rows) and CD8� T-cells (arrowheads). Thesecells express immunoreactivity for IL-1� (A)and iNOS (B) but not for activated caspase-3(C). Pancreatic �-cells undergoing apoptosis(thick arrows), in contrast, express immunore-activity for iNOS and activated caspase-3, butnot for IL-1�. The few infiltrating immune cellsin the islet of a diabetic Psammomys obesus

(D–F) are exclusively ED1� macrophages (ar-rows). These cells show no signs of immunore-activity for IL-1� (D), iNOS (E), or activatedcaspase-3 (F). �-Cells (thick arrows) of Psam-

momys showed signs of necrotic destructionincluding intra- and intercellular vacuolizationwithout expression of IL-1� (D), iNOS (E), oractivated caspase-3 (F). These �-cells showedno signs of nuclear heterochromatin condensa-tion. The same findings were made after 1 weekof a high-energy diet. ED1� macrophages, CD8�

T-cells, and pancreatic �-cells were identified bysequential sections immunostained with specificantibodies against the given cell type as previ-ously described (67,86) (data not shown). Orig-inal magnification �400.



Gunther, H.-J. Hedrich, D. Wedekind, M. Tiedge, S.L.,unpublished data), suggest the following sequence ofevents: 1) islets are initially infiltrated by macrophages,followed by CD8� and CD4� cells; 2) this infiltration isaccompanied by a high IL-1� (Fig. 4A) and TNF-� expres-sion in the invading immune cells (but not in the �-cells)and iNOS expression (Fig. 4B) in both immune cells and�-cells; and 3) the �-cells under attack progressivelyexpress procaspase-3 (Fig. 4C) and undergo apoptosis.These observations suggest that proinflammatory cyto-kines are synthesized and released by the activated infil-trating immune cells, but not by the �-cells themselves,leading to apoptotic �-cell death in a paracrine fashion.

Loss of pancreatic �-cell mass is slow in type 2 diabetes,and there is no evidence for mononuclear cell infiltration(2). This is well documented in a number of type 2 diabetesanimal models, including the Psammomys obesus (sandrat) (92) and the GK rat (93). When Psammomys areplaced on a high-carbohydrate diet, they rapidly evolve toa diabetic state because of the loss of endocrine pancreasfunction and �-cell destruction (48,67). In contrast to thetype 1 diabetes models (see above), �-cell demise occursmostly by necrosis (67). The necrotic cells are removed byscavenger macrophages, which, at variance from the type1 diabetes situation, are not activated and do not expressthe proinflammatory cytokines IL-1�, IFN-�, or TNF-�(Fig. 4D; A.J., S.L., unpublished data). Importantly, the�-cells from these animals do not express IL-1�, iNOS, orcaspase-3, as evaluated by immunohistochemistry (Fig. 4E

and F) and in situ PCR over the course of high-energydiet-induced metabolic changes (1–3 weeks; A.J., S.L.,unpublished data). As a positive control, IL-1� mRNAexpression was confirmed by in situ PCR using immunecells of pancreas draining lymph nodes.

The same sequence of events seems to take place in thephysiological situation, where �-cells undergoing apopto-sis during their cell renewal cycle are removed by nonac-tivated macrophages (94). Even when the �-cell turnoverrate is increased by administration of thyroid hormones(67,95), the increased demand for removal of apoptoticcells does not trigger macrophage activation or cytokineexpression (A.J., S.L., unpublished data). These observa-tions suggest a sequence of events that is different in type1 and type 2 diabetes models. Thus, �-cells die by necrosisor apoptosis in type 2 diabetes, but the cause of death isnot related to cytokine production by infiltrating mononu-clear cells or the �-cells themselves. The dead �-cellsattract scavenger macrophages, which in this case are theconsequence rather than the cause of �-cell death (Fig. 5).Analysis of the evidence for putative final commonpathways of �-cell death in type 1 and type 2 diabe-tes. It has been recently suggested that �-cells exposed invitro to high glucose produce IL-1�, thus activating NF-�Band Fas signaling and consequently triggering apoptosis(10,11). Another report indicated that FFAs also activateNF-�B in �-cells (96). Because both IL-1� and NF-�� arecrucial mediators of �-cell death in type 1 diabetes (4), theIL-1�–NF-�B pathway was suggested as a “common final

FIG. 5. Overview of the putative sequence of events leading to �-cell death in animal models of type 1 and type 2 diabetes. For additionalinformation on the mechanisms of �-cell apoptosis in type 1 diabetes, see Figs. 1 and 2. T1D, type 1 diabetes; T2D, type 2 diabetes.



pathway” for �-cell death in both forms of diabetes (11),providing a rationale for revising and unifying the classi-fication and treatment of diabetes (10).

As discussed above, exposure to IL-1� alone is notsufficient to kill human or rodent �-cells, and the signaltransduction of IFN-� is also required for �-cell demise. Toexclude that exposure of �-cells to high glucose or FFAsinduces the IFN-� pathway, we reviewed the results of fivedifferent microarray analyses of human or rodent isletsexposed to these nutrients (list of microarray studiesprovided upon request). We also contacted some of theauthors to make sure that small changes in gene expres-sion were not overlooked (D. Flamez, D. Melloul, G. Webb,personal communications). The data were compared withthe gene expression patterns in rat (14) or human (P.Ylipaasto, B. Kutlu, S. Raisilainen, J. Rasschaert, T. Teeri-joki, O. Korsgren, R. Lahesmaa, T. Hovi, D.L.E., T. Otokon-ski, M. Roivainen, unpublished data) islets exposed toIFN-�. The mRNAs whose expression was most aug-mented by IFN-� in �-cells were the transcription factorsSTAT-1, IRF-1, and IRF-7 and the chemokine CXCL 10(IP-10). Glucose or FFAs modified none of these genes in�-cells, practically excluding the IFN-�–STAT-1 pathwayas a mediator of glucotoxicity or lipotoxicity. We thereforefocused on IL-1�–NF-�B as the putative “common finalpathway” for �-cell death.

As mentioned above, there is strong evidence that IL-1�contributes to �-cell death in type 1 diabetes via activationof NF-�B. Which is the evidence that FFAs induce IL-1�production or NF-�B activation in �-cells? We (36) andothers (97) did not observe FFA-induced NF-�B activationin �-cells using three different techniques (gel shift, ELISA,and immunohistochemistry), and there are no reports ofFFA-induced IL-1� expression in these cells. Moreover,FFAs do not induce expression of the NF-�B–dependentgenes iNOS and MCP-1 in rodent �-cells (36,98). Whatabout high glucose? Most of the in vitro data supportingglucose-induced IL-1� production and NF-�B activationwere obtained by one group (rev. in 11). Based on theirobservations, this group initiated a clinical trial with theIL-1 receptor antagonist in type 2 diabetic patients (10). Ofconcern is that there is no in vivo evidence in animalmodels that blocking IL-1� protects �-cells against gluco-toxicity. In addition, it has been difficult to reproduce thekey findings of this “unifying hypothesis.” Thus, we couldnot detect glucose-induced NF-�B activation or IL-1�expression in rat islets or FACS-purified �-cells (Fig. 3;56). We examined whether this was due to a speciesdifference between rat (56) and human (10,11) islets.Exposure of five preparations of human islets to increas-ing glucose concentrations (11 and 28 vs. 5.6 mmol/l) didnot lead to the expression and release of IL-1� or otherNF-�B–dependent genes, such as I�B� or MCP-1 (99). Ofnote, the concentration of IL-1� released by human isletsexposed to 28 mmol/l glucose is negligible, i.e., 50-foldbelow the amount of IL-1� released by human monocytes(99). Moreover, there was no glucose-induced Fas mRNAexpression (99), the proposed NF-�B–dependent mecha-nism by which glucose causes �-cell death (11). In linewith our findings, islets isolated from mice deficient ineither the IL-1 receptor or Fas were not protected againsthigh glucose–induced �-cell death, and Fas was not de-tectable in wild-type mouse islets cultured at high glucose(100). As a whole, these observations argue against a rolefor IL-1�, NF-�B, or Fas in high glucose–induced �-celldeath.

In conclusion, the suggestion that �-cells are killed by asimilar mechanism in type 1 and type 2 diabetes isprobably an oversimplification, not supported by convinc-ing data. This oversimplification may bring confusion to adifficult and complex field and promote testing of noveltherapeutic approaches in humans without adequate ex-perimental support.

ACKNOWLEDGMENTS

Work by the authors was supported by the following: theJuvenile Diabetes Research Foundation Center for Preven-tion of �-Cell Destruction in Europe, under grant number4-2002-457 (D.L.E.); a European Foundation for the Studyof Diabetes/Johnson & Johnson Type 2 Diabetes ResearchGrant (M.C. and D.L.E.); the Fonds National de la Recher-che Scientifique (FNRS), Belgium (M.C., J.-C.J., andD.L.E.); Actions de Recherche Concertees of the BelgiumFrench Community (J.-C.J. and D.L.E.); Swedish MedicalResearch Council (72P-12995, 12X-11564, 12X-109) (N.W.);EFSD/Lilly European Diabetes Research Program (N.W.);the Swedish Diabetes Association (N.W.); the Family Ern-fors Fund (N.W.); the Deutsche ForschungsgemeinschaftGrant Jo 395/1-1/2 (A.J.); and National Institutes of HealthGrant 1R21AI55464–01 (S.L.).

We thank Dr. A.K. Cardozo for help in preparing Fig. 1and I. Kharroubi for performing the NF-�B immunostain-ing in Fig. 3.

NOTE ADDED IN PROOF

In agreement with the lack of IL-1� expression or releaseby human islets exposed to high glucose in vitro (asdiscussed in this review), recent data do not support a rolefor IL-1� in type 2 diabetes in vivo. Two studies, usingrespectively real-time RT-PCR and microarray analysis,demonstrate that IL-1� and Fas expression in islets iso-lated from type 2 diabetic patients is not increased ascompared with islets from nondiabetic controls (99,101).

REFERENCES

1. Expert Committee on the Diagnosis and Classification of DiabetesMellitus: Report of the Expert Committee on the Diagnosis and Classifi-cation of Diabetes Mellitus. Diabetes Care 20:1183–1197, 1997

2. Kloppel G, Lohr M, Habich K, Oberholzer M, Heitz PU: Islet pathology andthe pathogenesis of type 1 and type 2 diabetes mellitus revisited. Surv

Synth Pathol Res 4:110–125, 19853. Srikanta S, Ganda OP, Jackson RA, Gleason RE, Kaldany A, Garovoy MR,

Milford EL, Carpenter CB, Soeldner JS, Eisenbarth GS: Type I diabetesmellitus in monozygotic twins: chronic progressive � cell dysfunction.Ann Intern Med 99:320–326, 1983

4. Eizirik DL, Mandrup-Poulsen T: A choice of death: the signal-transductionof immune-mediated �-cell apoptosis. Diabetologia 44:2115–2133, 2001

5. Clark A, Wells CA, Buley ID, Cruickshank JK, Vanhegan RI, Matthews DR,Cooper GJ, Holman RR, Turner RC: Islet amyloid, increased A-cells,reduced B-cells and exocrine fibrosis: quantitative changes in the pan-creas in type 2 diabetes. Diabetes Res 9:151–159, 1988

6. Guiot Y, Sempoux C, Moulin P, Rahier J: No decrease of the �-cell massin type 2 diabetic patients. Diabetes 50 (Suppl. 1):S188, 2001

7. Sakuraba H, Mizukami H, Yagihashi N, Wada R, Hanyu C, Yagihashi S:Reduced �-cell mass and expression of oxidative stress-related DNAdamage in the islet of Japanese type II diabetic patients. Diabetologia

45:85–96, 20028. Butler AE, Janson J, Bonner-Weir S, Ritzel R, Rizza RA, Butler PC: �-Cell

deficit and increased �-cell apoptosis in humans with type 2 diabetes.Diabetes 52:102–110, 2003

9. Kahn SE: The relative contributions of insulin resistance and �-celldysfunction to the pathophysiology of type 2 diabetes. Diabetologia

46:3–19, 200310. Donath MY, Halban PA: Decreased �-cell mass in diabetes: significance,

mechanisms and therapeutic implications. Diabetologia 47:581–589, 2004



11. Donath MY, Storling J, Maedler K, Mandrup-Poulsen T: Inflammatorymediators and islet �-cell failure: a link between type 1 and type 2diabetes. J Mol Med 81:455–470, 2003

12. Hostens K, Pavlovic D, Zambre Y, Ling Z, Van Schravendijk C, Eizirik DL,Pipeleers DG: Exposure of human islets to cytokines can result indisproportionately elevated proinsulin release. J Clin Invest 104:67–72, 1999

13. Ohara-Imaizumi M, Cardozo AK, Kikuta T, Eizirik DL, Nagamatsu S: Thecytokine interleukin-1� reduces the docking and fusion of insulin gran-ules in pancreatic �-cells, preferentially decreasing the first phase ofexocytosis. J Biol Chem 279:41271–41274, 2004

14. Rasschaert J, Liu D, Kutlu B, Cardozo AK, Kruhoffer M, Orntoft TF, EizirikDL: Global profiling of double stranded RNA- and IFN-�-induced genes inrat pancreatic � cells. Diabetologia 46:1641–1657, 2003

15. Kutlu B, Cardozo AK, Darville MI, Kruhoffer M, Magnusson N, Orntoft T,Eizirik DL: Discovery of gene networks regulating cytokine-induceddysfunction and apoptosis in insulin-producing INS-1 cells. Diabetes

52:2701–2719, 200316. Cardozo AK, Heimberg H, Heremans Y, Leeman R, Kutlu B, Kruhoffer M,

Orntoft T, Eizirik DL: A comprehensive analysis of cytokine-induced andnuclear factor-�B-dependent genes in primary rat pancreatic �-cells.J Biol Chem 276:48879–48886, 2001

17. Cardozo AK, Kruhoffer M, Leeman R, Orntoft T, Eizirik DL: Identificationof novel cytokine-induced genes in pancreatic �-cells by high-densityoligonucleotide arrays. Diabetes 50:909–920, 2001

18. Eizirik DL, Kutlu B, Rasschaert J, Darville M, Cardozo AK: Use ofmicroarray analysis to unveil transcription factor and gene networkscontributing to � cell dysfunction and apoptosis. Ann N Y Acad Sci

1005:55–74, 200319. Smink LJ, Helton EM, Healy BC, Cavnor CC, Lam AC, Flamez D, Burren

OS, Wang Y, Dolman GE, Burdick DB, Everett VH, Glusman G, Laneri D,Rowen L, Schuilenburg H, Walker NM, Mychaleckyj J, Wicker LS, EizirikDL, Todd JA, Goodman N: T1DBase, a community web-based resourcefor type 1 diabetes research. Nucleic Acid Res 33:D544–D549, 2005

20. Giannoukakis N, Rudert WA, Trucco M, Robbins PD: Protection of humanislets from the effects of interleukin-1� by adenoviral gene transfer of anI�B repressor. J Biol Chem 275:36509–36513, 2000

21. Heimberg H, Heremans Y, Jobin C, Leemans R, Cardozo AK, Darville M,Eizirik DL: Inhibition of cytokine-induced NF-�B activation by adenovi-rus-mediated expression of a NF-�B super-repressor prevents �-cellapoptosis. Diabetes 50:2219–2224, 2001

22. Darville MI, Eizirik DL: Regulation by cytokines of the inducible nitricoxide synthase promoter in insulin-producing cells. Diabetologia 41:1101–1108, 1998

23. Norlin S, Ahlgren U, Edlund H: Nuclear factor-�B activity in �-cells isrequired for glucose-stimulated insulin secretion. Diabetes 54:125–132, 2005

24. Karlsen AE, Ronn SG, Lindberg K, Johannesen J, Galsgaard ED, Pociot F,Nielsen JH, Mandrup-Poulsen T, Nerup J, Billestrup N: Suppressor ofcytokine signaling 3 (SOCS-3) protects �-cells against interleukin-1�- andinterferon-�-mediated toxicity. Proc Natl Acad Sci U S A 98:12191–12196,2001

25. Flodstrom M, Tsai D, Fine C, Maday A, Sarvetnick N: Diabetogenicpotential of human pathogens uncovered in experimentally permissive�-cells. Diabetes 52:2025–2034, 2003

26. Karlsen AE, Heding PE, Frobose H, Ronn SG, Kruhoffer M, Orntoft TF,Darville M, Eizirik DL, Pociot F, Nerup J, Mandrup-Poulsen T, BillestrupN: Suppressor of cytokine signalling (SOCS)-3 protects � cells againstIL-1�-mediated toxicity through inhibition of multiple nuclear factor-�B-regulated proapoptotic pathways. Diabetologia 47:1998–2011, 2004

27. Bonny C, Oberson A, Negri S, Sauser C, Schorderet DF: Cell-permeablepeptide inhibitors of JNK: novel blockers of �-cell death. Diabetes

50:77–82, 200128. Larsen CM, Wadt KA, Juhl LF, Andersen HU, Karlsen AE, Su MS, Seedorf

K, Shapiro L, Dinarello CA, Mandrup-Poulsen T: Interleukin-1�-inducedrat pancreatic islet nitric oxide synthesis requires both the p38 andextracellular signal-regulated kinase 1/2 mitogen-activated protein ki-nases. J Biol Chem 273:15294–15300, 1998

29. Saldeen J, Lee JC, Welsh N: Role of p38 mitogen-activated protein kinase(p38 MAPK) in cytokine-induced rat islet cell apoptosis. Biochem Phar-

macol 61:1561–1569, 200130. Pavlovic D, Andersen NA, Mandrup-Poulsen T, Eizirik DL: Activation of

extracellular signal-regulated kinase (ERK)1/2 contributes to cytokine-induced apoptosis in purified rat pancreatic �-cells. Eur Cytokine Netw

11:267–274, 200031. Saldeen J, Tillmar L, Karlsson E, Welsh N: Nicotinamide- and caspase-

mediated inhibition of poly(ADP-ribose) polymerase are associated withp53-independent cell cycle (G2) arrest and apoptosis. Mol Cell Biochem

243:113–122, 2003

32. Schroder M, Kaufman RJ: ER stress and the unfolded protein response.Mutat Res 569:29–63, 2005

33. Harding HP, Zeng H, Zhang Y, Jungries R, Chung P, Plesken H, SabatiniDD, Ron D: Diabetes mellitus and exocrine pancreatic dysfunction inperk�/� mice reveals a role for translational control in secretory cellsurvival. Mol Cell 7:1153–1163, 2001

34. Oyadomari S, Takeda K, Takiguchi M, Gotoh T, Matsumoto M, Wada I,Akira S, Araki E, Mori M: Nitric oxide-induced apoptosis in pancreatic �cells is mediated by the endoplasmic reticulum stress pathway. Proc Natl

Acad Sci U S A 98:10845–10850, 200135. Cardozo AK, Ortis F, Storling J, Feng YM, Rasschaert J, Tonnesen M, Van

Eylen F, Mandrup-Poulsen T, Herchuelz A, Eizirik DL: Cytokines down-regulate the sarcoendoplasmic reticulum pump Ca2� ATPase 2b anddeplete endoplasmic reticulum Ca2�, leading to induction of endoplasmicreticulum stress in pancreatic �-cells. Diabetes 54:452–461, 2005

36. Kharroubi I, Ladriere L, Cardozo AK, Dogusan Z, Cnop M, Eizirik DL: Freefatty acids and cytokines induce pancreatic �-cell apoptosis by differentmechanisms: role of nuclear factor-�B and endoplasmic reticulum stress.Endocrinology 145:5087–5096, 2004

37. Maechler P, Wollheim CB: Mitochondrial function in normal and diabetic�-cells. Nature 414:807–812, 2001

38. Newmeyer DD, Ferguson-Miller S: Mitochondria: releasing power for lifeand unleashing the machineries of death. Cell 112:481–490, 2003

39. Friedlander RM: Apoptosis and caspases in neurodegenerative diseases.N Engl J Med 348:1365–1375, 2003

40. Barbu A, Welsh N, Saldeen J: Cytokine-induced apoptosis and necrosisare preceded by disruption of the mitochondrial membrane potential inpancreatic RINm5F cells: prevention by Bcl-2. Mol Cell Endocrinol

190:75–82, 200241. Iwahashi H, Hanafusa T, Eguchi Y, Nakajima H, Miyagawa J, Itoh N,

Tomita K, Namba M, Kuwajima M, Noguchi T, Tsujimoto Y, Matsuzawa Y:Cytokine-induced apoptotic cell death in a mouse pancreatic �-cell line:inhibition by Bcl-2. Diabetologia 39:530–536, 1996

42. Rabinovitch A, Suarez-Pinzon W, Strynadka K, Ju Q, Edelstein D, Brown-lee M, Korbutt GS, Rajotte RV: Transfection of human pancreatic isletswith an anti-apoptotic gene (bcl-2) protects �-cells from cytokine-induceddestruction. Diabetes 48:1223–1229, 1999

43. Barbu AR, Akusjarvi G, Welsh N: Adenoviral-induced islet cell cytotoxic-ity is not counteracted by Bcl-2 overexpression. Mol Med 8:733–741, 2002

44. Allison J, Thomas H, Beck D, Brady JL, Lew AM, Elefanty A, Kosaka H,Kay TW, Huang DC, Strasser A: Transgenic overexpression of humanBcl-2 in islet � cells inhibits apoptosis but does not prevent autoimmunedestruction. Int Immunol 12:9–17, 2000

45. Karlsen AE, Pavlovic D, Nielsen K, Jensen J, Andersen HU, Pociot F,Mandrup-Poulsen T, Eizirik DL, Nerup J: Interferon-� induces interleu-kin-1 converting enzyme expression in pancreatic islets by an interferonregulatory factor-1-dependent mechanism. J Clin Endocrinol Metab

85:830–836, 200046. Azevedo-Martins AK, Lortz S, Lenzen S, Curi R, Eizirik DL, Tiedge M:

Improvement of the mitochondrial antioxidant defense status preventscytokine-induced nuclear factor-�B activation in insulin-producing cells.Diabetes 52:93–101, 2003

47. U.K. Prospective Diabetes Study Group: U.K. Prospective Diabetes Study16. Overview of 6 years’ therapy of type II diabetes: a progressive disease.Diabetes 44:1249–1258, 1995

48. Kaiser N, Leibowitz G, Nesher R: Glucotoxicity and �-cell failure in type2 diabetes mellitus. J Pediatr Endocrinol Metab 16:5–22, 2003

49. Rhodes CJ: Type 2 diabetes: a matter of �-cell life and death? Science

307:380–384, 200550. Kosaka K, Kuzuya T, Akanuma Y, Hagura R: Increase in insulin response

after treatment of overt maturity-onset diabetes is independent of themode of treatment. Diabetologia 18:23–28, 1980

51. Robertson RP, Harmon J, Tran PO, Tanaka Y, Takahashi H: Glucosetoxicity in �-cells: type 2 diabetes, good radicals gone bad, and theglutathione connection. Diabetes 52:581–587, 2003

52. Malaisse WJ, Maggetto C, Leclercq-Meyer V, Sener A: Interference ofglycogenolysis with glycolysis in pancreatic islets from glucose-infusedrats. J Clin Invest 91:432–436, 1993

53. Weir GC, Laybutt DR, Kaneto H, Bonner-Weir S, Sharma A: �-Celladaptation and decompensation during the progression of diabetes.Diabetes 50 (Suppl. 1):S154–S159, 2001

54. Zhou YP, Marlen K, Palma JF, Schweitzer A, Reilly L, Gregoire FM, XuGG, Blume JE, Johnson JD: Overexpression of repressive cAMP responseelement modulators in high glucose and fatty acid-treated rat islets: acommon mechanism for glucose toxicity and lipotoxicity? J Biol Chem

278:51316–51323, 200355. Li LX, Yoshikawa H, Egeberg KW, Grill V: Interleukin-1� swiftly down-



regulates UCP-2 mRNA in �-cells by mechanisms not directly coupled totoxicity. Cytokine 23:101–107, 2003

56. Elouil H, Cardozo AK, Eizirik DL, Henquin JC, Jonas JC: High glucose andhydrogen peroxide increase c-Myc and haeme-oxygenase 1 mRNA levelsin rat pancreatic islets without activating NF-�B. Diabetologia 48:496–505, 2005

57. Eizirik DL, Flodstrom M, Karlsen AE, Welsh N: The harmony of thespheres: inducible nitric oxide synthase and related genes in pancreaticbeta cells. Diabetologia 39:875–890, 1996

58. Gurgul E, Lortz S, Tiedge M, Jorns A, Lenzen S: Mitochondrial catalaseoverexpression protects insulin-producing cells against toxicity of reac-tive oxygen species and proinflammatory cytokines. Diabetes 53:2271–2280, 2004

59. Fridlyand LE, Philipson LH: Does the glucose-dependent insulin secretionmechanism itself cause oxidative stress in pancreatic �-cells? Diabetes

53:1942–1948, 200460. Grill V, Bjorklund A: Overstimulation and �-cell function. Diabetes 50

(Suppl. 1):S122–S124, 200161. Brownlee M: A radical explanation for glucose-induced � cell dysfunc-

tion. J Clin Invest 112:1788–1790, 200362. Pertusa JA, Nesher R, Kaiser N, Cerasi E, Henquin JC, Jonas JC: Increased

glucose sensitivity of stimulus-secretion coupling in islets from Psammo-

mys obesus after diet induction of diabetes. Diabetes 51:2552–2560, 200263. Ling Z, Pipeleers DG: Prolonged exposure of human � cells to elevated

glucose levels results in sustained cellular activation leading to a loss ofglucose regulation. J Clin Invest 98:2805–2812, 1996

64. Khaldi MZ, Guiot Y, Gilon P, Henquin JC, Jonas JC: Increased glucosesensitivity of both triggering and amplifying pathways of insulin secretionin rat islets cultured for 1 wk in high glucose. Am J Physiol Endocrinol

Metab 287:E207–E217, 200465. Sempoux C, Guiot Y, Dubois D, Moulin P, Rahier J: Human type 2

diabetes: morphological evidence for abnormal �-cell function. Diabetes

50 (Suppl. 1):S172–S177, 200166. Khaldi MZ, Elouil H, Henquin JC, Jonas JC: Distinct effects of the

antioxidant N-acetyl-L-cystein on �-cell dysfunction induced by highglucose and hydrogen peroxide in cultured rat islets. Diabetologia 47(Suppl. 1):A173, 2004

67. Jorns A, Tiedge M, Ziv E, Shafrir E, Lenzen S: Gradual loss of pancreatic�-cell insulin, glucokinase and GLUT2 glucose transporter immunoreac-tivities during the time course of nutritionally induced type-2 diabetes inPsammomys obesus (sand rat). Virchows Arch 440:63–69, 2002

68. Jansson L, Eizirik DL, Pipeleers DG, Borg LA, Hellerstrom C, AnderssonA: Impairment of glucose-induced insulin secretion in human pancreaticislets transplanted to diabetic nude mice. J Clin Invest 96:721–726, 1995

69. Eizirik DL, Jansson L, Flodstrom M, Hellerstrom C, Andersson A:Mechanisms of defective glucose-induced insulin release in human pan-creatic islets transplanted to diabetic nude mice. J Clin Endocrinol Metab

82:2660–2663, 199770. Wagenknecht LE, Langefeld CD, Scherzinger AL, Norris JM, Haffner SM,

Saad MF, Bergman RN: Insulin sensitivity, insulin secretion, and abdom-inal fat: the Insulin Resistance Atherosclerosis Study (IRAS) FamilyStudy. Diabetes 52:2490–2496, 2003

71. Paolisso G, Tataranni PA, Foley JE, Bogardus C, Howard BV, Ravussin E:A high concentration of fasting plasma non-esterified fatty acids is a riskfactor for the development of NIDDM. Diabetologia 38:1213–1217, 1995

72. Zhou YP, Grill VE: Long-term exposure of rat pancreatic islets to fattyacids inhibits glucose-induced insulin secretion and biosynthesis througha glucose fatty acid cycle. J Clin Invest 93:870–876, 1994

73. Kashyap S, Belfort R, Gastaldelli A, Pratipanawatr T, Berria R, Pratipana-watr W, Bajaj M, Mandarino L, DeFronzo R, Cusi K: A sustained increasein plasma free fatty acids impairs insulin secretion in nondiabeticsubjects genetically predisposed to develop type 2 diabetes. Diabetes

52:2461–2474, 200374. Lee Y, Hirose H, Ohneda M, Johnson JH, McGarry JD, Unger RH: �-Cell

lipotoxicity in the pathogenesis of non-insulin-dependent diabetes melli-tus of obese rats: impairment in adipocyte-�-cell relationships. Proc Natl

Acad Sci U S A 91:10878–10882, 199475. Shimabukuro M, Ohneda M, Lee Y, Unger RH: Role of nitric oxide in

obesity-induced � cell disease. J Clin Invest 100:290–295, 199776. Cnop M, Hannaert JC, Hoorens A, Eizirik DL, Pipeleers DG: Inverse

relationship between cytotoxicity of free fatty acids in pancreatic isletcells and cellular triglyceride accumulation. Diabetes 50:1771–1777, 2001

77. Cnop M, Hannaert JC, Pipeleers DG: Troglitazone does not protect ratpancreatic � cells against free fatty acid-induced cytotoxicity. Biochem

Pharmacol 63:1281–1285, 200278. Briaud I, Kelpe CL, Johnson LM, Tran PO, Poitout V: Differential effects of

hyperlipidemia on insulin secretion in islets of Langerhans from hyper-glycemic versus normoglycemic rats. Diabetes 51:662–668, 2002

79. Maestre I, Jordan J, Calvo S, Reig JA, Cena V, Soria B, Prentki M, RocheE: Mitochondrial dysfunction is involved in apoptosis induced by serumwithdrawal and fatty acids in the �-cell line INS-1. Endocrinology

144:335–345, 200380. Shimabukuro M, Zhou YT, Lee Y, Unger RH: Troglitazone lowers islet fat

and restores � cell function of Zucker diabetic fatty rats. J Biol Chem

273:3547–3550, 199881. Rys-Sikora KE, Gill DL: Fatty acid-mediated calcium sequestration within

intracellular calcium pools. J Biol Chem 273:32627–32635, 199882. Ozcan U, Cao Q, Yilmaz E, Lee AH, Iwakoshi NN, Ozdelen E, Tuncman G,

Gorgun C, Glimcher LH, Hotamisligil GS: Endoplasmic reticulum stresslinks obesity, insulin action, and type 2 diabetes. Science 306:457–461, 2004

83. Nakatani Y, Kaneto H, Kawamori D, Yoshiuchi K, Hatazaki M, MatsuokaTA, Ozawa K, Ogawa S, Hori M, Yamasaki Y, Matsuhisa M: Involvement ofendoplasmic reticulum stress in insulin resistance and diabetes. J Biol

Chem 280:847–851, 200584. Eisenbarth GS: Type I Diabetes: Molecular, Cellular and Clinical Immu-

nology. New York, Oxford University, 200385. Lenzen S, Tiedge M, Elsner M, Lortz S, Weiss H, Jorns A, Kloppel G,

Wedekind D, Prokop CM, Hedrich HJ: The LEW. 1AR1/Ztm-iddm rat: anew model of spontaneous insulin-dependent diabetes mellitus. Diabeto-

logia 44:1189–1196, 200186. Jorns A, Kubat B, Tiedge M, Wedekind D, Hedrich HJ, Kloppel G, Lenzen

S: Pathology of the pancreas and other organs in the diabetic LEW.1AR1/Ztm-iddm rat, a new model of spontaneous insulin-dependentdiabetes mellitus. Virchows Arch 444:183–189, 2004

87. Mathis D, Vence L, Benoist C: �-Cell death during progression to diabetes.Nature 414:792–798, 2001

88. Kolb H, Worz-Pagenstert U, Kleemann R, Rothe H, Rowsell P, Scott FW:Cytokine gene expression in the BB rat pancreas: natural course andimpact of bacterial vaccines. Diabetologia 39:1448–1454, 1996

89. Rabinovitch A, Suarez-Pinzon W, El-Sheikh A, Sorensen O, Power RF:Cytokine gene expression in pancreatic islet-infiltrating leukocytes of BBrats: expression of Th1 cytokines correlates with �-cell destructiveinsulitis and IDDM. Diabetes 45:749–754, 1996

90. Reddy S, Young M, Ginn S: Immunoexpression of interleukin-1� inpancreatic islets of NOD mice during cyclophosphamide-accelerateddiabetes: co-localization in macrophages and endocrine cells and itsattenuation with oral nicotinamide. Histochem J 33:317–327, 2001

91. Cardozo AK, Proost P, Gysemans C, Chen MC, Mathieu C, Eizirik DL:IL-1� and IFN-� induce the expression of diverse chemokines and IL-15 inhuman and rat pancreatic islet cells, and in islets from pre-diabetic NODmice. Diabetologia 46:255–266, 2003

92. Shafrir E, Spielman S, Nachliel I, Khamaisi M, Bar-On H, Ziv E: Treatmentof diabetes with vanadium salts: general overview and amelioration ofnutritionally induced diabetes in the Psammomys obesus gerbil. Diabete

Metab Res Rev 17:55–66, 200193. Movassat J, Saulnier C, Portha B: �-Cell mass depletion precedes the

onset of hyperglycaemia in the GK rat, a genetic model of non-insulin-dependent diabetes mellitus. Diabete Metab 21:365–370, 1995

94. Bonner-Weir S: Life and death of the pancreatic � cells. Trends Endocri-

nol Metab 11:375–378, 200095. Jorns A, Tiedge M, Lenzen S: Thyroxine induces pancreatic � cell

apoptosis in rats. Diabetologia 45:851–855, 200296. Rakatzi I, Mueller H, Ritzeler O, Tennagels N, Eckel J: Adiponectin

counteracts cytokine- and fatty acid-induced apoptosis in the pancreatic�-cell line INS-1. Diabetologia 47:249–258, 2004

97. Buteau J, El-Assaad W, Rhodes CJ, Rosenberg L, Joly E, Prentki M:Glucagon-like peptide-1 prevents � cell glucolipotoxicity. Diabetologia

47:806–815, 200498. Carlsson C, Borg LA, Welsh N: Sodium palmitate induces partial mito-

chondrial uncoupling and reactive oxygen species in rat pancreatic isletsin vitro. Endocrinology 140:3422–3428, 1999

99. Welsh N, Cnop M, Kharroubi I, Bugliani M, Lupi R, Marchetti P, Eizirik DL:Is there a role for locally produced interkeukin-1 in the deleterious effectsof high glucose or the type 2 diabetes milieu to human pancreatic islets?Diabetes S4:3238–3244, 2005

100. Jamieson E, Thomas H, Kay TW: Glucose toxicity in �-cells is indepen-dent of IL-1 and Fas. Presented at the Immunology of Diabetes Society 7thConference, 28–31 March 2004, Cambridge, U.K. Program and Abstracts:135, 2004

101. Gunton JE, Kulkarni RN, Yim S, Okada T, Hawthorne WJ, Tseng YH,Roberson RS, Ricordi C, O’Connell PJ, Gonzalez FJ, Kahn CR: Loss ofARNT/HIF1� mediates altered gene expression and pancreatic-isletdysfunction in human type 2 diabetes. Cell 122:337–349, 2005



MOUNT SINAI JOURNAL OF MEDICINE 75:314–327, 2008 314

Immunology and Genetics ofType 1 Diabetes

Michael P. Morran,1 Gilbert S. Omenn2 and Massimo Pietropaolo1

1 Laboratory of Immunogenetics, Brehm Center for Type 1 Diabetes Research and Analysis, Division ofMetabolism, Endocrinology, and Diabetes, Department of Internal Medicine, University of Michigan Medical

School, Ann Arbor, MI2 Center for Computational Medicine and Biology, Departments of Internal Medicine and Human Genetics,

University of Michigan, Ann Arbor, MI

ABSTRACT

Type 1 diabetes is one of the most well-characterizedautoimmune diseases. Type 1 diabetes compro-mises an individual’s insulin production throughthe autoimmune destruction of pancreatic β-cells.Although much is understood about the mechanismsof this disease, multiple potential contributing factorsare thought to play distinct parts in triggering type1 diabetes. The immunological diagnosis of type 1diabetes relies primarily on the detection of autoan-tibodies against islet antigens in the serum of type 1diabetes mellitus patients. Genetic analyses of type1 diabetes have linked human leukocyte antigen,specifically class II alleles, to susceptibility to dis-ease onset. Environmental catalysts include variouspossible factors, such as viral infections, althoughthe evidence linking infections with type 1 diabetesremains inconclusive. Imbalances within the immunesystem’s system of checks and balances may pro-mote immune activation, while undermining immuneregulation. A lack of proper regulation and overac-tive pathogenic responses provide a framework forthe development of autoimmune abnormalities. Type1 diabetes is a predictable and potentially treat-able disease that still requires much research tofully understand and pinpoint the exact triggeringevents leading to autoimmune activation. In silicoresearch can aid the comprehension of the etiol-ogy of complex disease pathways, including Type Idiabetes, in order to and help predict the outcome

Address Correspondence to:

Massimo PietropaoloLaboratory of Immunogenetics

Brehm Center for Type 1Diabetes Research and Analysis

Ann Arbor, MIEmail: [email protected]

of therapeutic strategies aimed at preserving β-cellfunction. Mt Sinai J Med 75:314–327, 2008. 2008Mount Sinai School of Medicine

Key Words: adaptive immunity, autoimmunity,genetics in type 1 diabetes, innate immunity, type1 diabetes.

IMMUNE SYSTEM REVIEW

The immune system is the body’s natural defensesystem against invading pathogens. It protects thebody from infection and works to communicate anindividual’s well-being through a complex networkof interconnected cells and cytokines. This systemhas the power to initiate a wide gamut of cellularresponses with the ability to directly attack aninvading organism or signal cells to begin the healingprocess. Although this system is an associated hostdefense, an uncontrolled immune system has thepotential to trigger negative complications in the host.Therefore, well-controlled regulation of the immunesystem is necessary in order to prevent autoimmuneresponses from occurring.

In order to protect the body against for-eign pathogens, the immune system has developedthroughout evolution to recognize the differencebetween the self and nonself. The ability to becomeself-tolerant toward the body’s own proteins and anti-gens is critical to maintaining a properly functioningimmune system. An immune system that loses tol-erance to the self loses the ability to differentiatebetween friend and foe in immunological battles.Loss of tolerance leads the immune system towardautoimmune responses, in which the body attacksitself, causing substantial damage to the self, eveninflicting irreversible damage.

The immune system is composed of 2 uniquebranches, each with its own responsibilities. The

Published online in Wiley InterScience (www.interscience.wiley.com).DOI:10.1002/msj.20052

2008 Mount Sinai School of Medicine

MOUNT SINAI JOURNAL OF MEDICINE 315

T1D Development

Contributing Factors

T1D Protected

RegulationPathogenicity

Normal

Pathogenicity Regulation

T1D Prone

Pathogenicity Regulation

Fig 1. T1D development. A discrete balancing act is performed by theimmune system in order to attack foreign pathogens while protecting againstthe promotion of autoimmunity. Multiple contributing factors influence thedevelopment of T1D, including genetics, the environment, and cellular signals.In a normal individual the immune system is balanced, possessing the abilityto both promote and suppress cellular responses to pathogens. When animmune system is unbalanced and favors inflammation or excessive cellularresponses, the system is unbalanced and favors excesive regulation of cellularresponses, the system is protected against the development of T1D andautoimmunity. Abbreviations: T1D, type 1 diabetes. Adapted from AnnualReview of Immunology.102 [Color figure can be viewed in the online issue,which is available at www.interscience.wiley.com.].

innate immune system is the body’s first line ofdefense against invading pathogens. This system rec-ognizes common structural components of pathogensand elicits immune responses to signal the presenceof pathogens and infection.1 The adaptive immunesystem is the body’s secondary line of defense andspecifically targets identified pathogens. This sys-tem is antigen-specific and generates immunologicalmemory within the host, which allows for more effi-cient pathogen clearance upon repeat exposure tothe same pathogen.1 Although these 2 systems aretermed different branches of the immune system,they must work together as one unified system toprotect the body. The 2 branches of the immunesystem rely on each other to help properly performtheir jobs. If either branch fails to perform its job,the other branch suffers. Regulatory malfunctions ineither system can ultimately lead to the generation ofunwarranted or unregulated autoimmune responses.Underlying an immune response lies a delicate net-work of cell types and cytokines that communicate tobegin, generate, and end an immune response. If animmune system becomes unbalanced, this delicatenetwork, responsible for preventing autoimmunity,promotes it (Figure 1).

The immune system protects the host, butit also possesses the power to harm the hostas well. Numerous autoimmune diseases have

been characterized over time. This article providesan overview of how alterations in the immunesystem, due to genetics, cellular malfunctions, orcell signaling issues, lead to the development andpathogenesis of autoimmune diseases, specificallytype 1 diabetes (T1D). Understanding mechanisticallyhow T1D and autoimmune responses are triggeredis essential in order to develop strategies to combatthese diseases.

TYPE 1 DIABETES OVERVIEW

Diabetes mellitus is a group of diseases characterizedby the body’s inability to accurately maintain normalblood glucose levels, leading to multiple detrimentaleffects. Insulin is an important hormone in glucosemetabolism. When released, it signals cells to takeup glucose. If the body is unable to produceinsulin, blood glucose levels remain elevated, andthis is termed hyperglycemia. T1D is an autoimmunedisease in which the immune system targets anddestroys the insulin-producing β-cells found in theislets of Langerhans in the pancreas. Without insulin,individuals develop the clinical syndrome of T1D.

T1D is characterized by autoantibody produc-tion and progressive infiltration of immune cells into

DOI:10.1002/MSJ

316 M. P. MORRAN ET AL: IMMUNOLOGY AND GENETICS OF TYPE 1 DIABETES

the islets of the pancreas, followed by the destruc-tion of the islet cells.2 During the onset of T1D, cellsfrom both the innate and adaptive immune systemsinfiltrate islet lesions to produce insulitis. Studiesusing human and murine models of diabetes havedemonstrated that the autoimmune destructive pro-cess in T1D occurs in a cell-mediated organ-specificmanner and requires both CD+4 and CD+8 T cellsas well as macrophages.3–13 These cells accumulatein an islet lesion but are nondestructive.14–16 Anunknown triggering event occurs, which promotesthe autoimmune destruction of the β-cells.14–16 Clin-ical manifestations of diabetes occur after 90% of anindividual’s β-cell mass is destroyed.14

The exact trigger for the onset of T1D isstill unknown, as discussed later, although themechanism by which the insulin-producing β-cellsare destroyed is well understood. The destructiveprocess is T cell–mediated. Once the islets havebecome infiltrated and highly populated with Tcells and macrophages, a subsequent triggeringevent occurs to activate these cells.14–16 T cells, Bcells, and macrophages communicate via antigenpresentation and can, in turn, activate each othervia cytokines and direct cell communication throughsurface receptors.1 With such a high population ofimmune cells centered in one distinct area, activationsignals can travel fast, initiating a destructive cascadeeasily.

In general, macrophages play the role of antigen-presenting cells (APCs), and so they can directlypick up and present β-cell antigens. This leadsto the activation of β-cell–cytotoxic CD+8 T cellsand the generation of autoreactive CD+4 effectorT cells.15 Cytotoxic T cells can directly kill β-cells,whereas effector T cells can initiate the activation ofB cells, thus prompting autoantibody production.15

Once the immune system is triggered or activated,it begins to process and present self antigens toT cells, generating the autoimmune process thatleads progressively to the clinical manifestations ofT1D.

T cells possess the ability to directly destroy β-cells in a cytotoxic manner but also hold the power todirectly influence the induction of β-cell destructionthrough the release of cytotoxic molecules, includingcytokines, granzyme B, and perforin.16,17 T cells canalso signal β-cell death through the Fas pathway.16,17

β-Cell apoptosis is promoted by activation ofthe caspase pathway through multiple mechanisticpathways, including Fas and Fas ligand interactions,nitric oxide and oxygen-derived free radicals, andmembrane disruption due to perforin or granzymeB.18 Cytokine production by activated T cells may bea key factor in β-cell death, influential cytokines

released by T cells include: interleukin 1 (IL-1),interferon γ (IFNγ ), and tumor necrosis factorα (TNFα), all of which up-regulate Fas and theproduction of Fas ligand, nitric oxide, and free-radicals.18,19 Other notable cytokines are IL-2, IL-12,IL-17, and IL-18, all of which seem to promotean inflammatory state biased to T helper 1 typeresponses.18,19

AUTOANTIGENS IN TYPE 1 DIABETES

During the progression of T1D, multiple autoimmuneprocesses occur. Human and murine studies ofT1D have illustrated that, as the severity of diseaseincreases, so does the number of autoantigens andautoantibodies.20 Over the course of development,β-cells go through multiple divisions and changesin cell volume and increase in cell mass.21 Duringthis time, β-cells grow, develop, and eventually die;these dead cells may be engulfed and processed byAPCs. In an insulitis lesion, these APCs presentingprocessed β-cell components have the potentialto trigger the generation of β-cell autoreactive Tcells.14–16,21

Most well-known autoantigens are associatedwith β-cell components, including insulin, glutamicacid decarboxylase (GAD), and islet-cell antigen-2(IA-2).20 Insulin is the first antigenic target detectableduring the early progression of diabetes,22 althoughmost autoantibodies are targeted against the β-cells themselves and other β-cell–secreted proteins.20

Recently, ZnT-8, a pancreatic β-cell–specific zinctransporter, has been identified as a candidateautoantigen.23 Although research is limited on thismolecule, strong evidence already supports theassociation of ZnT-8 autoantibodies with T1D.23

During the progression of T1D, a process ofautoantigen epitope spreading occurs.24 Epitopespreading provides an explanation of how theimmune system is capable of recognizing increasingnumbers of autoantigens in correlation with increasedT1D disease severity.24 Epitope spreading begins withthe immune system recognizing and mounting animmune response against a single antigen, whichis recognized via a single epitope. Over time,new antigens can be recognized, and previouslyrecognized antigens can be differentially processedby APCs to generate multiple epitopes for a singleantigen.

One can think of this process as similar toa tree growing toward autoimmunity. In Figure 2,the tree stem symbolizes an immune system atbirth that lacks autoimmunity. As this tree grows

DOI:10.1002/MSJ


Autoimmune Disease Progression

Birth

GAD65

IA-2

Insulin Insulin

< 5 years

Auto-Antigen Epitope Spreading in T1D

Fig 2. Autoantigen epitope spreading in T1D. As the severity of symptoms associatedwith T1D increases over time, so does the number of autoantigens recognized by theimmune system. Epitope spreading begins once the immune system is triggered withinthe pancreas, leading to the processing and presentation of self antigens. As β-celldestruction takes place, multiple self antigens become targets of the immune system.During this process, insulin is the first antigenic target,22 and it is followed by otherβ-cell–associated components, such as GAD65 and IA−2.20 Over time, autoantigens areprocessed differently, creating various recognition epitopes for a given antigen. In Figure 1,the tree symbolizes an immune system at birth that lacks autoimmunity. As the tree growstoward autoimmune T1D, its limbs represent targeted self antigens that develop. As T1Dprogresses, multiple limbs grow off the tree, each from a different antigen. These growinglimbs next branch off, representing the unique epitopes recognized from differentialprocessing of similar self peptides. As T1D develops, the tree grows toward autoimmunityby increasing both the number of limbs and the number of branches on a given limb,representing the process of epitope spreading observed in T1D disease development.Abbreviations: GAD65, glutamic acid decarboxylase 65; IA-2, islet-cell antigen-2; T1D,type 1 diabetes.

toward autoimmune T1D, its limbs represent devel-oping targeted self antigens. Next, multiple limbsgrow from the stem, each targeting different anti-gens. These growing limbs branch off, representingthe unique epitopes recognized from differentialprocessing of similar peptides. It has been shownthat autoantibody production has been detectedup to 5 years prior to the development of hyper-glycemic events, and this indicates that autoanti-body production precedes the clinical manifesta-tion of T1D.24 As T1D develops, the tree growstoward autoimmunity by increasing both the num-ber of limbs (targeted antigens) and the numberof branches on a given limb (recognized epitopesfor a particular antigen). Although the progres-sion and mechanisms behind the development ofT1D are understood in general terms, many ques-tions remain to be answered concerning how theautoimmune disease state becomes triggered towardautoreactivity.

We have provided evidence suggesting that asubset of cytoplasmic islet-cell antibodies (ICAs)is related to a more rapid progression to insulin-requiring diabetes in glutamic acid decarboxylase65 (GAD65) and IA-2 antibody–positive relativesof proband patients versus relatives with GAD65and IA-2 antibodies without ICAs (Figure 3).25

The precise nature of the antigen(s) detectedby the indirect immunofluorescence ICA assay

remains enigmatic, but its utility to predict futuredevelopment of diabetes mellitus in individualswith other circulating ICAs has recently beenconfirmed.

WHAT ARE THE GENES ASSOCIATEDWITH TYPE 1 DIABETES?

Although it has been suggested that multiple genesplay a role in disease susceptibility, there is strongevidence for only 2 chromosomal regions thatare associated with T1D: the human leukocyteantigen (HLA) region on chromosome 6p21 [insulin-dependent diabetes mellitus 1 (IDDM1)] and theinsulin gene region on chromosome 11p15 [insulin-dependent diabetes mellitus 2 (IDDM2)]. Thecontributions of these 2 loci to familial inheritance areapproximately 42% for IDDM1 and 10% for IDDM2.As a result of genome-wide searches, many otherputative loci have been proposed to be related toT1D. These loci along with some potential candidategenes are listed in Tables 1 and 2. The fact that inhumans the highest risk-conferring locus linked tothe disease is the HLA cluster and, in particular,HLA genes encoding specific class II alleles stronglyindicates an important role of the immune cellsin both the development and activation of theautoimmune response leading to disease onset.26

DOI:10.1002/MSJ


1.0

0.8

0.6

0.4

0.2

0.0

Cum

ulat

ive

Pro

port

ion

of T

1DM

-fre

e

1.0

0.8

0.6

0.4

0.2

0.0

Cum

ulat

ive

Pro

port

ion

of T

1DM

-fre

e

1.0

0.8

0.6

0.4

0.2

0.0

Cum

ulat

ive

Pro

port

ion

of T

1DM

-fre

e

20100 155

Follow-up (years)ICA Negative: 124 120 111 89 25ICA Positive: 26 21 14 10 4

20100 155

Follow-up (years)ICA Negative: 26 26 26 23 8ICA Positive: 12 6 3 2

20100 155

Follow-up (years)ICA Negative: 96 92 85 70 23ICA Positive: 18 17 13 10 4

20100 155

Follow-up (years)ICA Negative: 14 14 14 11 2ICA Positive: 10 5 2 1

1.0

0.8

0.6

0.4

0.2

0.0C

umul

ativ

e P

ropo

rtio

nof

T1D

M-f

ree

(A) (B)

(C) (D)

Fig 3. The rate of progression to T1DM development in relatives carrying GAD65 AA, IA-2 AA, ora combination of both AA in the (–) absence or (- - -) presence of ICA. (A) Progression to insulin-requiring diabetes for relatives with GAD65 AA with respect to ICA positivity. (B) Progression toinsulin-requiring diabetes for relatives with IA-2 AA with respect to ICA positivity. (C) Progressionto insulin-requiring diabetes for relatives with either GAD65 or IA-2 AA with respect to ICApositivity (log rank: P = 0.01). (D) Remarkably, the cumulative risk of developing an insulinrequirement was 80% at 6.7 years and 90% at 12.9 years of follow-up in ICA-positive relatives;this is significantly higher than the cumulative risk of diabetes development in relatives whowere positive for GAD65 and IA-2 AA without ICA (log rank: P < 0.00001). Abbreviations: AA,auto-antibodies GAD65, glutamic acid decarboxylase 65; IA-2, islet-cell antigen-2; ICA, islet-cellantibody; T1DM, type 1 diabetes mellitus. Reprinted with permission from Pediatric Diabetes.25

Copyright 2005, International Society for Pediatric and Adolescent Diabetes.

Interestingly, the same HLA locus seems to have thecorresponding susceptibility influence in the primarymouse model of T1D, the nonobese diabetic (NOD)mouse (Table 3).27 The immune-mediated processesof β-cell destruction are mainly T cell–dependentand chronic in both mouse and rat models of T1D,and this makes it more likely to be the same inhumans.27 Comparative mapping of human (IDDM)and NOD mouse insulin-dependent diabetes (Idd)genes are shown in Table 3.

Genetic factors have long been thought tobe linked to the development of T1D. Althoughit has been hypothesized that in monozygotictwins a discordance rate greater than 50% couldbe explained by environmental factors of diseasedevelopment, studies using both monozygotic anddizygotic twins have suggested that environmental

factors have few causative roles in the developmentof islet autoimmunity, whereas genetic similarities indizygotic twins seem more important in determiningsusceptibility to diabetes.28 T1D is a polygenicdisease29 in which there presumably exist a smallnumber of genes with large effects, HLA being themain example, and a large number of genes withsmall effects overall.30

THE HUMAN LEUKOCYTE ANTIGENCOMPLEX

The short arm of human chromosome 6 (6p21)accommodates a ∼ 3.5−megabase genetic segmentcontaining a group of immune response genes

DOI:10.1002/MSJ


Table 1. Effect of Human Leukocyte Antigen Alleles on Type 1 DiabetesSusceptibility.

DQ Alleles Effect Associated DR

B1∗ 0302, A1∗ 0301 Susceptible DR4B1∗ 0201, A1∗ 0501 Susceptible DR3B1∗ 0501, A1∗ 0101 Susceptible DR1B1∗ 0201, A1∗ 0301 Susceptible (African Americans) DR7B1∗ 0502, A1∗ 0102 Susceptible (Sardinia) DR2 (DR16)B1∗ 0303, A1∗ 0301 Susceptible (Japanese) DR4B1∗ 0303, A1∗ 0301 Susceptible (Japanese) DR9B1∗ 0602, A1∗ 0102 Protective DR2 (DR15)B1∗ 0301, A1∗ 0501 Protective DR5B1∗ 0201, A1∗ 0201 Neutral DR7B1∗ 0303, A1∗ 0301 Neutral DR4B1∗ 0301, A1∗ 0301 Neutral DR4

Table 2. Genetic Risk Estimates for Human Leukocyte AntigenClass II in T1D.

High-Risk Genotype

Risk in anIndividualwith ThisGenotype

DQB1∗ 0302 (DQ3.2) 1 in 60DQ3.2/DQ2 (DR3) 1 in 25DQB1∗ 0302 and a family history of IDDM 1 in 10DQ3.2/DQ2 (DR3) and a family history of T1D 1 in 4

NOTE: This table was adapted from Annual Review ofMedicine.26

Abbreviations: IDDM, insulin-dependent diabetes mellitus;T1D, type 1 diabetes.

termed the major histocompatibility complex (MHC).The principal genes located within the MHC codefor HLAs, 2 molecular classes of cell surfaceglycoproteins differing in structure, function, andtissue distribution.

The class I HLA molecule exists as a heterodimer,consisting of a polymorphic 44-kDa MHC-encodedα or heavy chain in noncovalent association withβ2-microglobulin, a 12-kDa protein encoded by anonpolymorphic gene on chromosome 15. The classI molecule is anchored in the cell membrane onlyby the heavy chain. This chain contains 338 aminoacids and, beginning from the amino terminus, isfunctionally divided into 3 regions: an extracellularhydrophilic region, a transmembrane hydrophobicregion, and an intracytoplasmic hydrophilic region.The extracellular region is further subdivided into 3domains, designated α1, α2, and α3, each of whichhas approximately 90 amino acid residues. The α1

and α2 domains compose the peptide- or antigen-binding region of the molecule.

Class II HLA molecules consist of 2 glycoproteinchains, an α chain of approximately 34 kDa anda βchain of approximately 29 kDa, both encoded

within the MHC. As with the class I heavy chain,each class II chain can be divided into 3 regions(extracellular, transmembrane, and intracytoplasmic),but in contrast, both class II chains span thecell membrane. Each extracellular region of theclass II α and βchains has been further dividedinto 2 domains of approximately 90 amino acidresidues each, termed α1 and α2 and β1 and β2,respectively. The α1 and β1 domains form thepeptide-binding region of class II HLA molecules.Also of note, the class II α2 and β2 domains,class I α3 domain, and β2-microglobulin all showhomology to the constant region of immunoglobulinsand are therefore classified as members of theimmunoglobulin superfamily.

The genes that encode class I MHC are locatedat the HLA-A, HLA-B, and HLA-C loci, whereasclass II molecules are encoded by the DR, DQ,and DP genes. Other genes in this cluster includetransporters associated with antigen processing31,32

and low-molecular-weight proteins, both of whichare involved in antigen processing.33 A third region ofthe MHC, termed class III, encodes several moleculeshaving a variety of functions, such as complement

DOI:10.1002/MSJ


Table 3. Comparative Mapping of Human (IDDM)and Murine (Idd) Genes.

Human Locus (Chromosome)Marker/

Candidate Potential NOD ‘‘Idd’’ Homolog (Chromosome) Marker

IDDM1 (6p21) HLA ‘‘Idd1’’ (17) H2g7IDDM2 (11p15.5) INS/VNTR ? NOD.DR-2 (chromosome 7 region from C57L) ? Ins2IDDM3 (15q26) D15S107 ‘‘Idd2’’ (9) Cyp19IDDM4 (11q13) FGF3 None yet identified (distal 7)IDDM5 (6q25) ESR None yet identified (proximal 10)IDDM6 (18q) D18S64 ? ‘‘Idd5’’ (1) Bcl2IDDM7 (2q31-33) D2S326 ? ‘‘Idd5’’ (1) Il1r/Stat1IDDM8 (6q25-27) D6S264 None yet identified (proximal 10 or 17)IDDM9 (3q21-q25) D3S1303 None yet identified (middle 6)IDDM10 (10p11.2-q11.2) GAD2 None yet identified (proximal 2)IDDM11 (14q24.3-q31) D14S67 None yet identified (middle 12)IDDM12 (2q31-33) ? ‘‘Idd5.1’’ (1) Ctla4IDDM13 (2q34) IGFBP-2,5 ? ‘‘Idd5.2’’(1) Slc11a1 (Nramp)GCK (7p) GCK None yet identified (proximal 11)IDDM15 (6q21) D6S283 ? ‘‘Idd14’’ (13) D13Mit61IDDM16 ? (1p36.1-p35) NHE1 ? ‘‘Idd11’’ (4) Slc9a1 (Nhe1)

NOTE: This table was adapted from Molecular Pathology of Insulin Dependent Diabetes Mellitus.27 Idd nomenclatureis placed in quotation marks because there are multiple Idd loci on mouse chromosome 1.Abbreviations: GCK, glucokinase; HLA, human leukocyte antigen; Idd, insulin-dependent diabetes; IDDM,insulin-dependent diabetes mellitus; NOD, nonobese diabetic.

components (C4A, C4B, factor B, and C2), tumornecrosis factor α and tumor necrosis factor β, andthe 21-hydroxylase genes (CYP21P and CYP21).

Polymerase chain reaction studies34 have pro-vided researchers with a rapid means of estimatingT1D susceptibility in comparison with serologicaltechniques. Polymerase chain reaction amplificationof individuals’ HLA alleles in a variety of racial andethnic groups has revealed that the presence of aspecific human DQβchain variant encoding a neutralamino acid (alanine, valine, or serine), rather thanaspartic acid at position 57 (non-Asp-57), is stronglyassociated with T1D. In contrast, negatively chargedaspartic acid at position 57 of the DQβchain (Asp-57) appears to confer resistance to T1D progression.This association is much stronger than the associationbetween HLA-DR3 and HLA-DR4 and the presenceof the disease (Tables 1 and 2).35–37

Susceptibility to T1D is mainly conferred byspecific polymorphic regions within the MHC com-plex, such as HLA DR/DQ alleles.35,36 The genotypeassociated with the highest risk for T1D is the DR3/4-DQ8 (DQ8 is DQA1*0301, DQB1*0302) heterozygousgenotype. The HLA genotype DQB1*0602 confersdominant protection against T1D.

There is evidence that risk for islet autoimmunityincreases in DR3/4-DQ8 siblings who share both HLAhaplotypes with their diabetic proband sibling (63%by age 7 and 85% by age 15) in comparison withsiblings who do not share both HLA haplotypeswith their diabetic proband sibling.38 These findingsindicate that HLA genotyping at birth may identify

individuals at high risk of developing the diseasewith no detectable signs of islet autoimmunity, whocould then be enrolled in intervention trials aimed atpreventing overt disease.

The mechanisms by which the class II genes caninfluence susceptibility to or protection from T1Dare still the subject of discussion. Brown et al.39,40

characterized the structure of the crystallized HLAclass II molecule. One hypothesis is that effectiveantigen binding depends on the conformation ofthe antigen-binding site on the DQ dimer. The 2critical residues, DQα Arg-52 and DQβAsp-57, arelocated at opposite ends of the α-helices that formthe antigen-binding site of the DQ molecule. It hasbeen postulated that a substitution of an amino acidresidue at these positions of the DQ molecule leadsto conformational changes of the antigen-bindingsite and consequently to a modification of theaffinity of the class II molecule for the diabetogenicpeptide(s).37 As support for this hypothesis, it isknown that Asp-57 is involved in hydrogen and saltbonding with both the peptide main chain and theDRα Arg-76 side chain. Theoretically, modificationsin the DRα Arg-76 residue would also alter theantigen-binding site. It is noteworthy that studiesof the regulatory regions of the genes encodingDQ α and DQ βchains have shown that the levelof transcription of these genes may influence theamount of antigen binding. An increased level ofproduction of a class II chain may increase theavailability41 for dimerization.42 Studies by Demotzet al.43 have suggested that relatively few class II

DOI:10.1002/MSJ


heterodimers need to be present on the surface of anAPC to efficiently crosslink the T cell receptor (TCR)and initiate a T cell response.

ENVIRONMENTAL FACTORSINFLUENCING TYPE 1 DIABETES

Although T1D has long been considered to bedirectly linked to genetic factors, researchers havebeen searching for environmental factors capableof triggering T1D progression. Epidemiologists haveexamined trends for environmental triggers throughmajor geographical variations in T1D incidence, tem-poral trends in the incidence of T1D, and migrationstudies.44 The responsibility of environmental factorsfor T1D is strongly supported by statistical evidenceshowing that the incidence of newly diagnosed T1Dsubjects with high-risk HLA genotypes has decreasedover the last decades, whereas newly diagnosedT1D with low-risk or even protective HLA genotypeshas increased.45,46 Specific environmental factors thathave been investigated include dietary compounds,including cow’s milk, wheat gluten, soy products,fats, and even coffee or tea, along with vitamin defi-ciencies, N -nitroso compounds and other toxins, andviral infections.44 One specific hypothesis associatesthe development of T1D in Finland with seasonalchanges in climate, it being more common duringthe cold season.47,48 Although the genetic argumenthas long been considered the main factor in trigger-ing T1D, new studies are pointing to the importanceof the environment.

VIRAL TRIGGERS OF TYPE 1 DIABETES

Although multiple environmental factors have beenimplicated as possible triggers of T1D, viral diseaseinduction has long been argued as one of themain potential environmental triggers. Viruses ingeneral have been receiving much attention aspossible triggers in T1D because of their ability tomechanistically generate an active immune responsewhen encountered in the host. In the case of T1D,the host environment is primed with immune cellspoised to activate; all they need is an immunologicalstimulus. Researchers in the aforementioned seasonalpattern study on T1D induction even argue that theprimary culprit responsible for triggering inductionof disease during the winter season may actuallybe viral infections.47,48 Several viruses have beendirectly implicated as potential triggers of T1D,including enterovirus, adenovirus, Coxsackie B virus,

cytomegalovirus, hepatitis C virus, mumps virus,rotavirus, and rubella virus.49–56 The presence ofa viral infection can lead to immune cell activationthrough various possible mechanisms. Viruses maydirectly alter a host cell or tissue in such amanner that the immune system identifies it asan immunological target. Targeted host cells arelysed, releasing self peptides and fragments of thehost cell into the circulation, where they may beprocessed and presented via APCs.49,54 Viruses candirectly alter the immune system of the host byinducing polyclonal B cell activation, the release oflymphokines, the activation of immune cells, andthe disruption of the strict immune balance betweenT helper 1 and T helper 2 type responses, thuspromoting unwarranted immune activation.54 It hasbeen hypothesized that antiviral antibodies also canlead to the formation of anti-idiotypic antibodies,which can become autoreactive if the first antibodyis generated against the part of the virus that interactswith the host.54

Molecular mimicry is by far one of the mostwell studied processes associated with viral triggeringmechanisms and T1D disease induction. Virusesproduce proteins similar to those of the host.Although not all viral proteins share homology withthe host, certain viral components share a distincthomology with identified β-cell antigens targetedin an autoimmune response.44,54 Upon reacting toa viral infection, the immune system may processand present a homologous viral protein in such amanner that the epitope targeted by the immunesystem can interact with both self and viral proteins.This process becomes especially important withrespect to T1D disease induction if the homologousviral protein that is recognized and processedshares a distinct homology with β-cell–associatedproteins. Unintentional activation of an immuneresponse geared toward β-cell–associated proteinsput together with epitope spreading of antigens is avolatile combination for the host.

GAD is a well-defined example of an autoantigenin T1D that shares homology with viral proteins.Researchers have demonstrated that GAD peptidesshare mimicry with the P2-C viral sequence ofthe Coxsackie B virus and the major outer capsidprotein of rotavirus.55,56 Immunization studies of miceusing homologous Coxsackie viral sequences haveinduced T cell immune responses that cross-reactwith GAD peptides, showing that viral homologymay induce responses to self proteins.57 Furthermore,researchers have linked rotavirus infection andpancreatic islet autoimmunity in children on theHLA-DR4 background, linking both viral inductionand genetic susceptibility.56 Researching molecular

DOI:10.1002/MSJ


mimicry and viral induction of T1D provides a pivotalbridge between an environmental factor and animmunological system responsible for T1D. WhetherT1D is triggered environmentally, genetically, or incombination, it is still not known how exactly toprevent disease progression once it starts. Severalother unknown immunological factors may play arole in triggering T1D.

INNATE IMMUNE SYSTEM

The innate immune system is the body’s first line ofdefense against invading microbes and pathogens.Although this system does not generate long-termantigen-specific immunological memory, this systemdoes have a high degree of immunological specificitywhen encountering potential pathogens.8 The innateimmune system is composed of macrophages,dendritic cells, natural killer cells, neutrophils, andepithelial cells, all of which have their own uniquerole in the framework of an innate response. Therole of the innate immune system is to survey anddetect pathogens via host cell receptors that are ableto recognize common structural elements expressedby pathogens. These pattern recognition receptors(PRRs), though specific, are not clonal and lack theability for clonal expansion of cells, as seen in theadaptive immune system.58–60

Cells of the innate immune system use severaltypes of PRRs to help initiate an immune response.Toll-like receptors (TLRs) are an evolutionarilyconserved class of PRRs, which when activated causethe activation of the immune system.58,60,61 Thesereceptors recognize specific microbial membranecomponents, bacterial flagellin, and DNA and RNAfrom bacteria and viruses.63 Upon recognition ofa ligand, TLR activation brings about a cascadeof proinflammatory cellular responses, includingup-regulated production of cytokines, chemokines,and costimulatory molecules.58,62–65 TLRs recognizenumerous self-expressed mammalian molecules inaddition to the known nonself molecules.66 These selfantigens are usually indicators of stress and diseaseor are molecules that are modified because of thedisease.66,67

Once the innate immune system becomesactivated, it in turn has the ability to promoteactivation of the adaptive immune system.58,62–65

The activation of the innate immune systemis a prerequisite for the initiation of specificadaptive immune responses, such as T-helper 1type responses.59,67,68 It has been suggested that inT1D and other autoimmune diseases, TLRs may be

priming an unwarranted adaptive immune responsebecause of autoreactive processes directed againstself antigens.69,70 Innate immune cells activatedagainst self antigens may be responsible for a breakin tolerance. Once tolerance is broken, the body maypromote autoreactive immune responses.

TLRs are not the only PRRs used by cells ofthe innate immune system. Non-TLR PRRs includeNod-like Receptors (NLRs), Triggering ReceptorsExpressed on Myeloid Cells (TREMs), and C-typeLectin Receptors (CLRs) have not been directlylinked to autoimmunity or T1D, but represent anew class of receptors that may be responsiblefor triggering autoimmune responses or autoimmuneinitiation of the adaptive immune system.74–76

Any immunological trigger directed against self ispotentially dangerous. Innate immune cells triggeredvia self ligand binding to TLRs or non-TLR PRRs couldin turn activate T cells and an adaptive immuneresponse because of the production of cytokinesand other inflammatory signals. Once the balance ofthe immune system is flipped toward autoreactivity,negative outcomes are sure to follow.

The innate immune system has been well studiedin the context of T1D development in both humansand the NOD mouse. T1D disease progression hasbeen characterized in the 2 phase processes ofasymptomatic inflammation of the islets followed byautoimmune destruction of the islets. During phase1, macrophages infiltrate the pancreatic islets andmay be responsible for the asymptomatic inflam-mation that takes place there.8–13 Once activated,macrophages can secrete nitrogen and oxygen freeradicals, as well as inflammatory cytokines, into theirsurrounding microenvironment.74 These moleculescan either directly damage surrounding cells or ini-tiate cellular damage by in turn activating other celltypes to cause damage. It has even been illustratedthat without the presence of macrophages in NODmice, differentiation of β-cell–cytotoxic T cells doesnot take place.75–77 Restoring macrophages in thesetypes of depletion experiments restores the ability ofβ-cell–cytotoxic T cell generation,75–77 thus illustrat-ing the importance of macrophages in T1D diseaseprogression. The innate and adaptive immune sys-tems communicate using an intricate system of checksand balances. Once one end of the system becomesdisrupted, the whole system can falter.

ADAPTIVE IMMUNE SYSTEM

The adaptive immune system is the body’s secondline of defense against pathogens and disease,

DOI:10.1002/MSJ


activated by the innate immune system. This systemis antigen-specific, facilitating the generation ofimmunological memory, which serves to eliminatereoccurring pathogens more effectively upon repeatexposure to a given pathogen. This highly specificsystem uses receptor interaction between T cellsand APCs to determine self from nonself. Thehallmarks of an adaptive immune response arethe generation of long-term immunological memory,peptide presentation in the context of MHCs, and theproduction of antigen-specific antibodies.78

T lymphocytes and their specific TCRs are acrucial part of the adaptive immune system. Tlymphocytes are generated in the thymus, eachhaving its own structurally diverse and uniqueantigen-recognizing receptor known as the TCR.The TCR recognizes processed antigenic peptidespresented in the context of MHC via APCs. Oncerecognition of a peptide-bound MHC complex occursvia the TCR, a cascade of signaling events takesplace, leading to either a CD4+ T cell or CD8+ Tcell response, depending on which class of MHC,either MHC class I or MHC class II, was initiallyrecognized by the TCR. Ultimately, once a TCRrecognizes a processed peptide in the context ofan MHC, immunological events take place that bringabout either the direct targeted destruction of cellspresenting the peptide or the generation of anantibody response and further T cell activation.

In phase 2 of T1D disease progression, bothT and B lymphocytes accumulate in an islet lesion,where they can be activated against self antigensto trigger an immunological response toward theself. This autoreactive immune response then leadsto the destruction of the insulin-producing β-cellwithin the pancreas. Both MHC class I and MHCclass II restricted T cells are necessary for T1D diseaseprogression in both humans and NOD mice.79,80 Anti-insulin CD+4 T cell populations have been reportedin NOD mice. It is thought that interactions betweeninsulin peptides and MHC molecules can cause thedevelopment of an autoimmune T cell repertoire.81

Although this process is not totally understood, itis necessary to determine what exact factors arecausing the immune system to become unregulatedin such a manner as to promote an autoimmuneresponse.

REGULATION VIA T CELLS

Keeping the immune system tolerant is a strictbalancing act: excessive regulation restricts properimmune function, whereas a lack of regulation

permits overt immune function with the potential totarget the self. Regulating when an immune responsebegins, continues, and ends is a very important part ofthe immune system’s checks and balances. RegulatoryT cells (T regs) and natural killer T (NKT) cells are2 subtypes of T cells that are important regulators inthe progression of T1D and autoimmune diseases.

T regs are a unique population of T cells thatexpress the forkhead transcription factor forkheadbox P3 (FOXP3) and are CD+4 CD+25.91–93

T regs are well-known immunoregulators thatcan suppress proliferation of effector cells byshutting down IL-2 activation pathways.85 T regshelp the immune system maintain proper T cellhomeostasis by preventing T cell activation, and thisin turn leads to the development of inflammatoryresponses.86,87 T regs work to shut off an active Tcell response and are thought to prevent autoimmunedevelopment by regulating the expansion of T cellpopulations, T cell differentiation, and effector Tcell function.88 Although individuals with T1D havenumbers of T regs equal to those of normal healthyindividuals,89–91 T regs from T1D individuals displaydecreased suppressive characteristics, which suggesta defect in T reg function in people suffering fromT1D.89,90 Adoptive transfer experiments using T regsin NOD mice have effectively been demonstrated toprotect against the onset and progression of T1D.92

Although current use of T regs in NOD mice showspromising therapeutic qualities, the potential use ofT regs in humans is not sufficiently understood.

NKT cells share similar characteristics with bothnatural killer cells and T cells. These cells can causedirect cell lysis due to Fas-ligand interactions andinduce cytotoxic damage of cells due to IFNg, likeNatural Killer cells102–104, but they are also similarto T cells in that they are biased toward TCR usageand produce IL-4.102–104 When activated, NKT cellsproduce IL-4, which inhibits inflammatory T helper1 type responses and promotes inhibitory T helper2 type responses.96,97 Individuals suffering from T1Dshow a decreased number of NKT cells along witha decreased ability to produce and secrete IL−4.98

As seen with T regs, adoptive transfer experimentswith NKT cells effectively protect against the onsetand progression of T1D in NOD mice.99 Because ofthese 2 types of T cells, individuals that cannot inhibitor properly regulate T cell responses face increasedautoimmune problems.

CONCLUDING REMARKS

T1D results from autoimmune destruction/dysfunc-tion of insulin-secreting cells. Under physiologic

DOI:10.1002/MSJ


conditions, there is balance between pathogenicT cells that mediate disease such as T cells withmarked conservation of their TCRs (eg, insulin) andregulatory cells that control autoimmunity. In T1Dand other autoimmune disorders, there is an alteredbalance between pathogenic and T regs.

Autoantibodies are some of the most potent riskdeterminants for autoimmune diseases, with relativerisk exceeding 100 when all 4 autoantibodies arepresent in an asymptomatic child. The archetypicalmodel for the application of autoantibodies isT1D. Seminal studies have suggested that usinga combination of humoral immunological markersgives a higher predictive value for T1D progressionand greater sensitivity without a significant loss ofspecificity. There is a growing effort as well as alarge opportunity for exploring novel strategies aloneor in combination with immunomodulation with theultimate goal of finding the cure for T1D.

Emerging evidence indicates that in silicoresearch is complementary to current experimentalapproaches in T1D research and has the potentialnot only to assist researchers in designing labo-ratory experimentation but also to build a frame-work of data collection to predict the outcomeof therapeutic strategies aimed at halting the β-cell–specific autoimmune process. With support fromthe Juvenile Diabetes Research Foundation, the Insti-tute for Systems Biology has generated a T1Ddatabase (T1Dbase) with numerous data sets ofmolecular signatures of T1D (http://t1dbase.org/cgi-bin/dispatcher.cgi/welcome/display). Molecular sig-natures have revealed remarkable differentiationof organ-specific complications as reflected in thework of the Kretzler laboratory on transcriptionfactor binding site patterns in nephropathy asso-ciated with T2DM versus nephropathy due tolupus erythematosus, glomerulonephritis, or othercauses.100,101 Similar studies on T1D-associatedneuropathy, are being completed by our group atthe National Center for Integrative Biomedical Infor-matics (https://portal.ncibi.org/portal). Mathematicalmodeling of the imbalance of T cell subsets in T1D,in work led by Patrick Nelson, is expected to informour future work.

ACKNOWLEDGMENT

This work was supported by the National Institutes ofHealth (grants RO1 DK53456, DK56200, and NIDDKPA-04–081 to M.P.) and by the University of MichiganCenter for Computational Medicine and BiologyPilot Research Program. The authors gratefully

thank the Brehm Coalition for its support. Anyresearch involving human subjects was conducted inaccordance with the guidelines of the Declaration ofHelsinki and was approved by the institutional reviewboard; all subjects provided informed consent.

DISCLOSURES

Potential conflict of interest: Nothing to report.

REFERENCES

1. Janeway CA, Travis P, Walport M, Shlomchik M.Immunobiology. 6th ed. New York, NY: GarlandScience; 2007.

2. Bardsley JK, Want LL. Overview of diabetes. Crit CareNurse Q 2004; 27: 106–112.

3. Pietropaolo M, Barinas-Mitchell E, Kuller LH.Perspectives in diabetes heterogeneity of diabetesmellitus. Unraveling a dispute: is systemicinflammation related to islet autoimmunity? Diabetes2007; 56: 1189–1197.

4. Delovitch TL, Singh B. The nonobese diabeticmouse as a model of autoimmune diabetes: immunedysregulation gets the NOD. Immunity 1997; 7:727–738.

5. Katz JD, Benoist C, Mathis D. T helper cell subsetsin insulin-dependent diabetes. Science 1995; 268:1185–1188.

6. Rabinovitch A, Suarez-Pinzon WL, Sorensen O.Interleukin 12 mRNA expression in islets correlateswith beta-cell destruction in NOD mice. J Autoimmun1996; 9: 645–651.

7. Von Herrath MG, Oldstone MB. Interferon-gammais essential for destruction of beta-cells anddevelopment of insulin-dependent diabetes mellitus.J Exp Med 1997; 185: 531–539.

8. Wong FS, Janeway CA. The role of CD4 and CD8T cells in type 1 diabetes in the NOD mouse. ResImmunol 1997; 148: 327–332.

9. Yoon JW, Jun HS, Santamaria P. Cellular andmolecular mechanisms for the initiation andprogression of beta-cell destruction resulting fromthe collaboration between macrophages and T cells.Autoimmunity 1998; 127: 109–112.

10. Lee KU, Kim MK, Amano K, et al. Preferentialinfiltration of macrophages during early stages ofinsulitis in diabetes-prone BB rats. Diabetes 1989; 37:1053–1058.

11. Voorbij HA, Jeucken PH, Kabel PJ, et al. Dendriticcells and scavenger macrophages in pancreatic isletsof pre-diabetic BB rats. Diabetes 1989; 38: 1623–1629.

12. Jansen A, Homo-Delarche F, Hooijkaas H, et al.Immunohistochemical characterization of monocyte-macrophage and dendritic cells in the initiation ofinsulitis and beta cell destruction in NOD mice.Diabetes 1994; 43: 667–675.

13. Lee KU, Amano K, Yoon JW. Evidence for initialinvolvement of macrophages in development ofinsulitis in NOD mice. Diabetes 1988; 37: 989–991.

DOI:10.1002/MSJ


14. Kawasaki E, Gill RG, Eisenbarth GS. Type 1 DiabetesMellitus. Austin, TX: Landes Bioscience; 1999.

15. Yoon JW, Jun HS. Cellular and molecular pathogenicmechanisms of insulin-dependent diabetes mellitus.Ann N Y Acad Sci 2001; 928: 200–211.

16. Pearl-Yafe M, Kaminitz A, Yolcu ES, et al. Pancreaticislets under attack: cellular and molecular effectors.Curr Pharm Des 2007; 13: 749–760.

17. Estella E, McKenzie MD, Catterall, et al. Granzyme B-mediated death of pancreatic beta-cells requires theproapoptotic BH-3only molecule bid. Diabetes 2006;55: 2212–2219.

18. Kawasaki E, Abiru N, Eguchi K. Prevention of type1 diabetes: from the view point of beta cell damage.Diabetes Res Clin Pract 2004; 55(suppl 1): S27–S32.

19. Ichinose K, Kawasaki E, Eguchi K. Recentadvancement of understanding pathogenesis of type1 diabetes and potential relevance to diabeticnephropathy. Am J Nephrol 2007; 27: 554–564.

20. Atkinson MA, Eisenbarth GS. Type I diabetes: newperspectives on disease pathogenesis and treatment.Lancet 2001; 358: 221–229.

21. Finegood DT, Scaglia L, Bonner-Weir S. Dynamics ofβ-cell mass in the growing rat pancreas. Estimationwith a simple mathematical model. Diabetes 1995; 44:249–256.

22. Nakayama M, Abiru N, Moriyama H, et al. Prime rolefor an insulin epitope in the development of type 1diabetes in NOD mice. Nature 2005; 435: 220–223.

23. Hutton JC, Wenzlau JM, Juhl K, et al. The cation effluxtransporter ZnT8 (Slc30A8) is a major autoantigenin human type 1 diabetes. PNAS 2007; 104:17 040–17 045.

24. Herrath MV, Sanda S, Herold K. Type 1 diabetes asa relapsing-remitting disease? Nat Rev Immunol 2007;7: 988–994.

25. Pietropaolo M, Yu S, Libman IM, et al. Cytoplasmicislet cell antibodies remain valuable in defining therisk of progression to type 1 diabetes in subjectswith other autoantibodies. Pediatr Diabetes 2005; 6:184–192.

26. Nepom, GT. Class II antigens and diseasesusceptibility. Annu Rev Med 1995; 46: 17–25.

27. Serreze DV, Leiter EH. Molecular pathology of insulindependent diabetes mellitutus. Current Directions inAutoimmunity. 2001; 4: 31–67.

28. Redondo MJ, Fain PR, Krischer P, et al. Expressionof beta-cell autoimmunity does not differ betweenpotential dizygotic twins of siblings of patients withtype 1 diabetes. J Autoimmun 2004; 23: 275–279.

29. Thomson G. HLA DR antigens and susceptibility toinsulin-dependent diabetes mellitus. Am J Hum Genet1984; 36: 1309–1317.

30. Todd JA, Walker NM, Cooper JD, et al. Robustassociations of four new chromosome regions fromgenome-wide analyses of type 1 diabetes. Nat Genet2007; 39: 857–864.

31. Powis SH, Mockridge I, Kelly A, et al. Polymorphismin a second ABC transporter gene located within theclass II region of the human major histocompatibilitycomplex. Proc Natl Acad Sci U S A 1992; 89:1463–1467.

32. Colonna M, Bresnahan M, Bahram S, et al. Alleleicvariants of the human putative peptide transporterinvolved in antigen processing. Proc Natl Acad Sci US A 1992; 89: 3932–3936.

33. Germain RN. MHC-dependent antigen processingand peptide presentation: providing ligands for Tlymphocyte activation. Cell 1994; 76: 287–299.

34. Rudert WA, Trucco M. Rapid detection of sequencevariation using polymers of specific oligonucleotides.Nucleic Acid Res 1992; 5: 1146.

35. Todd JA, Bell JI, McDevitt HO. HLA-DQβgenecontributes to susceptibility and resistance to insulin-dependent diabetes mellitus. Nature 1987; 329:599–604.

36. Morel PA, Dorman JS, Todd JA, et al. Aspartic acid atposition 57 of the HLA-DQ beta chain protects againsttype I diabetes: a family study. Proc Natl Acad Sci US A 1988; 85: 8111–8115.

37. Trucco M. To be or not to be ASP 57, that is thequestion. Diabetes Care 1992; 15: 705–715.

38. Aly TA, Ide A, Jahromi MM, et al. Extreme geneticrisks for type 1A diabetes. Proc Natl Acad Sci U S A2006; 103: 14 074–14 079.

39. Brown JH, Jardetzky T, Saper MA, et al. A hypotheticalmodel of the foreign antigen binding site of classII histocompatibility molecules. Nature 1988; 332:845–850.

40. Brown JH, Jardetzky TS, Gorga JC, et al. Three-dimensional structure of the human class IIhistocompatibility antigen HLA-DR1. Nature 1993;364: 33–39.

41. Jardetzky TS, Brown JH, Gorga JC, et al.Three-dimensional structure of a human classII histocompatibility molecule complexed withsuperantigen. Nature 1994; 368: 711–718.

42. Turco E, Manfras BJ, Ge L, et al. The x boxes frompromoters of HLA class II B genes at different locido not compete for nuclear protein-specific binding.Immunogenetics 1990; 32: 117–128.

43. Demotz S, Grey HM, Sette A. The minimal numberof class II MHC-antigen complexes needed for T cellactivation. Science 1990; 249: 1028–1030.

44. Akerblom HK, Knip M. Putative environmental factorsin type 1 diabetes. Diabetes Metab Rev 1998; 14:31–67.

45. Herman R, Knip M, Veijola R, et al. The FINDIANEStudy Group: temporal changes in the frequenciesof HLA genotypes in patients with type 1 diabetes:indication of an increased environmental pressure?Diabetologia 2003; 46: 420–425.

46. Gillespie KM, Bain SC, Barnett AH, et al. The risingincidence of childhood type 1 diabetes and reducedcontribution of high-risk HLA haplotypes. Lancet2004; 364: 1645–1647.

47. Kimpimaki T, Kupila A, Hamalaninen AM, et al. Thefirst signs of B-cell autoimmunity appear in infancyin genetically susceptible children from the generalpopulation: the Finnish Type 1 Diabetes Predictionand Prevention Study. J Clin Endocrinol Metab 2001;86: 4782–4788.

48. Laron Z. Lessons from recent epidemiological studiesin type 1 childhood diabetes. J Pediatr EndocrinolMetab 1999; 12(suppl 3): 733–736.

49. Yoon JW. Role of viruses in the pathogenesis ofIDDM. Ann Med 1991; 23: 437–445.

50. Varela-Calvino R, Peakman M. Enteroviruses and type1 diabetes. Diabetes Metab Res Rev 2003; 19: 431–441.

51. Green J, Casabonne D, Newton R. Coxsackie B virusserology and type I diabetes mellitus: a systemic

DOI:10.1002/MSJ


review of published case-control studies. Diabet Med2004; 21: 507–514.

52. Barbu AR, Akusjarvi G, Welsh N. Adenoviral-mediatedtransduction of human pancreatic islets: importanceof adenoviral genome for cell viability and associationwith a deficient antiviral response. Endocrinology2005; 146: 2406–2414.

53. Chen LK, Chou YC, Tsai ST, et al. Hepatitis C virusinfection-related type 1 diabetes mellitus. Diabet Med2005; 22: 340–343.

54. Jun HS, Yoon JW. A new look at viruses in type 1diabetes. Diabetes Metab Res Rev 2002; 19: 8–31.

55. Honeyman MC, Coulson BS, Stone NL, et al.Association between rotavirus infection andpancreatic islet autoimmunity in children at riskof developing type 1 diabetes. Diabetes 2000; 49:1319–1324.

56. Atkinson, MA, Bowman MA, Campbell L, et al.Cellular immunity to a determinant common toglutamic decarboxylase and Coxsackie virus ininsulin-dependent diabetes. J Clin Invest 1994; 94:2125–2129.

57. Titan J, Lehmann PV, Kaufman DL. T cell cross-reactivity between Coxsackie virus and glutamatedecarboxylase is associated with a murine diabetessusceptibility allele. J Exp Med 1994; 180: 1979–1984.

58. Takeda K, Kaisho T, Akira S. Toll-like receptors. AnnuRev Immunol 2003; 21: 335–376.

59. Dabbagh K, Lewis DB. Toll-like receptors and T-helper-1/T-helper-2 responses. Curr Opin Infect Dis2003; 16: 199.

60. Pasare C, Medzhitov R. Toll-like receptors andacquired immunity. Semin Immunol 2004; 16: 23–26.

61. Akira S, Uematsu S, Takeuchi O. Pathogen recognitionand innate immunity. Cell 2006; 124: 783–801.

62. Godfrey DI, Hammond KJ, Poulton LD, et al. NKTcells: facts, functions and fallacies. Immunol Today2000; 21: 573–583.

63. Medzhitov R. Toll-like receptors and innate immunity.Nat Rev Immunol 2001; 1: 135–145.

64. Akira S, Takeda K. Toll-like receptor signaling. NatRev Immunol 2004; 4: 499–511.

65. Kawai T, Akira S. TLR signaling. Nature 2006; 13:816–825.

66. Johnson GB, Brunn GJ, Platt JL. Activation ofmammalian Toll-like receptors by endogenousagonists. Crit Rev Immunol 2003; 23: 15.

67. Termeer C, Benedix F, Sleeman, J, et al.Oligosaccharides of hyaluronan activate dendriticcells via toll-like receptor 4. J Exp Med 2002; 195:99.

68. Buttari B, Profumo E, Mattei V, et al. Oxidizedbeta2glycoprotein I induces human dendritic cellmaturation and promotes a T helper type 1 response.Blood 2005; 106: 3880.

69. McSorley SJ, Ehst BD, Yu Y, Gewirtz AT. Bacterialflagellin is an effective adjuvant for CD4+ T cellsin vivo. J Immunol 2002; 169: 3914–3919.

70. Martin DA, Elkon KB. Autoantibodies make a U-turn:the toll hypothesis for autoantibody specificity. J ExpMed 2005; 202: 1465–1469.

71. Fritz JH, Le Bourhis L Sellge G, et al. Nod-like proteinsin immunity, inflammation and disease. Nat Immunol2006; 12: 1250–1257.

72. Klesney-Tait J, Turnbull IR, Colonna M. The TREMreceptor family and signal integration. Nat Immunol2006; 12: 1266–1273.

73. Robinson MJ, Sancho D, Slack EC, et al. Myeloid C-type lectins in innate immunity. Nat Immunol 2006;12: 1258–1265.

74. Beyan H, Buckley LR, Yousaf N, et al. A role forinnate immunity in type 1 diabetes?. Diabetes MetabRes Rev 2003; 19: 89–100.

75. Jun HS, Yoon CS, Zbytnuik L, et al. The roleof macrophages in T-cell mediated autoimmunediabetes in nonobese diabetic mice. J Exp Med 1999;189: 347–358.

76. Oschilewski U, Kiesel U, Kolb H. Administration ofsilica prevents diabetes in BB rats. Diabetes 1985; 34:197–199.

77. Lee KU, Pak CY, Amano K, Yoon JW. Preventionof lymphocytic thyroditis and insulitis in diabetesprone BB rats by the depletion of macrophages.Diabetologia 1988; 31: 400–402.

78. Medzhitov R, Janeway CA Jr. Innate immunity: thevirtues of a nonclonal system of recognition. Cell1997; 91: 295–298.

79. Bottazzo GF, Dean BM, McNally JM, et al. In situcharacterization of autoimmune phenomena andexpression of HLA molecules in the pancreas indiabetic insulitis. New Engl J Med 1985; 313: 353–360.

80. Conrad B, Weidman E, Trucco G, et al. Evidencefor superantigen involvement in insulin-dependentdiabetes mellitus etiology. Nature 1994; 371: 351–355.

81. Levisetti MG, Suri A, Petzold SJ, Unanue ER. Theinsulin-specific T cells of nonobese diabetic micerecognize a weak MHC-binding segment in morethan one form. J Immunol 2007; 178: 6051–6057.

82. Khatrri R, Cox T, Yasayko SA, Ramsdell F. An essentialrole for scurfin in CD+4CD+25 T-regulatory cells. NatImmunol 2003; 4: 337–342.

83. Fontenot JD, Gavin MA, Rudensky AY. FoxP3programs the development and function of CD+4CD+25 regulatory T-cells. Nat Immunol 2003; 4:330–336.

84. Hori S, Nomura T, Sakaguchi S. Control of regulatoryT cell development by the transcription factor Foxp3.Science 2004; 299: 1057–1061.

85. Randolph DA, Fathman CG. CD+4CD+25 regulatoryT cells and their therapeutic potential. Annu Rev Med2006; 57: 381–402.

86. St. Clair EW, Turka LA, Saxon A, et al. New reagentson the horizon for immune tolerance. Annu Rev Med2007; 58: 329–346.

87. Bluestone JA, Tang Q. How do CD4+CD25+regulatory T cells control autoimmunity? Curr OpinImmunol 2005; 17: 638–642.

88. Tang Q, Bluestone JA. Regulatory T-cell physiologyand application to treat autoimmunity. Immunol Rev2006; 212: 217–237.

89. Brusko TM, Wasserfall CH, Clare-Salzler MJ, et al.Functional defects and the influence of age on thefrequency of CD+4CD+25 T-cells in type 1 diabetes.Diabetes 2005; 54: 1407–1414.

90. Lindley S, Dayan CM, Bishop A, et al. Defectivesuppressor function in CD+4CD+25 t cells frompatients with type 1 diabetes. Diabetes 2005; 54:92–99.

DOI:10.1002/MSJ


91. Putnam AL, Vendrame F, Dotta F, Gottlieb PA. CD+4CD+25 high regulatory T cells in human autoimmunediabetes. J Autoimmun 2005; 24: 55–62.

92. Mason D, Powrie F. Control of immune pathologyby regulatory T cells. Curr Opin Immunol 1998; 10:649–655.

93. Bendelac A. Mouse NK+1 T cells. Curr Opin Immunol1995; 7: 367.

94. Bendelac A, Rivera MN, Park SH, Roark JH. MouseCD1-specific NK1 T cells–development, specificity,and function. Ann Rev Immunol 1997; 15: 535–562.

95. Macdonald HR. NK1.1+ T cell receptor-α β+ cell–newclues to their origin, specificity, and function. J ExpMed 1995; 182: 633–638.

96. Burdin N, Brossay L, Kronenberg M. Immunizationwith α-galactosylceramyde polarizes CD1-reactive NKT-cells toward Th2 cytokine synthesis. Eur J Immunol1999; 29: 2014–2025.

97. Singh N, Hong S, Scherer DC, et al. Cuttingedge: activation of NK T-cells by CD1d and α-galactosylceramyde directs conventional T cells to

the acquisition of a Th2 phenotype. J Immunol 1999;163: 237–2377.

98. Wilson SB, Kent SC, Patton KT, et al. Extreme Th1bias of invariant Vα 24Jα Q T cells in type 1 diabetes.Nature 1998; 391: 177–181.

99. Hammond KJL, Poulton LD, Palimsano LJ, et al. α β-Tcell receptor (TCR)+ CD-4 CD-8 (NKT) thymocytesprevent insulin-dependent diabetes mellitus innonobese diabetic (NOD/Lt) mice by the influenceof interleukin-4 (IL-4) and/or IL-10. J Exp Med 1998;187: 1047–1056.

100. Smink LJ, Helton EM, Healy BC, et al. T1DBase, acommunity web-based resource for type 1 diabetesresearch. Nucleic Acids Res 2005; 1(33): D544–D549.

101. Cohen CD, Klingenhoff A, Boucherot A, et al.Comparative promoter analysis allows de novoidentification of specialized cell junction-associatedproteins. PNAS 2006; 103: 5682–5687.

102. Anderson MS, Bluestone JA. The NOD mouse: amodel of immune dysregulation. Annu Rev Immunol2005; 23: 447–485.

DOI:10.1002/MSJ

Type 1 Diabetes: Etiology, Immunology,and Therapeutic Strategies

TOM L. VAN BELLE, KEN T. COPPIETERS, AND MATTHIAS G. VON HERRATH

Center for Type 1 Diabetes Research, La Jolla Institute for Allergy and Immunology, La Jolla, California

I. Introduction 80II. The Genetics of Type 1 Diabetes 80

A. Rare monogenic forms 80B. HLA genes 81C. The insulin gene 81D. PTPN22 82E. IL2RA 82F. CTLA-4 82G. Genome-wide association studies and rare polymorphisms in the IFIH1 gene 82H. Parallels between human genetics and genetics of the NOD mouse 83

III. Precipitating Events 84A. Viral infections 84B. Bacteria 86C. Other environmental triggers 86

IV. Timelines for the Pathogenesis of Type 1 Diabetes 87V. Immune Events in Type 1 Diabetes 90

A. Metabolic changes: linking cause to immune events? 90B. B cells produce diabetes-associated anti-islet autoantibodies 90C. Islet-specific T cells in the periphery 90D. Immunological events in the human pancreas 91E. The honeymoon phase: do beta-cells transiently revive? 92F. The NOD mouse: an entirely distinct picture 92

VI. Identification of Prediabetic Individuals 93A. Genetic screening 93B. Diabetes-associated anti-islet autoantibodies 93C. Islet-specific T-cell assays 94D. Metabolomic screening 95E. Assessing �-cell mass 95

VII. Preventive Trials 95VIII. New-Onset Type 1 Diabetes Trials 96

A. Antigen-specific intervention trials in T1D 96B. Non-antigen-specific intervention trials in T1D 97C. Cell-based tolerogenic therapy 101D. Replacing beta-cell shortage 102E. Combination intervention trials in T1D 104

IX. Promising Bench-Side Therapeutics 105A. Insulin substitution 105B. Combination therapies with immune modulators and islet antigenic vaccines 105C. Combination therapies using immune modulators and compounds enhancing beta-cell mass

or function 105D. Cytokine-based therapeutics 106

X. Our Conclusions 106

Van Belle TL, Coppieters KT, von Herrath MG. Type 1 Diabetes: Etiology, Immunology, and TherapeuticStrategies. Physiol Rev 91: 79–118, 2011 doi:10.1152/physrev.00003.2010.—Type 1 diabetes (T1D) is a chronicautoimmune disease in which destruction or damaging of the beta-cells in the islets of Langerhans results in insulindeficiency and hyperglycemia. We only know for sure that autoimmunity is the predominant effector mechanism ofT1D, but may not be its primary cause. T1D precipitates in genetically susceptible individuals, very likely as a resultof an environmental trigger. Current genetic data point towards the following genes as susceptibility genes: HLA,

Physiol Rev 91: 79–118, 2011;doi:10.1152/physrev.00003.2010.

www.prv.org 790031-9333/11 Copyright © 2011 the American Physiological Society

on August 26, 2011

physrev.physiology.orgD

ownloaded from

http://physrev.physiology.org/

insulin, PTPN22, IL2Ra, and CTLA4. Epidemiological and other studies suggest a triggering role for enteroviruses,while other microorganisms might provide protection. Efficacious prevention of T1D will require detection of theearliest events in the process. So far, autoantibodies are most widely used as serum biomarker, but T-cell readoutsand metabolome studies might strengthen and bring forward diagnosis. Current preventive clinical trials mostlyfocus on environmental triggers. Therapeutic trials test the efficacy of antigen-specific and antigen-nonspecificimmune interventions, but also include restoration of the affected beta-cell mass by islet transplantation, neogenesisand regeneration, and combinations thereof. In this comprehensive review, we explain the genetic, environmental,and immunological data underlying the prevention and intervention strategies to constrain T1D.

I. INTRODUCTION

Type 1 and type 2 diabetes mellitus (T1D, T2D) havein common high blood glucose levels (hyperglycemia)that can cause serious health complications includingketoacidosis, kidney failure, heart disease, stroke, andblindness. Patients are often diagnosed with diabeteswhen they see a physician for clinical signs such as ex-cessive thirst, urination, and hunger. These symptomsresult from the underlying hyperglycemia that is in turncaused by insufficient insulin functionality. In T2D, whichis usually associated with obesity or older age, this ismostly the result of insulin resistance: the muscle oradipose cells do not respond adequately to normal levelsof insulin produced by intact beta-cells. T1D on the otherhand usually starts in people younger than 30 and istherefore also termed juvenile-onset diabetes, eventhough it can occur at any age. T1D is a chronic autoim-mune disorder that precipitates in genetically susceptibleindividuals by environmental factors (24). The body’s ownimmune system attacks the beta-cells in the islets ofLangerhans of the pancreas, destroying or damaging themsufficiently to reduce and eventually eliminate insulinproduction. On rare but increasing occasions, both T1Dand T2D are diagnosed in patients.

According to the American Center for Disease Con-trol, 23.6 million people, 7.8% of the population, have T1Dor T2D, and 1.6 million new cases of diabetes were diag-nosed in people aged 20 years or older in 2007. Theprevalence of T1D for residents of the United States aged0–19 years is 1.7/1,000. T1D incidence has been globallyrising during the past decades by as much as a 5.3%annually in the United States. If present trends continue,doubling of new cases of T1D in European childrenyounger than 5 years is predicted between 2005 and 2020,and prevalence of cases in individuals younger than 15years will rise by 70% (328), characteristic of a left shifttowards an earlier age (123). This suggests that whateverevent triggers the onset is increasingly affecting suscep-tible individuals (115, 165). The search for such triggeringfactors has been ongoing for many years and has so faronly yielded indirect evidence, predominantly implicatingcertain viral infections. It is now well established that aspecific genetic constitution is required for such an eventto cause diabetes. However, concordance rates betweenmonozygotic twins amount to only 50%, whereas between

dizygotic twins only �10% (236). With longer follow-up,the majority of discordant identical twins of patients withT1D eventually express anti-islet autoantibodies andprogress to diabetes, but anti-islet autoantibodies in thesecond twin may appear only 30 years after the first twindevelops diabetes (354, 355). Thus it seems that geneticsusceptibility persists for life, and progression to diabetesis usually preceded by a long prodrome of anti-islet auto-antibody expression measured in years. Nevertheless, al-though the concordance rate for monozygotic twins ishigher than previously thought, it is below unity, andthere are strong divergences in terms of the time it takesto develop T1D. This implies a strong environmental com-ponent to contribute to the development of T1D.

Since the early 1920s, diabetes has been treated byinsulin replacement, which, in the ideal case, will onlyshorten life expectancy by �10 years. This sets a highsafety bar for any immune-based intervention. Even moreso, recent technology (continuous blood glucose moni-tors, slow release insulin, etc.) can reduce the chance forlife-threatening hypoglycemic episodes from insulin over-doses. Therefore, immune-based interventions should ide-ally be effective, long-lasting, and have minimal side ef-fects to replace substitutive insulin treatment with a cure.Today, despite the many remaining challenges in the fieldof T1D immunotherapy, good progress has been made.

With this review, we provide a comprehensive over-view of the etiology and immunology of T1D and willdiscuss preventive or therapeutic biological strategiesthat have been tried or are currently undertaken.

II. THE GENETICS OF TYPE 1 DIABETES

A comprehensive overview of genetic data in mouseand human is beyond the scope of this article. Instead, wewill focus on how the various susceptibility genes andenvironmental triggers can fit in a mechanistic model forT1D etiology.

A. Rare Monogenic Forms

Autoimmune diabetes is only rarely caused by muta-tional defects in a single gene. These monogenic formsare typically accompanied by multiple other autoimmuneconditions due to the disruption of common regulatory

80 VAN BELLE, COPPIETERS, AND VON HERRATH

Physiol Rev • VOL 91 • JANUARY 2011 • www.prv.org

on August 26, 2011


ownloaded from


pathways. One such example is found in the IPEX syn-drome (immune dysregulation, polyendocrinopathy, en-teropathy, X-linked), in which mutations in the Foxp3transcription factor lead to the dysfunction of regulatoryT cells (Tregs) and wasting multiorgan autoimmunity (26,77, 474). Approximately 80% of affected children developautoimmune diabetes and generally succumb early due tooverwhelming autoimmunity. Another example is autoim-mune polyendocrine syndrome type 1 (APS-1, or APECEDfor autoimmune polyendocrinopathy-candidiasis-ectoder-mal dystrophy). Mutations in the transcription factorAIRE (autoimmune regulator) lead to severe autoimmuneconditions, and �20% of the cases develop T1D (455).Deficiencies in AIRE inhibit the expression of peripheralmolecules, for example, insulin, in the thymus. This re-duced expression allows autoreactive T cells to escapeinto the periphery, because it interferes with thymic de-letion (16, 246). These rare monogenic forms represent asmall minority of T1D cases, but highlight two main fea-tures related to its etiology. First, the observation thatseveral well-characterized autoimmune conditions de-velop in parallel highlights the common tolerance mech-anisms that prevent these diseases in healthy individuals.Second, although genetics and environment likely interactin a continuous spectrum in most autoimmune diseasepatients, the monogenic cause of IPEX and APS-1 illus-trates that genetic constitution can dominate in certainextreme cases.

B. HLA Genes

Early studies indicated that the HLA region on chro-mosome 6p21 (commonly termed IDDM1, for insulin-de-pendent diabetes mellitus locus) is a critical susceptibilitylocus for many human autoimmune diseases, includingT1D (305, 399). These initial findings revolutionized ourunderstanding of T1D etiology in two ways, as stated byNerup et al. (305) in conclusion of their 1974 report:1) T1D is a distinct disease entity, corroborating his-topathological evidence; and 2) an aberrant cellular im-mune response, potentially triggered by viral infection,instigates onset. Numerous new susceptibility loci haveemerged since, but none of them matches the strongassociation found with the HLA region. It is unlikely thatnew loci will ever be discovered that confer such a dra-matic risk to T1D development (96). In genetic studies,the odds ratio is the statistic used to calculate whether asingle nucleotide polymorphism (SNP) given is associatedwith the disease. An odds ratio of one implies that theevent is equally likely in both patient and control groups.Odds ratios of alleles predisposing to complex disordersare typically modest, often in the range of 1.2–1.3, andeven the HLA region has a predicted value of only 6.8.This suggests that if genetic predisposition is indeed a

dominant factor in T1D development, a vast amount ofcommon SNPs are still waiting to be discovered (96, 159).After several decades of continuous progress since the dis-covery of HLA association (for historical perspective, seeRef. 285), the class II genes remain the strongest geneticcontributor (138, 323, 429, 433, 439). Several HLA class IIgenes are pivotal as their alleles were found to determine asusceptibility hierarchy ranging from protection to stronglyat-risk (15, 73, 105, 134, 135, 237, 309, 393). The DRB1*1501-DQA1*0102-DQB1*0602 haplotype, found in �20% of thepopulation but only 1% of patients, confers dominant pro-tection against T1D (134). At the susceptible end of thisspectrum are individuals with the DR3/4-DQ8 heterozygoushaplotype (DR3 is DRB1*03-DQB1*0201, DR4 is DRB1*04-DQB1*0302, DQ8 is DQA1*0301, DQB1*0302). It is importantto note that only 30–50% of patients with T1D have theDR3/4-DQ2/8 genotype. A study in the Denver, Coloradoarea (15) identified this high-risk haplotype in 2.4% of new-borns and more than 20% of the children affected by T1D,and its presence marks a 55% risk of developing overt dia-betes by age 12. DR3/4-DQ2/8 siblings who are HLA identicalto a diabetic proband have a risk as high as 80% for persis-tent anti-islet autoantibodies and 60% for progression todiabetes by age 15 (15).

An equally consistent, albeit substantially less prom-inent, association has been found for class I alleles (140,192, 245, 302, 303, 308, 445, 446). A recent study by Ne-jentsev et al. (303) demonstrates that, after taking intoaccount the dominating influence of class II genes, mostof the residual association in the HLA region can beattributed to HLA-B and HLA-A genes (303). Most notablythe presence of the HLA-B*39 allele was found to be asignificant risk factor and is associated with a lower age atdiagnosis of T1D. Additionally, HLA-A*02 increases therisk in individuals possessing the high-risk class II DR3/4-DQ8 haplotype (140, 360). HLA-A*0201 is one of themost prevalent class I alleles, with a frequency of �60% inT1D patients. There is accumulating evidence for thepresence and functionality of HLA-A*02-restricted CD8 Tcells reacting against beta-cell antigens such as insulin,glutamate decarboxylase (GAD), and IAPP in T1D pa-tients and islet transplant recipients (325, 326, 337).Transgenic NOD mice have been generated expressinghuman HLA-A*02 molecules (271), and their accelerateddiabetes onset provides functional evidence for the in-volvement of this particular class I allele.

C. The Insulin Gene

A lesser genetic predisposition to T1D is conferred bythe IDDM2 locus on chromosome 11 containing the insu-lin gene region. A polymorphic region located 5= of theinsulin gene was first reported in 1984 to be associatedwith T1D in caucasoids (39). Now established as a pri-

TYPE 1 DIABETES: FROM CAUSE TO CURE 81


on August 26, 2011


ownloaded from


mary autoantigen in T1D, it is not at all surprising thatmutations in the insulin region could contribute to diseasesusceptibility. Detailed mapping showed that susceptibil-ity resides in a variable number of tandem repeat (VNTR)polymorphisms in the promoter region of the insulin gene(42, 214, 225, 253). The magnitude of risk correlates withthe number of these tandem repeats. VNTR type I (withshorter repeats) homozygous individuals in the highestrisk category and VNTR type III (longer repeats) protectscarriers against T1D. The predominant hypothesis is thatthese VNTR regulate the insulin expression levels in thethymus by affecting AIRE binding to its promoter region (16,342, 442). The importance of insulin expression by Aire-expressing medullary thymic epithelial cells (mTECs) wasrecently underscored by the observation that mTEC-specificinsulin deletion leads to diabetes in animals (137). As aconsequence, VNTR type I will induce lower transcription ofinsulin and its precursors in the thymus, leading to reducedtolerance and T1D development. Conversely, insulin-reac-tive T cells are more efficiently eliminated by negative se-lection in the thymus from individuals with the protectiveVNTR type III allelic variant.

D. PTPN22

A relatively new member of the T1D susceptibilitygene set is PTPN22, which encodes the lymphoid proteintyrosine phosphatase (LYP) (57, 402). The same allelicvariant mediates risk in several other autoimmune dis-eases, suggesting the involvement of a crucial signalingaxis (58). Indeed, the LYP protein is an important negativeregulator of T-cell receptor signaling by way of dephos-phorylation of Src family kinases Lck and Fyn, ITAMs ofthe TCR�/CD3 complex, as well as ZAP-70, Vav, valosin-containing protein, and other key signaling molecules(476). Explanations for the mechanism are contradicting.A loss-of-function mutation can cause a lower thresholdfor autoreactive T-cell activation in the periphery. In con-trast, a gain-of-function mutation that suppresses TCRsignaling during thymic development can allow autoreac-tive T cells to escape negative selection (451).

E. IL2RA

Allelic variation in the interleukin (IL)-2 receptor-�gene (IL2RA) region accounts for another genetic riskfactor implicated in T1D (252, 345, 453). The alpha chainof the IL-2 receptor complex (IL2R�, CD25) is an essentialmolecule expressed on T cells upon activation and onnatural Tregs at baseline. Tregs depend on IL-2 for theirgrowth and survival. The presence of the IL2R� subunitgreatly enhances the affinity of the IL-2 receptor (266). Inmultiple sclerosis (MS) (263) and other autoimmune con-ditions (5, 157, 166), increased levels of soluble IL2R�

(sIL2R�) are found in circulation. Given the indispensiblerole of IL-2 in Treg function and the potential for sIL2R�to neutralize IL-2, one could argue that IL2RA allelic riskvariants impair Treg functionality by upregulation ofsIL2R�. However, it was recently found that IL2RA sus-ceptibility genotypes in T1D are associated with lowerlevels of sIL2R� (252, 262). Furthermore, in vitro stimu-lated peripheral blood mononuclear cells (PBMCs) fromindividuals with T1D make less sIL2R� than those fromcontrol individuals (158). This could indicate a defect inthe cellular subset that is the source of cleaved IL2R�. Analternative explanation may be that, even in the presenceof normal Treg frequencies in T1D (70), IL2RA polymor-phisms account for functional defects in the Treg com-partment (72, 346). In conclusion, it seems that althoughgenetic variability in the IL2RA gene is associated withseveral autoimmune diseases including T1D, the mecha-nisms and extent to which sIL2R� levels mediate theseconditions differs significantly.

F. CTLA-4

Another confirmed T1D risk allele lies in the geneencoding cytotoxic T lymphocyte-associated protein 4(CTLA-4) in the IDDM12 region (307, 437). CTLA-4 is byall accounts a vital molecule for proper negative regu-lation of immune responses, as evidenced by the severelymphoproliferative disorders seen in knock-out mice(468). As with other regions, the risk association ofallelic variants in the CTLA-4 region is again not exclu-sively confined to T1D, but was replicated in severalother prevalent autoimmune disorders, including MS(283), systemic lupus erythematosus (SLE) (35), andrheumatoid arthritis (RA) (443). SNPs have been de-scribed in the human CTLA-4 promoter region and inexon 1. The A49G polymorphism is the only polymor-phism that changes the primary amino acid sequence ofCTLA-4. In vitro studies of A49G CTLA-4 have shownthat this mutant form of CTLA-4 is aberrantly processedin the endoplasmic reticulum, leading to reduced sur-face expression (17). Exactly how these polymorphismsaffect CTLA-4 function is still unclear. In addition toeffects on processing and intracellular trafficking, theymay affect oligomerization and surface retention (426).The predominant hypothesis in humans, however, is thatthe allelic variant lowers the mRNA levels of the solubleCTLA-4 splice variant (437).

G. Genome-Wide Association Studies and Rare

Polymorphisms in the IFIH1 Gene

The advent of genome-wide association (GWA)studies has enabled high-throughput analysis of SNPsacross the entire human genome at a resolution previ-



on August 26, 2011


ownloaded from


ously unattainable, in thousands of unrelated individu-als, in a non-hypothesis-driven manner (333). Publishedresults from multiple GWA studies and meta-analysesperformed upon them have confirmed the associationof the genes discussed above and identified a consid-erable number of new risk loci (75, 97, 164, 173, 174,404, 434). The most recent GWA study focusing on T1Dfound that over 40 loci affect risk of T1D, includingnewly identified coding regions for immunoregulatorymolecules such as IL-10 (36). It can be concluded fromsuch comprehensive studies that autoimmune diseasesindeed share many of the genetic risk factors, suggest-ing common underlying pathways. Further adding tothat argument was the very recent discovery of predis-posing SNP in the functional candidate gene TYK2,implicated in interferon (IFN)-�, IL-6, IL-10, and IL-12signaling. This association had been previously demon-strated in MS, ankylosing spondylitis, SLE, and autoim-mune thyroid disease (465). We will refrain from sum-marizing all these putative risk genes and speculatingon their etiological implications for T1D. Instead, wehighlight T1D-associated polymorphisms in the IFIH1region (403), because they link genetic constitution andputative environmental factors.

Interferon-induced helicase 1 (IFIH1) codes for anIFN-induced helicase that contributes to recognition ofdsRNA from picorna viruses. As such, IFIH1 serves as acytoplasmic sensor for viral infection (221, 291). As wediscuss below, one of the most prominent T1D-linkedviruses is coxsackievirus B (CVB), an enterovirus be-longing to the picornavirus family. Mouse studies con-firm that IFIH1 and its adaptor molecule MAVS arecritical for type I interferon responses to CVB (467),particularly during the late phase (495). Thus a geneticdefect in IFIH1 could potentially interfere with properdetection and clearance of viral infections and lead toan abnormal, diabetogenic immune response. An inde-pendent study confirmed the presence of T1D-associated polymorphisms in the IFIH1 gene andshowed that gene expression levels in PBMC are higherin individuals with the susceptible genotypes (248). Aplausible hypothesis is that in these individuals withhigher IFIH1 levels, viral infections may be primarilyrecognized by the IFIH1 pathway, leading to exacer-bated antiviral immunity and production of type I in-terferons. The identification of rare protective IFIH1variants in T1D is in line with this hypothesis (304). Oneof the protective variants is a nonsense mutation lead-ing to a truncated protein, while two other variantsprobably disrupt normal splicing of the IFIH1 tran-script. The initial prediction that these variants reducethe function of the IFIH1 protein, and thereby decreasethe risk of T1D, has since been experimentally con-firmed (395).

H. Parallels Between Human Genetics and

Genetics of the NOD Mouse

The majority of mechanistic data on T1D pathogen-esis and potential interventions have been derived frommouse studies. It is therefore important to understand thegenetic predisposition of the most widely used mousemodel, the NOD mouse. Similar to the human IDDM lociterminology, genetic regions that control progression toT1D in the NOD are designated as insulin-dependent dia-betes (Idd) loci. One approach for detailed functionalanalysis of the risk loci is by creating congenic mice, inwhich specific disease susceptibility loci are replacedwith protective genes derived from strains that are notdiabetes-prone. Such studies confirm that, as in humans,major histocompatibility complex (MHC) class II (Idd1)genes in particular are the dominant genetic contributorsto disease predisposition in NOD mice. In addition, over20 non-MHC Idd regions have been found to mediatedisease risk (264). We will focus on some of the suscep-tibility loci that are shared between human and mouse.

The first risk locus that shows correspondence be-tween human and mouse is Ctla4 (Idd5.1) (472). Humansexpress two major splice variants coding for membrane-bound and soluble forms of CTLA-4. Mice express anadditional variant lacking the B7 binding domain, termedligand-independent CTLA-4 (liCTLA-4) (437). Genetic pro-tection in the Idd5.1 congenic mouse strain was found tobe mediated by higher expression levels of the liCTLA-4isoform, which negatively modulates T-cell responses.This confirms the concept that polymorphisms affectingthe levels of alternatively spliced CTLA-4 isoforms con-tribute to T1D susceptibility (454).

Genetic susceptibility in humans and mice also sharevariations in the IL-2 signaling pathway. But, whereassusceptibility in humans relates to the IL2RA gene region,variants in the NOD are in the IL-2 gene (in Idd3) (259,339). Consistent with an indispensible role for proper IL-2signaling to prevent T1D, several reports have docu-mented the correlation of Idd3 with decreased IL-2 levelsleading to impaired tolerance induction and Treg func-tionality (160, 269, 272, 332, 390). Taken together, thisdegree of similarity between two evolutionary distinctspecies emphasizes the critical role of the IL-2 pathway inmaintaining self-tolerance. Although disruption occurs atdifferent parts of the signaling cascade, the outcomecould similarly be a breakdown of T-cell homeostasis andthe expansion of a diabetogenic T-cell subset.

Some observations suggest that the mouse orthologof PTPN22, Ptpn8, influences disease in the NOD, but thisassociation has yet to be confirmed (473). As of today, noIdd locus has been revealed that regulates thymic insulinexpression to the extent as demonstrated in humans.However, the concept of thymic insulin levels as a criticalregulatory threshold during negative selection has been



on August 26, 2011


ownloaded from


confirmed by studying NOD mice with different degreesof insulin gene-dosage deficiency (92). Apart from theinherent evolutionary divergence of both species, someother limitations arise with regard to translation of ge-netic data from the NOD model to humans. For instance,NOD mice exhibit much stronger insulitis than that foundin human islets (see also Fig. 3). Moreover, translation ofimmunomodulatory interventions from the NOD model toT1D patients is not straightforward (see below).

III. PRECIPITATING EVENTS

It is now established that clinical manifestation of T1Dreflects the consequence of an underlying, sustained auto-immune process. For instance, autoantibodies against isletantigens are detected before the clinical onset of T1D. Thissuggests that a sequence of inciting events precedes thehyperglycemia for at least months, but most likely severalyears. This wide gap between initiation and detection ofongoing diabetogenic events poses a cardinal problem inthe search for causative environmental triggers. Further-more, the possibility exists that inciting agents may be ofa “hit-and-run” type, leaving no detectable moleculartrace. Alternatively, it may take multiple environmentalinsults to unleash autoimmunity, and different patientsmay incur them in divergent combinations. Having dis-cussed the major genetic contributions to T1D suscepti-bility, we next discuss convincing epidemiological dataindicating that the rising incidence in T1D has to beattributed to environmental changes (51, 156).

A. Viral Infections

A search on PubMed using the keywords “virus” and“type 1 diabetes” produced 1,355 titles at the time ofwriting, mirroring the extensive efforts to elucidate therole of viral infections in T1D. Publications dating back asfar as 1926 document the seasonal variation of diabetesonset and make some link with viral infections (6). De-spite all these efforts, there is still no direct evidence fora particular viral strain being causative.

1. Enteroviruses

Extensive circumstantial data designate enterovi-ruses, and more specifically coxsackieviruses, as primeviral candidates that can cause precipitation of T1D (142).The first papers suggesting a link between coxsackievirusinfections and T1D were based on the finding of higherneutralizing antibody titers in serum from recent-onsetpatients versus healthy controls (152). These data werelater confirmed using PCR technology (95). Some studiesalso tested for antibodies against other viruses in parallel,but coxsackievirus was usually found to be most preva-

lent (32). The possibility of a causal relationship has sincebeen evaluated with varying success in both human stud-ies (149, 436) and animal models (483). The question wasapproached from an interesting angle in a 1971 study thatfollowed the diabetes incidence after an epidemic of Cox-sackievirus B4 (CVB4) infection on the isolated PribilofIslands. Five years after the epidemic, the incidence ofdiabetes in the CVB4-infected versus noninfected individ-uals was found to be similar, suggesting no link betweenCVB4 infection and T1D onset (117). Given our currentknowledge on genetic contribution, this is a textbookcase arguing that viral infection alone does not cause T1Din any given genetic background. Yoon et al. (484) laterprovided more direct support for CVB4 involvement bydemonstrating that the virus could infect beta-cells andcause insulitis and diabetes in susceptible mouse strains(484). Furthermore, the same group isolated a CVB4strain from a child with recent-onset T1D (482). Func-tional data show enhanced T-cell responses to CVB4 pro-teins in children with T1D after the onset of the disease(213). Among these CVB4-reactive cells, the effector/memory phenotype predominates around the time of di-agnosis (452). Last, sampling within the Finnish popula-tion revealed associations between enterovirus infectionand diabetes development in prospective studies (250,251; recently reviewed in Ref. 425). Although enterovirusinfection is unlikely to represent the exclusive cause forthe extremely high and rising incidence of T1D in Finland,it may well be a significant contributing factor.

In their 1987 landmark paper, Foulis et al. (145)reported on the presence of abundant levels of HLA classI and IFN-� in the islets of recent-onset diabetic children,further fueling interest into the role of viruses in T1D(145). It is conceivable that beta-cell-tropic infection up-regulates both HLA class I and IFN-� and leaves a molec-ular “viral signature.” This also could explain why theimmune response is directed specifically against the is-lets. A follow-up study probing for viral proteins in beta-cells disappointingly failed to detect any viral compo-nents (144), but the same group recently revised theirconclusions after reanalysis of the same cohort usingoptimized methodology (356). Evidence of enteroviruspresence was found in islets from 44 of 72 recent-onsetpatients versus 3 of 50 controls, offering the closest indi-cations to a causal relationship to date. However, islets of10 of 25 type 2 diabetics also showed traces of enteroviruspresence, and concerns were raised regarding the speci-ficity of the reagents that were used (363). Nevertheless,Dotta et al. (119) recently replicated the immunohisto-chemical detection of enterovirus protein and confirmedthe results by sequencing. Unique pancreas samples suchas those obtained by Foulis and co-workers (99) nowrarely become available, because of the greatly improvedclinical management of T1D. Therefore, replication of theresults depends on initiatives such as the Juvenile Diabe-



on August 26, 2011


ownloaded from


tes Research Fund-funded Network for Pancreatic OrganDonors (nPOD), aimed at nationwide procurement of tis-sue relevant to T1D research (481).

Whereas the presence of enterovirus particles in pan-creatic islets suggests that T1D is a consequence of selectiveviral infection of beta-cells, data favoring alternative mech-anisms have been reported. The striking sequence similari-ties between the 2C protein from coxsackievirus and GAD,a major autoantigen in T1D, lead to the postulation of viralmimicry in the etiology of T1D (223). Subsequent resultsargued both in favor (22, 430) and against (195, 357, 382)such a mechanism, and some suggested an important con-tribution of HLA-DR3 in susceptibility to viral mimicry (464).Despite the sequence similarity between GAD and hsp60(212), the association between autoimmunity to hsp60 andT1D turned out to be irreproducible (25, 357).

The timing of enterovirus infection in relation to T1Donset is a controversial issue. In addition to demonstrabletraces of infection in recent-onset individuals, enterovirusinfections have also been demonstrated before onset inprediabetic, autoantibody-positive children (374). Entero-virus infections during pregnancy were also identified asa risk factor for T1D (107, 126, 202). Together, these datasuggest that viral infection instigates the autoimmuneresponse. However, data from NOD mice show that pre-existing insulitis is required for coxsackievirus to inducediabetes (121, 196, 387). Translating this to the humansituation, susceptible individuals may have ongoing sub-clinical insulitis for years until viral challenge triggers anacceleration of beta-cell destruction and hyperglycemia.Another interesting observation in this regard is that en-teroviruses were recently found in intestinal biopsy sam-ples in 75% of T1D cases versus 10% of control patients.This might reflect a persistent enterovirus infection of gutmucosa (315) that serves as a previously unidentified viralreservoir that may spill into pancreatic territory at latertime points and cause insulitis.

CVB-induced upregulation of the chemokine CXCL10on pancreatic beta-cells was recently exposed as one ofthe earliest consequences of infection (43). Hyperexpres-sion and viral presence coincided in “fulminant” T1D, anextremely aggressive T1D variant found in the Japanesepopulation (419). Animal studies have delineated a crucialrole for CXCL10 in the islets during the recruitment ofCXCR3-expressing autoreactive T cells following viral in-fection (93).

Cumulatively, the available data suggest the involve-ment of CVB in at least a subset of T1D cases and wouldwarrant prevention of infection in susceptible individuals.Making matters complex, however, is the finding thatunder certain experimental conditions CVB can actuallyprotect from disease (435), supporting the “hygiene hy-pothesis” (27). Our group has shown that this protectionis mechanistically governed by two distinct pathways,including functional Treg enhancement and upregulation

of the coinhibitory receptor PD-L1 on lymphoid cells(141). These data shed light on the ambiguous role of viralinfections in the context of autoimmunity.

2. Other viruses

Other viral infections have been associated with T1D,but a causal relation has not been proven.

The potential association of T1D and rotaviruses, themost common cause of childhood gastroenteritis, is basedon possible molecular mimicry. Similarities were initiallyfound between T-cell epitopes in GAD and IA-2 and viralprotein (193). A subsequent Australian study found anassociation between rotavirus infection and islet autoan-tibody positivity in at-risk children (191), but a Finnishgroup failed to confirm this relationship (48). The sameFinnish group later unsuccessfully sought an associationbetween rotavirus-specific T-cell responses and the pres-ence of T1D-associated autoantibodies (265). Thus thepresent status of rotavirus infection in the etiology of T1Dremains unconfirmed. Likewise, many initial reports onthe potential role of other viruses in T1D, such as cyto-megalovirus (324), parvovirus (169, 220), and encephalo-myocarditis virus (103), were challenged or remain to beconfirmed in large patient populations.

A conceptually interesting case was made for theconvincing relationship between congenital rubella infec-tion and diabetes onset after birth, a topic recently re-viewed by Gale (150). Congenital rubella syndrome con-sists of a wide range of both physical and behavioraldisorders and hence is characterized as a multisystemdisease. Interestingly, progression to diabetes was asso-ciated with a higher frequency of the HLA-A1-B8-(DR3-DQ2) T1D susceptibility haplotype (289), but direct evi-dence for islet-specific autoimmunity is extremely scarce(329). An alternative mechanism that the virus interfereswith beta-cell mass development is now favored. As aconsequence of these atypical features, some clinical con-sensus guidelines list congenital rubella diabetes in adistinct category of “other specific types” of diabetes (74).But the most important argument that rubella infection isnot responsible in the rising global T1D incidence is thatthe virus has been largely eliminated in wealthier coun-tries since the introduction of an efficient vaccine in 1969.A similar argument excludes mumps infection despite arecent report documenting a case of “fulminant” T1D(161, 190).

Because vaccinations did not reduce T1D incidence,it was questioned whether widespread vaccination pro-grams could actually account for the observed rise. In-deed, introduction of general childhood immunizationsand the growing prevalence of T1D in developed coun-tries seemed to happen concurrently. However, multiplelarge studies found no support for any causal relationbetween childhood vaccination and T1D (47, 112, 136,



on August 26, 2011


ownloaded from


201), so the risk-to-benefit ratio remains strongly in favorof continued protection efforts against infectious diseasesby means of immunization.

B. Bacteria

The bacterial composition of the intestine has longbeen acknowledged as an important variable affectingT1D development. Direct evidence exists in rodents, asdiabetes is aggravated under specific pathogen-free con-ditions or upon administration of antibiotics (27). In otherstudies, however, administration of antibiotics preventsdiabetes (68, 384). Perhaps autoimmunity ensues when-ever the intricate microbial balance in the intestine isdisturbed. Additionally, the intestinal wall does not seemto have the same capacity to form a coherent barrierseparating luminal bacteria and the immune system inT1D models and patients versus controls. This so-called“leaky gut” phenotype is thought to enhance the exposureof bacterial antigens to the immune system (441). In theintestine of T1D patients, subclinical immune activation(380, 471) and evidence for an impaired Treg subset (431)were found. We have already mentioned the presence ofenteroviruses in intestinal specimens, which may be oneof the triggers of this inflammatory gut phenotype (315).Conversely, administration of specific probiotics may pre-vent islet autoimmunity under certain conditions (80),and clinical trials are ongoing to validate this hypothesis(249). Collectively, it appears that antibiotics and probi-otics may influence T1D development by altering the bal-ance of gut microbiota toward either a tolerogenic ornontolerogenic state, depending on constitution of theintestinal microflora at the time of administration (441).

A recent study by Wen et al. (469) sheds light on someof the mechanisms governing this delicate balancing act atthe intestinal level. NOD mice lacking the essential TLRsignaling component MyD88 were protected from diabetes.Furthermore, transferred CD4� T cells expressing the highlydiabetogenic TCR receptor BDC2.5 failed to expand in thepancreatic lymph node (LN) of MyD88�/� NOD mice. It waspostulated that abnormal recognition of certain gut bacteriamay be indispensible for diabetes development in the regu-lar NOD model, a pathway that is disrupted in the MyD88deletion variant. Administration of broad-spectrum antibiot-ics to MyD88�/� NOD mice reintroduced the potential todevelop disease, and germ-free MyD88�/� NOD miceshowed an increased risk to develop T1D compared withMyD88�/� NOD mice housed under SPF environment. Thelatter findings indicate that at least some members of themicrobial community in the intestine may protect from dia-betes independently of MyD88. Although the mechanisticpicture is far from complete, these results provide proof-of-principle in favor of a crucial role for intestinal immunehomeostasis in T1D prevention.

Another recently discovered bacterial risk factor maybe Mycobacterium avium subspecies paratuberculosis(MAP), which is the cause of paratuberculosis in rumi-nants. Of note, this bacterium is shed in the milk ofinfected cows and survives pasteurization. Clinically sig-nificant humoral responses to MAP antigens and wholecell lysates were detected in T1D patients (385). More-over, the presence of MAP was confirmed in T1D patientsby culture and was isolated from blood (367). It wassubsequently reported that a polymorphism within theSLC11A1 gene was associated with the presence of MAPDNA in T1D patients. Since MAP persists within macro-phages and is processed by dendritic cells, it was con-cluded that mutant forms of SLC11A1 may alter the pro-cessing or presentation of MAP antigens leading to dia-betogenic responses (109). We place the footnote that allthe studies referenced on this topic are from the samegroup documenting a relationship within the isolated Sar-dinian population. It remains to be seen whether thisfinding will be confirmed by others in different patientcohorts.

In conclusion, just as their viral counterparts, there issufficient indirect evidence warranting a focus on bacte-rial agents as potential triggers in T1D.

C. Other Environmental Triggers

We have pointed out the apparent vulnerability of theintestinal immune homeostasis in T1D, implying a poten-tial role for the bacterial flora. There are obviously manyother substances that may disturb physiological re-sponses at the site of the mucosal immune system, someof which have been under scrutiny as causal factors inT1D.

1. Cow’s milk

Cow’s milk, and in particular its albumin component,has been proposed to promote islet autoimmunity, be-cause cross-reactivity was found between serum antibod-ies to albumin and ICA-1 (p69), a beta-cell surface protein(218). Some studies suggested early introduction of cow’smilk as a predisposing factor as opposed to prolongedduration of breastfeeding during infancy (108, 233, 284).The emerging idea of cow’s milk ingestion as a contrib-uting parameter in T1D was soon challenged by reportsthat were unable to demonstrate any causal relevance (23,52, 311). This topic has since been subject to controversy,as even in NOD mice and the BB rat strain contradictingoutcomes were reported (110, 217, 267, 330). Support fora causal relationship in patients has predominantly comefrom the Finnish population. The TRIGR (Trial to ReduceIDDM in the Genetically at Risk) trial is particularly note-worthy in this respect. TRIGR will test whether hydro-lyzed infant formula compared with cow’s milk-based



on August 26, 2011


ownloaded from


formula decreases risk of developing T1D in children withincreased genetic susceptibility (1). Early data suggestthat hydrolyzed formula confers some protection and thatmaternal antibodies may offer breast-fed babies protec-tion against enteroviruses (10, 373). Another recent Finn-ish study suggests that the PTPN22 polymorphism affectsislet autoimmunity only if children are exposed to cow’smilk during early infancy, providing a possible explana-tion for the many contradictory findings (244). As itstands today, convincing arguments in favor of a patho-genic role for cow’s milk proteins in T1D are scarce. Aninteresting viewpoint put forward by Harrison and Hon-eyman (180) is that increased immunity to cow’s milkproteins, rather than being a unique risk factor, may re-flect a general impairment in mucosal immunity.

2. Wheat proteins

Albeit to a lesser extent than cow’s milk, wheat pro-teins and more specifically gluten have received someattention (85, 260, 457). A milk-free, wheat-based dietproduced higher diabetes frequency in BB rats and NODmice (38), and in the latter this diet induced a Th1-typecytokine bias in the gut (143). In T1D patients, increasedperipheral blood T-cell reactivity to wheat gluten wasmore frequently found than in healthy controls (232). Itwas further reported that timing of initial exposure tocereals in infancy may influence onset of islet autoimmu-nity in susceptible children (310, 497). This pattern isreminiscent of celiac disease (CD), an autoimmune disor-der triggered by the ingestion of gluten in susceptibleindividuals. Overlap between both diseases has long beenacknowledged, and the prevalence of (undiagnosed) CDwithin T1D and their relatives is higher than expected(199, 313, 414). Unlike in CD patients, however, there isno robust evidence to assume that gluten actually drivesthe initiation of autoimmunity in T1D (198). Analogous tocow’s milk proteins, immune reactivity against wheat pro-teins in T1D could be the general consequence of anaberrant mucosal response rather than the specific driverof islet autoimmunity.

3. Vitamin D

Some dietary components might increase T1D devel-opment, but a substantial body of knowledge points to-wards the protective properties of vitamin D in T1D.Vitamin D is not only taken up through nutrition, but alsosynthesized in the skin upon exposure to sunlight (re-viewed in Ref. 276). T1D has a clustered seasonal onset,and the monthly hours of sunshine and T1D incidence areinversely correlated (293). Vitamin D effectively inhibitsdendritic cell differentiation and immune activation. Vita-min D metabolite levels were lower in plasma from T1Dpatients around onset (247), and increased vitamin Dintake reduces incidence in mice (278) and humans (276).

The search for genetic polymorphisms in the vitamin Dreceptor (VDR) has yielded conflicting results on the cor-relation with T1D in the majority of studies (276), and arecent meta-analysis concluded that there was no suchevidence (170). Recent studies have confirmed the ab-sence of a relationship between VDR polymorphisms andbeta-cell autoimmunity (297), but did identify T1D-asso-ciated polymorphisms in genes encoding enzymes in-volved in vitamin D metabolism (29). Interestingly, inter-action was found between VDR and HLA alleles and ismediated by a vitamin D responsive element present inthe promoter region of the HLA-DRB1*0301 allele. It canbe reasonably envisioned that the absence of vitamin D inearly childhood may contribute to T1D development dueto poor expression of DRB1*0301 in the thymus (205). Insum, vitamin D can be considered an important environ-mental factor that is required for proper maintenance ofself-tolerance and protection against autoimmunity. Ofnote, therapeutic use of vitamin D metabolites in humansis hampered by its profound effects on calcium and bonemetabolism. Therefore, structural analogs have been de-signed that predominantly exert the immunomodulatoryeffects. Active treatment of T1D with vitamin D analogsremains a promising future avenue, and at-risk individualsshould probably avoid vitamin D deficiency (277).

A wide array of other dietary compounds and envi-ronmental triggers have been shown to affect diabetesdevelopment in animal models, and for some of thesesuch as omega-3 fatty acids (312), there is limited proof inhuman patients.

IV. TIMELINES FOR THE PATHOGENESIS OF

TYPE 1 DIABETES

Several timeline models have been put forward todepict the outcome of the interplay between the geneticand environmental factors. Figure 1 provides a visualoverview of some prominent hypotheses that have beenproposed over the years. The linear beta-cell decline hy-pothesis postulated by Eisenbarth in 1986 remains themost widely referenced benchmark model for T1D (124).According to this model (Fig. 1A), genetically susceptibleindividuals at some point encounter certain environmen-tal agents that initiate islet autoimmunity leading to alinear decay in beta-cell mass, development of autoanti-bodies, hyperglycemia, and eventually complete loss ofC-peptide. While this view provides a unifying explanationfor the sequence of events observed during the course ofT1D, it does not integrate factors contributing to thevariability along the time axis during the prediabeticphase. Some authors argue that disease progression inT1D is not a linear process, but rather proceeds at vari-able pace in individual patients (89). In the previous sec-tions, we have discussed the impact of specific genetic



on August 26, 2011


ownloaded from


polymorphisms on disease susceptibility. It seems accept-able that on the extreme end of the genetic risk spectrum(see, e.g., IPEX and APS-1, but perhaps also HLA haplo-type), the requirement for one or more environmentaltriggers is low, and patients will lose beta-cell mass lin-early regardless of such encounters (Fig. 1Bi). On theother end, subtle predisposing mutations may by them-

selves never lead to T1D (Fig. 1Bii), or require somedegree of environmental insult (e.g., IFIH1 and viral in-fection?) to culminate in hyperglycemia (Fig. 1B, iii andiv). Our group, in agreement with similar conclusions byBonifacio et al. (55), proposed a more detailed version ofthe nonlinear model depicting T1D as a “relapsing-remit-ting” disease (Fig. 1C) (462). Specifically, we suggest that

FIG. 1. Timelines for type 1 diabetes. A: model for linear beta-cell mass decay, as originally proposed by Eisenbarth (124). In the contextof genetic predisposition, an environmental trigger induces islet autoimmunity and beta-cell death leading to a sequence of prediabetic stagesand eventually clinical onset. B: it is widely acknowledged that the time window between initiation of autoimmunity and clinical onset is highlyvariable. Chatenoud and Bluestone (89) proposed some scenarios leading to the observed variability. Versatile interaction between geneticfactors and environmental challenges such as viral infections likely contribute to the fluctuations in beta-cell mass before onset. C: we haveintroduced the concept of T1D as a relapsing-remitting disease, dependent on cyclical disruption and restoration of the balance betweeneffector and Tregs and potentially counteracted by beta-cell proliferation (462). This model also provides a mechanistic rationale for itsvariable course. D: the fertile field hypothesis postulates the existence of a time window following viral infection during which at-riskindividuals may develop autoimmunity (463). Infection with a certain virus would temporarily create a fertile field. Whereas initial exposureto virus (e.g., via APC presentation; purple cells) will generate a normal antiviral response (green T cells), subsequent generation ofautoreactive cells (red T cells) may occur via cross-reactivity with viral antigens (molecular mimicry) or direct recognition of autoantigens(bystander activation). Bystander activation is thought to be mediated by APC that process and present self-antigens, with the potential toraise autoreactive T cells only in the presence of viral “danger” signals.



on August 26, 2011


ownloaded from


disequilibrium between autoreactive effector T cells andTregs could develop over time and eventually lead to adecline in beta-cell mass. Whereas the net balance shiftsto islet autoimmunity, this effect is temporarily counter-acted by the beta-cells’ proliferative response, perhapstranslating in a late transient phase of reduced insulin

requirement called “honeymoon phase” (also see below,Fig. 2). In an attempt to fit the role of infectious agentsinto this temporal T1D model, we introduced the “fertilefield” hypothesis (463). The fertile field (Fig. 1D) is de-scribed as a time window that follows virus infection, andwhich can vary depending on the type, anatomical loca-

FIG. 2. How T1D might arise. This figure represents the beta-cell mass or function (represented by the orange line) as well as the differentimmunological phases (columns with alphabetized tabs on top) that occur in the relevant anatomical sites (rows with numerical tabs on the right).Specific events will be referred to via alphanumerical coordinates in the following explanation. Once the orange line of beta-cell function falls intothe red zone, the individual is clinically diagnosed with type 1 diabetes. A complicated series of events precedes this and remains largely unnoticed.Initially, an unfortunate concurrence of genetic susceptibility (a1) and an environmental trigger (a2) sets an individual up for developing diabetesby causing two events. In the pancreas, beta-cells upregulate interferon (IFN)-� (b3) and subsequently MHC class I (c3). This exposes the beta-cellsto attack by autoreactive CD8 T cells with specificity for antigens in the pancreas (c3). Consequently, the released beta-cell antigens are picked upby resident antigen-presenting cells (APC) (c3) and transferred to the pancreas-draining lymph node (LN) (c2). Meanwhile in the periphery (c1), theenvironmental trigger has caused a metabolomic switch creating a proinflammatory environment that favors effector T-cell responses over Tregfunction. Beta-cell antigens presented in this proinflammatory context and with CD4 help (c2) initiate conversion of B cells into plasma cells (d2)and the appearance of insulin autoantibodies (seroconversion) (d1). Also, autoreactive CD8 T cells are stimulated to proliferate (d2) and migrateinto the pancreas (d3). The stress induced by this second wave of beta-cell killing (d3), which involves perforin, IFN-�, and tumor necrosis factor(TNF)-�, causes some beta-cells to halt insulin production (pseudoatrophy). The killing also causes the release of new beta-cell antigens that arepicked up by APCs, including migrated B cells (d3), and get shuttled to the pancreatic LN (d3-d2). This engages new specificities of CD4 and CD8T cells (e2) and B cells (e1) in a process called epitope spreading. A subsequent wave of beta-cell killing is therefore more severe and usually resultsin severe depletion of beta-cell function and mass (e3). Surprisingly, the autoimmune inflammation can also stimulate some beta-cell proliferation(f3), so that the beta-cell mass temporarily resurrects. Also, Tregs can sometimes overpower and dampen the effector response (f3). The fluctuationbetween destructive autoreactive responses and the alleviation by immune regulation and beta-cell proliferation possibly creates a nonstoprelapse-remitting profile of beta-cell mass (orange line). Eventually, the autoreactive response wins though, and T1D is diagnosed when only 10–30%of functional beta-cells remain. The remission after clinically diagnosed diabetes is termed the honeymoon phase (f3), a temporary state of relativeself-sufficient insulin production.



on August 26, 2011


ownloaded from


tion, and duration of the virus-induced inflammatory re-sponse. This fertile field would allow autoreactive T cellsto expand by mechanisms such as molecular mimicry orbystander activation and may lead to full-blown autoim-munity and T1D.

V. IMMUNE EVENTS IN TYPE 1 DIABETES

Several silent immune events occur before the clini-cal symptoms of type 1 diabetes become apparent. Mostimportantly, autoantibodies are produced and self-reac-tive lymphocytes become activated and infiltrate the pan-creas to destroy the insulin-producing beta-cells in theislets of Langerhans (56). This persistent, targeted de-struction may go undetected for many years, and the firstclinical symptoms only become apparent after a majorityof the beta-cells have been destroyed or rendered dys-functional, making the individual dependent on insulin forsurvival (Fig. 2). Therefore, high priority is given to thesearch for “biomarkers” as whistleblowers of an ongoingautoimmune response. We will highlight some importantimmunological events here. Additional information on im-mune cell cross-talk in T1D can be found elsewhere (243).

A. Metabolic Changes: Linking Cause to

Immune Events?

So far, seroconversion to autoantibody positivity isthe first detectable sign of an ongoing autoimmune re-sponse. But Oresic et al. (320) recently suggested thatmetabolic dysregulation precedes overt autoimmunity inT1D. Elevated serum concentrations of lysophosphatidyl-choline (lysoPC) precede the appearance of each isletautoantibody. In samples from the Finnish DIPP studycohort (235), characteristic changes in serum metaboliteswere found only in the children who later developed T1D.These changes included reduced serum succinate, PC,phospholipids, and ketoleucine, as well as elevated glu-tamic acid. These reactive lipid by-products are capableof activating proinflammatory molecules (286) that func-tion as a natural adjuvant for the immune system (215). Itremains unresolved whether these metabolic events trig-ger the initiation of the autoimmune period, or are justeasier to detect. Nevertheless, these findings create anopportunity for earlier diagnosis.

B. B Cells Produce Diabetes-Associated

Anti-islet Autoantibodies

The main autoantibodies in T1D are reactive to fourislet autoantigens (islet cell autoantibodies or ICA): insu-linoma-associated antigen-2 (I-A2, ICA512), insulin (microIAA or mIAA), glutamic acid decarboxylase 65 (GAD65),

and zinc transporter 8 (ZnT8) (470). The early presence ofautoantibodies implicates a role for antibody-producingplasma B cells in the initial immunological events. Indeed,B cells clearly contribute to the pathogenesis of humanT1D (334). In the NOD model, B cells infiltrate the pan-creas during the early stages of insulitis, and genetic orantibody-mediated ablation of B cells in NOD mice isprotective (197, 386). How, when, and where B cells con-tribute to diabetes onset is still debated and discussed infurther detail elsewhere (270, 397). In short, while Bcell-derived autoantibodies might reflect a prelude to au-toimmunity, B cells are likely active participants in theimmune response because of their capacity to presentantigen to diabetogenic CD4 and CD8 T cells.

C. Islet-Specific T Cells in the Periphery

A textbook definition of T1D could be “an autoim-mune disease in which CD4� and CD8� T cells infiltratethe islets of Langerhans, resulting in �-cell destruction”(361). Indeed, T cells are considered to be the final exec-utors of beta-cell destruction. This is evidenced by theprecipitation or prevention of diabetes by transfer orelimination of CD4 or CD8 T cells, respectively. CD8 Tcell-mediated beta-cell killing is likely a major mechanismof beta-cell destruction. CD8 T cells, found in insuliticlesions in NOD mice and in human (Fig. 3), can destroybeta cells upon activation via MHC class I expressed on betacells. Indeed, deficiency in MHC class I due to lack of beta-2microglobulin, or beta cell-restricted MHC class I deficiencyis sufficient to arrest diabetes development and preventbeta-cell destruction in NOD (176). Mechanistically, beta-cell destruction can involve the release by CD8 T cells ofcytolytic granules containing perforin and granzyme, orthrough Fas and Fas ligand-dependent interactions. CD4 Tcells mostly provide help to both B cells and CD8 T cells byproviding cytokines, such as IL-21, and a positive-feedbackloop via CD40L-CD40 interactions to antigen-presentingcells, culminating in a proficient autoreactive CD8 T-cellresponse. Their presence in insulitic lesions suggests a bea-con role and direct inflammatory properties.

The quest for autoreactive T-cell epitopes has beencentralized around the four proteins that are also themajor autoantibody targets: proinsulin (PI), GAD65, I-A2,and ZnT8 (114). So far, practical issues have made itdifficult to determine systematically when autoantibodiesand autoreactive T cells arise in the periphery relative toeach other. An excellent source on this topic is the reviewby Di Lorenzo et al. (114) that takes stock of the largeportfolio of epitopes identified to date. What is importantto know is that, in general, the main known CD4� T-cellepitopes are derived from GAD65, IA-2, and PI in bothhuman and mice. Additionally, smaller contributions fromheat shock protein (HSP)-60 and islet-specific glucose-6-



on August 26, 2011


ownloaded from


phosphatase catalytic subunit-related protein (IGRP), aswell as HSP70 in humans, complete the known range ofCD4 epitopes. On the other hand, autoreactive CD8epitopes in humans come mainly from preproinsulin sig-nal peptide (400), and to a lesser extent IA2, human isletamyloid polypeptide (IAPP) precursor protein (325),IGRP, cation efflux transporter ZnT8 (Slc30A8) (470), andGAD65 (326). In mice, the CD8 epitopes are mainly de-rived from IGRP and GAD65/67, and �30% from PI and10% from dystrophia myotonica kinase (DMK). Interest-ingly, CD8 peptide epitopes were identified from the hu-man preproinsulin signal peptide by elution from HLA-A2molecules (400). This has led to the concept that beta-cells unwittingly contribute to their own demise, becausethey are targeted even more when stimulated to producemore insulin. Together, these data have provided strongevidence that CD8 T-cell autoreactivity is associated withbeta-cell destruction in T1D in humans (337, 400).

D. Immunological Events in the Human Pancreas

We have recently reviewed the current status of ourknowledge on T1D histopathology (99). We concludedthat our fundamental understanding of what happens inthe pancreas is mostly based on old observations, datingback as far as the 1960s.

In 1965, Willy Gepts first identified lymphocytic infil-trations in pancreatic islets. These have since then be-come a hallmark feature of T1D termed “insulitis” (155).In their 1985 case report, Bottazzo et al. (56) showed

dramatic upregulation of MHC class I molecules and pin-pointed CD8� T cells as the predominant subset aroundthe islets. These observations were confirmed in a largecollection of samples by Foulis et al. (146). Several cor-nerstone insights were derived from subsequent analysesof these specimens. One such landmark paper docu-mented that MHC class I upregulation is a common prop-erty of diabetic islets around time of onset (145). Of note,significant levels of IFN-� were found exclusively in beta-cells. This fueled interest in a possible viral etiology,because IFN-� is typically induced as a cellular responseto viral infection (see above). Collectively, these findingscemented the idea of T1D being a multistep autoimmunedisease (Fig. 2).

It has been widely acknowledged that the autoim-mune process is quite variable, both between patients andwithin any given patient’s pancreas over time. Foulis es-timated the average number of lymphocytes associatedwith an inflamed islet to be �85. About 23% of insulin-positive islets were affected versus only 1% of the insulin-deficient islets (147). Japanese studies, using a biopsyapproach, found no (177) or very limited (2–62 mononu-clear cells in 3–33% of islets) (206) signs of insulitis. TheseJapanese studies suffer from a lack of accurate sampling,as only a very limited pancreatic region was assayed.

Willcox et al. (475) recently reexamined many of theolder findings using modern reagents on the samples col-lected by Foulis. CD8� T cells were confirmed to be acellular component of the insulitic lesions. Also, the num-bers of CD8s peak according to the degree of beta-cell

FIG. 3. The degree of pancreatic infiltration in T1D patients is limited compared with the insulitis in NOD mice around diabetes onset.Infiltration in or around the islets of Langerhans in the pancreas comprises CD8 T cells, CD4 T cells, but also B cells, macrophages and very fewdendritic cells. In this respect, i.e., the type of cell infiltrating the islets, pancreatic section from human and NOD mice correspond. But the degreeto which the pancreas and islets are inflamed, i.e., the number of infiltrating cells that can be found, is much more limited in humans than in NODmice. A: typical degree of infiltration in recent-onset type 1 diabetic pancreas in humans. Staining of human pancreas sections for insulin (in green)and CD8 T cells (in red) indicates that only low levels of CD8 T cells can be seen in the proximity of the islets. A similar low degree of infiltrationis observed for CD4 T cells and B cells in pancreatic sections of T1D patients (not shown). More information can be found in Reference 99. B: typicallevel of inflammation around diabetes onset in female NOD mice. Staining for insulin (in blue) and CD8 (in brown-red) shows a severe degree ofinfiltration by CD8 T cells in the islets. The type and degree of infiltration of CD4 T cells at this time point are usually similar (not shown).



on August 26, 2011


ownloaded from


decay and disappear from pseudoatrophic (i.e., insulin-deficient) islets. B cells were found to follow the samepattern. Macrophages, on the other hand, are present onlyin moderate numbers during the entire process. Interest-ingly, Tregs were detected in the islets of only one patient.This suggests that the default territory of Tregs is con-fined to the pancreatic lymph nodes or spleen. Alterna-tively, the local absence of Tregs is the reason why insu-litis could take place in the first place.

A logical consequence of beta-cell-directed autoim-munity is the induction of apoptosis. However, cross-sectional histology studies showed only very limited beta-cell apoptosis. The estimates ranged from none (296), 0.2beta-cells per islet (78), to 6% of all beta-cells (287). Theoverall low rate of beta-cell apoptosis at any given timelikely reflects a principal reason for the slow course ofT1D. A potential role for survivin, a molecule involved inprotection against apoptosis, was recently discovered inanimal models (211) and T1D patients (358).

Beta-cells have a robust capacity to regenerate byproliferation, likely in response to immune inflammation(11), at least in the NOD mouse. Similar evidence inhumans, using the Ki-67 proliferation marker, indicates no(79, 287), limited (204), or extensive levels (288) of beta-cell proliferation at various stages of disease. One of thepossible explanations of this discrepancy between spe-cies may be the relative lack of inflammation in humans.Characterization of T1D in the pancreas by immunohisto-chemistry has traditionally been performed on sectionsfrom recent-onset individuals. It would obviously be veryvaluable to have data on the early disease stages in pre-diabetic patients. Because the presence of insulitis andautoantibody status around onset closely correlate (203),serum screening of healthy individuals might identifyearly immunologic abnormalities around the islets in au-toantibody-positive, prediabetic patients. Only one studyhas systematically examined healthy autoantibody posi-tive donors. This study revealed a remarkably low inci-dence of insulitis in only 2 of 62 cases, in �10% of theislets (204). This outcome may reflect the very subtlenature of the diabetogenic process in most human cases,possibly reflecting its relapsing-remitting course (462).Alternatively, the data may illustrate the pronounced lob-ular progression, making it easy to miss out on the af-fected regions.

E. The Honeymoon Phase: Do Beta-Cells

Transiently Revive?

It is clear that patients are at the end stage of thedisease course at the time of clinical presentation. Some-where between 60 and 90% of the beta-cells are destroyedor dysfunctional. Exactly how many beta-cells remainaround onset is still unknown because of the lack of

accurate noninvasive imaging methods to quantify func-tional beta-cell mass in humans (98). In fact, assessmentof beta-cell mass via detection of insulin by immunohis-tochemistry may considerably underestimate the amountof beta-cells left. Studies in the NOD mouse have found asubstantial pool of nonfunctional, insulin-depleted beta-cells at the onset of hyperglycemia (394). Knowing howbig the pool of compromised beta-cells is is importantbecause it may be more achievable to revive than toregenerate beta-cells. An interesting window of opportu-nity in this respect is the honeymoon phase, a transientremission phase that occurs in up to 60% of T1D patientsafter initiation of insulin treatment (30) (Fig. 2). Thehoneymoon phase appears to occur more frequently withincreasing age at onset (2, 298) and can often last 3–6 mo,but may continue for 2 years. In this period, insulin dosescan be greatly reduced or even withdrawn completely (9,257). The mechanisms governing improved beta-cell func-tion are poorly understood, but it is thought that theconstant hyperglycemic stimulus exhausts the beta-cells.Initiation of insulin treatment relieves this stress factorand temporarily allows dysfunctional beta-cells to replen-ish their insulin content. Our group has studied the remis-sion phase from an immunological perspective. In a smallcohort of patients, we focused on antigen-specific cyto-kine responses from T cells in the peripheral blood (377).Surprisingly, we found lower FoxP3 levels in CD4�CD25�

Tregs and lower numbers of IL-10 producing cells inremission patients versus new-onset cases. In a limitedcross-sectional prospective study, we found that higherFoxP3 expression at diagnosis predicted worse glycemiccontrol, but higher mean numbers of IL-10 cells wereassociated with better future glucose control. Addition-ally, others reported lower IFN-� levels in remitters (12).Together, these data suggest that there may be an immu-nological component underlying the honeymoon phase,arguing for antigen-specific immunomodulation to curbautoimmunity soon after diagnosis. Finally, it remains tobe seen whether these findings in the peripheral bloodcorrelate with local events as they occur in the pancreas.This question is particularly difficult to address given theinaccessibility of patient samples.

F. The NOD Mouse: an Entirely Distinct Picture

The honeymoon phase does not occur in NOD mice.This already indicates the far more acute course of dis-ease in the NOD mouse. As a matter of fact, comparingthe typical histopathology from T1D patients and recent-onset NOD mice is like looking at two different diseases(Fig. 3).

Diabetes progression in the female NOD is charac-terized by nondestructive peri-insulitis, initially consistingof dendritic cells and macrophages, then followed by T



on August 26, 2011


ownloaded from


and B cells (209). This phase subsequently transgresses toa complete T-cell-mediated destruction of the beta-cellmass by 4–6 mo of age. The amount of inflammationeventually becomes so extensive that infiltrates developinto local tertiary lymphoid structures (254). These fea-tures are strikingly more aggressive than the subtle,chronic immune process in humans. Other distinctiveparameters of the NOD include the potent ability of beta-cells to proliferate in the presence of inflammation andthe considerable mass of nonfunctional beta-cells at on-set. It is not known whether this also occurs in humanT1D (394). Together, these differences may explain whythe many preventive and therapeutic successes in theNOD model translate so inefficiently to the clinic.

VI. IDENTIFICATION OF

PREDIABETIC INDIVIDUALS

Clinically, pre-T1D describes the period of ongoingislet beta-cell destruction in which sufficient beta-cellmass and functionality remains to preserve glucose ho-meostasis. Clinicians and researchers agree that silencingthe diabetogenic attack at very early stages of beta-celldestruction is the most desired time of treatment of T1D.This is because it might be possible to keep treatmentdoses low and therefore cause less side effects. Diagnos-tically, determining “ongoing beta-cell destruction” is dif-ficult, but individuals at high risk can nevertheless beidentified by a combination of tests (Table 1).

A. Genetic Screening

We have discussed above the genetic component ofT1D. The genetic susceptibility to T1D is determined by

genes related to immune function with the potential ex-ception of the insulin gene (434). The genetic susceptibil-ity component of T1D allows some targeting of primarypreventive care to family members of diagnosed T1Dpatients, but there is no complete inheritance of the dis-ease. Nevertheless, the risk for developing T1D comparedwith people with no family history is �10–15 timesgreater. Although �70% of individuals with T1D carrydefined risk-associated genotypes at the HLA locus, only3–7% of the carriers of such genetic risk markers developdiabetes (3).

Therefore, the focus is on relatives of diabetic indi-viduals, especially twins, and also on the genotype asso-ciated with the highest risk for T1D, namely, the DR3/4-DQ2/8 heterozygous genotype (15).

B. Diabetes-Associated Anti-islet Autoantibodies

The number of autoantibodies, rather than the spec-ificity of the autoantibody, is thought to be most predic-tive of progression to overt diabetes. In the BABYDIABstudy, almost no children expressing only one autoanti-body progress to diabetes (3, 383). On the other hand,almost all individuals expressing multiple diabetes-asso-ciated autoantibodies progress to diabetes with long-termfollow up (45, 353). Other cohort studies confirm thatexpression of two or more autoantibodies is associatedwith very high risk for type 1 diabetes and is rarelytransient (33, 208). However, autoantibodies can fluctuateor even completely disappear. An example can be foundin the American Diabetes Autoimmunity Study in theYoung (DAISY), which is biased toward young childrenbecause of following children from birth. About 95% of

TABLE 1. Available and future screening methods for T1D diagnosis

Screening Criterium Concerns

Genetic Relationship to diabetic individual (usually a first-degree relative)or identified to have high-risk HLA genotypes, like DR3/4-DQ2/8 (258)

Only 30-50% of T1D patients have the DR3/4-DQ2/8genotype

Only 50-80% of monozygous twins developdiabetes

Serologic Serum autoantibodies associated with islet beta-cells (ICA): I-A2,IAA or mIAA, GAD65, ZnT8 (219, 470)

Presence of autoantibodies can fluctuate

Metabolic First phase insulin production (by C-peptide) is low enough togive �50% risk for diabetes within the next 5 years and/orimpaired fasting glucose or impaired glucose tolerance (226,405)

Defective glucose control probably too late formost effective preventive treatment

T cell* Cellular immunoblot, Elispot and tetramers Reproducibility, standardization over multicenteris not yet complete enough

Metabolome* Elevated serum concentrations of lysoPCs Validation of specificity, access to massspectrometry

�-Cell mass* PET using radiolabeled IC2 Ab Validation of specificity

Based on results from the Diabetes Prevention Trial-1 (DPT-1) trial, it was determined that the combination of 1) the presence of two or moreautoantibodies, with 2) evidence of a defective first phase insulin response in 3) individuals that are first-degree relatives to a type 1 diabetes (T1D)patient, increased the risk of developing diabetes to over 75% within 5 yr (319). In younger individuals, the risk approaches 90% within 7 yr (319).However, other and earlier detection methods could increase detection and detect disease prodromes earlier. *Not in clinical use yet. ICA, islet cellautoantibodies; I-A2, insulinoma-associated antigen-2, ICA512; IAA, insulin autoantibodies; GAD65, glutamic acid decarboxylase 65; ZnT8, zinctransporter 8.



on August 26, 2011


ownloaded from


prediabetic children express anti-insulin autoantibodies,but at onset only 50% continue to express insulin autoan-tibodies. Similarly in the NOD mouse, insulin autoanti-bodies can be transient during progression to diabetes(488). So, multiple specificities of autoantibodies shouldbe checked. Screening for autoantibodies against ICA512,insulin, and GAD65 is accessible to primary care physi-cians, but ZnT8 is so far research or clinical trial use only.Further information can be found elsewhere (292).

C. Islet-Specific T-Cell Assays

Despite the strong contribution of T cells to diseasepathogenesis, the presence of autoreactive T cells is notroutinely assessed. An explanation for this is the poorknowledge of epitopes and the lack of robust assays todetect the low-affinity and/or low-frequency T cells (114).It is also uncertain that the peripheral blood providesaccess to the “right” T-cell pool. For instance, autoreac-tive T cells might selectively sequester to the islets orpancreatic lymph nodes, which makes consistent detec-tion in peripheral blood difficult. Moreover, the microen-vironment of the affected pancreas might alter the T-cellresponse. Recent data do indicate that cellular assaysperformed on peripheral blood have a high degree ofaccuracy (331, 389). T-cell assays can distinguish re-sponses from T1D patients and healthy control subjects,but reproducibility appears to be limited, especially forCD4 responses (183). In combination with autoantibodyassays however, the measurements reach a sensitivity of75% with 100% specificity in distinguishing diabetic pa-tients from nondiabetic controls (389). A possible expla-nation for this might be that T-cell responses to individualepitopes fluctuate over time, while autoreactivity as awhole persists in diabetic patients.

T-cell assays will become even more sensitive in thefuture, and systematic longitudinal data on epitope-spe-cific T-cell responses will become available. This willfacilitate efforts to determine the numbers and/or func-tion of autoreactive T cells and other immune markers inthe following settings.

1. Preventive screening

There is a vital need for biomarkers associated withT1D disease initiation and progression. Screening the T-cell self-specificities reveals the numbers, functional fea-tures, and specificities of the autoimmune reactivity andcan expose disease activity. 1) “Numbers” show whetherand how extensively autoreactive T cells are already in-volved. 2) “Functional features” are parameters such asproinflammatory versus immunoregulatory cytokines. Forinstance, HLA-DR4-matched subjects that contain IL-10-producing islet-reactive CD4� T cells develop disease 7years later than those individuals not containing such

cells (19). 3) “Specificities” indicate whether and howmany islet autoantigens for both CD4 and CD8 T cellshave become targeted, as well as whether secondaryepitopes (after epitope spreading) have emerged. Theoutcome of such T-cell profiling could allow tailoring ofstrategies aimed at redirecting immune responses. Theoptimal antigen for induction of Tregs can be selected, ormore systemic approaches can be added. In this respect,some treatments prevent diabetes when given very earlyin the autoimmune process, but can aggravate the auto-immune response when administered at the time of de-monstrable autoreactive CD8 T cells in the periphery(396, 448).

2. Transplantation

Islet transplantation and beta-cell replacement ther-apies provide unique opportunities to monitor recurrentautoimmune-mediated islet destruction within a relativelybrief time window. Tracking T-cell responses before andafter transplantation has shown that immune responses toislet allografts are associated with loss of beta-cell func-tion (362). Such T-cell tracking can also identify factorsinvolved in the attack (450) and distinguish the efficacy ofdifferent immune suppression protocols (361). Indeed, itis necessary to monitor closely the T-cell response in islettransplantation trials. For example, autoreactive T cellscan homeostatically expand after some immunosuppres-sion regimens used in islet transplantation (294), whichexplains why insulin independence is usually not sustain-able under the Edmonton protocol (392).

3. Primary and follow-up markers for trial outcome

We propose that the immunological efficacy andsafety of immune interventions is monitored by studyingchanges in T-cell autoreactivity. Assays are becomingavailable that allow sensitive, specific, and reproduciblemeasurement of the disappearance or functional silencingof islet-autoreactive T cells, or the appearance of regula-tory populations. T-cell ELISPOT analysis (ISL8Spot)showed that shifts, both in frequency and in immu-nodominance of CD8� T-cell responses, occur more rap-idly than the changes in autoantibody titers in human T1D(273). Recently, HLA-A2 tetramer technology was used toshow an increase in GAD65- and InsB-peptide reactiveCD8� T cells upon anti-CD3 treatment of T1D patients(86). These tools might also help identify patients that willexperience a honeymoon phase (377).

In conclusion, combining autoantibody and T-cell au-toreactivity readouts with a panel of biomarkers providesa more complete picture of disease activity. A betterevaluation of the patient’s response should benefit treat-ment outcome and safety.



on August 26, 2011


ownloaded from


D. Metabolomic Screening

Efficacious prevention of T1D will require detectionof the earliest events in the process. Autoimmunity islikely the predominant effector mechanism in T1D, but itis possibly not its primary cause. A recent interestingreport by Oresic et al. (320) (see sect. VA) showed thatelevated serum concentrations of lysophosphatidylcho-line precede the appearance of each islet autoantibody,and thus overt autoimmunity, in T1D. If these results arevalidated in other well-characterized cohorts, like theGerman BABYDIAB (4), the United States-based DAISY(37) and PANDA (84) studies, and the multinationalTEDDY study (172), metabolome screening could beadded to the screening panel to effectively identify predi-abetic individuals for preventive treatments.

E. Assessing �-Cell Mass

The number of islet beta-cells present at birth ismostly created by differentiation and proliferation of pan-creatic progenitor cells, in a process called neogenesis(61). After birth, only a small fraction of cycling beta-cellsremain capable of expanding. This might be sufficient tocompensate for increased insulin demands, but probablynot for regeneration after extensive tissue injury. Datafrom pathology samples indicate that only 10–30% of thebeta-cell mass is left in long-term T1D patients. It isunclear what amount of beta-cell mass remains at diabe-tes onset. The deficit in beta-cells necessitates measure-ment of beta-cell mass in vivo, but tools are still lacking todo this accurately and sensitively in a non- or mildlyinvasive way.

Noninvasive beta-cell imaging using modern diagnos-tic equipment could provide very high sensitivity in hu-mans and mice, but has some problems. First, single-cellresolution cannot yet be achieved to enable differentia-tion between scattered islets, single beta-cells, or sur-rounding tissue. Second, these imaging techniques rely ona specific molecular marker. Currently, the monoclonalIgM antibody IC2, which specifically binds to the surfaceof beta-cells, might be the only reliable marker for non-invasive imaging and quantification of native beta-cells(295, 378). Other candidates, like antibodies to ZnT-8 orthe ligand to the vesicular monoamine transporter type 2(VMAT-2), called dihydrotetrabenazine (DTBZ) and usedin clinical trial (NCT00771576), are not as specific (375).

Knowledge of the amount of beta-cell mass can helpwith decisions on the type of therapy and in treatmentfollow-up. For instance, drugs that stimulate beta-cellproliferation could be chosen when sufficient �-cell massremains, whereas very low remaining beta-cell masswould settle on islet transplantation, trans-differentiatingdrugs or stem cells. Such technology might also help

answering a basic question: What is the minimal fractionof the initial beta-cell mass required to preserve glucosehomeostasis? Lastly, early detection of beta-cell loss fol-lowing islet transplantation might help to adjust immunemodulation therapies in a timely and more targeted fash-ion.

VII. PREVENTIVE TRIALS

Successful prevention depends on 1) a good predic-tion/identification of at-risk individuals and 2) a very safeintervention that causes no harm in those individuals whowould have never developed T1D. Knowledge of the pri-mary cause of T1D might not be crucial, even at thepreventive stage. This statement is based on the fact thatimmune modulation appears to work in a variety of T1Dmodels and at different stages of the disease. However,many preventive trials are based on data from the NODmouse model which has improved our understanding ofdisease pathophysiology. A comprehensive analysis byShoda et al. (396) concluded that “some popular tenetsregarding NOD interventions were not confirmed: alltreatments do not prevent disease, treatment dose andtiming strongly influence efficacy, and several therapieshave successfully treated overtly diabetic mice.” So, thegood news is that some preventive strategies appear tohave a good chance to cure the disease, even during anadvanced status of beta-cell destruction. Examples ofsuccesful treatments in NOD mice are ATG, anti-CD3,hsp, and proinsulin DNA vaccine (91, 129, 398, 440). Ide-ally, the balance between therapeutic efficacy and diseasestage should be known prior to human trials.

A major problem with preventive trials is that it takesmany years before conclusions can be drawn. As can beseen in Table 2, preventive trials divide in two main classes.The first category mainly contains non-antigen-specific nu-tritional supplements: vitamin D3 (NCT00141986), hydro-lyzed cow’s milk (TRIGR; NCT00179777), and docosahexae-noic acid (DHA; NCT00333554). Other non-antigen-specificpreventive approaches had no or limited effect using keto-tifen (53), cyclosporine (83), nicotinamide (131, 151, 238), orcombinations thereof (340). For instance, cyclosporinecould delay onset, but not prevent T1D. The second categoryof preventive trials aims to induce antigen-specific oral tol-erance. During the disease process in animal models andhuman T1D, T-cell autoimmunity progressively spreads in-tra- and intermolecularly among �-cell autoantigens such asinsulin, GAD65, HSP, and IGRP (114, 322). Nevertheless,antigen-specific approaches use only one antigen (Table 2)to tolerize against multiple autoreactivities. The idea withthis approach is to induce “bystander suppression” by Tregsas follows. Tregs induced against one autoantigen prolifer-ate upon encountering their cognate autoantigen in the pan-creatic lymph node and maybe also in the islets. These Tregs



on August 26, 2011


ownloaded from


also secrete cytokines that can dampen immune function ormodulate APCs (421). The result is a localized antigen-spe-cific immunosuppression and resolution of the autoimmuneattack. However, translation to human T1D has been diffi-cult, because there are no functional biomarkers that wouldindicate that the correct amount, route, or antigen has beenused to achieve bystander suppression. For example, whilethe Diabetes Prevention Trial-1 (DPT-1) failed to demon-strate a benefit of oral (34, 401) or subcutaneous (120)insulin therapy in preventing T1D, post hoc subgroup anal-ysis indicated a potential delay in T1D subjects with highinsulin autoantibody titers who received oral insulin therapy(34, 319, 401, 406). A new clinical trial on the use of oralinsulin in relatives at risk is underway (NCT00419562). An-other example is the T1D Prediction and Prevention Study.This study could not demonstrate a beneficial effect of dailyintranasal insulin treatment in preventing or delaying diabe-tes, even when treatment began soon after seroconversion(301). In contrast, the Intranasal Insulin Trial did show thatintranasal insulin administration to individuals at high riskfor developing T1D increased antibodies and decreased T-cell responses to insulin, and more clinical trials are initi-ated.

VIII. NEW-ONSET TYPE 1 DIABETES TRIALS

Intervention trials are more affordable than preven-tion trials, because potential subjects are readily identi-fied and efficacy can be evaluated within a much shortertime frame. Current post-onset treatments are either sub-stitutive (e.g., the many formulations of insulin), palliative(long-term use of anti-inflammatory and/or immune sup-pressive agents), antigen specific (islet antigen-inducedTregs), or combinations thereof.

A. Antigen-Specific Intervention Trials in T1D

In general, the idea behind antigen-based therapies isto induce Treg responses (active tolerance) or anergizing/deleting pathogenic T cells (passive tolerance) withouthaving the side effects of long-term immune suppression.

Tolerizing against insulin or GAD65 has been rather ef-fective in experimental settings (128, 222, 432). This ap-proach has also led to success in T1D patients (87, 120,133)(Table 3). Future trials should also target other pep-tides (114), because some investigations in diabetes-prone mice suggest that ignored determinants of beta-cellantigens are a more optimal choice to inhibit late-stageautoimmune disease (317).

Several clinical intervention trials target insulin, be-cause it is the initiating antigen in the NOD model and itis also a major autoantigen in human T1D. A phase I trialhas confirmed the data in animal models that incompleteFreund’s adjuvant (IFA)-enhanced human insulin B-chainvaccination is safe and can induce insulin-specific Tregsfor up to 2 years after vaccination (299). A follow-up trialwill determine the effects on glycemic control (318). An-other approach uses a CpG-free proinsulin-based DNAplasmid vaccine BHT-3021 (Bayhill Therapeutics). Thisvaccine is designed to tolerize the immune system toproinsulin by combining DNA codons for an immunodom-inant peptide of insulin (440) and immunomodulatoryCpG oligonucleotides (189). Striking data in recent-onsetdiabetic NOD mice suggested that BHT-3021 induces pro-insulin-specific Tregs (440) that can act as bystander sup-pressors. In recent-onset patients, the vaccine can pre-serve C-peptide and reduce HbA1c (162). This demon-strates the increased efficacy of the BayHill plasmidcompared with the first generation of insulin B-expressingpCMV plasmids (94).

Another target for antigen-specific therapy is GAD65.GAD-Alum is an aluminum hydroxide (Alum) formulationof full-length recombinant human GAD65 (Diamyd Ther-apeutics). It was shown to be safe and to preserve resid-ual insulin secretion in subjects with late-onset autoim-mune diabetes of adulthood (LADA) (7). A subsequentphase II trial in recent-onset T1D showed significant pres-ervation of residual insulin secretion and a GAD-specificimmune response, both humoral and cell mediated (8).Currently, phase III studies are ongoing in Europe and theUnited States. The patients in all these trials were se-lected on the basis of elevated GAD65 autoantibodies.The formulation is crucial to Dyamid’s GAD drug because

TABLE 2. Prevention trials in T1D

Agent Target/Mechanism Development Phase, Clinical ID, Organizer

Vitamin D3 Treg induction Pilot, NCT00141986*, CDAOmega-3 fatty acids Anti-inflammatory Phase II, NCT00333554*, NIDDKHydrolyzed cow’s milk Abnormal handling of intact foreign protein TRIGR, Phase II, NCT00179777*, CHEOOral insulin Ag-specific tolerance (oral) Phase II, NCT00419562*, NIDDKNasal insulin Ag-specific tolerance (mucosal) Phase III, NCT00223613*, University of Turku

Phase II, NCT00850161*, CPEX PharmaPhase II, NCT00336674*, Melbourne Health

* More information is available at http://www.clinicaltrials.gov/. CDA, Canadian Diabetes Association; CHEO, Children’s Hospital of EasternOntario; NIDDK, National Institute of Diabetes and Digestive and Kidney Diseases.



on August 26, 2011


ownloaded from


1) adjuvant reduces the required quantity of antigen and2) aluminum salts preferentially induce a humoral ratherthan cellular immune response (255). Immune readoutsshow an increase in FoxP3 and transforming growth fac-tor-� in cells from GAD-Alum-treated patients comparedwith placebo after 15 mo (255).

Despite these promising studies, it is too early tojudge whether GAD65 and/or insulin are optimal targetantigens to induce Tregs that can modulate the course ofhuman T1D. Combination therapies with a short-termcourse of a suitable immune modulator are considered toenhance efficacy in recent-onset patients.

B. Non-antigen-Specific Intervention Trials in T1D

Since autoimmunity is the main effector mechanismin T1D, many intervention trials have used drug regimensto silence and/or modulate the immune response, prefer-ably without negative effects on Tregs (Table 4). These

immune-suppressive regimens will also prove valuable, ifnot critical, for the success of islet transplants and/or�-cell regenerative therapy.

The first immunosuppressive agent used in a placebo-controlled, double-blind clinical trial for T1D was cyclospor-ine A. Cyclosporine A inhibits calcineurin, which is respon-sible for the activation of IL-2 transcription. The lack of IL-2and other cytokines reduces the function of effector T cells,but unfortunately also of Tregs. Cyclosporine treatment in-duced remission of T1D, but its chronic use was suspendedbecause of unacceptable side effects (21, 59, 210). However,this limited success indicated that immunosuppression canreduce the autoimmune inflammation in T1D.

DiaPep277 was originally used in an antigen-specifictherapy. The idea was that this peptide from HSP60 be-comes an autoantigen in T1D because of cross-reactivity(127, 129). More recent insights indicate that Diap277 is asystemic modulator: HSP60 induces Treg via the Toll-likereceptor (TLR)-2 (417, 490). A phase II trial showed that

TABLE 3. Intervention trials and Ag-specific monotherapy

Agent Target or Mechanism Phase, ID, Organizer Details Reference Nos.

Diap277 Induction of Treg viaTLRs

Phase III (adults) Phase I: preserved C-peptideout to 18 mo

351

NCT00615264 Phase II: no effect in T1Dadults or children

240

NCT00644501 Phase III: recruiting 381Andromeda Biotech

Ins B in IFA Tolerance vaccination toinsulin B chain

Phase I/II Ongoing 111, 299NCT00057499ITN

NBI-6024 Tolerance vaccination toinsulin

Phase I /II Phase I: shift Th1 toprotective Th2

14

(APL of insulin) NCT00873561 Phase II: no effect on residual�-cell mass and insulinneeds

466

Neurocrine

BHT-3021 Tolerance vaccination toinsulin

Phase I/II Reduced insulin autoAb titers,preserved C-peptide andreduced HbA1c

162

(Proinsulin vaccine) NCT00453375 No adverse effectsBayHill Therapeutics

GAD65-Alum Tolerance to GAD65,skewing Th1 to Th2

Phase II (LADA) Preservation of residualinsulin secretion, GAD-specific humoral andcellular immune response,but no protective effect oftreatment �6 mo afterdiagnosis, no change infasting C-peptide after 15mo

7

NCT00456027 8

Diamyd 256

Phase II (T1D) OngoingNCT00435981Phase III Ongoing, EuropeNCT00723411Phase III Ongoing, USA (DIAPREVENT)NCT00751842

Islet autoAg-derived peptideseluted from human HLAclass II

Regulated immuneresponse

Phase I N/A 400

DVDC

APL, altered peptide ligand; IFA, incomplete Freund’s adjuvant; ITN, immune tolerance network, DVDC, Diabetes Vaccine Development Center;PI, proinsulin.



on August 26, 2011


ownloaded from


DiaPep277 preserved C-peptide up to 18 mo in adult new-onset T1D patients (350). Beta-cell preservation was associ-ated with IL-10 production before treatment and a decreasein autoantigen-specific T-cell proliferation after treatment

(200). But so far, no DiaPep277-specific Tregs have beencharacterized in mice or humans. However, while a phase IIIstudy is ongoing in adults, no treatment effect was observedin children with T1D (240).

TABLE 4. Intervention trials and Ag-nonspecific monotherapy

Agent Target/Mechanism Phase, ID, Organizer Details Reference Nos.

Calmette-Guerin (BCG) Hygiene hypothesis Phase I N/A new trial is recruiting 13NCT00607230 Trial 10 yr ago: no effectMGH

Diazoxide Counteract chronic overstimulationof �-cells

Phase IV No effect on �-cell function 46, 168, 321NCT00131755 Side effects (eliminated by lowering

dose)V. Grill, MDIngested IFN-� T1D is a type 1 IFN

immunodeficiency syndromePhase II Safe 66, 369NCT00005665 Lower dose is more efficacious in

preserving C-peptideNCCR Need better dosing

Cyclosporin A Immune suppression Completed Remission successful duringtreatment, but severe side effects

21, 59

Anti-CD3 (hOKT3)g1(Ala-Ala)

T-cell immunomodulation and Treggeneration by FcR-nonbindinganti-CD3 mAb

Phase I/II Remission out to 24 mo 182, 184-186

MGA031 Completed 36 mo positive effects on C-peptidelevels

Teplizumab ITN (Herold)(US based) Phase I/II Test: single 14-day course of Ab

administered to recent-onsetpatients 4-12 mo postdiagnosis

NCT00378508JDRF/NIDDKPhase II AbATE Multidose: 2 courses of Ab

administered 1 yr apartNCT00129259 Read-out: C-peptide levelsITN (Herold)Phase II/III

Protégé

Read out: insulin requirement,HbA1c

NCT00385697Macrogenics

Anti-CD3 mAb(ChAglyCD3)

T-cell immunomodulation and Treggeneration by FcR-nonbindinganti-CD3 mAb

Phase II 6-day Tx: better maintenance ofC-peptide levels, reduced insulinrequirement out to 18 mo

91, 228

TRX4 Completed 2005 During Tx: headaches, nausea, bodyaches, and other flu-likesymptoms; 6wk postTx: flare-upof EBV, sore throat, and swollenglands (temporarily and mild)

Otelixizumab

(Europe based)Phase II TTEDD Test: multidose regimenNCT00451321JDRF/TolerRxPhase III

DEFEND-1

Test: 8-day treatment

NCT00678886 RecruitingJDRF/TolerRxPhase I Test: subcutaneous administrationNCT00946257GSK

ATG Thymoglobulin/Atgam T-cell depletion, generate Tregpopulation

Phase II START Could cause cytokine releasesyndrome

398

NCT00515099NIAID/ITN

Anti-CD20 mAb B-cell depletion Phase II/III C-peptide AUC in MMTT better �12mo

197, 334

Rituximab NCT00279305 From 3-6 mo onwards: C-peptideAUC rate of decline similar toplacebo

TrialNetCTLA4-Ig Abatacept,

BelataceptCostimulation blockade Phase II N/A, recruiting 50, 239

NCT00505375NIDDK



on August 26, 2011


ownloaded from


TABLE 4.—Continued


IL-1 antagonists Anakinra Anti-inflammatory and improve �-cell survival

Phase II/III

(AIDA)

N/A recruiting 116, 335

NCT00711503 http://www.aidastudy.orgSTENO/JDRFPhase I/II N/A, although completedNCT00645840UTSMC

Canakinumab Phase II N/A, pendingNCT00947427NIDDK

Xoma 052 Phase I/II N/A, pendingNCT00998699Zurich University

Rilonacept IL-1 beta trap Phase II N/A, pending 427Arcalyst NCT00962026

UTSMCTNF-� blockade

etanerceptAnti-inflammatory Phase I/II Lower HbA1C and insulin need,

increased C-peptide AUC274

NCT00730392 No adverse effects, but TNFinhibitors have been associatedwith an increased risk ofreactivation of latent tuberculosis

31, 94, 480

AmgenGCSF (granulocyte colony

stimulating factor)Neulasta

Enhance T regulatory cell numbers Phase I/II N/A, recruitingNCT00662519JDRF/NIH

Byetta, exendin-4, orexenatide

GLP-1 analog, stimulates insulinsecretion

Phase IV N/A, ongoing. 171, 370, 478,479NCT00456300 In T2D: short half-life, GI side

effects, development ofantibodies

Baylor CollegeLiraglutide GLP-1 analog, stimulates insulin

secretionPhase II/III N/A, ongoing 76, 104, 290,

341NCT00993720 In T2D: HbA1c lower, fewer sideeffects than exenatide

Hvidovre U.Hospital

Sitagliptin Inhibitor of DPP-4, GLP-1-degrading enzyme

Phase I N/A, recruiting (immune function) 230, 231NCT00813228NIDDKPhase IV N/A, recruiting (glucose effects)NCT00978796BDC

Epidermal growth factor(EGF)

Islet neogenesis, �-cellproliferation

Phase II (E1-INT) Daytime insulin reduced 35-75% in 3of 4 patients. Reductions ofdaytime insulin usage evidentafter 28-day treatment and peak1-2 mo posttreatment (stable BGcontrol)

Company

website

TransitionTherapeutics

Pioglitazone PPAR-� stimulation Phase I N/A, recruiting (testing course ofT1D)

44, 224, 491

NCT00545857Stony Brook Univ

INGAP peptides Islet cell regeneration Phase II Increase in C-peptide secretion inT1DM patients

122, 338, 366

NCT00995540 HbA1c decreased (�0.4%)Exsulin

Corporation

ATG, anti-thymocyte globulin; AUC, area under curve (for C-peptide levels in glucose tolerance test); BDC, Barbara Davis Center; DCCT, DiabetesControl and Complications Trial; DENIS, Deutsche Nicotinamide Intervention Study; EBV, Epstein-Barr virus, a herpesvirus that causes mononu-cleosis; ENDIT, European Nicotinamide Diabetes Intervention Trial; GSK, GlaxoSmithKline; ITN, Immune Tolerance Network; mAb, monoclonalantibody; MGH, Massachusetts General Hospital; MMF, Mycophenolate mofetil; NCCR, National Center for Research Resources; NIAID, NationalInstitute of Allergy and Infectious Diseases; UTSMC, University of Texas Southwestern Medical Center.



on August 26, 2011


ownloaded from


An important class of biological immune modulatorscomprises antibodies that target receptors on T cells. Forexample, FcR-non-binding anti-CD3 monoclonal antibod-ies (mAbs) have shown the most promising results so farin T1D therapy. Anti-CD3 mAb acts at various levels. Itcauses a short-term internalization of the TCR-CD3 com-plex that makes the cell “blind” to antigen (88). Also, italters TCR-mediated signal transduction so that anergy orapoptosis is induced preferentially in activated Th1 cells(91). The apoptosis is partially mediated by CD95-CD95Linteractions with neighboring T cells. This might explainwhy effector T-cell death is most dramatic where theT-cell density is highest, i.e., at the site of inflammation.Moreover, anti-CD3 treatment also results in Treg devel-opment (486). It is thought that the Tregs may protectagainst damage by effector T cells long after the drug hasbeen eliminated from the body. To obtain all these effects,optimal dosage is crucial. Too little mAb causes insuffi-cient modulation and Treg generation, whereas too muchmAb could lead to stimulation of the effector T cells andcytokine release. Multiple clinical trials have been initi-ated and were based on two different antibodies, bothfully humanized IgG1, nonmitogenic and specific to hu-man CD3 (Table 4): Teplizumab (United States trials) andTRX4/Otelixizumab (European trials) (54, 477). Tepli-zumab halted progression of recent-onset T1D for morethan 1 year in most patients (phase II). Three years aftertreatment, the patients continued to have better preser-vation of C-peptide levels and a lower use of insulincompared with control groups (185, 186). TRX4/Otelixi-zumab also preserved beta-cell function very efficientlyand decreased the insulin requirements drastically, even18 mo after single course treatment (228). However, noneof these treatments achieved euglycemia. The Europeanstudy also revealed two side effects. First, anti-idiotypicantibodies were detected 2–3 wk after injection of thedrug. This should only become a problem when repeatedtreatment is needed. Second, there was reactivation ofEpstein-Barr virus, but this was transient, self-limiting,and isolated (227).

Other approaches using polyclonal anti-T cell anti-body (ALS) (314) and anti-thymocyte globulin (ATG)might be able to temporarily eliminate a larger proportionof T cells from the bloodstream. In NOD mice, murineATG can prevent diabetes in a late stage and can induceTregs (398). It is unclear whether ATG will be as efficientas anti-CD3 (327). Treatment of T1D patients with rATG(ATG-Fresenius) prolonged the honeymoon period andimproved the stimulated C-peptide levels up to 12 mo intothe study (379). Consequently, ATG monotherapy is nowtested in a phase II trial, the Study of Thymoglobulin toArrest T1D (START). But ATG treatment also carriesrisks. ATG can cause cytokine release syndrome andmaybe even a lymphopenia-induced outgrowth of autore-active T cells, as was shown for other depletion-based

immunosuppression (294, 392). Combination treatmentwith equine ATG and prednisone, a steroid that can coun-teract the cytokine release syndrome, led to a prolongedhoneymoon phase in new-onset T1D (125). But mostpromising in the ATG trials was that some subjects wentinto complete remission and were insulin independent forat least 1 mo.

An immunomodulatory drug adopted from the trans-plantation arena is the CD20-specific mAb rituximab (Rit-uxan; Genentech and Biogen Idec). Rituximab aims todeplete a potentially potent antigen-presenting populationof B cells without affecting long-lived antibody-producingplasma cells (197). Data from NOD mice clearly show thatB cells are required for T1D (388). A phase II clinical trialshowed some preservation of C-peptide levels for 3–6 mo(334), showing that B cells also contribute to pathogene-sis in human T1D. However, the efficacy is small com-pared with anti-CD3 treatments (186, 228). Perhaps Ocre-lizumab, a humanized anti-CD20 antibody successfullyused in RA (154), might increase efficacy of B-cell deple-tion in T1D.

Anti-CD25 mAbs such as basiliximab (chimericmouse-human monoclonal antibody) and daclizumab (hu-manized IgG1 mAb) do not cause a cytokine release syn-drome. Therefore, they are increasingly being used inplace of ATG as an induction therapy. However, in T1D,anti-CD25 mAbs are only used in combination therapies(see below).

Another class of targets consists of costimulatorymolecules. CTLA-4-Ig (Abatacept), CTLA-4 fused to animmunoglobulin chain, interferes with costimulation of Tcells. Classically, CD28/B7 interactions mediate costimu-lation and significantly enhance peripheral T-cell re-sponses. In contrast, CTLA-4, interacting with the sameB7 molecules, dampens T-cell activity. So, CTLA-4-Iglikely mediates its profound effects by preventing positivecostimulation of CD28 by B7 during activation. This re-sults in limited clonal expansion, induction of passive celldeath, and IDO production in APCs (reviewed in Refs. 50,376). The safety profile of CTLA-4-Ig treatment might bebetter than other immunosuppressive agents, becauseCTLA-4-Ig does not deplete T cells. However, because ofthe role of CD28 in Treg development and survival (376),CTLA-4-Ig may negatively affect Tregs. That said, CTLA-4-Ig therapy did not affect Tregs in renal transplantation(140). CTLA-4-Ig monotherapy is currently in a phase IIclinical trial (data not published). Moreover, a high-affin-ity variant of CTLA4-Ig (LEA29Y, belatacept) (239) isbeing tested in islet transplantation in a phase I/II trial(NCT00501709). And the LEA29Y Emory Edmonton Pro-tocol (LEEP, NCT00468403) phase II clinical trial com-bines CTLA-4-Ig with daclizumab or basiliximab (againstacute transplant rejection) and mycophenolate mofetil(maintenance immunosuppressive therapy).



on August 26, 2011


ownloaded from


Manipulation of cytokines has seen a revival recently.Both the cytokines that affect T-cell responses (e.g., IL-2,IL-15) and the cytokines that play a role in inflammationand beta-cell death are considered as targets. It is knownthat IL-1 is selectively cytotoxic to rodent and humanbeta-cells in vitro. Anti-IL-1 therapies can reduce diabetesincidence in animal prevention models (41, 268, 279). Asin RA, multiple clinical trials assess whether anti-IL1 ther-apy may be useful in the treatment of T1D. Results froma completed phase I/II trial using IL-1RA (anakinra, Kine-ret by Amgen) (116) in newly diagnosed T1D have notbeen released yet (335). In an ongoing phase II/III, pa-tients will inject themselves with Anakinra once a day fortwo years, which requires a big commitment on their part.Also, three phase II trials are pending. One will testCanakinumab, a fully human anti-IL-1� monoclonal anti-body, in new-onset T1D patients. Another one will test theanti-IL-1 mAb Xoma52 in established (�2 year), well-controlled T1D. The RID-1 study will test Rilonacept, adimeric fusion protein that acts as cytokine trap for IL-1�,after satisfactory safety data in gouty arthritis (427).

The cytokine tumor necrosis factor (TNF)-� is a“master regulator” of the inflammatory response in manyorgan systems (139). TNF-� antagonists, such as etaner-cept (a soluble TNF-receptor) and infliximab (an anti-body), are already used with success in RA. In T1D, phaseII trial data showed that treatment with etanercept low-ered HbA1C and increased endogenous insulin produc-tion, indicative of the preservation of beta-cell function(274). However, the story is not that simple. TNF-� mightplay a dual role in T1D. For instance, TNF and a TNFR2agonist can selectively kill human autoreactive CD8 Tcells (31). In animal models, TNF-� propels or lessens thediabetogenic response early or late in the T1D process,respectively (94, 207, 241, 480). Adding to the confusionare some clinical case reports documenting the develop-ment of T1D in arthritis (JIA or RA) patients followingetanercept treatment (49, 60, 418), but also the resolutionof T1D in patients requiring anti-TNF-� therapy for RA(18). These opposing outcomes need to be clarified soon.Of note, Bacille Calmette-Guerin (BCG) vaccinationraises the systemic levels of TNF-� and was used in anunsuccessful interventional trial. A follow up study willdetermine better timing and dosage (13, 130).

Others have hypothesized that autoimmunity is dueto lack of type I IFN. Type I IFNs can counteract type IIIFN, which is likely a central factor in autoimmune in-flammation (65). Clinical trials have so far shown thatlow-dose ingested rhIFN-� is safe and more efficacious atpreserving C-peptide levels compared with high doses(66, 369). Mechanistically, ingested rhIFN-� reducedTNF-� levels in MS patients, indicating a link with TNF-�blockade therapy (67). However, this whole hypothesis iscontroversial. Others suggest that activation of TLRs bydouble-stranded RNA or poly I:C (viral mimic) through

induction of IFN-� may activate or accelerate immune-mediated beta-cell destruction (113).

Granulocyte colony stimulating factor (GCSF), a neu-trophil mobilizing agent, prevents diabetes in NOD miceby induction of both tolerogenic dendritic cells and Tregs(216, 371). Safety and C-peptide preservation upon GCSFtherapy (Neulasta) are currently being tested in a phaseI/II clinical trial. A combination trial with ATG is alsounderway (see below).

Another angle of research aims to delay the demise ofbeta-cells by reducing the amount of insulin they secrete.This is anticipated to reduce beta-cell stress associatedwith the diabetic state and might also reduce the pre-sented autoantigens, such as (pro)insulin. Diazoxide, anATP-sensitive potassium channel opener, showed somepreservation of residual insulin production in recent-on-set T1D patients, but also substantial side effects (321). Arecent phase IV trial indicated that doses that do notcause side effects are also inefficacious at preservingbeta-cell function (168).

C. Cell-Based Tolerogenic Therapy

Propelled by evidence from animal models, cell-based tolerogenic immunotherapy has gained momentum(Table 5). The idea is to compensate a presumed defi-ciency by transferring cell types with immunomodulatorycapacity.

Cellular immunotherapy with autologous Treg repre-sents an attractive and feasible approach for curing T1D (69,71). This was first indicated by the reestablished immunetolerance after adoptive transfer of autoantigen-specificTreg or Tr1 into NOD mice (422, 424, 485). Planned clinicaltrials aim to treat T1D by isolating the patient’s Tregs forexpansion outside the body and reinfusion of larger num-bers (167). Experts in the field acknowledge the numeroustechnical problems that are likely to be encountered: a bonafide set of markers for “pure” human Tregs [currently set atCD4�CD127low/minusCD25� (343, 420)], a low frequency inthe circulation (�5–7% of CD4� T cells)(28), the number ofcells to be transferred, the frequency of transfers, in vitroexpansion methods, the survival of these cells in vivo, cor-rect homing to the target tissue, the inability to eliminate thetransferred cells, and instability of the regulatory function(300, 496). Therefore, the field is divided into believers andnonbelievers. Some see cell-based tolerogenic therapy as aviable, routine clinical approach. Others prefer to targetbeta-cell antigens in conjunction with small molecules ormAb to augment islet-specific immunoregulatory cells di-rectly in vivo.

Immunoregulatory dendritic cells (iDC) can also pre-vent diabetes in NOD mice (179, 261). In current clinicaltrials in Pittsburgh, autologous monocyte-derived DCs aretreated ex vivo with antisense phosphorothioate-modified



on August 26, 2011


ownloaded from


oligonucleotides that target the primary transcripts of theCD40, CD80, and CD86 costimulatory molecules. Oneconcern is that instability of the antisense knockdownmight allow reexpression of the targeted molecules. Also,the therapy rationale promotes the production of IL-7 byiDC as survival factor for Treg. But IL-7 is also importantfor naive and memory T cells (118, 179, 415), so presum-ably also for autoreactive T cells. Of note, B cells with aregulatory phenotype were augmented in some of thepatients who received the immunomodulatory DCs(Massimo Trucco, personal communication).

D. Replacing Beta-Cell Shortage

The autoimmune attack in T1D reduces beta-cell massand/or function to a critical point at which clinical diabetesbecomes apparent (Figs. 1, A and B, and 2). Some treat-ments aim to directly compensate for this loss in beta-cells.In its simplest and most commonly used form, this is doneby insulin injections. Other treatments increase beta-cellmass or function by exploiting the regenerative capacity thatbeta-cells display in response to the autoimmune attack(394) or nonautoimmune stimuli (11, 306). Five angles areconsidered to replace beta-cell shortage (Table 4, bottom):1) stimulation of insulin secretion, 2) islet neogenesis fromprogenitor cells, 3) islet regeneration from existing beta-cellmass, 4) islet transplantation, and 5) transplantation of stemcells. So far, all methods of beta-cell mass restoration en-counter immunological problems. This is because the newlygenerated or transplanted beta-cells continue to provideantigens that are equally susceptible to autoimmune attacks(462). Islet transplantation and allogeneic stem cell ap-proaches have the additional problem of an alloimmuneresponse. This is why most islet grafts are lost within thefirst 4–5 years, despite the use of immunosuppressive cock-tails (178, 372). These approaches will benefit from im-

proved strategies to control autoimmunity long-term, likelyvia combination therapies.

1. Stimulation of insulin secretion

Adapted from T2D treatment, analogs of the incretinhormone glucagon-like peptide-1 (GLP-1) (341) with suffi-cient half-life stimulate insulin secretion in the remainingbeta-cells. Examples are Exenatide (Byetta by Amylin Phar-maceuticals and Eli Lilly), a synthetic version of the ex-endin-4 hormone found in the saliva of the Gila monster, andLiraglutide (Victoza by NovoNordisk), a GLP-1 analog thatbinds to albumin for slow release (76, 104). GLP-1 receptoractivation modestly delayed diabetes onset in NOD mice(171). Mechanistically, exenatide not only stimulates insulinsecretion, but might also enhance beta-cell replication andneogenesis in rats (478), protect against IFN-�-mediatedbeta-cell death (102), and increase Treg frequency in NODmice (479). However, the effects on beta-cell mass and theimmune system are controversial. Exenatide monotreat-ment is already in phase IV trial, but a combination trial ofexenatide with daclizumab yielded disappointing results(see below and Ref. 370). Liraglutide on the other handsupports the engraftment and function of syngeneic islettransplants in NOD mice (290), which has led a phase II/IIItrial testing Liraglutide monotherapy.

Another approach to increase insulin secretion is toslow down the physiological degradation of GLP-1. Sitaglip-tin inhibits the enzyme dipeptidyl peptidase-4 (DPP-4) that isresponsible for the destruction of GLP-1. Sitagliptin canprolong islet graft survival (230, 231) and can partially re-verse diabetes in NOD mice (408). Clinical trials will exam-ine sitagliptin, either as stand-alone treatment or in combi-nation with islet transplantation or antigen-specific therapy(see below). A phase I trial was also initiated recently tostudy the impact of DPP-4 inhibitors on the immune system(280).

TABLE 5. Intervention trials and cell-based monotherapy


T regulatory cell therapy Augment Treg numbers Phase I (pending) Risk of conversion Treg to Teff (496) 343, 344, 422Adult human mesenchymal

stem cells (Prochymal)Reduce inflammation and assist

in tissue repairPhase II

NCT00690066Ongoing

OsirisAutologous umbilical cord

blood infusionEnhance T regulatory cell

numbersPhase I/II No effect on C-peptide preservationFlorida Univ, JDRF No infusion-related adverse eventsNCT00305344

Hematopoietic stem cells Immune resetting effect Phase I/II Some insulin independence, C-peptide AUC increased, HbA1Clower

100, 101, 458

NCT00315133 Bilateral nosocomial pneumonia, lateendocrine dysfunction,oligospermia, but no mortality

Sao Paulo UnivAutologous dendritic cell

therapyTolerogenic vaccine Phase I Ongoing 179, 261

NCT00445913Pittsburgh Univ



on August 26, 2011


ownloaded from


2. Beta-cell neogenesis

In animal models, gastrin stimulates beta-cell neo-genesis without increasing proliferation and hypertro-phy, and without reducing beta-cell death (364, 365).Also, the beta-cell mitogenic properties of epidermalgrowth factor (EGF) can aid in restoring beta-cell mass.In a phase II trial, the EGF analog E1-INT (TransitionTherapeutics) reduced daytime insulin usage by 35–75%and helped maintain stable blood glucose control asmeasured by HbA1c in some of the T1D patients. Ac-cording to the company, the results were similar inNOD mice. However, others have shown that gastrinand EGF need to be combined to increase beta-cellmass and reverse hyperglycemia in recent-onset NODmice (352, 412, 413). A phase I study (E1 � G1 INT,NCT00853151) is ongoing.

3. Islet cell regeneration

Islet neogenesis associated protein (INGAP) peptidetherapy induces islet cell regeneration from progenitor cellsresiding in the pancreas in a manner that recapitulates isletdevelopment during normal embryogenesis (348). INGAPpeptide can increase beta-cell mass and reverse hyperglyce-mia in animal models (338, 366). In phase I and II trials,INGAP peptide injections are safe and can increase C-pep-tide secretion, but hardly decrease HbA1c levels (122). Adose optimizing trial is ongoing.

4. Islet transplantation

In most developed countries, pancreas transplantationis the only accepted procedure to achieve normoglycemia(178). The Edmonton case series demonstrated that approx-imately two-thirds of the recipients enjoy insulin indepen-dence for 1 year after receiving their final islet infusion (391).The long-term results are unfortunately less encouraging.Islet function decreases over time so that by 5 years post-transplant �10% of the recipients remain insulin indepen-dent (178, 372). Why is this? In short, allosensitizationagainst transplants from multiple donors can be controlledusing immune suppression, but the current regimens mightinadvertently propel autoreactivity in the long term and evennegatively affect beta-cell functionality. In the original Edm-onton protocol, patients infused with pancreatic islets frommultiple cadaveric donors simultaneously received immunesuppression in the form of a humanized anti-CD25 mAb(daclizumab) and continuous administration of low-doserapamycin (sirolimus), which inhibits the response to IL-2,and FK-506 (tacrolimus), a calcineurin inhibitor blockingIL-2 production. However, this regimen was shown to causelymphopenia and an elevation of the levels of homeostaticcytokines that drive the expansion of autoreactive CD8 Tcells (294, 447). Consequently, the Edmonton protocol hasbeen modified in several ways. For example, ATG plus et-

anercept (TNF-� blockade) and maintenance immunosup-pression using cyclosporine and everolimus (a derivative ofsirolimus) rendered five of six recipients insulin indepen-dent at 1 yr, and four of six for an additional 3 yr (40, 391).A recent phase I/II trial will assess whether treatment withanti-CD3 mAb, sirolimus, and low-dose tacrolimus can pre-vent islet transplant rejection. However, animal studies havealready shown that anti-CD3 no longer induces tolerancewhen tacrolimus was coadministered, even though it con-tinues to immune suppress (90). Another phase II trialwill test efficacy of a steroid-free, calcineurin inhibitor-free immunosuppression protocol for islet transplanta-tion (NCT00315627), based on sirolimus, MMF, andCampath-1. Assessment of certain (auto)immune pa-rameters before transplantation might also increase thesuccess rate of islet transplantation. Indeed, T1D pa-tients receiving intraportal islet cells under ATG-tacrolimus-MMF therapy have lower graft function ifautoreactive T cells were detected before transplanta-tion (187).

Current immune suppressive drugs can also interferewith beta-cell function (306, 359). More specifically, rapamy-cin (sirolimus) impairs engraftment (492), interferes withangiogenesis (81), induces insulin resistance (148), and in-hibits �-cell replication (489). Rapamycin also, like cortico-steroids, tacrolimus (306), and MMF, decreases insulin tran-scription (reviewed in Ref. 359). Finally, a recent studysuggests that MMF also inhibits beta-cell neogenesis (153).

Human islet isolation techniques are still unsatisfactory(336, 407) to yield the �12,000 islet equivalents per kilogrambody weight required to restore insulin-independent normo-glycemia in recipients (391). Only �2,000 subjects in theUnited States can benefit from an islet transplant each year,because only half of the isolation efforts yield islets suitablefor transplantation and recipients usually require islets frommultiple donors (392). The use of xenogeneic islets, mostlyfrom pigs or transgenic pigs (82, 181, 449), can fill the gapbetween supply and demand in islet transplantation. Porcineislets are recognized as the most physiologically compatiblexenogeneic insulin-producing cells. Their xenogeneicnature likely requires immunoprotection in capsulesthat allow the inward passage of nutrients and glucoseand the outward passage of insulin (see below).

5. Stem cells

Generation of beta-cells provides an exciting approachtowards curing T1D (Table 5). This can be done by differ-entiation of embryonic stem (ES) cells (234) or pluripotentstem (IPS) cells (487), or the “reprogramming” of cells fromtheir initial phenotype into beta-like cells (494). Stem cellscan regenerate the beta-cell mass in vivo, as shown for bonemarrow-derived stem cell transfers in immunodeficient micewith chemically induced pancreatic damage (498). But in aphase I/II trial, stem cells from umbilical cord blood assisted



on August 26, 2011


ownloaded from


the preservation of C-peptide levels poorly (175). Autolo-gous nonmyeloablative hematopoietic stem cell transplanta-tion (HSCT) yielded better results in newly diagnosed T1D(100, 459). Insulin independence was achieved, but was lostafter 4–5 yr in most recipients, and side effects discouragethe use of this approach for universal therapy (100).

Stem cells can potentially also be differentiated to beta-cells in vitro. The use of autologous stem cells avoids allo-immune, but not autoimmune, responses to the transplantedbeta-like cells. A lot of effort is invested in encapsulationtechniques to protect the transplanted cells or islets afterimplantation. For instance, Viacyte differentiates hESCs intopancreatic endoderm and has plans to subsequently encap-sulate these cells in a protective permeable device for trans-plantation (Encaptra, see website) (106, 234). The idea isthat isletlike structures form and become functionally re-sponsive to glucose in vivo. A comparable encapsulationapproach had been shown to protect human fibroblast allo-grafts from rejection in rhesus monkeys (423). Also, implan-tation of primary murine beta-cells encapsulated in a similardevice could ameliorate diabetes in NOD mice (242). Mostimportantly, Viacyte, formerly NovoCell, published that theirdifferentiated hESCs generate glucose-responsive insulin-producing cells that can protect against streptozotocin-in-duced diabetes in mice (234). However, recent data exam-ining this approach in athymic nude rats could not com-pletely confirm this. Islet-like structures did develop fromhESC differentiated to pancreatic endoderm, but the extentof endocrine cell formation and secretory function was con-sidered insufficient to be clinically relevant (282). An alter-native, the alginate microcapsule containing porcine isletsfrom DiabeCell, allows some insulin production up to 9.5years postimplant (132) and is currently being tested inphase I/II (NCT00940173, Living Cell Technologies). Like-wise, the Cell Pouch is a device subcutaneously implantedprior to delivery of transplanted cells to allow tissue andblood vessel formation (Sernova). Last, the Sertolin technol-ogy codelivers Sertoli cells to provide an immune-protectedenvironment for (islet) cell transplants (Sernova). A small-scale study in humans showed long-term transplant survivaland positive effects on metabolic control (444), but exten-sive clinical trials have not yet begun. Although promising,two problems could arise: 1) clogging can limit influx ofnutrients and glucose and efflux of insulin, and 2) solublemediators that elicit beta-cell death can still reach the trans-plant.

E. Combination Intervention Trials in T1D

Like cancer treatment, T1D therapy might benefit fromcombining approaches that synergize to reverse autoimmu-nity, establish tolerance, and limit side effects. We repeathere our previous proposal that combination therapiesshould achieve three major goals (64): 1) the “freezing” of

the active immune response and dampening of any autore-active response without strong side effects, 2) generation ofTregs that can maintain long-term tolerance, and 3) theregeneration of a critical beta-cell mass to maintain eugly-cemia without repetitive insulin injections.

We are convinced that combination treatments willbecome integral to T1D therapy and should therefore tran-scend licensing issues between the manufacturers of theindividual compounds. Because of issues with regulatoryaffairs, initial steps will most logically use a combination oftreatments with either a documented effect as monotherapyin T1D, or a combination treatment that has proven efficacyin another (auto)immune disease (Table 6).

To date, only two clinical trials of a combination ther-apy have released results, and they are disappointing. Therecent phase III trial testing anti-CD25 mAb (daclizumab) incombination with MMF (mofetil, CellCept by Roche) re-ported no preservation of beta-cell function (163). Anti-CD25 mAbs block the IL-2 signaling pathway in activated Tcells, but do not interfere with Tregs (456). MMF is anadjuvant drug that selectively inhibits T- and B-lymphocyteproliferation by suppressing the de novo purine synthesis. InBB rats, MMF and anti-CD25 mAb alone or in combinationwere shown to delay and prevent diabetes (438). Both sys-temic immunosuppressants have also proven efficacy inother autoimmune diseases and, in combination, in prevent-ing acute graft rejection (275). Another unsuccessful trial,testing exenatide combined with daclizumab, showed noimproved function of the remaining beta-cells in patientswith long-standing T1D (21.3 � 10.7 yr)(370).

A new Proleukin and Rapamune phase I trial will testthe combination of IL-2 and rapamycin (which inhibits theresponse to IL-2). The rationale is that IL-2 promotes theexpansion of Treg in favor of effector T cells if the expan-sion of effectors is simultaneously blocked by rapamycin(347). As a result, deletion of autoreactive Th1 cells causesa shift from Th1- to Th2- and Th3-type cytokine-producingcells. This approach is supported by promising preclinicalresults showing the prevention of spontaneous T1D onset inNOD (347).

A privately funded clinical trial (Helmsley Trust) willassess whether a combination of ATG and GCSF reversesnew-onset diabetes in humans, based on data from NODmice (327). The rationale is as follows: ATG temporarilyreduces T cells in the bloodstream, while GCSF mobilizesgranulocytes and hematopoietic stem cells from the bonemarrow (428) and induces tolerogenic dendritic cells (216,371). Additionally, both ATG and GCSF induce a Treg pop-ulation to ensure long-term protection (216, 398).

The Diamyd-Sitagliptin-Lansoprazole phase II clinicaltrial is recruiting patients to test the combination of antigen-specific tolerance induction and beta-cell regeneration.GAD-Alum can prevent immune destruction and delay orprevent diabetes onset in NOD mice (see above) (7, 8, 256,349). Sitagliptin indirectly stimulates insulin secretion (see



on August 26, 2011


ownloaded from


above). Lanzoprazole is a proton-pump inhibitor that pro-vides partial reversal of diabetes in NOD mice, but strongreversal when combined with a DDP-4 inhibitor (408).

IX. PROMISING BENCH-SIDE THERAPEUTICS

Insights from preclinical research are crucial to therapydevelopment, but translation is often intricate. A vast num-ber of interventions have been tested preclinically, manywith beneficial outcomes (396), but few have started clinicaltrials.

A. Insulin Substitution

The many formulations of insulin on the market havedramatically increased the T1D patient’s quality of life. In-sulin is not a cure, but it will nevertheless remain the majortreatment in the short term and probably will be used assupplementation to other therapies in the long term. Con-tinuous blood glucose monitors or closed-loop insulin-deliv-ery systems, like the artificial pancreas, make diabetes man-agement easier, but some problematic issues remain (188).A noteworthy approach with favorable preclinical data isSmartInsulin that consists of a layered, biocompatible, andbiodegradable polymer-therapeutic that is bound to an en-gineered glucose-binding molecule. Insulin is released onlywhen it is unbound by the presence of a specific glucoseconcentration.

B. Combination Therapies With Immune

Modulators and Islet Antigenic Vaccines

We have shown that suboptimal doses of the FcR-non-binding anti-CD3 F(ab’)2 in conjunction with intranasal ad-ministration of proinsulin peptide can reverse diabetes intwo mouse models of diabetes (63). Insulin-specific Tregsare induced, secrete immunomodulatory cytokines likeTGF-� and IL-10, and confer dominant tolerance upon trans-fer to recent-onset diabetic recipient mice. Follow-up exper-iments showed the broader applicability of this approach:anti-CD3 in conjunction with a GAD65 plasmid vaccinationcould synergize strongly in a RIP-LMCV model of T1D (62).However, success depended on the genetic background,possibly due to how antigen is presented to Tregs (62).

C. Combination Therapies Using Immune

Modulators and Compounds Enhancing

Beta-Cell Mass or Function

Anti-CD3 in conjunction with exenatide combines in-duction of immune tolerance by FcR-non-binding anti-CD3mAb with stimulation of insulin secretion of the remainingbeta-cells. Given that this approach addresses two parts ofthe T1D problem, it is anticipated that this approach has ahigher chance of success than the beta-cell regenerativeagents gastrin and exenatide trials, which lack the immuno-modulatory arm required to stop autoimmune attack of thepancreatic beta-cells. That said, favorable results in diabetic

TABLE 6. Intervention trials and combination therapy


Exenatide �daclizumab

Stimulation of insulin secretionand blockade of IL-2 signalingpathway

Phase II Combination of intensified insulintherapy, exenatide, anddaclizumab did not induceimproved function of remainingbeta-cells

370

NCT00064714

NIDDK

IL-2 � rapamycinProleukin and

Rapamune

Downregulate T effector whilesparing T regulatory functionPromote Treg

Phase I N/A (recruiting) 347, 420NCT00525889NIAID, ITN

MMF � anti-CD25 Selectively inhibits T- and B-cellproliferation/blockade of IL-2signaling pathway

Phase III No preservationof beta-cellfunction

163Zenapax,

Daclizumab

NCT00100178NIDDK/TrialNet

Thymoglobulin �Neulasta

Induce Treg and tolerogenic DCs,recruit granulocytes/HSCs

Phase II

Helmsley TrustEGF and gastrin Beta-cell regeneration and islet

neogenesisPhase I Completed, results not released 409, 412, 413

(E1 �G1 INT) NCT00239148Transition

TherapeuticsGAD alum (Diamyd),

Lansoprazole,sitagliptin

GAD-specific immunomodulation,proton pump inhibitor

Phase II N/A (recruiting) 8, 230, 231, 256

DPP-4 inhibitor NCT00837759NIDDK

ATG, anti-thymocyte globulin; AUC, area under curve (for C-peptide levels in glucose tolerance test); BDC, Barbara Davis Center; DCCT, DiabetesControl and Complications Trial; DENIS, Deutsche Nicotinamide Intervention Study; EBV, Epstein-Barr virus, a herpesvirus that causes mononu-cleosis; ENDIT, European Nicotinamide Diabetes Intervention Trial; GSK, GlaxoSmithKline; ITN, Immune Tolerance Network.



on August 26, 2011


ownloaded from


NOD mice have led to a phase II trial of gastrin � exenatide(410, 411), while an anti-CD3 � exenatide trial is anticipated.

D. Cytokine-Based Therapeutics

Interference with cytokines in immune intervention is acomplex matter, as underscored by a broad range of efficacyand dangerous side effects. The majority of cytokine manip-ulations also revealed a dual effect depending on the timing,dose, and route of administration. For instance, both IL-18and TNF-� accelerate or inhibit T1D in NOD when admin-istrated early or late, respectively (94, 316, 368). This level ofcomplexity obstructs the clinical translation of cytokine tar-geting strategies. One strategy considered for clinical trial isthe combined administration of rapamycin with agonisticIL-2-Fc and antagonistic IL-15-Fc fusion proteins, which hasbeen shown to provide long-term engraftment/tolerance inallotransplant models (493). The idea is to limit effectorT-cell activation and expansion (by blocking IL-15 signals)while promoting Tregs (IL-2 and rapamycin). Noteworthy isthe essential role of the IL21/IL21R axis to autoimmunediabetes in NOD (416) and its part in genetic susceptibility toT1D (20).

X. OUR CONCLUSIONS

The incidence of T1D increases rapidly, especially inthe developed world, and the time of onset shifts towards ayounger age. T1D most likely results from an unfortunatecombination of genetic susceptibility and exposure to anenvironmental trigger. The main effector mechanism isclearly an autoimmune reaction, which is also evident attime of clinical diagnosis. This implies two concepts forclinical treatment: 1) knowledge of the cause is more criticalfor prevention than for at-onset therapy, and 2) at-onset ornear-onset therapy requires an immune silencing or modu-lating component. Our current opinion is that we shouldconduct trials with treatments like anti-CD3 that are effec-tive at onset (alone or in combination), but advance theirinitiation based on early detection instead of waiting untilovert hyperglycemia. Early detection is also required tomaximally preserve the remaining beta-cell mass, becausethe ability to secrete even small amounts of insulin can makedisease control easier and help minimize the complicationsof chronic inadequate glycemic control (194, 229). Screeningefforts would ideally unify results from genetic (HLD-DR3/4-DQ2/8, family), metabolomic (lysophosphatidylcholine),and C-peptide release tests with autoantibody titers (insulin,GAD65, ICA512, ZnT08) and autoreactive T-cell assays (Pro-insulin, GAD65, I-A2, and ZnT8).

Much of our current understanding of T1D comes fromthe NOD mouse model. For a more translational focus, it isnecessary to look beyond the NOD mouse to take full ad-vantage of the additional models available (461). And even

then, the majority of T1D treatment discovered in mousemodels have not yet translated to viable treatments in hu-mans. The landscape of possible treatment has beenchanged by the prospect that T1D progression may beblocked by the active stimulation of tolerance induced by(auto)antigen-specific immunization to generate Treg. Suc-cess will depend on correctly hitting the following fourfactors on target: 1) the choice of protein or peptide that isdelivered, 2) the dosage, 3) the disease stage, and 4) theroute of administration. A combination of computer-drivenbiosimulation (in silico) and “wet lab” experiments couldincrease the chance and reduce the time to reveal the sweetspot(s) of immune therapy.

The ultimate goal of autoimmune therapy is to silencethe immune attack against self without sacrificing the pa-tient’s protective immune response to infections. This ismost likely achieved by a therapy that combines a nonspe-cific immune suppressant (e.g., anti-CD3), antigen-specificinduction of Tregs (e.g., Proinsulin, GAD65), and a suitablecompound that increases beta-cell mass or function. Forpatients with C-peptide levels �0.5 (nM), a two-compoundtherapy might suffice, because preclinical research suggeststhat “natural” beta-cell regeneration still occurs (228). Pa-tients with C-peptide levels in the 0.2–0.5 range might re-quire additional beta-cell regenerative compounds, like gas-trin and exenatide. C-peptide levels �0.2 indicate the needfor pancreatic islet transplantation.

The unfortunate reality is that combination therapiesusing one or more nonapproved drugs are difficult to license(460). The current viewpoint is that each compound of acombination treatment needs to be efficacious and licensedon its own first (281). This should make place for safetytrials of individual compounds and efficacy trials for thecombination therapy. Also, competing interests obstruct thecombination of drugs from different companies. So, majorefforts on several fronts are still required to fully realize thebenefits of the technological and scientific advances in au-toimmune diabetes research.

ACKNOWLEDGMENTS

Address for reprint requests and other correspondence: M.von Herrath, Center for Type 1 Diabetes Research, La JollaInstitute for Allergy and Immunology, 9420 Athena Circle, LaJolla, CA 92037 (e-mail: [email protected]).

GRANTS

T. Van Belle was funded through a Mentor-Based Postdoc-toral Fellowship Award from the American Diabetes Associa-tion. This work was funded by the Juvenile Diabetes ResearchFoundation (16–2007-370), The Brehm Center for T1D Researchand Analysis, and National Institutes of Health Grants U01-DK-78013 and P01-AI-58105.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declaredby the author(s).



on August 26, 2011


ownloaded from


REFERENCES

1. Study design of the Trial to Reduce IDDM in the Genetically

at Risk (TRIGR). Pediatr Diabetes 8: 117–137, 2007.2. Abdul-Rasoul M, Habib H, Al-Khouly M. “The honeymoon

phase” in children with type 1 diabetes mellitus: frequency, dura-tion, and influential factors. Pediatr Diabetes 7: 101–107, 2006.

3. Achenbach P, Bonifacio E, Koczwara K, Ziegler AG. Naturalhistory of type 1 diabetes. Diabetes 54 Suppl 2: S25–31, 2005.

4. Achenbach P, Koczwara K, Knopff A, Naserke H, Ziegler AG,

Bonifacio E. Mature high-affinity immune responses to (pro)insu-lin anticipate the autoimmune cascade that leads to type 1 diabetes.J Clin Invest 114: 589–597, 2004.

5. Adachi K, Kumamoto T, Araki S. Interleukin-2 receptor levelsindicating relapse in multiple sclerosis. Lancet 1: 559–560, 1989.

6. Adams SF. The seasonal variation in the incidence of acute dia-betes. Arch Int Med 37: 861, 1926.

7. Agardh CD, Cilio CM, Lethagen A, Lynch K, Leslie RD. Clinicalevidence for the safety of GAD65 immunomodulation in adult-onset autoimmune diabetes. J Diabetes Complications 19: 238–246, 2005.

8. Agardh CD, Lynch KF, Palmer M, Link K, Lernmark A. GAD65vaccination: 5 years of follow-up in a randomised dose-escalatingstudy in adult-onset autoimmune diabetes. Diabetologia 52: 1363–1368, 2009.

9. Agner T, Damm P, Binder C. Remission in IDDM: prospectivestudy of basal C-peptide and insulin dose in 268 consecutive pa-tients. Diabetes Care 10: 164–169, 1987.

10. Akerblom HK, Virtanen SM, Ilonen J, Savilahti E, Vaarala O.

Dietary manipulation of beta cell autoimmunity in infants at in-creased risk of type 1 diabetes: a pilot study. Diabetologia 48:829–837, 2005.

11. Akirav E, Kushner JA, Herold KC. Beta-cell mass and type 1diabetes: going, going, gone? Diabetes 57: 2883–2888, 2008.

12. Alizadeh BZ, Hanifi-Moghaddam P, Eerligh P, van der Slik AR,

Kolb H. Association of interferon-gamma and interleukin 10 geno-types and serum levels with partial clinical remission in type 1diabetes. Clin Exp Immunol 145: 480–484, 2006.

13. Allen HF, Klingensmith GJ, Jensen P, Simoes E, Hayward A,

Chase HP. Effect of Bacillus Calmette-Guerin vaccination on new-onset type 1 diabetes. A randomized clinical study. Diabetes Care

22: 1703–1707, 1999.14. Alleva DG, Maki RA, Putnam AL, Robinson JM, Kipnes MS.

Immunomodulation in type 1 diabetes by NBI-6024, an alteredpeptide ligand of the insulin B epitope. Scand J Immunol 63:59–69, 2006.

15. Aly TA, Ide A, Jahromi MM, Barker JM, Fernando MS. Ex-treme genetic risk for type 1A diabetes. Proc Natl Acad Sci USA

103: 14074–14079, 2006.16. Anderson MS, Venanzi ES, Klein L, Chen Z, Berzins SP. Pro-

jection of an immunological self shadow within the thymus by theaire protein. Science 298: 1395–1401, 2002.

17. Anjos S, Nguyen A, Ounissi-Benkalha H, Tessier MC, Poly-

chronakos C. A common autoimmunity predisposing signal pep-tide variant of the cytotoxic T-lymphocyte antigen 4 results ininefficient glycosylation of the susceptibility allele. J Biol Chem

277: 46478–46486, 2002.18. Arif S, Cox P, Afzali B, Lombardi G, Lechler RI, Peakman M,

Mirenda V. Anti-TNFalpha therapy–killing two birds with onestone? Lancet 375: 2278.

19. Arif S, Tree TI, Astill TP, Tremble JM, Bishop AJ. Autoreac-tive T cell responses show proinflammatory polarization in diabe-tes but a regulatory phenotype in health. J Clin Invest 113: 451–463, 2004.

20. Asano K, Ikegami H, Fujisawa T, Nishino M, Nojima K. Mo-lecular scanning of interleukin-21 gene and genetic susceptibility totype 1 diabetes. Hum Immunol 68: 384–391, 2007.

21. Assan R, Feutren G, Debray-Sachs M, Quiniou-Debrie MC,

Laborie C. Metabolic and immunological effects of cyclosporin inrecently diagnosed type 1 diabetes mellitus. Lancet 1: 67–71, 1985.

22. Atkinson MA, Bowman MA, Campbell L, Darrow BL, Kaufman

DL, Maclaren NK. Cellular immunity to a determinant common to

glutamate decarboxylase and coxsackie virus in insulin-dependentdiabetes. J Clin Invest 94: 2125–2129, 1994.

23. Atkinson MA, Bowman MA, Kao KJ, Campbell L, Dush PJ.

Lack of immune responsiveness to bovine serum albumin in insu-lin-dependent diabetes. N Engl J Med 329: 1853–1858, 1993.

24. Atkinson MA, Eisenbarth GS. Type 1 diabetes: new perspectiveson disease pathogenesis and treatment. Lancet 358: 221–229, 2001.

25. Atkinson MA, Holmes LA, Scharp DW, Lacy PE, Maclaren NK.

No evidence for serological autoimmunity to islet cell heat shockproteins in insulin dependent diabetes. J Clin Invest 87: 721–724,1991.

26. Bacchetta R, Passerini L, Gambineri E, Dai M, Allan SE.

Defective regulatory and effector T cell functions in patients withFOXP3 mutations. J Clin Invest 116: 1713–1722, 2006.

27. Bach JF. The effect of infections on susceptibility to autoimmuneand allergic diseases. N Engl J Med 347: 911–920, 2002.

28. Baecher-Allan C, Brown JA, Freeman GJ, Hafler DA.

CD4�CD25 high regulatory cells in human peripheral blood. J

Immunol 167: 1245–1253, 2001.29. Bailey R, Cooper JD, Zeitels L, Smyth DJ, Yang JH. Associa-

tion of the vitamin D metabolism gene CYP27B1 with type 1 dia-betes. Diabetes 56: 2616–2621, 2007.

30. Baker L, Kaye R, Root AW. The early partial remission of juvenilediabetes mellitus. The roles of insulin and growth hormone. J

Pediatr 71: 825–831, 1967.31. Ban L, Zhang J, Wang L, Kuhtreiber W, Burger D, Faustman

DL. Selective death of autoreactive T cells in human diabetes byTNF or TNF receptor 2 agonism. Proc Natl Acad Sci USA 105:13644–13649, 2008.

32. Banatvala JE, Bryant J, Schernthaner G, Borkenstein M,

Schober E. Coxsackie B, mumps, rubella, and cytomegalovirusspecific IgM responses in patients with juvenile-onset insulin-de-pendent diabetes mellitus in Britain, Austria, and Australia. Lancet

1: 1409–1412, 1985.33. Barker JM. Clinical review: type 1 diabetes-associated autoimmu-

nity: natural history, genetic associations, and screening. J Clin

Endocrinol Metab 91: 1210–1217, 2006.34. Barker JM, McFann KK, Orban T. Effect of oral insulin on

insulin autoantibody levels in the Diabetes Prevention Trial Type 1oral insulin study. Diabetologia 50: 1603–1606, 2007.

35. Barreto M, Santos E, Ferreira R, Fesel C, Fontes MF. Evi-dence for CTLA4 as a susceptibility gene for systemic lupus ery-thematosus. Eur J Hum Genet 12: 620–626, 2004.

36. Barrett JC, Clayton DG, Concannon P, Akolkar B, Cooper JD.

Genome-wide association study and meta-analysis find that over 40loci affect risk of type 1 diabetes. Nat Genet 41: 703–707, 2009.

37. Baschal EE, Aly TA, Babu SR, Fernando MS, Yu L. HLA-DPB1*0402 protects against type 1A diabetes autoimmunity in thehighest risk DR3-DQB1*0201/DR4-DQB1*0302 DAISY population.Diabetes 56: 2405–2409, 2007.

38. Beales PE, Elliott RB, Flohe S, Hill JP, Kolb H. A multi-centre,blinded international trial of the effect of A(1) and A(2) beta-caseinvariants on diabetes incidence in two rodent models of spontane-ous Type I diabetes. Diabetologia 45: 1240–1246, 2002.

39. Bell GI, Horita S, Karam JH. A polymorphic locus near thehuman insulin gene is associated with insulin-dependent diabetesmellitus. Diabetes 33: 176–183, 1984.

40. Bellin MD, Kandaswamy R, Parkey J, Zhang HJ, Liu B. Pro-longed insulin independence after islet allotransplants in recipientswith type 1 diabetes. Am J Transplant 8: 2463–2470, 2008.

41. Bendtzen K, Mandrup-Poulsen T, Nerup J, Nielsen JH, Din-

arello CA, Svenson M. Cytotoxicity of human pI 7 interleukin-1for pancreatic islets of Langerhans. Science 232: 1545–1547, 1986.

42. Bennett ST, Lucassen AM, Gough SC, Powell EE, Undlien DE.

Susceptibility to human type 1 diabetes at IDDM2 is determined bytandem repeat variation at the insulin gene minisatellite locus. Nat

Genet 9: 284–292, 1995.43. Berg AK, Korsgren O, Frisk G. Induction of the chemokine

interferon-gamma-inducible protein-10 in human pancreatic isletsduring enterovirus infection. Diabetologia 49: 2697–2703, 2006.

44. Bhat R, Bhansali A, Bhadada S, Sialy R. Effect of pioglitazonetherapy in lean type 1 diabetes mellitus. Diabetes Res Clin Pract 78:349–354, 2007.



on August 26, 2011


ownloaded from


45. Bingley PJ, Gale EA, European Nicotinamide Diabetes Inter-

vention Trial. Progression to type 1 diabetes in islet cell antibody-positive relatives in the European Nicotinamide Diabetes Interven-tion Trial: the role of additional immune, genetic and metabolicmarkers of risk. Diabetologia 49: 881–890, 2006.

46. Bjork E, Berne C, Kampe O, Wibell L, Oskarsson P, Karlsson

FA. Diazoxide treatment at onset preserves residual insulin secre-tion in adults with autoimmune diabetes. Diabetes 45: 1427–1430,1996.

47. Blom L, Nystrom L, Dahlquist G. The Swedish childhood diabe-tes study. Vaccinations and infections as risk determinants fordiabetes in childhood. Diabetologia 34: 176–181, 1991.

48. Blomqvist M, Juhela S, Erkkila S, Korhonen S, Simell T.

Rotavirus infections and development of diabetes-associated auto-antibodies during the first 2 years of life. Clin Exp Immunol 128:511–515, 2002.

49. Bloom BJ. Development of diabetes mellitus during etanercepttherapy in a child with systemic-onset juvenile rheumatoid arthri-tis. Arthritis Rheum 43: 2606–2608, 2000.

50. Bluestone JA, St Clair EW, Turka LA. CTLA4Ig: bridging thebasic immunology with clinical application. Immunity 24: 233–238,2006.

51. Bodansky HJ, Staines A, Stephenson C, Haigh D, Cartwright

R. Evidence for an environmental effect in the aetiology of insulindependent diabetes in a transmigratory population. BMJ 304: 1020–1022, 1992.

52. Bodington MJ, McNally PG, Burden AC. Cow’s milk and type 1childhood diabetes: no increase in risk. Diabet Med 11: 663–665,1994.

53. Bohmer KP, Kolb H, Kuglin B, Zielasek J, Hubinger A. Linearloss of insulin secretory capacity during the last six months pre-ceding IDDM. No effect of antiedematous therapy with ketotifen.Diabetes Care 17: 138–141, 1994.

54. Bolt S, Routledge E, Lloyd I, Chatenoud L, Pope H. Thegeneration of a humanized, non-mitogenic CD3 monoclonal anti-body which retains in vitro immunosuppressive properties. Eur J

Immunol 23: 403–411, 1993.55. Bonifacio E, Scirpoli M, Kredel K, Fuchtenbusch M, Ziegler

AG. Early autoantibody responses in prediabetes are IgG1 domi-nated and suggest antigen-specific regulation. J Immunol 163: 525–532, 1999.

56. Bottazzo GF, Dean BM, McNally JM, MacKay EH, Swift PG,

Gamble DR. In situ characterization of autoimmune phenomenaand expression of HLA molecules in the pancreas in diabeticinsulitis. N Engl J Med 313: 353–360, 1985.

57. Bottini N, Musumeci L, Alonso A, Rahmouni S, Nika K. Afunctional variant of lymphoid tyrosine phosphatase is associatedwith type I diabetes. Nat Genet 36: 337–338, 2004.

58. Bottini N, Vang T, Cucca F, Mustelin T. Role of PTPN22 in type1 diabetes and other autoimmune diseases. Semin Immunol 18:207–213, 2006.

59. Bougneres PF, Landais P, Boisson C, Carel JC, Frament N.

Limited duration of remission of insulin dependency in childrenwith recent overt type I diabetes treated with low-dose cyclosporin.Diabetes 39: 1264–1272, 1990.

60. Boulton JG, Bourne JT. Unstable diabetes in a patient receivinganti-TNF-alpha for rheumatoid arthritis. Rheumatology 46: 178–179, 2007.

61. Bouwens L, Rooman I. Regulation of pancreatic beta-cell mass.Physiol Rev 85: 1255–1270, 2005.

62. Bresson D, Fradkin M, Manenkova Y, Rottembourg D, von

Herrath M. Genetic-induced variations in the GAD65 T-cell reper-toire governs efficacy of anti-CD3/GAD65 combination therapy innew-onset type 1 diabetes. Mol Ther 18: 307–316, 2010.

63. Bresson D, Togher L, Rodrigo E, Chen Y, Bluestone JA,

Herold KC, von Herrath M. Anti-CD3 and nasal proinsulin com-bination therapy enhances remission from recent-onset autoim-mune diabetes by inducing Tregs. J Clin Invest 116: 1371–1381,2006.

64. Bresson D, von Herrath M. Moving towards efficient therapies intype 1 diabetes: to combine or not to combine? Autoimmun Rev 6:315–322, 2007.

65. Brod SA. Autoimmunity is a type I interferon-deficiency syndromecorrected by ingested type I IFN via the GALT system. J Interferon

Cytokine Res 19: 841–852, 1999.66. Brod SA, Atkinson M, Lavis VR, Brosnan PG, Hardin DS.

Ingested IFN-alpha preserves residual beta cell function in type 1diabetes. J Interferon Cytokine Res 21: 1021–1030, 2001.

67. Brod SA, Nguyen M, Hood Z, Shipley GL. Ingested (oral) IFN-alpha represses TNF-alpha mRNA in relapsing-remitting multiplesclerosis. J Interferon Cytokine Res 26: 150–155, 2006.

68. Brugman S, Klatter FA, Visser JT, Wildeboer-Veloo AC,

Harmsen HJ, Rozing J, Bos NA. Antibiotic treatment partiallyprotects against type 1 diabetes in the Bio-Breeding diabetes-pronerat. Is the gut flora involved in the development of type 1 diabetes?Diabetologia 49: 2105–2108, 2006.

69. Brusko T, Bluestone J. Clinical application of regulatory T cellsfor treatment of type 1 diabetes and transplantation. Eur J Immu-

nol 38: 931–934, 2008.70. Brusko T, Wasserfall C, McGrail K, Schatz R, Viener HL. No

alterations in the frequency of FOXP3� regulatory T-cells in type 1diabetes. Diabetes 56: 604–612, 2007.

71. Brusko TM, Putnam AL, Bluestone JA. Human regulatory Tcells: role in autoimmune disease and therapeutic opportunities.Immunol Rev 223: 371–390, 2008.

72. Brusko TM, Wasserfall CH, Clare-Salzler MJ, Schatz DA,

Atkinson MA. Functional defects and the influence of age on thefrequency of CD4� CD25� T-cells in type 1 diabetes. Diabetes 54:1407–1414, 2005.

73. Bugawan TL, Klitz W, Alejandrino M, Ching J, Panelo A. Theassociation of specific HLA class I and II alleles with type 1 diabe-tes among Filipinos. Tissue Antigens 59: 452–469, 2002.

74. Burgess MA, Forrest JM. Congenital rubella and diabetes melli-tus. Diabetologia 52: 369–370, 2009.

75. Burton PR, Clayton DG, Cardon LR, Craddock N, Deloukas P.

Association scan of 14,500 nonsynonymous SNPs in four diseasesidentifies autoimmunity variants. Nat Genet 39: 1329–1337, 2007.

76. Buse JB, Rosenstock J, Sesti G, Schmidt WE, Montanya E.

Liraglutide once a day versus exenatide twice a day for type 2diabetes: a 26-week randomised, parallel-group, multinational,open-label trial (LEAD-6). Lancet 374: 39–47, 2009.

77. Bussone G, Mouthon L. Autoimmune manifestations in primaryimmune deficiencies. Autoimmun Rev 8: 332–336, 2009.

78. Butler AE, Galasso R, Meier JJ, Basu R, Rizza RA, Butler PC.

Modestly increased beta cell apoptosis but no increased beta cellreplication in recent-onset type 1 diabetic patients who died ofdiabetic ketoacidosis. Diabetologia 50: 2323–2331, 2007.

79. Butler PC, Meier JJ, Butler AE, Bhushan A. The replication ofbeta cells in normal physiology, in disease and for therapy. Nat

Clin Pract Endocrinol Metab 3: 758–768, 2007.80. Calcinaro F, Dionisi S, Marinaro M, Candeloro P, Bonato V.

Oral probiotic administration induces interleukin-10 productionand prevents spontaneous autoimmune diabetes in the non-obesediabetic mouse. Diabetologia 48: 1565–1575, 2005.

81. Cantaluppi V, Biancone L, Romanazzi GM, Figliolini F, Bel-

tramo S. Antiangiogenic and immunomodulatory effects of rapa-mycin on islet endothelium: relevance for islet transplantation. Am

J Transplant 6: 2601–2611, 2006.82. Cardona K, Korbutt GS, Milas Z, Lyon J, Cano J. Long-term

survival of neonatal porcine islets in nonhuman primates by tar-geting costimulation pathways. Nat Med 12: 304–306, 2006.

83. Carel JC, Boitard C, Eisenbarth G, Bach JF, Bougneres PF.

Cyclosporine delays but does not prevent clinical onset in glucoseintolerant pre-type 1 diabetic children. J Autoimmun 9: 739–745,1996.

84. Carmichael SK, Johnson SB, Baughcum A, North K, Hopkins

D. Prospective assessment in newborns of diabetes autoimmunity(PANDA): maternal understanding of infant diabetes risk. Genet

Med 5: 77–83, 2003.85. Catassi C, Guerrieri A, Bartolotta E, Coppa GV, Giorgi PL.

Antigliadin antibodies at onset of diabetes in children. Lancet 2:158, 1987.

86. Cernea S, Herold KC. Monitoring of antigen-specific CD8 T cellsin patients with type 1 diabetes treated with antiCD3 monoclonalantibodies. Clin Immunol 134: 121–129, 2009.



on August 26, 2011


ownloaded from


87. Chaillous L, Lefevre H, Thivolet C, Boitard C, Lahlou N. Oralinsulin administration and residual beta-cell function in recent-onset type 1 diabetes: a multicentre randomised controlled trial.Diabete Insuline Orale group Lancet 356: 545–549, 2000.

88. Chatenoud L, Baudrihaye MF, Kreis H, Goldstein G, Schin-

dler J, Bach JF. Human in vivo antigenic modulation induced bythe anti-T cell OKT3 monoclonal antibody. Eur J Immunol 12:979–982, 1982.

89. Chatenoud L, Bluestone JA. CD3-specific antibodies: a portal tothe treatment of autoimmunity. Nat Rev Immunol 7: 622–632, 2007.

90. Chatenoud L, Primo J, Bach JF. CD3 antibody-induced dominantself tolerance in overtly diabetic NOD mice. J Immunol 158: 2947–2954, 1997.

91. Chatenoud L, Thervet E, Primo J, Bach JF. Anti-CD3 antibodyinduces long-term remission of overt autoimmunity in nonobesediabetic mice. Proc Natl Acad Sci USA 91: 123–127, 1994.

92. Chentoufi AA, Polychronakos C. Insulin expression levels in thethymus modulate insulin-specific autoreactive T-cell tolerance: themechanism by which the IDDM2 locus may predispose to diabetes.Diabetes 51: 1383–1390, 2002.

93. Christen U, McGavern DB, Luster AD, von Herrath MG, Old-

stone MB. Among CXCR3 chemokines, IFN-gamma-inducible pro-tein of 10 kDa [CXC chemokine ligand (CXCL) 10] but not mono-kine induced by IFN-gamma (CXCL9) imprints a pattern for thesubsequent development of autoimmune disease. J Immunol 171:6838–6845, 2003.

94. Christen U, Wolfe T, Mohrle U, Hughes AC, Rodrigo E. A dualrole for TNF-alpha in type 1 diabetes: islet-specific expressionabrogates the ongoing autoimmune process when induced late butnot early during pathogenesis. J Immunol 166: 7023–7032, 2001.

95. Clements GB, Galbraith DN, Taylor KW. Coxsackie B virusinfection and onset of childhood diabetes. Lancet 346: 221–223,1995.

96. Concannon P, Rich SS, Nepom GT. Genetics of type 1A diabetes.N Engl J Med 360: 1646–1654, 2009.

97. Cooper JD, Smyth DJ, Smiles AM, Plagnol V, Walker NM.

Meta-analysis of genome-wide association study data identifiesadditional type 1 diabetes risk loci. Nat Genet 40: 1399–1401, 2008.

98. Coppieters K, von Herrath M. Taking a closer look at the pan-creas. Diabetologia 51: 2145–2147, 2008.

99. Coppieters KT, von Herrath MG. Histopathology of type 1 dia-betes: old paradigms and new insights. Rev Diabet Stud 6: 85–96,2009.

100. Couri CE, Oliveira MC, Stracieri AB, Moraes DA, Pieroni F.

C-peptide levels and insulin independence following autologousnonmyeloablative hematopoietic stem cell transplantation in newlydiagnosed type 1 diabetes mellitus. JAMA 301: 1573–1579, 2009.

101. Couri CE, Voltarelli JC. Stem cell therapy for type 1 diabetesmellitus: a review of recent clinical trials. Diabetol Metab Syndr 1:19, 2009.

102. Couto FM, Minn AH, Pise-Masison CA, Radonovich M, Brady

JN. Exenatide blocks JAK1-STAT1 in pancreatic beta cells. Metab-

olism 56: 915–918, 2007.103. Craighead JE, McLane MF. Diabetes mellitus: induction in mice

by encephalomyocarditis virus. Science 162: 913–914, 1968.104. Croom KF, McCormack PL. Liraglutide: a review of its use in type

2 diabetes mellitus. Drugs 69: 1985–2004, 2009.105. Cucca F, Muntoni F, Lampis R, Frau F, Argiolas L. Combina-

tions of specific DRB1, DQA1, and DQB1 haplotypes are associatedwith insulin-dependent diabetes mellitus in Sardinia. Hum Immu-

nol 37: 85–94, 1993.106. D’Amour KA, Agulnick AD, Eliazer S, Kelly OG, Kroon E,

Baetge EE. Efficient differentiation of human embryonic stemcells to definitive endoderm. Nat Biotechnol 23: 1534–1541, 2005.

107. Dahlquist G, Frisk G, Ivarsson SA, Svanberg L, Forsgren M,

Diderholm H. Indications that maternal coxsackie B virus infec-tion during pregnancy is a risk factor for childhood-onset IDDM.Diabetologia 38: 1371–1373, 1995.

108. Dahlquist G, Savilahti E, Landin-Olsson M. An increased levelof antibodies to beta-lactoglobulin is a risk determinant for early-onset type 1 (insulin-dependent) diabetes mellitus independent ofislet cell antibodies and early introduction of cow’s milk. Diabeto-

logia 35: 980–984, 1992.

109. Dai YD, Marrero IG, Gros P, Zaghouani H, Wicker LS, Sercarz

EE. Slc11a1 enhances the autoimmune diabetogenic T-cell re-sponse by altering processing and presentation of pancreatic isletantigens. Diabetes 58: 156–164, 2009.

110. Daneman D, Fishman L, Clarson C, Martin JM. Dietary triggersof insulin-dependent diabetes in the BB rat. Diabetes Res 5: 93–97,1987.

111. Daniel D, Wegmann DR. Protection of nonobese diabetic micefrom diabetes by intranasal or subcutaneous administration ofinsulin peptide B-(9–23). Proc Natl Acad Sci USA 93: 956–960,1996.

112. DeStefano F, Mullooly JP, Okoro CA, Chen RT, Marcy SM.

Childhood vaccinations, vaccination timing, and risk of type 1diabetes mellitus. Pediatrics 108: E112, 2001.

113. Devendra D, Eisenbarth GS. Interferon alpha–a potential link inthe pathogenesis of viral-induced type 1 diabetes and autoimmu-nity. Clin Immunol 111: 225–233, 2004.

114. Di Lorenzo TP, Peakman M, Roep BO. Translational mini-reviewseries on type 1 diabetes: systematic analysis of T cell epitopes inautoimmune diabetes. Clin Exp Immunol 148: 1–16, 2007.

115. Diamond Study Group. Incidence and trends of childhood Type 1diabetes worldwide 1990–1999. Diabet Med 23: 857–866, 2006.

116. Dinarello CA. The many worlds of reducing interleukin-1. Arthri-

tis Rheum 52: 1960–1967, 2005.117. Dippe SE, Bennett PH, Miller M, Maynard JE, Berquist KR.

Lack of causal association between Coxsackie B4 virus infectionand diabetes. Lancet 1: 1314–1317, 1975.

118. Dooms H, Wolslegel K, Lin P, Abbas AK. Interleukin-2 enhancesCD4� T cell memory by promoting the generation of IL-7R alpha-expressing cells. J Exp Med 204: 547–557, 2007.

119. Dotta F, Censini S, van Halteren AG, Marselli L, Masini M.

Coxsackie B4 virus infection of beta cells and natural killer cellinsulitis in recent-onset type 1 diabetic patients. Proc Natl Acad Sci

USA 104: 5115–5120, 2007.120. DPT Type I Diabetes Study Group. Effects of insulin in relatives

of patients with type 1 diabetes mellitus. N Engl J Med 346: 1685–1691, 2002.

121. Drescher KM, Kono K, Bopegamage S, Carson SD, Tracy S.

Coxsackievirus B3 infection and type 1 diabetes development inNOD mice: insulitis determines susceptibility of pancreatic islets tovirus infection. Virology 329: 381–394, 2004.

122. Dungan KM, Buse JB, Ratner RE. Effects of therapy in type 1and type 2 diabetes mellitus with a peptide derived from isletneogenesis associated protein (INGAP). Diabetes Metab Res Rev

25: 558–565, 2009.123. Ehehalt S, Dietz K, Willasch AM, Neu A. Epidemiological per-

spectives on type 1 diabetes in childhood and adolescence inGermany: 20 years of the DIARY registry. Diabetes Care 2009.

124. Eisenbarth GS. Type I diabetes mellitus. A chronic autoimmunedisease. N Engl J Med 314: 1360–1368, 1986.

125. Eisenbarth GS, Srikanta S, Jackson R, Rabinowe S, Dolinar

R, Aoki T, Morris MA. Anti-thymocyte globulin and prednisoneimmunotherapy of recent onset type 1 diabetes mellitus. Diabetes

Res 2: 271–276, 1985.126. Elfving M, Svensson J, Oikarinen S, Jonsson B, Olofsson P.

Maternal enterovirus infection during pregnancy as a risk factor inoffspring diagnosed with type 1 diabetes between 15 and 30 yearsof age. Exp Diabetes Res 2008: 271958, 2008.

127. Elias D, Avron A, Tamir M, Raz I. DiaPep277 preserves endog-enous insulin production by immunomodulation in type 1 diabetes.Ann NY Acad Sci 1079: 340–344, 2006.

128. Elias D, Cohen IR. Peptide therapy for diabetes in NOD mice.Lancet 343: 704–706, 1994.

129. Elias D, Reshef T, Birk OS, van der Zee R, Walker MD, Cohen

IR. Vaccination against autoimmune mouse diabetes with a T-cellepitope of the human 65-kDa heat shock protein. Proc Natl Acad

Sci USA 88: 3088–3091, 1991.130. Elliott JF, Marlin KL, Couch RM. Effect of bacille Calmette-

Guerin vaccination on C-peptide secretion in children newly diag-nosed with IDDM. Diabetes Care 21: 1691–1693, 1998.

131. Elliott RB, Chase HP. Prevention or delay of type 1 (insulin-dependent) diabetes mellitus in children using nicotinamide. Dia-

betologia 34: 362–365, 1991.



on August 26, 2011


ownloaded from


132. Elliott RB, Escobar L, Tan PL, Muzina M, Zwain S, Buchanan

C. Live encapsulated porcine islets from a type 1 diabetic patient9.5 yr after xenotransplantation. Xenotransplantation 14: 157–161,2007.

133. Ergun-Longmire B, Marker J, Zeidler A, Rapaport R, Raskin

P. Oral insulin therapy to prevent progression of immune-mediated(type 1) diabetes. Ann NY Acad Sci 1029: 260–277, 2004.

134. Erlich H, Valdes AM, Noble J, Carlson JA, Varney M. HLADR-DQ haplotypes and genotypes and type 1 diabetes risk: analysisof the type 1 diabetes genetics consortium families. Diabetes 57:1084–1092, 2008.

135. Erlich HA, Zeidler A, Chang J, Shaw S, Raffel LJ. HLA class IIalleles and susceptibility and resistance to insulin dependent dia-betes mellitus in Mexican-American families. Nat Genet 3: 358–364,1993.

136. EURODiab. Infections and vaccinations as risk factors for child-hood type I (insulin-dependent) diabetes mellitus: a multicentrecase-control investigation. EURODIAB Substudy 2 Study Group.Diabetologia 43: 47–53, 2000.

137. Fan Y, Rudert WA, Grupillo M, He J, Sisino G, Trucco M.

Thymus-specific deletion of insulin induces autoimmune diabetes.EMBO J 28: 2812–2824, 2009.

138. Farid NR, Sampson L, Noel P, Barnard JM, Davis AJ, Hillman

DA. HLA-D-related (DRw) antigens in juvenile diabetes mellitus.Diabetes 28: 552–557, 1979.

139. Feldmann M, Maini RN. Lasker Clinical Medical Research Award.TNF defined as a therapeutic target for rheumatoid arthritis andother autoimmune diseases. Nat Med 9: 1245–1250, 2003.

140. Fennessy M, Metcalfe K, Hitman GA, Niven M, Biro PA,

Tuomilehto J, Tuomilehto-Wolf E. A gene in the HLA class Iregion contributes to susceptibility to IDDM in the Finnish popu-lation Childhood Diabetes in Finland (DiMe) Study Group. Diabe-

tologia 37: 937–944, 1994.141. Filippi CM, Estes EA, Oldham JE, von Herrath MG. Immuno-

regulatory mechanisms triggered by viral infections protect fromtype 1 diabetes in mice. J Clin Invest 119: 1515–1523, 2009.

142. Filippi CM, von Herrath MG. Viral trigger for type 1 diabetes:pros and cons. Diabetes 57: 2863–2871, 2008.

143. Flohe SB, Wasmuth HE, Kerad JB, Beales PE, Pozzilli P. Awheat-based, diabetes-promoting diet induces a Th1-type cytokinebias in the gut of NOD mice. Cytokine 21: 149–154, 2003.

144. Foulis AK, Farquharson MA, Cameron SO, McGill M, Schonke

H, Kandolf R. A search for the presence of the enteroviral capsidprotein VP1 in pancreases of patients with type 1 (insulin-depen-dent) diabetes and pancreases and hearts of infants who died ofcoxsackieviral myocarditis. Diabetologia 33: 290–298, 1990.

145. Foulis AK, Farquharson MA, Meager A. Immunoreactive alpha-interferon in insulin-secreting beta cells in type 1 diabetes mellitus.Lancet 2: 1423–1427, 1987.

146. Foulis AK, Liddle CN, Farquharson MA, Richmond JA, Weir

RS. The histopathology of the pancreas in type 1 (insulin-depen-dent) diabetes mellitus: a 25-year review of deaths in patients under20 years of age in the United Kingdom. Diabetologia 29: 267–274,1986.

147. Foulis AK, McGill M, Farquharson MA. Insulitis in type 1 (in-sulin-dependent) diabetes mellitus in man–macrophages, lympho-cytes, and interferon-gamma containing cells. J Pathol 165: 97–103,1991.

148. Fraenkel M, Ketzinel-Gilad M, Ariav Y, Pappo O, Karaca M.

mTOR inhibition by rapamycin prevents beta-cell adaptation tohyperglycemia and exacerbates the metabolic state in type 2 dia-betes. Diabetes 57: 945–957, 2008.

149. Frisk G, Friman G, Tuvemo T, Fohlman J, Diderholm H.

Coxsackie B virus IgM in children at onset of type 1 (insulin-dependent) diabetes mellitus: evidence for IgM induction by arecent or current infection. Diabetologia 35: 249–253, 1992.

150. Gale EA. Congenital rubella: citation virus or viral cause of type 1diabetes? Diabetologia 51: 1559–1566, 2008.

151. Gale EA, Bingley PJ, Emmett CL, Collier T, European Nico-

tinamide Diabetes Intervention Trial. European NicotinamideDiabetes Intervention Trial (ENDIT): a randomised controlled trialof intervention before the onset of type 1 diabetes. Lancet 363:925–931, 2004.

152. Gamble DR, Kinsley ML, FitzGerald MG, Bolton R, Taylor

KW. Viral antibodies in diabetes mellitus. Br Med J 3: 627–630,1969.

153. Gao R, Ustinov J, Korsgren O, Otonkoski T. Effects of immu-nosuppressive drugs on in vitro neogenesis of human islets: myco-phenolate mofetil inhibits the proliferation of ductal cells. Am J

Transplant 7: 1021–1026, 2007.154. Genovese MC, Kaine JL, Lowenstein MB, Giudice JD, Baldas-

sare A. Ocrelizumab, a humanized anti-CD20 monoclonal anti-body, in the treatment of patients with rheumatoid arthritis: aphase I/II randomized, blinded, placebo-controlled, and dose-rang-ing study. Arthritis Rheum 58: 2652–2661, 2008.

155. Gepts W. Pathologic anatomy of the pancreas in juvenile diabetesmellitus. Diabetes 14: 619–633, 1965.

156. Gillespie KM, Bain SC, Barnett AH, Bingley PJ, Christie MR,

Gill GV, Gale EA. The rising incidence of childhood type 1 dia-betes and reduced contribution of high-risk HLA haplotypes. Lan-

cet 364: 1699–1700, 2004.157. Giordano C, Galluzzo A, Marco A, Panto F, Amato MP, Ca-

ruso C, Bompiani GD. Increased soluble interleukin-2 receptorlevels in the sera of type 1 diabetic patients. Diabetes Res 8:135–138, 1988.

158. Giordano C, Panto F, Caruso C, Modica MA, Zambito AM.

Interleukin 2 and soluble interleukin 2-receptor secretion defect invitro in newly diagnosed type I diabetic patients. Diabetes 38:310–315, 1989.

159. Goldstein DB. Common genetic variation and human traits. N

Engl J Med 360: 1696–1698, 2009.160. Gordon EJ, Wicker LS, Peterson LB, Serreze DV, Markees

TG. Autoimmune diabetes and resistance to xenograft transplan-tation tolerance in NOD mice. Diabetes 54: 107–115, 2005.

161. Goto A, Takahashi Y, Kishimoto M, Nakajima Y, Nakanishi K,

Kajio H, Noda M. A case of fulminant type 1 diabetes associatedwith significant elevation of mumps titers. Endocr J 55: 561–564,2008.

162. Gottlieb PA. BHT-3021 DNA vaccine trial in type 1 diabetes. In:ADA Conference June 2009, New Orleans LA.

163. Gottlieb PA, Quinlan S, Krause-Steinrauf H, Greenbaum CJ,

Wilson DM. Failure to preserve beta-cell function with mycophe-nolate mofetil and daclizumab combined therapy in patients withnew-onset type 1 diabetes. Diabetes Care 33: 826–832.

164. Grant SF, Qu HQ, Bradfield JP, Marchand L, Kim CE. Fol-low-up analysis of genome-wide association data identifies novelloci for type 1 diabetes. Diabetes 58: 290–295, 2009.

165. Green A, Patterson CC. Trends in the incidence of childhood-onset diabetes in Europe 1989–1998. Diabetologia 44 Suppl 3:B3–B8, 2001.

166. Greenberg SJ, Marcon L, Hurwitz BJ, Waldmann TA, Nelson

DL. Elevated levels of soluble interleukin-2 receptors in multiplesclerosis. N Engl J Med 319: 1019–1020, 1988.

167. Gregori S, Bacchetta R, Passerini L, Levings MK, Roncarolo

MG. Isolation, expansion, and characterization of human naturaland adaptive regulatory T cells. Methods Mol Biol 380: 83–105, 2007.

168. Grill V, Radtke M, Qvigstad E, Kollind M, Bjorklund A. Ben-eficial effects of K-ATP channel openers in diabetes: an update onmechanisms and clinical experiences. Diabetes Obes Metab 11Suppl 4: 143–148, 2009.

169. Guberski DL, Thomas VA, Shek WR, Like AA, Handler ES.

Induction of type I diabetes by Kilham’s rat virus in diabetes-resistant BB/Wor rats. Science 254: 1010–1013, 1991.

170. Guo SW, Magnuson VL, Schiller JJ, Wang X, Wu Y, Ghosh S.

Meta-analysis of vitamin D receptor polymorphisms and type 1diabetes: a HuGE review of genetic association studies. Am J

Epidemiol 164: 711–724, 2006.171. Hadjiyanni I, Baggio LL, Poussier P, Drucker DJ. Exendin-4

modulates diabetes onset in nonobese diabetic mice. Endocrinol-

ogy 149: 1338–1349, 2008.172. Hagopian WA, Lernmark A, Rewers MJ, Simell OG, She JX.

TEDDY–The Environmental Determinants of Diabetes in theYoung: an observational clinical trial. Ann NY Acad Sci 1079:320–326, 2006.



on August 26, 2011


ownloaded from


173. Hakonarson H, Grant SF, Bradfield JP, Marchand L, Kim CE.

A genome-wide association study identifies KIAA0350 as a type 1diabetes gene. Nature 448: 591–594, 2007.

174. Hakonarson H, Qu HQ, Bradfield JP, Marchand L, Kim CE. Anovel susceptibility locus for type 1 diabetes on Chr12q13 identifiedby a genome-wide association study. Diabetes 57: 1143–1146, 2008.

175. Haller MJ, Wasserfall CH, McGrail KM, Cintron M, Brusko

TM. Autologous umbilical cord blood transfusion in very youngchildren with type 1 diabetes. Diabetes Care 32: 2041–2046, 2009.

176. Hamilton-Williams EE, Palmer SE, Charlton B, Slattery RM.

Beta cell MHC class I is a late requirement for diabetes. Proc Natl

Acad Sci USA 100: 6688–6693, 2003.177. Hanafusa T, Miyazaki A, Miyagawa J, Tamura S, Inada M.

Examination of islets in the pancreas biopsy specimens from newlydiagnosed type 1 (insulin-dependent) diabetic patients. Diabetolo-

gia 33: 105–111, 1990.178. Harlan DM, Kenyon NS, Korsgren O, Roep BO, Immunology of

Diabetes Study. Current advances and travails in islet transplan-tation. Diabetes 58: 2175–2184, 2009.

179. Harnaha J, Machen J, Wright M, Lakomy R, Styche A. Inter-leukin-7 is a survival factor for CD4� CD25� T-cells and is ex-pressed by diabetes-suppressive dendritic cells. Diabetes 55: 158–170, 2006.

180. Harrison LC, Honeyman MC. Cow’s milk and type 1 diabetes: thereal debate is about mucosal immune function. Diabetes 48: 1501–1507, 1999.

181. Hering BJ, Wijkstrom M, Graham ML, Hardstedt M, Aasheim

TC. Prolonged diabetes reversal after intraportal xenotransplanta-tion of wild-type porcine islets in immunosuppressed nonhumanprimates. Nat Med 12: 301–303, 2006.

182. Herold KC, Bluestone JA, Montag AG, Parihar A, Wiegner A,

Gress RE, Hirsch R. Prevention of autoimmune diabetes withnonactivating anti-CD3 monoclonal antibody. Diabetes 41: 385–391,1992.

183. Herold KC, Brooks-Worrell B, Palmer J, Dosch HM, Peakman

M. Validity and reproducibility of measurement of islet autoreac-tivity by T-cell assays in subjects with early type 1 diabetes. Dia-

betes 58: 2588–2595, 2009.184. Herold KC, Burton JB, Francois F, Poumian-Ruiz E, Glandt

M, Bluestone JA. Activation of human T cells by FcR nonbindinganti-CD3 mAb, hOKT3gamma1(Ala-Ala). J Clin Invest 111: 409–418, 2003.

185. Herold KC, Gitelman SE, Masharani U, Hagopian W, Bisikir-

ska B. A single course of anti-CD3 monoclonal antibodyhOKT3gamma1(Ala-Ala) results in improvement in C-peptide re-sponses and clinical parameters for at least 2 years after onset oftype 1 diabetes. Diabetes 54: 1763–1769, 2005.

186. Herold KC, Hagopian W, Auger JA, Poumian-Ruiz E, Taylor L.

Anti-CD3 monoclonal antibody in new-onset type 1 diabetes melli-tus. N Engl J Med 346: 1692–1698, 2002.

187. Hilbrands R, Huurman VA, Gillard P, Velthuis JH, De Waele

M. Differences in baseline lymphocyte counts and autoreactivityare associated with differences in outcome of islet cell transplan-tation in type 1 diabetic patients. Diabetes 58: 2267–2276, 2009.

188. Hirsch IB. Clinical review: realistic expectations and practical useof continuous glucose monitoring for the endocrinologist. J Clin

Endocrinol Metab 94: 2232–2238, 2009.189. Ho PP, Fontoura P, Ruiz PJ, Steinman L, Garren H. An immu-

nomodulatory GpG oligonucleotide for the treatment of autoimmu-nity via the innate and adaptive immune systems. J Immunol 171:4920–4926, 2003.

190. Honeyman M. How robust is the evidence for viruses in theinduction of type 1 diabetes? Curr Opin Immunol 17: 616–623,2005.

191. Honeyman MC, Coulson BS, Stone NL, Gellert SA, Goldwater

PN. Association between rotavirus infection and pancreatic isletautoimmunity in children at risk of developing type 1 diabetes.Diabetes 49: 1319–1324, 2000.

192. Honeyman MC, Harrison LC, Drummond B, Colman PG, Tait

BD. Analysis of families at risk for insulin-dependent diabetesmellitus reveals that HLA antigens influence progression to clinicaldisease. Mol Med 1: 576–582, 1995.

193. Honeyman MC, Stone NL, Harrison LC. T-cell epitopes in type1 diabetes autoantigen tyrosine phosphatase IA-2: potential formimicry with rotavirus and other environmental agents. Mol Med 4:231–239, 1998.

194. Hood KK, Peterson CM, Rohan JM, Drotar D. Associationbetween adherence and glycemic control in pediatric type 1 diabe-tes: a meta-analysis. Pediatrics 124: 1171–1179, 2009.

195. Horwitz MS, Bradley LM, Harbertson J, Krahl T, Lee J,

Sarvetnick N. Diabetes induced by Coxsackie virus: initiation bybystander damage and not molecular mimicry. Nat Med 4: 781–785,1998.

196. Horwitz MS, Fine C, Ilic A, Sarvetnick N. Requirements forviral-mediated autoimmune diabetes: beta-cell damage and im-mune infiltration. J Autoimmun 16: 211–217, 2001.

197. Hu CY, Rodriguez-Pinto D, Du W, Ahuja A, Henegariu O.

Treatment with CD20-specific antibody prevents and reverses au-toimmune diabetes in mice. J Clin Invest 117: 3857–3867, 2007.

198. Hummel M, Bonifacio E, Naserke HE, Ziegler AG. Eliminationof dietary gluten does not reduce titers of type 1 diabetes-associ-ated autoantibodies in high-risk subjects. Diabetes Care 25: 1111–1116, 2002.

199. Hummel M, Bonifacio E, Stern M, Dittler J, Schimmel A,

Ziegler AG. Development of celiac disease-associated antibodiesin offspring of parents with type I diabetes. Diabetologia 43: 1005–1011, 2000.

200. Huurman VA, van der Meide PE, Duinkerken G, Willemen S,

Cohen IR, Elias D, Roep BO. Immunological efficacy of heatshock protein 60 peptide DiaPep277 therapy in clinical type Idiabetes. Clin Exp Immunol 152: 488–497, 2008.

201. Hviid A, Stellfeld M, Wohlfahrt J, Melbye M. Childhood vacci-nation and type 1 diabetes. N Engl J Med 350: 1398–1404, 2004.

202. Hyoty H, Hiltunen M, Knip M, Laakkonen M, Vahasalo P. Aprospective study of the role of coxsackie B and other enterovirusinfections in the pathogenesis of IDDM. Childhood Diabetes inFinland (DiMe) Study Group. Diabetes 44: 652–657, 1995.

203. Imagawa A, Hanafusa T, Tamura S, Moriwaki M, Itoh N.

Pancreatic biopsy as a procedure for detecting in situ autoimmunephenomena in type 1 diabetes: close correlation between serolog-ical markers and histological evidence of cellular autoimmunity.Diabetes 50: 1269–1273, 2001.

204. In’t Veld P, Lievens D, De Grijse J, Ling Z, Van der Auwera B.

Screening for insulitis in adult autoantibody-positive organ donors.Diabetes 56: 2400–2404, 2007.

205. Israni N, Goswami R, Kumar A, Rani R. Interaction of vitamin Dreceptor with HLA DRB1 0301 in type 1 diabetes patients fromNorth India. PLoS One 4: e8023, 2009.

206. Itoh N, Hanafusa T, Miyazaki A, Miyagawa J, Yamagata K.

Mononuclear cell infiltration and its relation to the expression ofmajor histocompatibility complex antigens and adhesion mole-cules in pancreas biopsy specimens from newly diagnosed insulin-dependent diabetes mellitus patients. J Clin Invest 92: 2313–2322,1993.

207. Jacob CO, Aiso S, Michie SA, McDevitt HO, Acha-Orbea H.

Prevention of diabetes in nonobese diabetic mice by tumor necro-sis factor (TNF): similarities between TNF-alpha and interleukin 1.Proc Natl Acad Sci USA 87: 968–972, 1990.

208. Jaeger C, Hatziagelaki E, Stroedter A, Becker F, Scherer S.

The Giessen-Bad Oeynhausen family study: improved prediction oftype I diabetes in a low incidence population of relatives usingcombinations of islet autoantibodies in a dual step model. Exp Clin

Endocrinol Diabetes 107: 496–505, 1999.209. Jansen A, Homo-Delarche F, Hooijkaas H, Leenen PJ, Dard-

enne M, Drexhage HA. Immunohistochemical characterization ofmonocytes-macrophages and dendritic cells involved in the initia-tion of the insulitis and beta-cell destruction in NOD mice. Diabetes

43: 667–675, 1994.210. Jenner M, Bradish G, Stiller C, Atkison P. Cyclosporin A treat-

ment of young children with newly-diagnosed type 1 (insulin-de-pendent) diabetes mellitus. London Diabetes Study Group. Diabe-

tologia 35: 884–888, 1992.211. Jiang Y, Nishimura W, Devor-Henneman D, Kusewitt D, Wang

H. Postnatal expansion of the pancreatic beta-cell mass is depen-dent on survivin. Diabetes 57: 2718–2727, 2008.



on August 26, 2011


ownloaded from


212. Jones DB, Hunter NR, Duff GW. Heat-shock protein 65 as a betacell antigen of insulin-dependent diabetes. Lancet 336: 583–585,1990.

213. Juhela S, Hyoty H, Roivainen M, Harkonen T, Putto-Laurila

A, Simell O, Ilonen J. T-cell responses to enterovirus antigens inchildren with type 1 diabetes. Diabetes 49: 1308–1313, 2000.

214. Julier C, Hyer RN, Davies J, Merlin F, Soularue P. Insulin-IGF2region on chromosome 11p encodes a gene implicated in HLA-DR4-dependent diabetes susceptibility. Nature 354: 155–159, 1991.

215. Kabarowski JH, Zhu K, Le LQ, Witte ON, Xu Y. Lysophosphati-dylcholine as a ligand for the immunoregulatory receptor G2A.Science 293: 702–705, 2001.

216. Kared H, Masson A, Adle-Biassette H, Bach JF, Chatenoud L,

Zavala F. Treatment with granulocyte colony-stimulating factorprevents diabetes in NOD mice by recruiting plasmacytoid den-dritic cells and functional CD4(�)CD25(�) regulatory T-cells. Di-

abetes 54: 78–84, 2005.217. Karges W, Hammond-McKibben D, Cheung RK, Visconti M,

Shibuya N, Kemp D, Dosch HM. Immunological aspects of nutri-tional diabetes prevention in NOD mice: a pilot study for the cow’smilk-based IDDM prevention trial. Diabetes 46: 557–564, 1997.

218. Karjalainen J, Martin JM, Knip M, Ilonen J, Robinson BH. Abovine albumin peptide as a possible trigger of insulin-dependentdiabetes mellitus. N Engl J Med 327: 302–307, 1992.

219. Karjalainen JK. Islet cell antibodies as predictive markers forIDDM in children with high background incidence of disease. Di-

abetes 39: 1144–1150, 1990.220. Kasuga A, Harada R, Saruta T. Insulin-dependent diabetes mel-

litus associated with parvovirus B19 infection. Ann Intern Med 125:700–701, 1996.

221. Kato H, Takeuchi O, Sato S, Yoneyama M, Yamamoto M.

Differential roles of MDA5 and RIG-I helicases in the recognition ofRNA viruses. Nature 441: 101–105, 2006.

222. Kaufman DL, Clare-Salzler M, Tian J, Forsthuber T, Ting GS.

Spontaneous loss of T-cell tolerance to glutamic acid decarboxyl-ase in murine insulin-dependent diabetes. Nature 366: 69–72, 1993.

223. Kaufman DL, Erlander MG, Clare-Salzler M, Atkinson MA,

Maclaren NK, Tobin AJ. Autoimmunity to two forms of gluta-mate decarboxylase in insulin-dependent diabetes mellitus. J Clin

Invest 89: 283–292, 1992.224. Kawano Y, Irie J, Nakatani H, Yamada S. Pioglitazone might

prevent the progression of slowly progressive type 1 diabetes.Intern Med 48: 1037–1039, 2009.

225. Kennedy GC, German MS, Rutter WJ. The minisatellite in thediabetes susceptibility locus IDDM2 regulates insulin transcription.Nat Genet 9: 293–298, 1995.

226. Keskinen P, Korhonen S, Kupila A, Veijola R, Erkkila S.

First-phase insulin response in young healthy children at geneticand immunological risk for Type I diabetes. Diabetologia 45: 1639–1648, 2002.

227. Keymeulen B, Candon S, Fafi-Kremer S, Ziegler A, Leruez-

Ville M. Transient Epstein Barr virus reactivation in CD3 mono-clonal antibody-treated patients. Blood 115: 1145–1155, 2010.

228. Keymeulen B, Vandemeulebroucke E, Ziegler AG, Mathieu C,

Kaufman L. Insulin needs after CD3-antibody therapy in new-onset type 1 diabetes. N Engl J Med 352: 2598–2608, 2005.

229. Kilpatrick ES, Rigby AS, Atkin SL. The Diabetes Control andComplications Trial: the gift that keeps giving. Nat Rev Endocrinol

5: 537–545, 2009.230. Kim SJ, Nian C, Doudet DJ, McIntosh CH. Dipeptidyl peptidase

IV inhibition with MK0431 improves islet graft survival in diabeticNOD mice partially via T-cell modulation. Diabetes 58: 641–651,2009.

231. Kim SJ, Nian C, Doudet DJ, McIntosh CH. Inhibition of dipep-tidyl peptidase IV with sitagliptin (MK0431) prolongs islet graftsurvival in streptozotocin-induced diabetic mice. Diabetes 57:1331–1339, 2008.

232. Klemetti P, Savilahti E, Ilonen J, Akerblom HK, Vaarala O.

T-cell reactivity to wheat gluten in patients with insulin-dependentdiabetes mellitus. Scand J Immunol 47: 48–53, 1998.

233. Kostraba JN, Cruickshanks KJ, Lawler-Heavner J, Jobim LF,

Rewers MJ. Early exposure to cow’s milk and solid foods in

infancy, genetic predisposition, and risk of IDDM. Diabetes 42:288–295, 1993.

234. Kroon E, Martinson LA, Kadoya K, Bang AG, Kelly OG. Pan-creatic endoderm derived from human embryonic stem cells gen-erates glucose-responsive insulin-secreting cells in vivo. Nat Bio-

technol 26: 443–452, 2008.235. Kupila A, Muona P, Simell T, Arvilommi P, Savolainen H.

Feasibility of genetic and immunological prediction of type I dia-betes in a population-based birth cohort. Diabetologia 44: 290–297,2001.

236. Kyvik KO, Green A, Beck-Nielsen H. Concordance rates ofinsulin dependent diabetes mellitus: a population based study ofyoung Danish twins. BMJ 311: 913–917, 1995.

237. Lambert AP, Gillespie KM, Thomson G, Cordell HJ, Todd JA,

Gale EA, Bingley PJ. Absolute risk of childhood-onset type 1diabetes defined by human leukocyte antigen class II genotype: apopulation-based study in the United Kingdom. J Clin Endocrinol

Metab 89: 4037–4043, 2004.238. Lampeter EF, Klinghammer A, Scherbaum WA, Heinze E,

Haastert B, Giani G, Kolb H. The Deutsche Nicotinamide Inter-vention Study: an attempt to prevent type 1 diabetes. DENIS Group

Diabetes 47: 980–984, 1998.239. Larsen CP, Pearson TC, Adams AB, Tso P, Shirasugi N. Ratio-

nal development of LEA29Y (belatacept), a high-affinity variant ofCTLA4-Ig with potent immunosuppressive properties. Am J Trans-

plant 5: 443–453, 2005.240. Lazar L, Ofan R, Weintrob N, Avron A, Tamir M. Heat-shock

protein peptide DiaPep277 treatment in children with newly diag-nosed type 1 diabetes: a randomised, double-blind phase II study.Diabetes Metab Res Rev 23: 286–291, 2007.

241. Lee LF, Xu B, Michie SA, Beilhack GF, Warganich T, Turley S,

McDevitt HO. The role of TNF-alpha in the pathogenesis of type 1diabetes in the nonobese diabetic mouse: analysis of dendritic cellmaturation. Proc Natl Acad Sci USA 102: 15995–16000, 2005.

242. Lee SH, Hao E, Savinov AY, Geron I, Strongin AY, Itkin-

Ansari P. Human beta-cell precursors mature into functional in-sulin-producing cells in an immunoisolation device: implicationsfor diabetes cell therapies. Transplantation 87: 983–991, 2009.

243. Lehuen A, Diana J, Zaccone P, Cooke A. Immune cell crosstalkin type 1 diabetes. Nat Rev Immunol 10: 501–513.

244. Lempainen J, Vaarala O, Makela M, Veijola R, Simell O. Inter-play between PTPN22 C1858T polymorphism and cow’s milk for-mula exposure in type 1 diabetes. J Autoimmun 33: 155–164, 2009.

245. Lie BA, Todd JA, Pociot F, Nerup J, Akselsen HE. The predis-position to type 1 diabetes linked to the human leukocyte antigencomplex includes at least one non-class II gene. Am J Hum Genet

64: 793–800, 1999.246. Liston A, Lesage S, Wilson J, Peltonen L, Goodnow CC. Aire

regulates negative selection of organ-specific T cells. Nat Immunol

4: 350–354, 2003.247. Littorin B, Blom P, Scholin A, Arnqvist HJ, Blohme G. Lower

levels of plasma 25-hydroxyvitamin D among young adults at diag-nosis of autoimmune type 1 diabetes compared with control sub-jects: results from the nationwide Diabetes Incidence Study inSweden (DISS). Diabetologia 49: 2847–2852, 2006.

248. Liu S, Wang H, Jin Y, Podolsky R, Reddy MV. IFIH1 polymor-phisms are significantly associated with type 1 diabetes and IFIH1gene expression in peripheral blood mononuclear cells. Hum Mol

Genet 18: 358–365, 2009.249. Ljungberg M, Korpela R, Ilonen J, Ludvigsson J, Vaarala O.

Probiotics for the prevention of beta cell autoimmunity in childrenat genetic risk of type 1 diabetes–the PRODIA study. Ann NY Acad

Sci 1079: 360–364, 2006.250. Lonnrot M, Korpela K, Knip M, Ilonen J, Simell O. Enterovirus

infection as a risk factor for beta-cell autoimmunity in a prospec-tively observed birth cohort: the Finnish Diabetes Prediction andPrevention Study. Diabetes 49: 1314–1318, 2000.

251. Lonnrot M, Salminen K, Knip M, Savola K, Kulmala P. Entero-virus RNA in serum is a risk factor for beta-cell autoimmunity andclinical type 1 diabetes: a prospective study. Childhood Diabetes inFinland (DiMe) Study Group. J Med Virol 61: 214–220, 2000.

252. Lowe CE, Cooper JD, Brusko T, Walker NM, Smyth DJ. Large-scale genetic fine mapping and genotype-phenotype associations



on August 26, 2011


ownloaded from


implicate polymorphism in the IL2RA region in type 1 diabetes. Nat

Genet 39: 1074–1082, 2007.253. Lucassen AM, Julier C, Beressi JP, Boitard C, Froguel P,

Lathrop M, Bell JI. Susceptibility to insulin dependent diabetesmellitus maps to a 4.1 kb segment of DNA spanning the insulin geneand associated VNTR. Nat Genet 4: 305–310, 1993.

254. Ludewig B, Odermatt B, Landmann S, Hengartner H, Zinker-

nagel RM. Dendritic cells induce autoimmune diabetes and main-tain disease via de novo formation of local lymphoid tissue. J Exp

Med 188: 1493–1501, 1998.255. Ludvigsson J. Therapy with GAD in diabetes. Diabetes Metab Res

Rev 25: 307–315, 2009.256. Ludvigsson J, Faresjo M, Hjorth M, Axelsson S, Cheramy M.

GAD treatment and insulin secretion in recent-onset type 1 diabe-tes. N Engl J Med 359: 1909–1920, 2008.

257. Ludvigsson J, Heding LG. Beta-cell function in children withdiabetes. Diabetes 27 Suppl 1: 230–234, 1978.

258. Ludvigsson J, Lindblom B. Human lymphocyte antigen DR typesin relation to early clinical manifestations in diabetic children.Pediatr Res 18: 1239–1241, 1984.

259. Lyons PA, Armitage N, Argentina F, Denny P, Hill NJ. Con-genic mapping of the type 1 diabetes locus, Idd3, to a 780-kb regionof mouse chromosome 3: identification of a candidate segment ofancestral DNA by haplotype mapping. Genome Res 10: 446–453,2000.

260. MacFarlane AJ, Burghardt KM, Kelly J, Simell T, Simell O,

Altosaar I, Scott FW. A type 1 diabetes-related protein fromwheat (Triticum aestivum). cDNA clone of a wheat storage glob-ulin, Glb1, linked to islet damage. J Biol Chem 278: 54–63, 2003.

261. Machen J, Harnaha J, Lakomy R, Styche A, Trucco M, Gian-

noukakis N. Antisense oligonucleotides down-regulating costimu-lation confer diabetes-preventive properties to nonobese diabeticmouse dendritic cells. J Immunol 173: 4331–4341, 2004.

262. Maier LM, Anderson DE, Severson CA, Baecher-Allan C,

Healy B. Soluble IL-2RA levels in multiple sclerosis subjects andthe effect of soluble IL-2RA on immune responses. J Immunol 182:1541–1547, 2009.

263. Maier LM, Lowe CE, Cooper J, Downes K, Anderson DE.

IL2RA genetic heterogeneity in multiple sclerosis and type 1 dia-betes susceptibility and soluble interleukin-2 receptor production.PLoS Genet 5: e1000322, 2009.

264. Maier LM, Wicker LS. Genetic susceptibility to type 1 diabetes.Curr Opin Immunol 17: 601–608, 2005.

265. Makela M, Oling V, Marttila J, Waris M, Knip M, Simell O,

Ilonen J. Rotavirus-specific T cell responses and cytokine mRNAexpression in children with diabetes-associated autoantibodies andtype 1 diabetes. Clin Exp Immunol 145: 261–270, 2006.

266. Malek TR, Bayer AL. Tolerance, not immunity, crucially dependson IL-2. Nat Rev Immunol 4: 665–674, 2004.

267. Malkani S, Nompleggi D, Hansen JW, Greiner DL, Mordes JP,

Rossini AA. Dietary cow’s milk protein does not alter the fre-quency of diabetes in the BB rat. Diabetes 46: 1133–1140, 1997.

268. Mandrup-Poulsen T. The role of interleukin-1 in the pathogenesisof IDDM. Diabetologia 39: 1005–1029, 1996.

269. Mangada J, Pearson T, Brehm MA, Wicker LS, Peterson LB.

Idd loci synergize to prolong islet allograft survival induced bycostimulation blockade in NOD mice. Diabetes 58: 165–173, 2009.

270. Marino E, Grey ST. A new role for an old player: do B cellsunleash the self-reactive CD8� T cell storm necessary for thedevelopment of type 1 diabetes? J Autoimmun 31: 301–305, 2008.

271. Marron MP, Graser RT, Chapman HD, Serreze DV. Functionalevidence for the mediation of diabetogenic T cell responses byHLA-A2.1 MHC class I molecules through transgenic expression inNOD mice. Proc Natl Acad Sci USA 99: 13753–13758, 2002.

272. Martinez X, Kreuwel HT, Redmond WL, Trenney R, Hunter K.

CD8� T cell tolerance in nonobese diabetic mice is restored byinsulin-dependent diabetes resistance alleles. J Immunol 175:1677–1685, 2005.

273. Martinuzzi E, Novelli G, Scotto M, Blancou P, Bach JM. Thefrequency and immunodominance of islet-specific CD8� T-cellresponses change after type 1 diabetes diagnosis and treatment.Diabetes 57: 1312–1320, 2008.

274. Mastrandrea L, Yu J, Behrens T, Buchlis J, Albini C, Fourtner

S, Quattrin T. Etanercept treatment in children with new-onsettype 1 diabetes: pilot randomized, placebo-controlled, double-blindstudy: response to peters. Diabetes Care 32: e154, 2009.

275. Mastrobuoni S, Ubilla M, Cordero A, Herreros J, Rabago G.

Two-dose daclizumab, tacrolimus, mycophenolate mofetil, steroid-free regimen in de novo cardiac transplant recipients: early expe-rience. Transplant Proc 39: 2163–2166, 2007.

276. Mathieu C, Gysemans C, Giulietti A, Bouillon R. Vitamin D anddiabetes. Diabetologia 48: 1247–1257, 2005.

277. Mathieu C, van Etten E, Gysemans C, Decallonne B, Bouillon

R. Seasonality of birth in patients with type 1 diabetes. Lancet 359:1248, 2002.

278. Mathieu C, Waer M, Laureys J, Rutgeerts O, Bouillon R.

Prevention of autoimmune diabetes in NOD mice by 1,25 dihy-droxyvitamin D3. Diabetologia 37: 552–558, 1994.

279. Mathis D, Vence L, Benoist C. �-Cell death during progression todiabetes. Nature 414: 792–798, 2001.

280. Matteucci E, Giampietro O. Dipeptidyl peptidase-4 (CD26):knowing the function before inhibiting the enzyme. Curr Med

Chem 16: 2943–2951, 2009.281. Matthews JB, Staeva TP, Bernstein PL, Peakman M, von

Herrath M, the ITNJTDCTAG. Developing combination immu-notherapies for type 1 diabetes: recommendations from the ITN-JDRF Type 1 Diabetes Combination Therapy Assessment Group.Clin Exp Immunol 160: 176–184, 2010.

282. Matveyenko AV, Georgia S, Bhushan A, Butler PC. Inconsis-tent formation and non function of insulin positive cells frompancreatic endoderm derived from human embryonic stem cells inathymic nude rats. Am J Physiol Endocrinol Metab 2010.

283. Maurer M, Ponath A, Kruse N, Rieckmann P. CTLA4 exon 1dimorphism is associated with primary progressive multiple scle-rosis. J Neuroimmunol 131: 213–215, 2002.

284. Mayer EJ, Hamman RF, Gay EC, Lezotte DC, Savitz DA,

Klingensmith GJ. Reduced risk of IDDM among breast-fed chil-dren. The Colorado IDDM Registry. Diabetes 37: 1625–1632, 1988.

285. McDevitt HO. Discovering the role of the major histocompatibilitycomplex in the immune response. Annu Rev Immunol 18: 1, 2000.

286. Mehta D. Lysophosphatidylcholine: an enigmatic lysolipid. Am J

Physiol Lung Cell Mol Physiol 289: L174–L175, 2005.287. Meier JJ, Bhushan A, Butler AE, Rizza RA, Butler PC. Sus-

tained beta cell apoptosis in patients with long-standing type 1diabetes: indirect evidence for islet regeneration? Diabetologia 48:2221–2228, 2005.

288. Meier JJ, Lin JC, Butler AE, Galasso R, Martinez DS, Butler

PC. Direct evidence of attempted beta cell regeneration in an89-year-old patient with recent-onset type 1 diabetes. Diabetologia

49: 1838–1844, 2006.289. Menser MA, Forrest JM, Honeyman MC, Burgess JA. Letter:

diabetes, HL-A antigens, and congenital rubella. Lancet 2: 1508–1509, 1974.

290. Merani S, Truong W, Emamaullee JA, Toso C, Knudsen LB,

Shapiro AM. Liraglutide, a long-acting human glucagon-like pep-tide 1 analog, improves glucose homeostasis in marginal mass islettransplantation in mice. Endocrinology 149: 4322–4328, 2008.

291. Meylan E, Tschopp J, Karin M. Intracellular pattern recognitionreceptors in the host response. Nature 442: 39–44, 2006.

292. Miao D, Yu L, Eisenbarth GS. Role of autoantibodies in type 1diabetes. Front Biosci 12: 1889–1898, 2007.

293. Mohr SB, Garland CF, Gorham ED, Garland FC. The associa-tion between ultraviolet B irradiance, vitamin D status and inci-dence rates of type 1 diabetes in 51 regions worldwide. Diabetolo-

gia 51: 1391–1398, 2008.294. Monti P, Scirpoli M, Maffi P, Ghidoli N, De Taddeo F. Islet

transplantation in patients with autoimmune diabetes induces ho-meostatic cytokines that expand autoreactive memory T cells. J

Clin Invest 118: 1806–1814, 2008.295. Moore A, Bonner-Weir S, Weissleder R. Noninvasive in vivo

measurement of beta-cell mass in mouse model of diabetes. Dia-

betes 50: 2231–2236, 2001.296. Moriwaki M, Itoh N, Miyagawa J, Yamamoto K, Imagawa A.

Fas and Fas ligand expression in inflamed islets in pancreas sec-



on August 26, 2011


ownloaded from


tions of patients with recent-onset Type I diabetes mellitus. Dia-

betologia 42: 1332–1340, 1999.297. Mory DB, Rocco ER, Miranda WL, Kasamatsu T, Crispim F,

Dib SA. Prevalence of vitamin D receptor gene polymorphismsFokI and BsmI in Brazilian individuals with type 1 diabetes andtheir relation to beta-cell autoimmunity and to remaining beta-cellfunction. Hum Immunol 70: 447–451, 2009.

298. Muhammad BJ, Swift PG, Raymond NT, Botha JL. Partialremission phase of diabetes in children younger than age 10 years.Arch Dis Child 80: 367–369, 1999.

299. Muir A, Peck A, Clare-Salzler M, Song YH, Cornelius J. Insulinimmunization of nonobese diabetic mice induces a protective in-sulitis characterized by diminished intraislet interferon-gammatranscription. J Clin Invest 95: 628–634, 1995.

300. Murai M, Turovskaya O, Kim G, Madan R, Karp CL, Cheroutre

H, Kronenberg M. Interleukin 10 acts on regulatory T cells tomaintain expression of the transcription factor Foxp3 and suppres-sive function in mice with colitis. Nat Immunol 10: 1178–1184,2009.

301. Nanto-Salonen K, Kupila A, Simell S, Siljander H, Salonsaari

T. Nasal insulin to prevent type 1 diabetes in children with HLAgenotypes and autoantibodies conferring increased risk of disease:a double-blind, randomised controlled trial. Lancet 372: 1746–1755,2008.

302. Nejentsev S, Gombos Z, Laine AP, Veijola R, Knip M. Non-classII HLA gene associated with type 1 diabetes maps to the 240-kbregion near HLA-B. Diabetes 49: 2217–2221, 2000.

303. Nejentsev S, Howson JM, Walker NM, Szeszko J, Field SF.

Localization of type 1 diabetes susceptibility to the MHC class Igenes HLA-B and HLA-A. Nature 450: 887–892, 2007.

304. Nejentsev S, Walker N, Riches D, Egholm M, Todd JA. Rarevariants of IFIH1, a gene implicated in antiviral responses, protectagainst type 1 diabetes. Science 324: 387–389, 2009.

305. Nerup J, Platz P, Andersen OO, Christy M, Lyngsoe J. HL-Aantigens and diabetes mellitus. Lancet 2: 864–866, 1974.

306. Nir T, Melton DA, Dor Y. Recovery from diabetes in mice by betacell regeneration. J Clin Invest 117: 2553–2561, 2007.

307. Nistico L, Buzzetti R, Pritchard LE, Van der Auwera B, Gio-

vannini C. The CTLA-4 gene region of chromosome 2q33 is linkedto, associated with, type 1 diabetes. Belgian Diabetes Registry.Hum Mol Genet 5: 1075–1080, 1996.

308. Noble JA, Valdes AM, Bugawan TL, Apple RJ, Thomson G,

Erlich HA. The HLA class I A locus affects susceptibility to type 1diabetes. Hum Immunol 63: 657–664, 2002.

309. Noble JA, Valdes AM, Cook M, Klitz W, Thomson G, Erlich

HA. The role of HLA class II genes in insulin-dependent diabetesmellitus: molecular analysis of 180 Caucasian, multiplex families.Am J Hum Genet 59: 1134–1148, 1996.

310. Norris JM, Barriga K, Klingensmith G, Hoffman M, Eisen-

barth GS, Erlich HA, Rewers M. Timing of initial cereal exposurein infancy and risk of islet autoimmunity. JAMA 290: 1713–1720,2003.

311. Norris JM, Beaty B, Klingensmith G, Yu L, Hoffman M. Lack ofassociation between early exposure to cow’s milk protein andbeta-cell autoimmunity. Diabetes Autoimmunity Study in the Young(DAISY). JAMA 276: 609–614, 1996.

312. Norris JM, Yin X, Lamb MM, Barriga K, Seifert J. Omega-3polyunsaturated fatty acid intake and islet autoimmunity in chil-dren at increased risk for type 1 diabetes. JAMA 298: 1420–1428,2007.

313. Not T, Tommasini A, Tonini G, Buratti E, Pocecco M. Undiag-nosed coeliac disease and risk of autoimmune disorders in subjectswith Type I diabetes mellitus. Diabetologia 44: 151–155, 2001.

314. Ogawa N, Minamimura K, Kodaka T, Maki T. Short administra-tion of polyclonal anti-T cell antibody (ALS) in NOD mice withextensive insulitis prevents subsequent development of autoim-mune diabetes. J Autoimmun 26: 225–231, 2006.

315. Oikarinen M, Tauriainen S, Honkanen T, Oikarinen S, Vuori

K. Detection of enteroviruses in the intestine of type 1 diabeticpatients. Clin Exp Immunol 151: 71–75, 2008.

316. Oikawa Y, Shimada A, Kasuga A, Morimoto J, Osaki T. Sys-temic administration of IL-18 promotes diabetes development inyoung nonobese diabetic mice. J Immunol 171: 5865–5875, 2003.

317. Olcott AP, Tian J, Walker V, Dang H, Middleton B. Antigen-based therapies using ignored determinants of beta cell antigenscan more effectively inhibit late-stage autoimmune disease in dia-betes-prone mice. J Immunol 175: 1991–1999, 2005.

318. Orban T, Farkas K, Jalahej H, Kis J, Treszl A. Autoantigen-specific regulatory T cells induced in patients with type 1 diabetesmellitus by insulin B-chain immunotherapy. J Autoimmun 34: 408–415, 2010.

319. Orban T, Sosenko JM, Cuthbertson D, Krischer JP, Skyler JS.

Pancreatic islet autoantibodies as predictors of type 1 diabetes inthe Diabetes Prevention Trial-Type 1 (DPT-1). Diabetes Care 32:2269–2274, 2009.

320. Oresic M, Simell S, Sysi-Aho M, Nanto-Salonen K, Seppanen-

Laakso T. Dysregulation of lipid and amino acid metabolism pre-cedes islet autoimmunity in children who later progress to type 1diabetes. J Exp Med 205: 2975–2984, 2008.

321. Ortqvist E, Bjork E, Wallensteen M, Ludvigsson J, Aman J.

Temporary preservation of beta-cell function by diazoxide treat-ment in childhood type 1 diabetes. Diabetes Care 27: 2191–2197,2004.

322. Ott PA, Dittrich MT, Herzog BA, Guerkov R, Gottlieb PA. Tcells recognize multiple GAD65 and proinsulin epitopes in humantype 1 diabetes, suggesting determinant spreading. J Clin Immunol

24: 327–339, 2004.323. Owerbach D, Lernmark A, Platz P, Ryder LP, Rask L, Peterson

PA, Ludvigsson J. HLA-D region beta-chain DNA endonucleasefragments differ between HLA-DR identical healthy and insulin-dependent diabetic individuals. Nature 303: 815–817, 1983.

324. Pak CY, Eun HM, McArthur RG, Yoon JW. Association of cyto-megalovirus infection with autoimmune type 1 diabetes. Lancet 2:1–4, 1988.

325. Panagiotopoulos C, Qin H, Tan R, Verchere CB. Identificationof a beta-cell-specific HLA class I restricted epitope in type 1diabetes. Diabetes 52: 2647–2651, 2003.

326. Panina-Bordignon P, Lang R, van Endert PM, Benazzi E, Felix

AM. Cytotoxic T cells specific for glutamic acid decarboxylase inautoimmune diabetes. J Exp Med 181: 1923–1927, 1995.

327. Parker MJ, Xue S, Alexander JJ, Wasserfall CH, Campbell-

Thompson ML. Immune depletion with cellular mobilization im-parts immunoregulation and reverses autoimmune diabetes innonobese diabetic mice. Diabetes 58: 2277–2284, 2009.

328. Patterson CC, Dahlquist GG, Gyurus E, Green A, Soltesz G,

Group ES. Incidence trends for childhood type 1 diabetes inEurope during 1989–2003 and predicted new cases 2005–20: amulticentre prospective registration study. Lancet 373: 2027–2033,2009.

329. Patterson K, Chandra RS, Jenson AB. Congenital rubella, insu-litis, diabetes mellitus in an infant. Lancet 1: 1048–1049, 1981.

330. Paxson JA, Weber JG, Kulczycki A Jr. Cow’s milk-free diet doesnot prevent diabetes in NOD mice. Diabetes 46: 1711–1717, 1997.

331. Peakman M. CD8 and cytotoxic T cells in type 1 diabetes. Novartis

Found Symp 292: 113–129, 2008.332. Pearson T, Weiser P, Markees TG, Serreze DV, Wicker LS.

Islet allograft survival induced by costimulation blockade in NODmice is controlled by allelic variants of Idd3. Diabetes 53: 1972–1978, 2004.

333. Pearson TA, Manolio TA. How to interpret a genome-wide asso-ciation study. JAMA 299: 1335–1344, 2008.

334. Pescovitz MD, Greenbaum CJ, Krause-Steinrauf H, Becker

DJ, Gitelman SE. Rituximab, B-lymphocyte depletion, and pres-ervation of beta-cell function. N Engl J Med 361: 2143–2152, 2009.

335. Pickersgill LM, Mandrup-Poulsen TR. The anti-interleukin-1 intype 1 diabetes action trial–background and rationale. Diabetes

Metab Res Rev 25: 321–324, 2009.336. Pileggi A, Ricordi C, Kenyon NS, Froud T, Baidal DA. Twenty

years of clinical islet transplantation at the Diabetes ResearchInstitute–University of Miami. Clin Transplant 177–204, 2004.

337. Pinkse GG, Tysma OH, Bergen CA, Kester MG, Ossendorp F.

Autoreactive CD8 T cells associated with beta cell destruction intype 1 diabetes. Proc Natl Acad Sci USA 102: 18425–18430, 2005.

338. Pittenger GL, Taylor-Fishwick DA, Johns RH, Burcus N, Ko-

suri S, Vinik AI. Intramuscular injection of islet neogenesis-asso-



on August 26, 2011


ownloaded from


ciated protein peptide stimulates pancreatic islet neogenesis inhealthy dogs. Pancreas 34: 103–111, 2007.

339. Podolin PL, Wilusz MB, Cubbon RM, Pajvani U, Lord CJ.

Differential glycosylation of interleukin 2, the molecular basis forthe NOD Idd3 type 1 diabetes gene? Cytokine 12: 477–482, 2000.

340. Pozzilli P, Visalli N, Boccuni ML, Baroni MG, Buzzetti R.

Randomized trial comparing nicotinamide and nicotinamide pluscyclosporin in recent onset insulin-dependent diabetes (IMDIAB 1).The IMDIAB Study Group. Diabetes Med 11: 98–104, 1994.

341. Pratley RE, Gilbert M. Targeting incretins in type 2 diabetes: roleof GLP-1 receptor agonists and DPP-4 inhibitors. Rev Diabet Stud

5: 73–94, 2008.342. Pugliese A, Zeller M, Fernandez A Jr, Zalcberg LJ, Bartlett

RJ. The insulin gene is transcribed in the human thymus andtranscription levels correlated with allelic variation at the INSVNTR-IDDM2 susceptibility locus for type 1 diabetes. Nat Genet 15:293–297, 1997.

343. Putnam AL, Brusko TM, Lee MR, Liu W, Szot GL. Expansion ofhuman regulatory T-cells from patients with type 1 diabetes. Dia-

betes 58: 652–662, 2009.344. Putnam AL, Vendrame F, Dotta F, Gottlieb PA. CD4�CD25high

regulatory T cells in human autoimmune diabetes. J Autoimmun

24: 55–62, 2005.345. Qu HQ, Montpetit A, Ge B, Hudson TJ, Polychronakos C.

Toward further mapping of the association between the IL2RAlocus and type 1 diabetes. Diabetes 56: 1174–1176, 2007.

346. Qu HQ, Verlaan DJ, Ge B, Lu Y, Lam KC. A cis-acting regulatoryvariant in the IL2RA locus. J Immunol 183: 5158–5162, 2009.

347. Rabinovitch A, Suarez-Pinzon WL, Shapiro AM, Rajotte RV,

Power R. Combination therapy with sirolimus and interleukin-2prevents spontaneous and recurrent autoimmune diabetes in NODmice. Diabetes 51: 638–645, 2002.

348. Rafaeloff R, Pittenger GL, Barlow SW, Qin XF, Yan B. Cloningand sequencing of the pancreatic islet neogenesis associated pro-tein (INGAP) gene and its expression in islet neogenesis in ham-sters. J Clin Invest 99: 2100–2109, 1997.

349. Ramiya VK, Lan MS, Wasserfall CH, Notkins AL, Maclaren

NK. Immunization therapies in the prevention of diabetes. J Auto-

immun 10: 287–292, 1997.350. Raz I, Avron A, Tamir M, Metzger M, Symer L. Treatment of

new-onset type 1 diabetes with peptide DiaPep277 is safe andassociated with preserved beta-cell function: extension of a ran-domized, double-blind, phase II trial. Diabetes Metab Res Rev 23:292–298, 2007.

351. Raz I, Elias D, Avron A, Tamir M, Metzger M, Cohen IR.

Beta-cell function in new-onset type 1 diabetes and immunomodu-lation with a heat-shock protein peptide (DiaPep277): a random-ised, double-blind, phase II trial. Lancet 358: 1749–1753, 2001.

352. Reddy S, Cheung CC, Chai RC, Rodrigues JA. Persistence ofresidual beta cells and islet autoimmunity during increasing dura-tion of diabetes in NOD mice and experimental approaches towardreversing new-onset disease with bioactive peptides. Ann NY Acad

Sci 1150: 171–176, 2008.353. Redondo MJ, Babu S, Zeidler A, Orban T, Yu L. Specific human

leukocyte antigen DQ influence on expression of antiislet autoan-tibodies and progression to type 1 diabetes. J Clin Endocrinol

Metab 91: 1705–1713, 2006.354. Redondo MJ, Jeffrey J, Fain PR, Eisenbarth GS, Orban T.

Concordance for islet autoimmunity among monozygotic twins. N

Engl J Med 359: 2849–2850, 2008.355. Redondo MJ, Yu L, Hawa M, Mackenzie T, Pyke DA, Eisen-

barth GS, Leslie RD. Heterogeneity of type I diabetes: analysis ofmonozygotic twins in Great Britain and the United States. Diabe-

tologia 44: 354–362, 2001.356. Richardson SJ, Willcox A, Bone AJ, Foulis AK, Morgan NG.

The prevalence of enteroviral capsid protein vp1 immunostainingin pancreatic islets in human type 1 diabetes. Diabetologia 52:1143–1151, 2009.

357. Richter W, Mertens T, Schoel B, Muir P, Ritzkowsky A,

Scherbaum WA, Boehm BO. Sequence homology of the diabetes-associated autoantigen glutamate decarboxylase with coxsackieB4–2C protein and heat shock protein 60 mediates no molecularmimicry of autoantibodies. J Exp Med 180: 721–726, 1994.

358. Roberto Gianani MCT, Suparna Sarkar A, Wasserfall C,

Pugliese A, Kent C, Hering B, West E, Steck A, Bonner-Weir

S, Atkinson MA, Coppieters K, von Herrath M, Eisenbarth

GS. Dimorphic Histopathology of Long Standing Childhood-OnsetDiabetes. Diabetologia. In press.

359. Robertson RP. Islet transplantation as a treatment for diabetes: awork in progress. N Engl J Med 350: 694–705, 2004.

360. Robles DT, Eisenbarth GS, Wang T, Erlich HA, Bugawan TL.

Millennium award recipient contribution. Identification of childrenwith early onset and high incidence of anti-islet autoantibodies.Clin Immunol 102: 217–224, 2002.

361. Roep BO. The role of T-cells in the pathogenesis of Type 1 diabe-tes: from cause to cure. Diabetologia 46: 305–321, 2003.

362. Roep BO, Atkinson MA, van Endert PM, Gottlieb PA, Wilson

SB, Sachs JA. Autoreactive T cell responses in insulin-dependent(Type 1) diabetes mellitus. Report of the first international work-shop for standardization of T cell assays. J Autoimmun 13: 267–282, 1999.

363. Roivainen M, Klingel K. Role of enteroviruses in the pathogene-sis of type 1 diabetes. Diabetologia 52: 995–996, 2009.

364. Rooman I, Bouwens L. Combined gastrin and epidermal growthfactor treatment induces islet regeneration and restores normogly-caemia in C57Bl6/J mice treated with alloxan. Diabetologia 47:259–265, 2004.

365. Rooman I, Lardon J, Bouwens L. Gastrin stimulates beta-cellneogenesis and increases islet mass from transdifferentiated butnot from normal exocrine pancreas tissue. Diabetes 51: 686–690,2002.

366. Rosenberg L, Lipsett M, Yoon JW, Prentki M, Wang R. Apentadecapeptide fragment of islet neogenesis-associated proteinincreases beta-cell mass and reverses diabetes in C57BL/6J mice.Ann Surg 240: 875–884, 2004.

367. Rosu V, Ahmed N, Paccagnini D, Gerlach G, Fadda G. Specificimmunoassays confirm association of Mycobacterium avium

Subsp. paratuberculosis with type-1 but not type-2 diabetes melli-tus. PLoS One 4: e4386, 2009.

368. Rothe H, Hausmann A, Casteels K, Okamura H, Kurimoto M.

IL-18 inhibits diabetes development in nonobese diabetic mice bycounterregulation of Th1-dependent destructive insulitis. J Immu-

nol 163: 1230–1236, 1999.369. Rother KI, Brown RJ, Morales MM, Wright E, Duan Z. Effect of

ingested interferon-alpha on beta-cell function in children withnew-onset type 1 diabetes. Diabetes Care 32: 1250–1255, 2009.

370. Rother KI, Spain LM, Wesley RA, Digon BJ, Baron A 3rd.

Effects of exenatide alone and in combination with daclizumab onbeta-cell function in long-standing type 1 diabetes. Diabetes Care

32: 2251–2257, 2009.371. Rutella S, Bonanno G, Pierelli L, Mariotti A, Capoluongo E.

Granulocyte colony-stimulating factor promotes the generation ofregulatory DC through induction of IL-10 and IFN-alpha. Eur J

Immunol 34: 1291–1302, 2004.372. Ryan EA, Paty BW, Senior PA, Bigam D, Alfadhli E. Five-year

follow-up after clinical islet transplantation. Diabetes 54: 2060–2069, 2005.

373. Sadeharju K, Knip M, Virtanen SM, Savilahti E, Tauriainen S.

Maternal antibodies in breast milk protect the child from entero-virus infections. Pediatrics 119: 941–946, 2007.

374. Sadeharju K, Lonnrot M, Kimpimaki T, Savola K, Erkkila S.

Enterovirus antibody levels during the first two years of life inprediabetic autoantibody-positive children. Diabetologia 44: 818–823, 2001.

375. Saisho Y, Harris PE, Butler AE, Galasso R, Gurlo T, Rizza RA,

Butler PC. Relationship between pancreatic vesicular monoaminetransporter 2 (VMAT2) and insulin expression in human pancreas.J Mol Histol 39: 543–551, 2008.

376. Salomon B, Bluestone JA. Complexities of CD28/B7: CTLA-4costimulatory pathways in autoimmunity and transplantation.Annu Rev Immunol 19: 225–252, 2001.

377. Sanda S, Roep BO, von Herrath M. Islet antigen specific IL-10�immune responses but not CD4�CD25�FoxP3� cells at diagnosispredict glycemic control in type 1 diabetes. Clin Immunol 127:138–143, 2008.



on August 26, 2011


ownloaded from


378. Saudek F, Brogren CH, Manohar S. Imaging the beta-cell mass:why and how. Rev Diabet Stud 5: 6–12, 2008.

379. Saudek F, Havrdova T, Boucek P, Karasova L, Novota P,

Skibova J. Polyclonal anti-T-cell therapy for type 1 diabetes mel-litus of recent onset. Rev Diabet Stud 1: 80–88, 2004.

380. Savilahti E, Ormala T, Saukkonen T, Sandini-Pohjavuori U,

Kantele JM. Jejuna of patients with insulin-dependent diabetesmellitus (IDDM) show signs of immune activation. Clin Exp Im-

munol 116: 70–77, 1999.381. Schloot NC, Meierhoff G, Lengyel C, Vandorfi G, Takacs J.

Effect of heat shock protein peptide DiaPep277 on beta-cell func-tion in paediatric and adult patients with recent-onset diabetesmellitus type 1: two prospective, randomized, double-blind phase IItrials. Diabetes Metab Res Rev 23: 276–285, 2007.

382. Schloot NC, Willemen SJ, Duinkerken G, Drijfhout JW, de

Vries RR, Roep BO. Molecular mimicry in type 1 diabetes mellitusrevisited: T-cell clones to GAD65 peptides with sequence homologyto Coxsackie or proinsulin peptides do not crossreact with homol-ogous counterpart. Hum Immunol 62: 299–309, 2001.

383. Schlosser M, Banga JP, Madec AM, Binder KA, Strebelow M.

Dynamic changes of GAD65 autoantibody epitope specificities inindividuals at risk of developing type 1 diabetes. Diabetologia 48:922–930, 2005.

384. Schwartz RF, Neu J, Schatz D, Atkinson MA, Wasserfall C.

Comment on: Brugman S et al. (2006) Antibiotic treatment partiallyprotects against type 1 diabetes in the Bio-Breeding diabetes-pronerat Is the gut flora involved in the development of type 1 diabetes?Diabetologia 49: 2105–2108. Diabetologia 50: 220–221, 2007.

385. Sechi LA, Rosu V, Pacifico A, Fadda G, Ahmed N, Zanetti S.

Humoral immune responses of type 1 diabetes patients to Myco-

bacterium avium subsp. paratuberculosis lend support to the in-fectious trigger hypothesis. Clin Vaccine Immunol 15: 320–326,2008.

386. Serreze DV, Chapman HD, Varnum DS, Hanson MS, Reifsny-

der PC. B lymphocytes are essential for the initiation of T cell-mediated autoimmune diabetes: analysis of a new “speed congenic”stock of NOD. Ig mu null mice. J Exp Med 184: 2049–2053, 1996.

387. Serreze DV, Ottendorfer EW, Ellis TM, Gauntt CJ, Atkinson

MA. Acceleration of type 1 diabetes by a coxsackievirus infectionrequires a preexisting critical mass of autoreactive T-cells in pan-creatic islets. Diabetes 49: 708–711, 2000.

388. Serreze DV, Silveira PA. The role of B lymphocytes as keyantigen-presenting cells in the development of T cell-mediatedautoimmune type 1 diabetes. Curr Dir Autoimmun 6: 212–227,2003.

389. Seyfert-Margolis V, Gisler TD, Asare AL, Wang RS, Dosch

HM. Analysis of T-cell assays to measure autoimmune responses insubjects with type 1 diabetes: results of a blinded controlled study.Diabetes 55: 2588–2594, 2006.

390. Sgouroudis E, Albanese A, Piccirillo CA. Impact of protectiveIL-2 allelic variants on CD4� Foxp3� regulatory T cell function insitu and resistance to autoimmune diabetes in NOD mice. J Im-

munol 181: 6283–6292, 2008.391. Shapiro AM, Lakey JR, Ryan EA, Korbutt GS, Toth E. Islet

transplantation in seven patients with type 1 diabetes mellitususing a glucocorticoid-free immunosuppressive regimen. N Engl J

Med 343: 230–238, 2000.392. Shapiro AM, Ricordi C, Hering BJ, Auchincloss H, Lindblad R.

International trial of the Edmonton protocol for islet transplanta-tion. N Engl J Med 355: 1318–1330, 2006.

393. Sheehy MJ, Scharf SJ, Rowe JR, Neme de Gimenez MH,

Meske LM, Erlich HA, Nepom BS. A diabetes-susceptible HLAhaplotype is best defined by a combination of HLA-DR and -DQalleles. J Clin Invest 83: 830–835, 1989.

394. Sherry NA, Kushner JA, Glandt M, Kitamura T, Brillantes

AM, Herold KC. Effects of autoimmunity and immune therapy onbeta-cell turnover in type 1 diabetes. Diabetes 55: 3238–3245, 2006.

395. Shigemoto T, Kageyama M, Hirai R, Zheng J, Yoneyama M,

Fujita T. Identification of loss of function mutations in humangenes encoding RIG-I and MDA5: implications for resistance totype I diabetes. J Biol Chem 284: 13348–13354, 2009.

396. Shoda LK, Young DL, Ramanujan S, Whiting CC, Atkinson

MA. A comprehensive review of interventions in the NOD mouseand implications for translation. Immunity 23: 115–126, 2005.

397. Silveira PA, Serreze DV, Grey ST. Invasion of the killer B’s intype 1 diabetes. Front Biosci 12: 2183–2193, 2007.

398. Simon G, Parker M, Ramiya V, Wasserfall C, Huang Y. Murineantithymocyte globulin therapy alters disease progression in NODmice by a time-dependent induction of immunoregulation. Diabetes

57: 405–414, 2008.399. Singal DP, Blajchman MA. Histocompatibility (HL-A) antigens,

lymphocytotoxic antibodies and tissue antibodies in patients withdiabetes mellitus. Diabetes 22: 429–432, 1973.

400. Skowera A, Ellis RJ, Varela-Calvino R, Arif S, Huang GC.

CTLs are targeted to kill beta cells in patients with type 1 diabetesthrough recognition of a glucose-regulated preproinsulin epitope. J

Clin Invest 118: 3390–3402, 2008.401. Skyler JS, Krischer JP, Wolfsdorf J, Cowie C, Palmer JP.

Effects of oral insulin in relatives of patients with type 1 diabetes:The Diabetes Prevention Trial–Type 1. Diabetes Care 28: 1068–1076, 2005.

402. Smyth D, Cooper JD, Collins JE, Heward JM, Franklyn JA.

Replication of an association between the lymphoid tyrosine phos-phatase locus (LYP/PTPN22) with type 1 diabetes: evidence for itsrole as a general autoimmunity locus. Diabetes 53: 3020–3023,2004.

403. Smyth DJ, Cooper JD, Bailey R, Field S, Burren O. A genome-wide association study of nonsynonymous SNPs identifies a type 1diabetes locus in the interferon-induced helicase (IFIH1) region.Nat Genet 38: 617–619, 2006.

404. Smyth DJ, Howson JM, Payne F, Maier LM, Bailey R. Analysisof polymorphisms in 16 genes in type 1 diabetes that have beenassociated with other immune-mediated diseases. BMC Med Genet

7: 20, 2006.405. Srikanta S, Ganda OP, Gleason RE, Jackson RA, Soeldner JS,

Eisenbarth GS. Pre-type I diabetes. Linear loss of beta cell re-sponse to intravenous glucose. Diabetes 33: 717–720, 1984.

406. Staeva-Vieira T, Peakman M, von Herrath M. Translationalmini-review series on type 1 diabetes: Immune-based therapeuticapproaches for type 1 diabetes. Clin Exp Immunol 148: 17–31,2007.

407. Stock PG, Bluestone JA. Beta-cell replacement for type I diabe-tes. Annu Rev Med 55: 133–156, 2004.

408. Suarez-Pinzon WL, Cembrowski GS, Rabinovitch A. Combina-tion therapy with a dipeptidyl peptidase-4 inhibitor and a protonpump inhibitor restores normoglycaemia in non-obese diabeticmice. Diabetologia 52: 1680–1682, 2009.

409. Suarez-Pinzon WL, Lakey JR, Brand SJ, Rabinovitch A. Com-bination therapy with epidermal growth factor and gastrin inducesneogenesis of human islet �-cells from pancreatic duct cells and anincrease in functional �-cell mass. J Clin Endocrinol Metab 90:3401–3409, 2005.

410. Suarez-Pinzon WL, Lakey JR, Rabinovitch A. Combinationtherapy with glucagon-like peptide-1 and gastrin induces beta-cellneogenesis from pancreatic duct cells in human islets transplantedin immunodeficient diabetic mice. Cell Transplant 17: 631–640,2008.

411. Suarez-Pinzon WL, Power RF, Yan Y, Wasserfall C, Atkinson

M, Rabinovitch A. Combination therapy with glucagon-like pep-tide-1 and gastrin restores normoglycemia in diabetic NOD mice.Diabetes 57: 3281–3288, 2008.

412. Suarez-Pinzon WL, Rabinovitch A. Combination therapy withepidermal growth factor and gastrin delays autoimmune diabetesrecurrence in nonobese diabetic mice transplanted with syngeneicislets. Transplant Proc 40: 529–532, 2008.

413. Suarez-Pinzon WL, Yan Y, Power R, Brand SJ, Rabinovitch A.

Combination therapy with epidermal growth factor and gastrinincreases beta-cell mass and reverses hyperglycemia in diabeticNOD mice. Diabetes 54: 2596–2601, 2005.

414. Sumnik Z, Kolouskova S, Malcova H, Vavrinec J, Venhacova J,

Lebl J, Cinek O. High prevalence of coeliac disease in siblings ofchildren with type 1 diabetes. Eur J Pediatr 164: 9–12, 2005.

415. Surh CD, Boyman O, Purton JF, Sprent J. Homeostasis ofmemory T cells. Immunol Rev 211: 154–163, 2006.



on August 26, 2011


ownloaded from


416. Sutherland APR, Van Belle T, Wurster AL, Suto A, Michaud

M. Interleukin-21 is required for the development of type 1 diabetesin NOD mice. Diabetes 58: 1144–1155, 2009.

417. Sutmuller RP, den Brok MH, Kramer M, Bennink EJ, Toonen

LW. Toll-like receptor 2 controls expansion and function of regu-latory T cells. J Clin Invest 116: 485–494, 2006.

418. Tack CJ, Kleijwegt FS, Van Riel PL, Roep BO. Development oftype 1 diabetes in a patient treated with anti-TNF-alpha therapy foractive rheumatoid arthritis. Diabetologia 52: 1442–1444, 2009.

419. Tanaka S, Nishida Y, Aida K, Maruyama T, Shimada A. Entero-virus infection, CXC chemokine ligand 10 (CXCL10), and CXCR3circuit: a mechanism of accelerated beta-cell failure in fulminanttype 1 diabetes. Diabetes 58: 2285–2291, 2009.

420. Tang Q, Adams JY, Penaranda C, Melli K, Piaggio E. Centralrole of defective interleukin-2 production in the triggering of isletautoimmune destruction. Immunity 28: 687–697, 2008.

421. Tang Q, Bluestone JA. The Foxp3� regulatory T cell: a jack of alltrades, master of regulation. Nat Immunol 9: 239–244, 2008.

422. Tang Q, Henriksen KJ, Bi M, Finger EB, Szot G. In vitro-expanded antigen-specific regulatory T cells suppress autoimmunediabetes. J Exp Med 199: 1455–1465, 2004.

423. Tarantal AF, Lee CC and Itkin-Ansari P. Real-time biolumines-cence imaging of macroencapsulated fibroblasts reveals allograftprotection in rhesus monkeys (Macaca mulatta). Transplantation

88: 38–41, 2009.424. Tarbell KV, Yamazaki S, Olson K, Toy P, Steinman RM. CD25�

CD4� T cells, expanded with dendritic cells presenting a singleautoantigenic peptide, suppress autoimmune diabetes. J Exp Med

199: 1467–1477, 2004.425. Tauriainen S, Oikarinen S, Oikarinen M, Hyoty H. Enterovi-

ruses in the pathogenesis of type 1 diabetes. Semin Immuno-

pathol. In press.426. Teft WA, Kirchhof MG, Madrenas J. A molecular perspective of

CTLA-4 function. Annu Rev Immunol 24: 65–97, 2006.427. Terkeltaub R, Sundy JS, Schumacher HR, Murphy F, Book-

binder S. The interleukin 1 inhibitor rilonacept in treatment ofchronic gouty arthritis: results of a placebo-controlled, monose-quence crossover, non-randomised, single-blind pilot study. Ann

Rheum Dis 68: 1613–1617, 2009.428. Thomas J, Liu F, Link DC. Mechanisms of mobilization of hema-

topoietic progenitors with granulocyte colony-stimulating factor.Curr Opin Hematol 9: 183–189, 2002.

429. Thorsby E, Ronningen KS. Particular HLA-DQ molecules play adominant role in determining susceptibility or resistance to type 1(insulin-dependent) diabetes mellitus. Diabetologia 36: 371–377,1993.

430. Tian J, Lehmann PV, Kaufman DL. T cell cross-reactivity be-tween coxsackievirus and glutamate decarboxylase is associatedwith a murine diabetes susceptibility allele. J Exp Med 180: 1979–1984, 1994.

431. Tiittanen M, Westerholm-Ormio M, Verkasalo M, Savilahti E,

Vaarala O. Infiltration of forkhead box P3-expressing cells in smallintestinal mucosa in coeliac disease but not in type 1 diabetes. Clin

Exp Immunol 152: 498–507, 2008.432. Tisch R, Yang XD, Singer SM, Liblau RS, Fugger L, McDevitt

HO. Immune response to glutamic acid decarboxylase correlateswith insulitis in non-obese diabetic mice. Nature 366: 72–75, 1993.

433. Todd JA, Bell JI, McDevitt HO. HLA-DQ beta gene contributesto susceptibility and resistance to insulin-dependent diabetes mel-litus. Nature 329: 599–604, 1987.

434. Todd JA, Walker NM, Cooper JD, Smyth DJ, Downes K. Ro-bust associations of four new chromosome regions from genome-wide analyses of type 1 diabetes. Nat Genet 39: 857–864, 2007.

435. Tracy S, Drescher KM, Chapman NM, Kim KS, Carson SD.

Toward testing the hypothesis that group B coxsackieviruses(CVB) trigger insulin-dependent diabetes: inoculating nonobesediabetic mice with CVB markedly lowers diabetes incidence. J

Virol 76: 12097–12111, 2002.436. Tuvemo T, Dahlquist G, Frisk G, Blom L, Friman G, Landin-

Olsson M, Diderholm H. The Swedish childhood diabetes studyIII: IgM against coxsackie B viruses in newly diagnosed type 1(insulin-dependent) diabetic children: no evidence of increasedantibody frequency. Diabetologia 32: 745–747, 1989.

437. Ueda H, Howson JM, Esposito L, Heward J, Snook H. Associ-ation of the T-cell regulatory gene CTLA4 with susceptibility toautoimmune disease. Nature 423: 506–511, 2003.

438. Ugrasbul F, Moore WV, Tong PY, Kover KL. Prevention ofdiabetes: effect of mycophenolate mofetil and anti-CD25 on onsetof diabetes in the DRBB rat. Pediatr Diabetes 9: 596–601, 2008.

439. Undlien DE, Friede T, Rammensee HG, Joner G, Dahl-Jor-

gensen K. HLA-encoded genetic predisposition in IDDM: DR4subtypes may be associated with different degrees of protection.Diabetes 46: 143–149, 1997.

440. Urbanek-Ruiz I, Ruiz PJ, Paragas V, Garren H, Steinman L,

Fathman CG. Immunization with DNA encoding an immunodom-inant peptide of insulin prevents diabetes in NOD mice. Clin Im-

munol 100: 164–171, 2001.441. Vaarala O, Atkinson MA, Neu J. The “perfect storm” for type 1

diabetes: the complex interplay between intestinal microbiota, gutpermeability, and mucosal immunity. Diabetes 57: 2555–2562, 2008.

442. Vafiadis P, Bennett ST, Todd JA, Nadeau J, Grabs R. Insulinexpression in human thymus is modulated by INS VNTR alleles atthe IDDM2 locus. Nat Genet 15: 289–292, 1997.

443. Vaidya B, Pearce SH, Charlton S, Marshall N, Rowan AD. Anassociation between the CTLA4 exon 1 polymorphism and earlyrheumatoid arthritis with autoimmune endocrinopathies. Rheuma-

tology 41: 180–183, 2002.444. Valdes-Gonzalez RA, Dorantes LM, Garibay GN, Bracho-

Blanchet E, Mendez AJ. Xenotransplantation of porcine neonatalislets of Langerhans and Sertoli cells: a 4-year study. Eur J Endo-

crinol 153: 419–427, 2005.445. Valdes AM, Erlich HA, Noble JA. Human leukocyte antigen class

I B and C loci contribute to Type 1 Diabetes (T1D) susceptibilityand age at T1D onset. Hum Immunol 66: 301–313, 2005.

446. Valdes AM, Wapelhorst B, Concannon P, Erlich HA, Thomson

G, Noble JA. Extended DR3-D6S273-HLA-B haplotypes are asso-ciated with increased susceptibility to type 1 diabetes in US Cau-casians. Tissue Antigens 65: 115–119, 2005.

447. Van Belle T, von Herrath M. Immunosuppression in islet trans-plantation. J Clin Invest 118: 1625–1628, 2008.

448. Van Belle TL, Juntti T, Liao J, von Herrath MG. Pre-existingautoimmunity determines type 1 diabetes outcome after Flt3-ligandtreatment. J Autoimmun 34: 445–452, 2010.

449. Van der Windt DJ, Bottino R, Casu A, Campanile N, Smetanka

C. Long-term controlled normoglycemia in diabetic non-humanprimates after transplantation with hCD46 transgenic porcine is-lets. Am J Transplant 9: 2716–2726, 2009.

450. Van Kampen CA, van de Linde P, Duinkerken G, van Schip JJ,

Roelen DL. Alloreactivity against repeated HLA mismatches ofsequential islet grafts transplanted in non-uremic type 1 diabetespatients. Transplantation 80: 118–126, 2005.

451. Vang T, Congia M, Macis MD, Musumeci L, Orru V. Autoim-mune-associated lymphoid tyrosine phosphatase is a gain-of-func-tion variant. Nat Genet 37: 1317–1319, 2005.

452. Varela-Calvino R, Ellis R, Sgarbi G, Dayan CM, Peakman M.

Characterization of the T-cell response to coxsackievirus B4: evi-dence that effector memory cells predominate in patients with type1 diabetes. Diabetes 51: 1745–1753, 2002.

453. Vella A, Cooper JD, Lowe CE, Walker N, Nutland S. Localiza-tion of a type 1 diabetes locus in the IL2RA/CD25 region by use oftag single-nucleotide polymorphisms. Am J Hum Genet 76: 773–779, 2005.

454. Vijayakrishnan L, Slavik JM, Illes Z, Greenwald RJ, Rainbow

D. An autoimmune disease-associated CTLA-4 splice variant lack-ing the B7 binding domain signals negatively in T cells. Immunity

20: 563–575, 2004.455. Villasenor J, Benoist C, Mathis D. AIRE and APECED: molecu-

lar insights into an autoimmune disease. Immunol Rev 204: 156–164, 2005.

456. Vlad G, Ho EK, Vasilescu ER, Fan J, Liu Z. Anti-CD25 treatmentand FOXP3-positive regulatory T cells in heart transplantation.Transpl Immunol 18: 13–21, 2007.

457. Volta U, Bonazzi C, Pisi E, Salardi S, Cacciari E. Antigliadinand antireticulin antibodies in coeliac disease and at onset ofdiabetes in children. Lancet 2: 1034–1035, 1987.



on August 26, 2011


ownloaded from


458. Voltarelli JC, Couri CE. Stem cell transplantation for type 1diabetes mellitus. Diabetol Metab Syndr 1: 4, 2009.

459. Voltarelli JC, Couri CE, Stracieri AB, Oliveira MC, Moraes

DA. Autologous nonmyeloablative hematopoietic stem cell trans-plantation in newly diagnosed type 1 diabetes mellitus. JAMA 297:1568–1576, 2007.

460. Von Herrath M, Chan A. How can we improve the translationallandscape for a faster cure of type 1 diabetes? J Clin Invest 119:1061–1065, 2009.

461. Von Herrath M, Nepom GT. Remodeling rodent models to mimichuman type 1 diabetes. Eur J Immunol 39: 2049–2054, 2009.

462. Von Herrath M, Sanda S, Herold K. Type 1 diabetes as a relaps-ing-remitting disease? Nat Rev Immunol 7: 988–994, 2007.

463. Von Herrath MG, Fujinami RS, Whitton JL. Microorganismsand autoimmunity: making the barren field fertile? Nat Rev Micro-

biol 1: 151–157, 2003.464. Vreugdenhil GR, Geluk A, Ottenhoff TH, Melchers WJ, Roep

BO, Galama JM. Molecular mimicry in diabetes mellitus: thehomologous domain in coxsackie B virus protein 2C and isletautoantigen GAD65 is highly conserved in the coxsackie B-likeenteroviruses and binds to the diabetes associated HLA-DR3 mol-ecule. Diabetologia 41: 40–46, 1998.

465. Wallace C, Smyth DJ, Maisuria-Armer M, Walker NM, Todd

JA, Clayton DG. The imprinted DLK1-MEG3 gene region on chro-mosome 14q32.2 alters susceptibility to type 1 diabetes. Nat Genet

42: 68–71, 2009.466. Walter M, Philotheou A, Bonnici F, Ziegler AG, Jimenez R,

Group NBIS. No effect of the altered peptide ligand NBI-6024 onbeta-cell residual function and insulin needs in new-onset type 1diabetes. Diabetes Care 32: 2036–2040, 2009.

467. Wang JP, Cerny A, Asher DR, Kurt-Jones EA, Bronson RT,

Finberg RW. MDA5 and MAVS mediate type I interferon responsesto coxsackie B virus. J Virol 84: 254–260.

468. Waterhouse P, Penninger JM, Timms E, Wakeham A, Shahin-

ian A. Lymphoproliferative disorders with early lethality in micedeficient in Ctla-4. Science 270: 985–988, 1995.

469. Wen L, Ley RE, Volchkov PY, Stranges PB, Avanesyan L.

Innate immunity and intestinal microbiota in the development ofType 1 diabetes. Nature 455: 1109–1113, 2008.

470. Wenzlau JM, Juhl K, Yu L, Moua O, Sarkar SA. The cationefflux transporter ZnT8 (Slc30A8) is a major autoantigen in humantype 1 diabetes. Proc Natl Acad Sci USA 104: 17040–17045, 2007.

471. Westerholm-Ormio M, Vaarala O, Pihkala P, Ilonen J,

Savilahti E. Immunologic activity in the small intestinal mucosa ofpediatric patients with type 1 diabetes. Diabetes 52: 2287–2295,2003.

472. Wicker LS, Chamberlain G, Hunter K, Rainbow D, Howlett S.

Fine mapping, gene content, comparative sequencing, and expres-sion analyses support Ctla4 and Nramp1 as candidates for Idd5.1and Idd5 2 in the nonobese diabetic mouse. J Immunol 173: 164–173, 2004.

473. Wicker LS, Clark J, Fraser HI, Garner VE, Gonzalez-Munoz A.

Type 1 diabetes genes and pathways shared by humans and NODmice. J Autoimmun 25 Suppl: 29–33, 2005.

474. Wildin RS, Ramsdell F, Peake J, Faravelli F, Casanova JL.

X-linked neonatal diabetes mellitus, enteropathy and endocrinop-athy syndrome is the human equivalent of mouse scurfy. Nat Genet

27: 18–20, 2001.475. Willcox A, Richardson SJ, Bone AJ, Foulis AK, Morgan NG.

Analysis of islet inflammation in human type 1 diabetes. Clin Exp

Immunol 155: 173–181, 2009.476. Wu J, Katrekar A, Honigberg LA, Smith AM, Conn MT. Iden-

tification of substrates of human protein-tyrosine phosphatasePTPN22. J Biol Chem 281: 11002–11010, 2006.

477. Xu D, Alegre ML, Varga SS, Rothermel AL, Collins AM. In vitrocharacterization of five humanized OKT3 effector function variantantibodies. Cell Immunol 200: 16–26, 2000.

478. Xu G, Stoffers DA, Habener JF, Bonner-Weir S. Exendin-4stimulates both beta-cell replication and neogenesis, resulting inincreased beta-cell mass and improved glucose tolerance in dia-betic rats. Diabetes 48: 2270–2276, 1999.

479. Xue S, Wasserfall CH, Parker M, Brusko TM, McGrail S.

Exendin-4 therapy in NOD mice with new-onset diabetes increasesregulatory T cell frequency. Ann NY Acad Sci 1150: 152–156, 2008.

480. Yang XD, McDevitt HO. Role of TNF-alpha in the development ofautoimmunity and the pathogenesis of insulin-dependent diabetesmellitus in NOD mice. Circ Shock 43: 198–201, 1994.

481. Yao Q, Nishiuchi R, Kitamura T, Kersey JH. Human leukemiaswith mutated FLT3 kinase are synergistically sensitive to FLT3 andHsp90 inhibitors: the key role of the STAT5 signal transductionpathway. Leukemia 19: 1605–1612, 2005.

482. Yoon JW, Austin M, Onodera T, Notkins AL. Isolation of a virusfrom the pancreas of a child with diabetic ketoacidosis. N Engl J

Med 300: 1173–1179, 1979.483. Yoon JW, London WT, Curfman BL, Brown RL, Notkins AL.

Coxsackie virus B4 produces transient diabetes in nonhuman pri-mates. Diabetes 35: 712–716, 1986.

484. Yoon JW, Onodera T, Notkins AL. Virus-induced diabetes mel-litus. XV. Beta cell damage and insulin-dependent hyperglycemia inmice infected with coxsackie virus B4. J Exp Med 148: 1068–1080,1978.

485. You S, Chen C, Lee WH, Brusko T, Atkinson M, Liu CP.

Presence of diabetes-inhibiting, glutamic acid decarboxylase-spe-cific, IL-10-dependent, regulatory T cells in naive nonobese diabeticmice. J Immunol 173: 6777–6785, 2004.

486. You S, Leforban B, Garcia C, Bach JF, Bluestone JA, Chaten-

oud L. Adaptive TGF-beta-dependent regulatory T cells controlautoimmune diabetes and are a privileged target of anti-CD3 anti-body treatment. Proc Natl Acad Sci USA 104: 6335–6340, 2007.

487. Yu J, Hu K, Smuga-Otto K, Tian S, Stewart R, Slukvin II,

Thomson JA. Human induced pluripotent stem cells free of vectorand transgene sequences. Science 324: 797–801, 2009.

488. Yu L, Robles DT, Abiru N, Kaur P, Rewers M, Kelemen K,

Eisenbarth GS. Early expression of antiinsulin autoantibodies ofhumans and the NOD mouse: evidence for early determination ofsubsequent diabetes. Proc Natl Acad Sci USA 97: 1701–1706, 2000.

489. Zahr E, Molano RD, Pileggi A, Ichii H, Jose SS. Rapamycinimpairs in vivo proliferation of islet beta-cells. Transplantation 84:1576–1583, 2007.

490. Zanin-Zhorov A, Nussbaum G, Franitza S, Cohen IR, Lider O.

T cells respond to heat shock protein 60 via TLR2: activation ofadhesion and inhibition of chemokine receptors. FASEB J 17:1567–1569, 2003.

491. Zdravkovic V, Hamilton JK, Daneman D, Cummings EA. Pio-glitazone as adjunctive therapy in adolescents with type 1 diabetes.J Pediatr 149: 845–849, 2006.

492. Zhang N, Su D, Qu S, Tse T, Bottino R. Sirolimus is associatedwith reduced islet engraftment and impaired beta-cell function.Diabetes 55: 2429–2436, 2006.

493. Zheng XX, Sanchez-Fueyo A, Sho M, Domenig C, Sayegh MH,

Strom TB. Favorably tipping the balance between cytopathic andregulatory T cells to create transplantation tolerance. Immunity

19: 503–514, 2003.494. Zhou Q, Brown J, Kanarek A, Rajagopal J, Melton DA. In vivo

reprogramming of adult pancreatic exocrine cells to beta-cells.Nature 455: 627–632, 2008.

495. Zhou S, Cerny AM, Zacharia A, Fitzgerald KA, Kurt-Jones EA,

Finberg RW. Induction and inhibition of the type I IFN responsesby distinct components of Lymphocytic Choriomeningitis Virus(LCMV). J Virol 84: 9452–9462, 2010.

496. Zhou X, Bailey-Bucktrout SL, Jeker LT, Penaranda C, Mar-

tinez-Llordella M. Instability of the transcription factor Foxp3leads to the generation of pathogenic memory T cells in vivo. Nat

Immunol 10: 1000–1007, 2009.497. Ziegler AG, Schmid S, Huber D, Hummel M, Bonifacio E. Early

infant feeding and risk of developing type 1 diabetes-associatedautoantibodies. JAMA 290: 1721–1728, 2003.

498. Zorina TD, Subbotin VM, Bertera S, Alexander AM, Haluszc-

zak C. Recovery of the endogenous beta cell function in the NODmodel of autoimmune diabetes. Stem Cells 21: 377–388, 2003.



on August 26, 2011


ownloaded from


doi:10.1152/physrev.00003.2010 91:79-118, 2011.Physiol RevTom L. Van Belle, Ken T. Coppieters and Matthias G. Von HerrathTherapeutic StrategiesType 1 Diabetes: Etiology, Immunology, and

You might find this additional info useful...

488 articles, 166 of which can be accessed free at:This article cites http://physrev.physiology.org/content/91/1/79.full.html#ref-list-1

2 other HighWire hosted articlesThis article has been cited by

[PDF] [Full Text] [Abstract]

, August 2, 2011; 108 (31): 12857-12862.PNASYing-Hua Yang, Rajesh Rajaiah, Erkki Ruoslahti and Kamal D. MoudgilPeptides targeting inflamed synovial vasculature attenuate autoimmune arthritis

[PDF] [Full Text] [Abstract], August 12, 2011; 366 (1575): 2307-2311.Phil. Trans. R. Soc. B

D. A. MeltonUsing stem cells to study and possibly treat type 1 diabetes

including high resolution figures, can be found at:Updated information and services http://physrev.physiology.org/content/91/1/79.full.html

can be found at:Physiological Reviewsabout Additional material and information http://www.the-aps.org/publications/prv

This infomation is current as of August 26, 2011.

website at http://www.the-aps.org/.MD 20814-3991. Copyright © 2011 by the American Physiological Society. ISSN: 0031-9333, ESSN: 1522-1210. Visit ourpublished quarterly in January, April, July, and October by the American Physiological Society, 9650 Rockville Pike, Bethesda

provides state of the art coverage of timely issues in the physiological and biomedical sciences. It isPhysiological Reviews

on August 26, 2011


ownloaded from

http://physrev.physiology.org/content/91/1/79.full.html#ref-list-1

http://rstb.royalsocietypublishing.org/content/366/1575/2307.abstract.html

http://rstb.royalsocietypublishing.org/content/366/1575/2307.full.html

http://rstb.royalsocietypublishing.org/content/366/1575/2307.full.pdf

http://www.pnas.org/content/108/31/12857.abstract.html

http://www.pnas.org/content/108/31/12857.full.html

http://www.pnas.org/content/108/31/12857.full.pdf

http://physrev.physiology.org/content/91/1/79.full.html


Autoimmune Destruction of Pancreatic b Cells

Ji-Won Yoon* and Hee-Sook Jun

Type 1 diabetes results from the destruction of insulin-producing pancreatic b cells by a b cell–specificautoimmune process. b Cell autoantigens, macrophages, dendritic cells, B lymphocytes, and Tlymphocytes have been shown to be involved in the pathogenesis of autoimmune diabetes. b Cellautoantigens are thought to be released from b cells by cellular turnover or damage and are processedand presented to T helper cells by antigen-presenting cells. Macrophages and dendritic cells are the firstcell types to infiltrate the pancreatic islets. Naive CD4+ T cells that circulate in the blood and lymphoidorgans, including the pancreatic lymph nodes, may recognize major histocompatibility complex and bcell peptides presented by dendritic cells and macrophages in the islets. These CD4+ T cells can beactivated by interleukin (IL)-12 released frommacrophages and dendritic cells. While this process takesplace, b cell antigen-specific CD8+ T cells are activated by IL-2 produced by the activated TH1 CD4+

Tcells, differentiate into cytotoxic Tcells and are recruited into the pancreatic islets. These activated TH1CD4+ Tcells and CD8+ cytotoxic Tcells are involved in the destruction of b cells. In addition, b cells canalso be damaged by granzymes and perforin released from CD8+ cytotoxic T cells and by solublemediators such as cytokines and reactive oxygen molecules released from activated macrophages in theislets. Thus, activated macrophages, TH1 CD4+ T cells, and b cell-cytotoxic CD8+ T cells act syner-gistically to destroy b cells, resulting in autoimmune type 1 diabetes.

Keywords: autoimmune diabetes, destruction of b cells, b cell autoantigens, T cells, macrophages/den-dritic cells

INTRODUCTION

Type 1 diabetes results from insulin deficiency causedby the loss of insulin-producing pancreatic b cells,1–3

generally develops in the young, and accounts forapproximately 5%–10% of the diabetic populationworldwide. The development of type 1 diabetes isthe consequence of progressive b cell destruction byautoimmune processes during an asymptomatic periodthat often extends over many years. Genetic suscepti-bility is believed to be a prerequisite for the de-velopment of type 1 diabetes.4,5 However, studies onthe development of the disease in identical twins

showed that their concordance rate is only 40%,6

suggesting that environmental or nongenetic factorscontribute to the development of the disease. Histo-logic analysis of the pancreas from patients withrecent-onset type 1 diabetes revealed an infiltrationof the islets of Langerhans by mononuclear cells,7

which were later identified as T and B lymphocytes,monocytes/macrophages, and natural killer (NK)cells.8,9 As well, circulating islet-reactive autoanti-bodies10 and islet-reactive T cells have been found inpatients with type 1 diabetes.11–14

The pathogenesis of autoimmune type 1 diabeteshas been extensively studied using 2 animal models,the nonobese diabetic (NOD) mouse and the diabetes-prone BioBreeding (DP-BB) rat, which have greatlyenhanced our understanding of the pathogenesis ofhuman type 1 diabetes. Although the exact mecha-nisms involved in the initiation and progression ofb cell destruction are still not clear, it is generallybelieved that b cell autoantigens, macrophages, dendritic

Rosalind Franklin Comprehensive Diabetes Center, Department ofPathology, Chicago Medical School, North Chicago, IL.*Address for correspondence: Rosalind Franklin ComprehensiveDiabetes Center, Chicago Medical School, 3333 Green Bay Road,North Chicago, IL 60064. E-mail: [email protected]

American Journal of Therapeutics 12, 580–591 (2005)

1075-2765 ! 2005 Lippincott Williams & Wilkins

cells, B lymphocytes, and T lymphocytes are involvedin the b cell–specific autoimmune process.15–19 b Cellautoantigens are processed by macrophages, dendriticcells, or B cells in the pancreatic islets and presented toautoreactive CD4+ T cells in the peripheral lymphoidsystem. These autoreactive CD4+ T cells are activatedand secrete cytokines, which can activate b cell–specificcytotoxic CD8+ T cells. The activated T cells are re-cruited to the pancreatic islets and produce cytokines,which further activate macrophages and other T cells,contributing to the destruction of b cells (Fig. 1). In thisreview, we briefly discuss the possible roles of islet cellautoantigens, macrophages, and T cells in the autoim-mune destruction of pancreatic b cells.

ISLET CELL AUTOANTIGENS

Pancreatic b cell autoantigens are the targets of immune-mediated destruction of b cells. Therefore, antigen-specific immune reactions are believed to be involved inthe b cell destruction. One of the most common immu-nologic markers of humans and animals with autoim-mune diabetes is the presence of autoantibodies and

autoreactive T cells directed against b cell autoantigens.These autoantigens include insulin,20 glutamic acid decar-boxylase (GAD)65,21 tyrosine phosphatase, insulinomaantigen (IA)-2 and IA-2b,22,23 carboxypeptidase-H,24

islet cell antigen (ICA)-69,25 GM gangliosides,26 a 38-kdautoantigen,27 and SOX13.28 Autoantibodies againstthese b cell autoantigens are not believed to have apathogenic role. However, the risk of developing dia-betes is strongly related to the number of autoantibodymarkers, that is, the presence of 2 or more autoanti-bodies gives a higher probability of developing thedisease than the presence of a single autoantibody.29

Glutamic acid decarboxylase (GAD)

GAD, which is a synthetic enzyme of the inhibitoryneurotransmitter g-aminobutyric acid, has been themost extensively studied b cell autoantigen. Sera fromtype 1 diabetic patients was found to precipitate a 64-kd protein, which was later identified as GAD. GAD65antibody is found in 70%–75% of type 1 diabetespatients and 1%–2% of healthy individuals.30 The anti-GAD autoantibodies in type 1 diabetes patients arepredominantly directed to a conformational epitope of

FIGURE 1. Etiology of type 1 dia-betes. Genetic predisposition ap-pears to be a prerequisite for thedevelopment of type 1 diabetes.Environmental factors such as vi-ruses, toxins, and diet may be in-volved in the clinical expression ofgenetic susceptibility. Once b cell–specific autoimmunity has devel-oped, autoimmune-mediated de-struction of b cells results in theonset of type 1 diabetes. Autoan-tigens released from b cells areprocessed by antigen-presentingcells (APCs)andpresented tohelperT cells (Th cells) in associationwith MHC class II molecules. IL-12released from APCs activates TH1-type CD4+ T cells, causing the im-mune balance between effectorand regulatory cells to breakdown.TH1 cells produce IL-2, which acti-vates b cell–specific precytotoxicT cells (Pre CTL) to become cyto-toxic (CTL), and IFN-g, which may

cause macrophages (MØ) to become cytotoxic. These cytotoxic macrophages release b cell–cytotoxic cytokines includingIL-1b, TNF-a, and IFN-g, and free radicals. TH1 cells also secrete cytokines that are directly cytotoxic to b cells. b Cell antigen-specific CD8+ cytotoxic T cells (CTL) recognize antigens expressed on b cells in associationwithMHC class Imolecules. TheseCTLs release granzyme and perforin (cytolysin), which are toxic to b cells. In addition, Fas- and TNFR-mediated apoptosis areinvolved inb cell destruction. In thisway,macrophages, T cells, and cytokines synergistically act todestroyb cells, resulting inthe development of autoimmune type 1 diabetes.

Autoimmune Destruction of Pancreatic b Cells 581

American Journal of Therapeutics (2005) 12(6)

GAD, and the major antigenic region in humans hasbeen identified as the middle and carboxyterminalregion of GAD65.31,32

Although the presence of anti-GAD65 antibodies isa highly predictive marker for the development of type1 diabetes in humans, contradictory results have beenreported regarding the correlation between the pres-ence of anti-GAD antibodies and diabetes in NODmice. One study found that anti-GAD65 antibodieswere detected in the early stages of the disease,33

whereas another study found that the presence of anti-GAD65 antibodies was not a prerequisite for thedevelopment of diabetes in NOD/Lt and NOD/Wehimice, which have a higher and lower incidence ofdiabetes, respectively, than NOD mice.34 The latterstudy suggested that a strong humoral response toGAD may actually be associated with less destructivepathology, as indicated by the negative correlation be-tween insulitis and anti-GAD antibody levels.

It was found that the initial immune response directedagainst pancreatic islets in NOD mice is a TH1 responseagainst a confined region of GAD (amino acids 509–528and 524–543) and that later responses are directedagainst another region of GAD and other autoanti-gens.35 Immunization of NOD mice with purified GADresulted in the tolerization of GAD-reactive T cells andblocked the T cell response to other b cell antigens, thuspreventing diabetes.35 There is also direct evidence thatGAD-reactive T cells are diabetogenic in NOD mice.36,37

GAD-specific CD4+ T cells have also been observed inrecent-onset type 1 diabetic patients and their relativesat risk to develop diabetes.38–40 In addition to CD4+

T cells, MHC class I human lymphocyte antigen (HLA)-A*0201-restricted CD8+ cytotoxic T cells reactive againstGAD were identified in recently diagnosed diabeticpatients and in high-risk subjects, but not in healthycontrol subjects expressing HLA-A*0201.41 These resultssuggest that GAD may be an important target antigenand GAD-reactive T cells may play a pathogenic role inthe destruction of pancreatic b cells in animal modelsand in human type 1 diabetes. However, contradictoryresults regarding the role of GAD in the pathogenesis oftype 1 diabetes have been reported. Some GAD-reactiveT cells do not have the ability to induce diabetes, andTcell responses to GAD-derived peptides were observedin mice resistant to type 1 diabetes,42,43 suggesting thatperipheral tolerance to GAD is not associated withprotection from diabetes. In addition, molecular mim-icry between GAD and Coxsackie B4 virus has beenhypothesized for the development of type 1 diabetes,but controversies still exist.44,45

To investigate the role of GAD in the pathogenesis ofautoimmune diabetes, several lines of transgenic micehave been established in which the expression of GAD

has been manipulated. Hyperexpression of GAD65 inb cells of NOD mice resulted in a lower incidence ofdiabetes in one line of transgenic mice and no dif-ference in the incidence of diabetes in another trans-genic line, as compared with nontransgenic controlNOD mice. A quantitative difference in the expressionof GAD between the 2 lines might have accounted forprevention of diabetes.46 A transgenic NODmouse linethat expresses GAD65 in all tissues showed anaccelerated onset and increased incidence of diabetesas compared with control NOD mice.47 Interestingly,b cell–specific suppression of GAD65 and GAD67expression prevented insulitis and diabetes in anti-sense-GAD transgenic mice backcrossed with NODmice.48,49 These results suggest that the expression ofGAD in pancreatic b cells is involved in the modulationof b cell–specific autoimmunity. Recently, it wasreported that expression of a modified form of GADunder the control of the invariant chain promoter inNOD mice induced tolerance to GAD65, but failed toprevent insulitis and diabetes,50 suggesting that in-hibition of the generation of GAD65-reactive T cellsmay not be the mechanism for the prevention ofdiabetes in GAD-suppressed NOD mice. However,GAD67-reactive T cells were not tolerized in this study.Because GAD67 is the dominant form in mice and anylevel of GAD expression in NOD mice results in thedevelopment of diabetes, further study on the toleriza-tion of both GAD65 and GAD67-reactive T cellsremains to be done to determine the precise role ofGAD-reactive T cells in the development of diabetes.

Another possibility regarding the prevention of b celldestruction in GAD-suppressed NOD mice is that sup-pression of GAD in b cells may have rendered theb cells more resistant to destruction by T cells and/ormacrophages because normal b cells expressing GADshow expression of tissue transglutaminase, which isknown to promote apoptosis of cells, whereas tissuetransglutaminase is suppressed in GAD-suppressedb cells. A further possibility is that a diabetes-resistantgene from the strain of origin might have been trans-mitted to the transgenic offspring, as these antisenseGAD transgenic mice were produced using eggs from(SJL 3 C57BL/6) F2 mice, which are diabetes resistant.Systemic GAD65 knockout mice backcrossed withNOD mice for 4 generations still developed diabetesand insulitis similar to wild-type NOD mice.51 How-ever, it is difficult to draw any definite conclusionsfrom this study, as mouse b cells predominantlyexpress GAD67 and very low levels of GAD65, andthese GAD65 knockout mice still express GAD67.Therefore, b cell–specific conditional GAD65/67knockout NOD mice are essential to find whetherthe expression of GAD in b cells truly plays

582 Yoon and Jun


a critical role in the initiation of b cell–specific auto-immune diabetes.

Insulin

Insulin was the first diabetes-related autoantigen tobe discovered. Autoantibodies to insulin are found in50%–70% of type 1 diabetic children52 and are the firstsign of an ongoing autoimmune process.53 It was alsoreported that the early appearance of insulin auto-antibody predicts early development of diabetes inNODmice.54 Insulin B chain-specific CD4+ Tcell clonesaccelerate diabetes in NOD.scid mice.55 Transgenicexpression of mouse proinsulin II under a MHC class IIgene promoter prevented the development of diabetesin NOD mice.56 These results suggest that insulin auto-antigen plays an important role in the developmentof type 1 diabetes. However, the pathogenic role of in-sulin autoantibody and insulin-reactive T cells needsfurther investigation.

IA-2/IA-2b

IA-2 and IA-2b are newly discovered members of theprotein tyrosine phosphatase family and are consideredto be major autoantigens of type 1 diabetes. Initially,autoantibodies directed against a 37/40-kd trypticfragment were detected in sera of type 1 diabetespatients.57 Later, IA-2 was identified as the precursor ofthe 40-kd fragment,22,58 and IA-2b (also known asphogrin) was identified as the precursor of the 37-kdfragment.23,59 IA-2 and IA-2b have a high degree ofhomology and are located in secretory granules, but therole of these molecules in b cells is not clear. Autoanti-bodies to IA-2 are present in 70%–80% of type 1 diabeticchildren,60 making them useful serological markersfor human type 1 diabetes. In addition, T cells fromdiabetic patients respond to IA-2 antigen.11,12 However,the precise role of the IA-2 and IA-2b antigens in thepathogenesis of type 1 diabetes is unknown.

MACROPHAGES

Macrophages and dendritic cells are among the firstcell types to infiltrate the pancreatic islets during thedisease process in NOD mice and BB rats.61–63 Thisinfiltration precedes invasion of the islets by T lympho-cytes, NK cells, and B lymphocytes.64 It was reportedthat B cells also play an important role as antigen-presenting cells (APCs) for b cell antigens in the processof the development of b cell–specific autoimmunity inNOD mice.65 The activation of b cell–reactive T cellsmay be initiated from the processing and presentationof b cell autoantigens by APCs in the pancreatic islets.APCs that present b cell autoantigenic peptides in

conjunction with MHC molecules migrate to the pan-creatic lymph nodes where they are recognized bycirculating naive b cell–autoreactive T cells. The naiveT cells are then activated, migrate to the pancreaticislets, and are further activated by re-encountering thecognate b cell antigens (Fig. 2).Inactivation of macrophages in NOD mice or BB rats

significantly prevented the development of diabetes.66–68

Further studies showed that macrophages are requiredfor the development of effector T cells that destroyb cells in NOD mice69 (Fig. 3). T cells in macrophage-depleted NOD mouse recipients did not destroy trans-planted NOD islets, indicating that T cells in a macro-phage-depleted environment lose their ability todifferentiate into cytotoxic T cells. The level of IL-4secreted from TH2 cells was increased, whereas thelevel of IFN-g secreted from TH1 cells was decreasedin macrophage-depleted NOD mice, indicating thatdown-regulation of the TH1 immune response and up-regulation of the TH2 immune response may be factorsin the loss of the ability of T cells in a macrophage-depleted environment to kill b cells.69 T cell prolif-eration to islet antigens or GAD was significantlydecreased when splenocytes from macrophage-depleted NOD mice were used as APCs, suggesting

FIGURE2. Activationofb cell–autoreactiveTcells byb cellautoantigens and APCs in autoimmune type 1 diabetes.Naive b cell–autoreactive T cells may circulate through theblood and lymphoid organs. b Cell autoantigens areprocessed and presented by APCs (dendritic cells [DC],macrophages, and B cells) in the pancreatic islets. MatureAPCs, which present b cell autoantigen peptides inconjunctionwithMHCmolecules, migrate to the pancreaticlymph nodes. Naive b cell–autoreactive T cells in thecirculation recognize the MHC/b cell autoantigen peptidecomplex on APCs and become activated. These activated bcell–reactive T cells access the pancreatic islets and re-encounter cognate b cell antigens and become reactivated.The activated b cell–reactive T cells can then kill b cells.



that depletion of macrophages results in the reductionof APC function and subsequently down-regulation ofantigen-specific CD4+ T cell activation (Fig. 3).

Splenic Tcells frommacrophage-depleted NODmiceshowed a significant decrease in the expression of FasLand perforin as comparedwith macrophage-containingNODmice.69 This result suggests that macrophages arerequired for the activation of cytotoxic T cells that candestroy pancreatic b cells. Taken together, we suggestthat IL-12 secreted by macrophages may activate TH1-type CD4+ T cells and subsequently the IL-2 and IFN-gproduced by these activated CD4+ T cells may assistin maximizing the activation of CD8+ T cells. Down-regulation of islet cell–specific T cell activation may beanother factor contributing to the impairment of thecapability of T cells to kill b cells in macrophage-depleted NOD mice.

Macrophages can also damage b cells by productionof soluble mediators such as oxygen free radicals andother cytokines including IL-1b, TNF-a, and IFN-g70–73

(Fig. 3). The toxic effect produced by activated macro-phages on b cells is thought to be mediated by thesuperoxide anion and hydrogen peroxide. The expres-sion of IL-1b, TNF-a, and IFN-g is detected duringearly insulitis in NOD mice and BB rats.74,75 It has beenreported that IL-1 is selectively cytotoxic to pancreaticb cells in vitro,76 and IL-1 receptors are present onb cells.77 IL-1 and TNF-a inhibit insulin secretion fromisolated islets.78,79 The destruction of b cells induced byIL-1 is potentiated by TNF-a. In vitro studies revealthat the cytotoxic effect of IL-6, TNF-a, lymphotoxin,and IFN-g on the islets is cumulative. These mediators,

acting alone or synergistically, destroy islets. Thus,treatment with IL-1 receptor antagonist, soluble IL-1receptor, anti-IL-6 antibody, anti-TNF-a antibody,or anti-IFN-g antibody decreases the incidenceof diabetes.80

Activated macrophages produce reactive oxygenspecies including oxygen radicals, hydroxyl radicals,and NO species (Fig. 3). Studies suggest that IL-1b– andTNF-a–stimulated endogenous iNOS activity inducesNO production in b cells, which contributes to b cellinjury.81 The generation of NO in response to cytokinesin b cells82 and the generation of peroxynitrite wasdetected in b cells of NOD mice.83 The b cells are verysensitive to free radicals because b cells exhibit very lowfree radical scavenging activity.84,85 Thus, overexpres-sion of cytosolic superoxide dismutase (Cu/Zn-SOD) inb cells prevents diabetes.86,87 In addition, SOD mimeticsprotect from diabetes induced by the diabetogenic Tcellclone, BDC2.5, in NOD.scid mice.88 Reactive oxygenspecies can damage b cells by causing DNA strandbreaks, which induce the DNA repair enzyme poly(ADP)ribose polymerase (PARP). PARP uses and depletesnicotinamide adenosine dinucleotide (NAD), leadingto necrosis of b cells. In line with this, PARP-deficientmice are resistant to diabetes induced by a low doseof streptozotocin.89

T CELLS

It is known that both MHC class II–restricted CD4+

T cells and MHC class I–restricted CD8+ T cells playa critical role in the pathogenesis of type 1 diabetes inNOD mice. Athymic NOD and NOD.scid mice do notdevelop insulitis or diabetes.90,91 In addition, treatmentof NOD mice with anti-CD3 antibodies inhibits thedevelopment of diabetes.92 However, the precise role ofCD4+ and CD8+ T cells in the pathogenesis of auto-immune type 1 diabetes is not clearly understood.

Both CD4+ and CD8+ T cells are required to transferdiabetes, and CD4+ T cells transfer insulitis, but notdiabetes to NOD.scid mice.91 CD4+ T cells are requiredfor the recruitment of b cell–cytotoxic CD8+ T cells intothe islets,93 but some CD8+ T cell clones from diabeticNOD mice can transfer diabetes without the help ofCD4+ T cells.94,95 However, other results showed thatsome CD4+ T cell clones can induce diabetes in theabsence of CD8+ T cells,96 and splenocytes from dia-betic donors can transfer disease into recipients lackingMHC class I expression on their islets,97 suggesting thatCD4+ T cells alone have the ability to induce diabetes.

We have cloned many CD4+ and CD8+ islet-reactiveT cells from lymphocytes infiltrating the pancreaticislets of NOD mice98 (Table 1). Islet-specific CD8+ Tcell

FIGURE 3. Role of macrophages in the development ofautoimmune type 1 diabetes. Macrophages are involvedin b cell antigen processing and presentation and theinduction of the TH1 immune response, contributing to thecreation of microenvironment for the development andactivation of b cell cytotoxic T cells. In addition, macro-phages produce solublemediators that are toxic to b cells.This results in the destruction of b cells and subsequentdevelopment of autoimmune diabetes.

584 Yoon and Jun


clones selectively destroyed b cells in vitro (Figs. 4 and5), whereas CD4+ T cell clones did not destroy b cellsbut attached closely to them84 (Fig. 4). Further studiesusing cloned CD4+ and CD8+ T cells revealed thatMHC class I–restricted cytotoxic CD8+ T lymphocytesplay an important role as final effectors in b celldestruction in vivo and that CD4+ T cells are requiredfor the activation of CD8+ T cells and their recruitmentinto the pancreatic islets.93,98 Among the CD4+ andCD8+ T cell clones produced, NY8.3 (CD8+ T cell clone)and NY4.1 (CD4+ T cell clone) were used to establishT cell receptor transgenic mice. TCR-b transgenic NODmice with TCR-b rearrangements of NY8.3 showeda 10-fold increase in frequency of precursors of b cell–

specific cytotoxic T lymphocytes and an acceleratedonset, but not an increased incidence, of diabetes.99

TCR-ab transgenic mice, with both a- and b-TCRrearrangements of NY8.3, showed a 400-fold increasein the peripheral frequency of b cell–specific cytotoxiclymphocytes and a dramatically accelerated onset ofdiabetes without an increase in disease incidence.100

TCR-ab transgenic mice with both a- and b-TCR re-arrangements of the NY4.1 clone98 also showed anaccelerated onset of diabetes as a result of a more rapidprogression of islet inflammation.100 However, micewith only the TCR-b rearrangement of NY4.1 becamediabetic later than the mice with the TCR-a and -brearrangement of NY4.1, indicating that the acceleratedonset of diabetes in these mice required the coex-pression of both the TCR-a and -b transgenes.Cytokines produced by Tcells also play an important

role in the pathogenesis of autoimmune type 1 diabetes(Figs. 1 and 6). Many different approaches have beentried to study the role of cytokines in the pathologyof type 1 diabetes, including systemic administration ofcytokines, neutralization of cytokines, and knockout ofcytokine genes.101 In general, TH1 cytokines (IL-2, IFN-g,TNF-b) cause the development of the disease, whereasTH2 or Th3 cytokines (IL-4, IL-10, TGF-b) prevent thedisease. However, the role of cytokines in the patho-genesis of autoimmune type 1 diabetes is complex.Treatment of NOD mice with anti-IFN-g antibodyprevented the development of diabetes,102 and trans-genic expression of IFN-g resulted in the developmentof diabetes in diabetes-resistant mice.103 However, thegenetic absence of IFN-g in NODmice results in a delayin the development of diabetes but does not prevent itsdevelopment.104 Systemic administration of IL-4105 orIL-10106 prevented type 1 diabetes in NOD mice, andthe transgenic expression of IL-4 in b cells also pre-vented the disease.107 However, local expression ofIL-10 in the islets accelerated the development ofdiabetes in NOD mice, and IL-4 knockout NOD mice

Table 1. CD4+ and CD8+ T cell clones and theirreactivity characteristics.

Clones* PhenotypeDestructionof b cells†

Cytotoxicityto NOD islets

(51Cr-release assay,T/E = 1/20, %)‡

NY1.1 CD4 + 6.3NY3.1 CD4 + 5.7NY3.2 CD4 2 1.2NY4.1 CD4 + 5.9NY4. CD4 2 1.1NY2.3 CD8 +++ 26.3NY5.2 CD8 ++ 19.5NY6.3 CD8 ++ 20.6NY8.3 CD8 +++ 28.7NY9.3 CD8 ++ 22.1

*CD4+ and CD8+ T cells were cloned from the islet infiltrates ofacutely diabetic female NOD mice.†Destruction of b cells was evaluated by morphological exam-ination using light and electron microscopy.‡Cytotoxicity to islet cells was evaluated by morphologicalexamination using phase-contrast microscopy and 51Cr-releasecytotoxicity assay.T/E, target/effector.

FIGURE 4. Distinct difference in in-teraction with b cells between CD4+

andCD8+Tcellclones. Islet-derivedTcell clones NY4.1 (CD4+) (A) andNY2.3 (CD8+) (B) were incubatedwith islets from silica-treated NODmice. After 2.5 hours, tissues wereexaminedbyelectronmicroscopy.A:CD4+ T cells (NY4.1) attached closelyto b cells but did not cause cell lysis.B: CD8+ T cells (NY2.3) attachedclosely to b cells and extendedpseudopods into them, causing celldamage (loss of plasma membraneand electron-dense cytoplasm).



did not show accelerated disease onset.108 Therefore,the interactions of the many different cytokines in theimmune system are complex and the development ofdiabetes may depend on which way the finely tunedimmune balance of immunoregulatory T cells is tipped(Figs. 1 and 6).

Effector T cells destroy b cells through direct contactwith surface ligands of apoptosis-inducing receptors,such as FasL and membrane-bound TNF-a, or throughthe secretion of perforin molecules, which facilitatethe passage of protease granzymes (Figs. 1 and 6).Granzymes activate nucleases in the cells and kill them.In addition, proinflammatory cytokines produced fromT cells contribute to the apoptosis of b cells. SystemicFas-deficient NOD mice did not develop insulitis ordiabetes,109 and the expression of Fas increases during

the development of diabetes, suggesting that Fas playsan important role in b cell destruction.However, a recentstudy reported that b cell–specific Fas-deficient mice, inwhich CD4+ T cells with a transgenic TCR specific forinfluenza hemagglutinin (HA) cause diabetes in micethat express HA under control of the rat insulin pro-moter, developed autoimmune diabetes with slightlyaccelerated kinetics, indicating that Fas-dependent apo-ptosis of b cells is a dispensable mode of b cell death inautoimmune type 1 diabetes.110 TNF receptor 1-deficientNOD mice failed to develop diabetes,111 and NODmice lacking perforin expression developed insulitis butnot diabetes,112 suggesting that the TNF and perforinpathway appears to play an important role forb cell killing.

CONCLUSION

The cause of autoimmune type 1 diabetes is multifac-torial, and both genetic and environmental factors areinvolved in the initiation and progression of b celldestruction. Thus, the elucidation of pathogenic mech-anisms involved in the etiology of the disease isdifficult. Based on experimental results from studiesusing NOD mice and BB rats over the past 3 decades,we have reviewed the possible pathogenic mechanismsinvolved in the autoimmune destruction of pancreaticb cells in type 1 diabetes and summarized them in themodel shown in Figure 1. This model shows thepossible interactions between b cell autoantigens andimmunocytes such as macrophages, dendritic cells,T cells, and their secretory products in connection withMHC class I and II molecules, based on our researchand that of others primarily using animal models oftype 1 diabetes. This model may not encompass allaspects of the pathogenicmechanisms involved in auto-immune diabetes in humans; nevertheless, it mayprovide helpful information with respect to the syner-gistic destruction of pancreatic b cells by immunocytesand their cytokines and a basis for the formation of newhypotheses for further investigation.

ACKNOWLEDGMENTS

The authors gratefully acknowledge the editorial assis-tance of Dr. Ann Kyle. This work was supportedby grant MA9584 from the Canadian Institutes ofHealth Research and the American Diabetes Associa-tion to J.W.Y.

FIGURE 5. Selective destruction of b cells by NY2.3(CD8+) T cells.NY2.3Tcellswere incubatedwith islets fromsilica-treated NOD mice. After 2.5 hours, tissues wereexamined under the electron microscope. b cells (b) wereextensively damaged, whereas a cells (a) (A) and d cells (d)(B) were spared. L, CD8+ cytotoxic T lymphocyte. Horizon-tal bar = 0.1 mm.

586 Yoon and Jun


REFERENCES

1. Lernmark A, Falorni A. Immune phenomena and eventsin the islets in insulin-dependent diabetes mellitus. In:Pickup JC, Williams G, eds. Textbook of Diabetes, 2nd ed.Oxford: Blackwell Science; 1997:15.1–15.23.

2. Atkinson MA, McLaren NK. The pathogenesis ofinsulin-dependent diabetes mellitus. N Engl J Med.1994;31:1428–1436.

3. Schranz DB, Lernmark A. Immunology in diabetes: anupdate. Diabetes Metab Rev. 1998;14:3–29.

4. She JX, Marron MP. Genetic susceptibility factors in type1 diabetes: linkage, disequilibrium and functionalanalyses. Curr Opin Immunol. 1998;10:682–689.

5. Onengut-Gumuscu S, Concannon P. Mapping genes forautoimmunity in humans: type 1 diabetes as a model.Immunol Rev. 2002;190:182–194.

6. Barnett AH, Eff C, Leslie RD, et al. Diabetes in iden-tical twins. A study of 200 pairs. Diabetologia. 1981;20:87–93.

7. GeptsW, Lecompte PM. The pancreatic islets in diabetes.Am J Med. 1981;70:105–115.

8. Hanninen A, Jalkanen S, Salmi M, et al. Macrophages,T cell receptor usage, and endothelial cell activation inthe pancreas at the onset of insulin-dependent diabetesmellitus. J Clin Invest. 1992;90:1901–1910.

9. Itoh N, Hanafusa T, Miyazaki A, et al. Mononuclear cellinfiltration and its relation to the expression of major

histocompatibility complex antigens and adhesion mole-cules in pancreas biopsy specimens from newly diag-nosed insulin-dependent diabetes mellitus patients.J Clin Invest. 1993;92:2313–2322.

10. Baekkeskov S, LandinM, Kristensen JK, et al. Antibodiesto a 64,000 Mr human islet cell antigen precede theclinical onset of insulin-dependent diabetes. J Clin Invest.1987;79:926–934.

11. Hawkes CJ, Schloot NC, Marks J, et al. T-cell linesreactive to an immunodominant epitope of the tyrosinephosphatase-like autoantigen IA-2 in type 1 diabetes.Diabetes. 2000;49:356–366.

12. Dromey JA, Weenink SM, Peters GH, et al. Mappingof epitopes for autoantibodies to the type 1 diabetesautoantigen IA-2 by peptide phage display and molec-ular modeling: overlap of antibody and T cell determi-nants. J Immunol. 2004;172:4084–4090.

13. Neophytou PI, Roep BO, Arden SD, et al. T-cell epitopeanalysis using subtracted expression libraries (TEASEL):application to a 38-kDA autoantigen recognized byT cells from an insulin-dependent diabetic patient. ProcNatl Acad Sci U S A. 1996;93:2014–2018.

14. Durinovic-Bello I, Steinle A, Ziegler AG, et al. HLA-DQ-restricted, islet-specific T-cell clones of a type I diabeticpatient. T-cell receptor sequence similarities to insulitis-inducing T-cells of nonobese diabetic mice. Diabetes.1994;43:1318–1325.

15. Tisch R, McDevitt H. Insulin-dependent diabetes melli-tus. Cell. 1996;85:291–297.

FIGURE 6. Schematic representa-tion of the collaboration betweenmacrophages and T cells in thedestruction of pancreatic b cells. bCell autoantigens may be releasedfrom the b cells during spontane-ous turnover of b cells. The anti-gens are then processed bydendritic cells (DC) and/or macro-phages (MØ) and presented tohelper T cells (CD4+ TH1 cell) inconjunction with MHC class IImolecules. The activated macro-phages secrete IL-12, which acti-vates CD4+ TH1-type T cells (1). TheCD4+ T cells secrete cytokines suchas IFN-g, TNF-a, TNF-b, and IL-2.While this process is takingplace,bcell–specific precytotoxic T cells(CD8+ Tc1 T cell) may be recruitedinto the islets. These precytotoxic T

cellsmaybeactivatedby IL-2 andother cytokines releasedbyCD4+TH1cells todifferentiate intoCD8+effectorTcells (2). IFN-g released by CD4+ TH1 cells and cytotoxic CD8+ T cells may cause macrophages to become cytotoxic (3). These cytotoxicmacrophages release substantial amounts of b cell–cytotoxic cytokines (including IL-1b, TNF-a, and IFN-g). Cytokinesreleased frommacrophages andT cellsmay induce the expressionof Fasonpancreaticb cells.bCells are destroyedbyFas-mediatedapoptosis (4, 5) and/orgranzymesandperforin (cytolysin),whichare toxic tob cells (6).Oxygen free radicals (nitricoxide, H2O2) secreted from activated macrophages can also kill b cells (7).



16. Delovitch TL, Singh B. The nonobese diabetic mouse asa model of autoimmune diabetes: immune dysregula-tion gets the NOD. Immunity. 1997;7:727–738.

17. Yoon JW, Jun HS. Insulin-dependent diabetes mellitus.In: Roitt IM, Delves PJ, eds. Encyclopedia of Immunology,2nd ed. London: Academic Press; 1998:1390–1398.

18. Rossini AA, Greiner DL, Friedman HP, et al. Immuno-pathogenesis of diabetes mellitus. Diabetes Rev. 1993;1:43–75.

19. Bach JF. Insulin-dependent diabetes mellitus as a b celltargeted disease of immunoregulation. J Autoimmun. 1995;8:439–463.

20. Palmer JP, Asplin CM, Clemons P, et al. Insulin anti-bodies in insulin-dependent diabetics before insulintreatment. Science. 1983;222:1337–1339.

21. Baekkeskov S, Aanstoot HJ, Christgau S, et al. Identi-fication of the 64K autoantigen in insulin-dependentdiabetes as the GABA-synthesizing enzyme glutamicacid decarboxylase. Nature. 1990;347:151–156.

22. Bonifacio E, Lampasona V, Genovese S, et al. Identifi-cation of protein tyrosine phosphatase-like IA2 (islet cellantigen 512) as the insulin-dependent diabetes-related37/40K autoantigen and a target of islet-cell antibodies.J Immunol. 1995;155:5419–5426.

23. Lu J, Li Q, Xie H, et al. Identification of a secondtransmembrane protein tyrosine phosphatase, IA-2beta,as an autoantigen in insulin-dependent diabetes melli-tus: precursor of the 37-kDa tryptic fragment. Proc NatlAcad Sci U S A. 1996;93:2307–2311.

24. Castano L, Russo E, Zhou L, et al. Identification andcloning of a granule autoantigen (carboxypeptidase-H)associated with type I diabetes. J Clin Endocrinol Metab.1991;73:1197–1201.

25. Pietropaolo M, Castano L, Babu S, et al. Islet cell auto-antigen 69 kD (ICA69).Molecular cloning and characteri-zation of a novel diabetes-associated autoantigen. J ClinInvest. 1993;92:359–371.

26. Nayak RC, Omar MA, Rabizadeh A, et al. ‘‘Cytoplas-mic’’ islet cell antibodies. Evidence that the targetantigen is a sialoglycoconjugate. Diabetes. 1985;34:617–619.

27. Arden SD, Roep BO, Neophytou PI, et al. Imogen 38:a novel 38-kD islet mitochondrial autoantigen recog-nized by T cells from a newly diagnosed type 1 diabeticpatient. J Clin Invest. 1996;97:551–561.

28. Kasimiotis H, Myers MA, Argentaro A, et al. Sex-determining region Y-related protein SOX13 is a diabetesautoantigen expressed in pancreatic islets.Diabetes. 2000;49:555–561.

29. Verge CF, Gianani R, Kawasaki E, et al. Prediction of typeI diabetes in first-degree relatives using a combinationof insulin, GAD, and ICA512bdc/IA-2 autoantibodies.Diabetes. 1996;45:926–933.

30. Sanjeevi CB, Falorni A, Robertson J, et al. Glutamic aciddecarboxylase in IDDM. Diabetes Nutr Metab. 1996;9:167–182.

31. Velloso LA, Kampe O, Hallberg A, et al. Demonstration ofGAD-65 as the main immunogenic isoform of glutamate

decarboxylase in type 1 diabetes and determination ofautoantibodies using a radioligand produced by eukary-otic expression. J Clin Invest. 1993;91:2084–2090.

32. Richter W, Shi Y, Baekkeskov S. Autoreactive epitopesdefined by diabetes-associated human monoclonal anti-bodies are localized in themiddle andC-terminal domainsof the smaller form of glutamate decarboxylase. ProcNatl Acad Sci U S A. 1993;90:2832–2836.

33. Tisch R, Yang XD, Singer SM, et al. Immune response toglutamic acid decarboxylase correlates with insulitis innon-obese diabetic mice. Nature. 1993;366:72–75.

34. DeAizpurua HJ, French MB, Chosich N, et al. Naturalhistory of humoral immunity to glutamic acid decarbox-ylase in non-obese diabetic (NOD) mice. J Autoimmun.1994;7:643–653.

35. Kaufman DL, Clare-Salzler M, Tian J, et al. Spontaneousloss of T-cell tolerance to glutamic acid decarboxylasein murine insulin-dependent diabetes. Nature. 1993;366:69–72.

36. Zekzer D, Wong FS, Ayalon O, et al. GAD-reactive CD4+

Th1 cells induce diabetes in NOD/SCID mice. J ClinInvest. 1998;101:68–73.

37. Quim A, MacInerney MF, Sercarz EE. MHC class I-restricted determinants on the glutamic acid decarbox-ylase 65 molecular induce spontaneous CTL activity.J Immunol. 2001;167:1748–1757.

38. Atkinson MA, Kaufman DL, Campbell L, et al. Responseof peripheral blood mononuclear cells to glutamatedecarboxylase in insulin-dependent diabetes. Lancet.1992;339:458–459.

39. Endl J, Otto H, Jung G, et al. Identification of naturallyprocessed T cell epitopes from glutamic acid decarbox-ylase presented in the context of HLA-DR alleles byT lymphocytes of recent onset IDDM patients. J ClinInvest. 1997;99:2405–2415.

40. Honeyman MC, Cram DS, Harrison LC. Glutamic aciddecarboxylase 67-reactive T cells: a maker of insulin-dependent diabetes. J Exp Med. 1993;177:535–540.

41. Panina-Bordignon P, Lang R, van Endert PM, et al.Cytotoxic Tcells specific for glutamic acid decarboxylasein autoimmune diabetes. J Exp Med. 1995;181:1923–1927.

42. Schloot NC, Daniel D, Norbury-Glaser M, et al.Peripheral T cell clones from NOD mice specific forGAD65 peptides: lack of islet responsiveness ordiabetogenicity. J Autoimmun. 1996;9:357–363.

43. Chen SL, Whiteley PJ, Freed DC, et al. Responses ofNOD congenic mice to glutamic acid decarboxylase-derived peptide. J Autoimmun. 1994;7:635–641.

44. Atkinson MA, Bowman MA, Campbell L, et al. Cellularimmunity to an epitope common to glutamate decarbox-ylase and Coxsackie virus in insulin-dependent diabetes.J Clin Invest. 1994;94:2125–2129.

45. Richter W, Mertens T, Schoel B, et al. Muir Sequencehomology of the diabetes-associated autoantigen gluta-mate decarboxylase with Coxsackie B4-2C protein andheat shock protein 60 mediates no molecular mimicry ofautoantibodies. J Exp Med. 1994;180:721–726.

588 Yoon and Jun


46. Bridgett M, Cetkovic-Cvrlje M, O’Rourke R, et al.Differential protection in two transgenic lines of NOD/Ltmice hyperexpressing the autoantigen GAD65 in pan-creatic beta-cells. Diabetes. 1998;47:1848–1856.

47. Geng L, Solimena M, Flavell RA, et al. Widespreadexpression of an autoantigen-GAD65 transgene does nottolerize non-obese diabetic mice and can exacerbatedisease. Proc Natl Acad Sci U S A. 1998;95:10055–10060.

48. Yoon JW, Yoon CS, Lim HW, et al. Control of auto-immune diabetes in NOD mice by GAD expression orsuppression in beta cells. Science. 1999;284:1183–1187.

49. Yoon JW, Sherwin RS, Kwon H, et al. Has GAD a centralrole in type 1 diabetes? J Autoimmun. 2000;15:273–278.

50. Jaeckel E, Klein L, Martin-Orozco N, et al. Normal in-cidence of diabetes in NOD mice tolerant to glutamicacid decarboxylase. J Exp Med. 2003;197:1635–1644.

51. Kash SF, Condie BG, Baekkeskov S. Glutamate decar-boxylase and GABA in pancreatic islets: lessons fromknock-out mice. Horm Metab Res. 1999;31:340–344.

52. Greenbaum CJ, Palmer JP, Kuglin B, et al. Insulinautoantibodies measured by radioimmunoassay meth-odology are more related to insulin-dependent diabetesmellitus than thosemeasured by enzyme-linked immuno-sorbent assay: results of the Fourth International Work-shop on the Standardization of Insulin AutoantibodyMeasurement. J Clin Endocrinol Metab. 1992;74:1040–1044.

53. Ziegler AG, HummelM, SchenkerM, et al. Autoantibodyappearance and risk for development of childhood dia-betes in offspring of parents with type 1 diabetes: the2-year analysis of the German BABYDIAB Study.Diabetes. 1999;48:460–468.

54. Yu L, Robles DT, Abiru N, et al. Early expression ofantiinsulin autoantibodies of humans and the NODmouse: evidence for early determination of subsequentdiabetes. Proc Natl Acad Sci U S A. 2000;97:1701–1706.

55. Daniel D, Gill RG, Schloot N, et al. Epitope specificity,cytokine production profile and diabetogenic activity ofinsulin-specific T cell clones isolated from NOD mice.Eur J Immunol. 1995;25:1056–1062.

56. French MB, Allison J, Cram DS, et al. Transgenic expres-sion of mouse proinsulin II prevents diabetes in non-obese diabetic mice. Diabetes. 1997;46:34–39.

57. Christie MR, Vohra G, Champagne P, et al. Distinctantibody specificities to a 64-kD islet cell antigen in type1 diabetes as revealed by trypsin treatment. J Exp Med.1990;172:789–794.

58. Lan MS, Lu J, Goto Y, et al. Molecular cloning andidentification of a receptor-type protein tyrosine phos-phatase, IA-2, from human insulinoma. DNA Cell Biol.1994;13:505–514.

59. Wasmeier C, Hutton JC. Molecular cloning of phogrin,a protein-tyrosine phosphatase homologue localized toinsulin secretory granule membranes. J Biol Chem. 1996;271:18161–18170.

60. Gorus FK, Goubert P, Semakula C, et al. IA-2-autoanti-bodies complement GAD65-autoantibodies in new-onset IDDM patients and help predict impending

diabetes in their siblings. The Belgian Diabetes Registry.Diabetologia. 1997;40:95–99.

61. Kolb H, Kantwerk G, Treichel U, et al. Prospectiveanalysis of islet lesions in BB rats. Diabetologia. 1986;29(suppl 1):A559.

62. Voorbij HA, Jeucken PH, Kabel PJ, et al. Dendritic cellsand scavenger macrophages in pancreatic islets of pre-diabetic BB rats. Diabetes. 1989;38:1623–1629.

63. Jansen A, Homo-Delarche F, Hooijkaas H, et al. Immuno-histochemical characterization of monocytes-macrophagesand dendritic cells involved in the initiation of theinsulitis and beta-cell destruction in NODmice. Diabetes.1994;43:667–675.

64. Amano K, Yoon JW. Studies on autoimmunity forinitiation of beta-cell destruction. V. Decrease ofmacrophage-dependent T lymphocytes and naturalkiller cytotoxicity in silica-treated BB rats. Diabetes.1990;39:590–596.

65. Serreze DV, Fleming SA, Chapman HD, et al. B lympho-cytes are critical antigen-presenting cells for the initia-tion of T cell-mediated autoimmune diabetes in non-obese diabetic mice. J Immunol. 1998;161:3912–3918.

66. Lee KU, Amano K, Yoon JW. Evidence for initial involve-ment of macrophage in development of insulitis in NODmice. Diabetes. 1988;37:989–991.

67. Oschilewski U, Kiesel U, Kolb H. Administration ofsilica prevents diabetes in BB-rats. Diabetes. 1985;34:197–199.

68. Lee KU, Kim MK, Amano K, et al. Preferential infil-tration of macrophages during early stages of insulitis indiabetes-prone BB rats. Diabetes. 1988;37:1053–1058.

69. Jun HS, Yoon CS, Zbytnuik L, et al. The role ofmacrophages in T cell-mediated autoimmune diabetesin nonobese diabetic mice. J Exp Med. 1999;189:347–358.

70. Pankewycz OG, Guan JX, Benedict JF. Cytokines asmediators of autoimmune diabetes and diabetic com-plications. Endocr Rev. 1995;16:164–176.

71. Mandrup-Poulsen T, Bendtzen K, Dinarello CA, et al.Human tumor necrosis factor potentiates human inter-leukin 1-mediated rat pancreatic beta-cell cytotoxicity.J Immunol. 1987;139:4077–4082.

72. Appels B, Burkart V, Kantwerk-Funke G, et al. Spontane-ous cytotoxicity of macrophages against pancreatic isletcells. J Immunol. 1989;142:3803–3808.

73. Pukel C, Baquerizo H, Rabinovitch A. Destruction of ratislet cell monolayers by cytokines. Synergistic interactionsof interferon-gamma, tumor necrosis factor, lymphotoxin,and interleukin 1. Diabetes. 1988;37:133–136.

74. Jiang Z, Woda BA. Cytokine gene expression in the isletsof the diabetic Biobreeding/Worcester rat. J Immunol.1991;146:2990–2994.

75. Toyoda H, Formby B, Magalong D, et al. In situ isletcytokine gene expression during development of type Idiabetes in the non-obese diabetic mouse. Immunol Lett.1994;39:283–288.

76. Bendtzen K, Mandrup-Poulsen T, Nerup J, et al.Cytotoxicity of human pI 7 interleukin-1 for pancreaticislets of Langerhans. Science. 1986;232:1545–1547.



77. Eizirik DL, Tracey DE, Bendtzen K, et al. An interleukin-1 receptor antagonist protein protects insulin-producingbeta cells against suppressive effects of interleukin-1beta. Diabetologia. 1991;34:445–448.

78. Hughes JH, Colca JR, Easom RA, et al. Interleukin 1inhibits insulin secretion from isolated rat pancreaticislets by a process that requires gene transcription andmRNA translation. J Clin Invest. 1990;86:856–863.

79. Southern C, Schulster D, Green IC. Inhibition of insulinsecretion by interleukin-1 beta and tumour necrosisfactor-alpha via an L-arginine-dependent nitric oxide gen-erating mechanism. FEBS Lett. 1990;276:42–44.

80. Eizirik DL, Mandrup-Poulsen T. A choice of death–thesignal-transduction of immune-mediated beta-cell apo-ptosis. Diabetologia. 2001;44:2115–2133.

81. Corbett JA, Wang JL, Sweetland MA, et al. Interleukin 1beta induces the formation of nitric oxide by beta-cellspurified from rodent islets of Langerhans. Evidence forthe beta-cell as a source and site of action of nitric oxide.J Clin Invest. 1992;90:2384–2391.

82. Corbett JA, Lancaster JR Jr, Sweetland MA, et al.Interleukin-1 beta-induced formation of EPR-detectableiron-nitrosyl complexes in islets of Langerhans. Role ofnitric oxide in interleukin-1 beta-induced inhibition ofinsulin secretion. J Biol Chem. 1991;266:21351–21354.

83. Suarez-Pinzon WL, Szabo C, Rabinovitch A. Develop-ment of autoimmune diabetes in NODmice is associatedwith the formation of peroxynitrite in pancreatic isletbeta-cells. Diabetes. 1997;46:907–911.

84. Faust A, Kleemann R, Rothe H, et al. Role of macro-phages and cytokines in b-cell death. In: Shafrir E, ed.Lessons From Animal Diabetes VI. Boston: Birhauser; 1996:47–56.

85. Asayama K, Kooy NW, Burr IM. Effect of vitamin Edeficiency and selenium deficiency on insulin secretoryreserve and free radical scavenging systems in islets:decrease of islet manganosuperoxide dismutase. J LabClin Med. 1986;107:459–464.

86. Kubisch HM, Wang J, Luche R, et al. Transgeniccopper/zinc superoxide dismutase modulates suscepti-bility to type I diabetes. Proc Natl Acad Sci U S A. 1994;91:9956–9959.

87. Kubisch HM, Wang J, Bray TM, et al. Targetedoverexpression of Cu/Zn superoxide dismutase protectspancreatic beta-cells against oxidative stress. Diabetes.1997;46:1563–1566.

88. Piganelli JD, Flores SC, Cruz C, et al. Ametalloporphyrin-based superoxide dismutase mimic inhibits adoptivetransfer of autoimmune diabetes by a diabetogenic T-cellclone. Diabetes. 2002;51:347–355.

89. Burkart V, Wang ZQ, Radons J, et al. Mice lacking thepoly(ADP-ribose) polymerase gene are resistant to pan-creatic beta-cell destruction and diabetes developmentinduced by streptozocin. Nat Med. 1999;5:314–319.

90. Ogawa M, Maruyama T, Hasegawa T, et al. Theinhibitory effect of neonatal thymectomy on the in-cidence of insulitis in non-obese diabetes (NOD) mice.Biomed Res. 1985;6:103–106.

91. Christianson SW, Shultz LD, Leiter EH. Adoptivetransfer of diabetes into immunodeficient NOD-scid/scid mice. Relative contributions of CD4+ and CD8+

T-cells from diabetic versus prediabetic NOD.NON-Thy-1a donors. Diabetes. 1993;42:44–55.

92. Hayward AR, Shreiber M. Neonatal injection of CD3antibody into nonobese diabetic mice reduces the inci-dence of insulitis and diabetes. J Immunol. 1989;143:1555–1559.

93. Nagata M, Santamaria P, Kawamura T, et al. Evidencefor the role of CD8+ cytotoxic T cells in the destruc-tion of pancreatic beta-cells in nonobese diabetic mice.J Immunol. 1994;152:2042–2050.

94. Wong FS, Visintin I, Wen L, et al. CD8 T cell clones fromyoung nonobese diabetic (NOD) islets can transfer rapidonset of diabetes in NOD mice in the absence of CD4cells. J Exp Med. 1996;183:67–76.

95. Graser RT, DiLorenzo TP, Wang F, et al. Identification ofa CD8 T cell that can independently mediate auto-immune diabetes development in the complete absenceof CD4 Tcell helper functions. J Immunol. 2000;164:3913–3918.

96. Kurrer MO, Pakala SV, Hanson HL, et al. Beta cellapoptosis in T cell-mediated autoimmune diabetes. ProcNatl Acad Sci U S A. 1997;94:213–218.

97. Serreze DV, Chapman HD, Varnum DS, et al. Initiationof autoimmune diabetes in NOD/Lt mice is MHC classI-dependent. J Immunol. 1997;158:3978–3986.

98. Nagata M, Yoon JW. Studies on autoimmunity for T-cell-mediated beta-cell destruction. Distinct difference inbeta-cell destruction between CD4+ and CD8+ T-cellclones derived from lymphocytes infiltrating the islets ofNOD mice. Diabetes. 1992;41:998–1008.

99. Verdaguer J, Yoon JW, Anderson B, et al. Acceleration ofspontaneous diabetes in TCR-beta-transgenic nonobesediabetic mice by beta-cell cytotoxic CD8+ T cellsexpressing identical endogenous TCR-alpha chains.J Immunol. 1996;157:4726–4735.

100. Verdaguer J, Schmidt D, Amrani A, et al. Spontaneousautoimmune diabetes in monoclonal Tcell nonobese dia-betic mice. J Exp Med. 1997;186:1663–1676.

101. Rabinovitch A, Suarez-Pinzon WL. Role of cytokines inthe pathogenesis of autoimmune diabetes mellitus. RevEndocr Metab Disord. 2003;4:291–299.

102. Debray-Sachs M, Carnaud C, Boitard C, et al. Pre-vention of diabetes in NOD mice treated with anti-body to murine IFN gamma. J Autoimmun. 1991;4:237–248.

103. Sarvetnick N, Liggitt D, Pitts SL, et al. Insulin-dependentdiabetes mellitus induced in transgenic mice by ectopicexpression of class II MHC and interferon-gamma. Cell.1988;52:773–782.

104. Hultgren B, Huang X, Dybdal N, et al. Genetic absenceof gamma-interferon delays but does not preventdiabetes in NOD mice. Diabetes. 1996;45:812–817.

105. Cameron MJ, Arreaza GA, Zucker P, et al. IL-4 preventsinsulitis and insulin-dependent diabetes mellitus innonobese diabetic mice by potentiation of regulatory

590 Yoon and Jun


T helper-2 cell function. J Immunol. 1997;159:4686–4692.

106. Pennline KJ, Roque-Gaffney E, Monahan M. Recombi-nant human IL-10 prevents the onset of diabetes in thenonobese diabetic mouse. Clin Immunol Immunopathol.1994;71:169–175.

107. Mueller R, Krahl T, Sarvetnick N. Pancreatic expressionof interleukin-4 abrogates insulitis and autoimmunediabetes in nonobese diabetic (NOD) mice. J Exp Med.1996;184:1093–1099.

108. Wang B, Gonzalez A, Hoglund P, et al. Interleukin-4deficiency does not exacerbate disease in NOD mice.Diabetes. 1998;47:1207–1211.

109. Itoh N, Imagawa A, Hanafusa T, et al. Requirement ofFas for the development of autoimmune diabetes innonobese diabetic mice. J Exp Med. 1997;186:613–618.

110. Apostolou I, Hao Z, Rajewsky K, et al. Effective destruc-tion of Fas-deficient insulin-producing beta cells in type1 diabetes. J Exp Med. 2003;198:1103–1106.

111. Kurrer MO, Pakala SV, Hanson HL, et al. Beta cellapoptosis in T cell-mediated autoimmune diabetes. ProcNatl Acad Sci U S A. 1997;94:213–218.

112. Kagi D, Odermatt B, Seiler P, et al. Reduced incidenceand delayed onset of diabetes in perforin-deficientnonobese diabetic mice. J Exp Med. 1997;186:989–997.



DIABETES/METABOLISM RESEARCH AND REVIEWS R E V I E W A R T I C L EDiabetes Metab Res Rev 2007; 23: 169–183.Published online 14 November 2006 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/dmrr.695

Viral infections as potential triggers of type 1diabetes

Nienke van der WerfFrans G.M. KroeseJan Rozing*Jan-Luuk Hillebrands

Department of Cell Biology,Immunology Section, UniversityMedical Center Groningen, Universityof Groningen, The Netherlands

*Correspondence to: Jan Rozing,Department of Cell Biology,Immunology section, UniversityMedical Center Groningen,University of Groningen, A.Deusinglaan 1, NL-9713 AV,Groningen, The Netherlands.E-mail: [email protected]

Received: 14 June 2006Revised: 13 September 2006Accepted: 19 September 2006

Summary

During the last decades, the incidence of type 1 diabetes (T1D) has increasedsignificantly, reaching percentages of 3% annually worldwide. This increasesuggests that besides genetical factors environmental perturbations (includingviral infections) are also involved in the pathogenesis of T1D. T1D has beenassociated with viral infections including enteroviruses, rubella, mumps,rotavirus, parvovirus and cytomegalovirus (CMV). Although correlationsbetween clinical presentation with T1D and the occurrence of a viral infectionthat precedes the development of overt disease have been recognized,causalities between viruses and the diabetogenic process are still elusiveand difficult to prove in humans. The use of experimental animal models istherefore indispensable, and indeed more insight in the mechanism by whichviruses can modulate diabetogenesis has been provided by studies in rodentmodels for T1D such as the biobreeding (BB) rat, nonobese diabetic (NOD)mouse or specific transgenic mouse strains. Data from experimental animalsas well as in vitro studies indicate that various viruses are clearly able tomodulate the development of T1D via different mechanisms, including directβ-cell lysis, bystander activation of autoreactive T cells, loss of regulatory Tcells and molecular mimicry. Data obtained in rodents and in vitro systemshave improved our insight in the possible role of viral infections in thepathogenesis of human T1D. Future studies will hopefully reveal which humanviruses are causally involved in the induction of T1D and this knowledge mayprovide directions on how to deal with viral infections in diabetes-susceptibleindividuals in order to delay or even prevent the diabetogenic process.Copyright 2006 John Wiley & Sons, Ltd.

Keywords type 1 diabetes; virus; autoimmunity; infection; mechanisms

Introduction

Diabetes mellitus

Type 1 diabetes (T1D) accounts for ∼5–10% of all cases of diabetes, varyingfrom 0.7/100 000 per year in China (Shanghai) to 30/100 000 in Finland.The incidence of childhood T1D is, however, increasing worldwide. If theincidences of T1D from 37 populations are taken together, an overall increasein incidence of 3.0% per year is observed [1]. T1D is believed to result from aT-cell–mediated autoimmune process directed against the insulin-producingβ-cells in the pancreas. The majority of diabetic subjects suffer from type 2diabetes (T2D). In contrast to T1D, T2D is non-autoimmune-mediated andresults from a combination of insulin resistance and a lack of the pancreaticβ-cells to compensate this insensitivity.

In T1D, 60–80% of the β-cells must be destroyed before clinical symptomssuch as polyuria, polydipsia, weight loss and ketosis occur [2]. Other

Copyright 2006 John Wiley & Sons, Ltd.

170 N. van der Werf et al.

islet cells, which secrete glucagon, somatostatin andpancreatic polypeptide are generally spared during theautoimmune attack against the β-cells. Histologically,T1D is characterized by an immune infiltrate withinor around the pancreatic islets (insulitis), consisting oflarge numbers of mononuclear cells and CD8+ T cells[3]. Owing to the decreased β-cell mass and consequentlyvirtual absence of endogenous insulin production, subjectswith T1D are dependent on the administration ofexogenous insulin to maintain their blood glucose levelsat normoglycaemic values. Since the introduction of theEdmonton Protocol, islet transplantation has also becomea feasible strategy to treat diabetic patients with T1D[4], although this therapy is still restricted to patientswith brittle diabetes or problems with hypoglycemia.Poor glycaemic control in patients with T1D increases thelong-term risk of diabetes-related complications that canbe divided into microvascular (nephropathy, retinopathyand neuropathy) and macrovascular (cardiovasculardisease) disorders [5]. The development of these diabetes-related complications accounts for most of the increasedmorbidity and mortality of the disease.

There is accumulating evidence indicating that geneticfactors are important for the development of T1D.First, proband concordance rates for T1D are higher ingenetically identical monozygotic twins than in dizygotictwins (43% versus 7%) [6]. Second, the frequency ofdiabetic patients is higher in close relatives of diabeticsubjects than in the general (nondiabetic) population.A large Finnish population-based study showed thatthe frequency of T1D in siblings of T1D patients is onaverage 5.5% at 30 years of age [7] versus 0.4% in thegeneral population at the same age [8]. Using linkage andassociation analyses, more than 20 T1D susceptibility locihave been identified, of which the HLA region is the mostpotent determinant, which contributes up to 40–50% ofthe genetic risk [9]. Diabetes-predisposing HLA allelesinclude the HLA class II loci HLA-DRB1, DQB1 and DQA1,also collectively referred to as IDDM1 [8]. In addition tothe HLA loci, two other loci (insulin-VNTR [IDDM2] andCTLA-4 [IDDM12]) were identified as candidate genespredisposing for T1D. The latter two loci confer ∼15% ofthe genetic susceptibility [10].

Besides genetic factors, environmental factors alsoappear to play a pivotal role in the pathogenesis ofT1D. As already mentioned above, concordance ratesfor T1D in monozygotic twins are ‘only’ 43–53% [6,11],indicating that environmental triggers must be involvedin the pathogenesis of T1D. Putative triggers of T1Dinclude environmental toxins, food antigens (e.g. cow’smilk proteins early in life, cereals or gluten) and especiallyviral infections [12]. However, a competing hypothesisstates that environmental factors are of little importancein the induction of T1D [13]. Instead, according tothis hypothesis, stochastic, random processes that occurwithin the immune system itself are proposed to initiateautoimmune disease [13].

Accumulating evidence does, however, point to a role ofviral infections in the induction of T1D. In this review, we

focus on the current knowledge about the role of viruses inthe pathogenesis of T1D. Correlative data obtained fromclinical studies as well as more mechanistic data obtainedfrom in vivo animal experimental work and in vitro studiesare summarized and the possible mechanisms by whichviruses may trigger β-cell autoimmunity are discussed.

Animal models for type 1 diabetes

Whereas (prospective) studies in humans can revealcorrelative associations between, for example, anti-viralantibodies indicative of a viral infection and autoanti-bodies, studies in animal models are indispensable toreveal causal relationships between a viral infection anddevelopment of T1D. Owing to the ethical and practicalconstraints of studying the role of viral infections in thepathogenesis of T1D in humans, animal models are used,which can be infected and analyzed at fixed time points.Also, to study the role of inheritance in the pathogenesisof T1D, animals with known genetic predisposition can beused and their response to environmental perturbationscan be tested under standardized conditions [14]. Animalmodels also allow analyses of the underlying mechanismsinvolved. The most extensively studied animal modelsin which T1D-like autoimmune disease develops are thediabetes-prone biobreeding (BBDP) rat and the nonobesediabetic (NOD) mouse.

Biobreeding (BB) rat model for T1D

The BBDP rat was discovered in 1974 by the ChappelBrothers in a colony of outbred Wistar rats at theBioBreeding Laboratories in Ottawa, Canada [15,16].Inbred BBDP rats develop spontaneous pancreaticinsulitis, which is morphologically similar to that observedin human T1D [14]. Initial insulitis is followed by rapidautoimmune destruction of the pancreatic β-cells anddevelopment of overt insulin-dependent diabetes at theage of 70–120 days [14]. T cells play a pivotal role inthe pathogenesis of T1D in BBDP rats since neonatalthymectomy prevents development of spontaneous T1D.Furthermore, transfer of ConA-activated splenocytesobtained from recent-onset diabetic BBDP rats to youngprediabetic BBDP rats accelerates diabetes onset [17].Diabetes in BBDP rats is lethal unless treated withexogenous insulin. Not all rats develop diabetes and thecumulative incidence of BBDP rats is between 80 and90% in the colony kept at our institution (UniversityMedical Center Groningen, The Netherlands), which isderived from the Worcester colony maintained at BRMInc., Worcester, MA, USA. Clinical symptoms presentedby recent-onset diabetic BBDP rats include weight loss,hyperglycaemia, hypoinsulinemia, ketonuria, polyuria,polydipsia and the presence of autoantibodies directedagainst islet cells [14,17]. In addition to insulitis, BBDPrats also develop lymphocytic thyroiditis [18]. Thepresence of an immunopathological phenotype, that is,severe lymphopenia, is required for the expression of

Copyright 2006 John Wiley & Sons, Ltd. Diabetes Metab Res Rev 2007; 23: 169–183.DOI: 10.1002/dmrr

Viral Triggers of Type 1 Diabetes 171

T1D in BBDP rats [19]. Lymphopenia in BBDP rats ischaracterized by decreased absolute T-cell numbers [20]with a nearly complete absence of CD8+ T cells [21]and CD4+ART2+ T cells. The latter T-cell subset hasbeen shown to possess regulatory potential [22–24].Severe T-lymphopenia in BBDP rats develops due toa mutation in a recessive gene designated Ian4L1 orIDDM2 [25,26]. The disruption of Ian4 in BBDP T cells(Ian4−/−) causes mitochondrial dysfunction, followed byspontaneous apoptosis. Activation of BBDP T cells andcaspase 8 inhibition are able to rescue the T cells fromdying of apoptosis [27].

Diabetes-resistant BB (BBDR) rats are derived fromBBDP forebears and do not develop T1D spontaneously.In contrast to BBDP rats, BBDR rats are Ian4+/+ andtherefore not lymphopenic. BBDR rats have normalnumbers of circulating CD4+, CD8+ and ART2+ T cellsand do not develop diabetes if housed under virus-free conditions. However, diabetes can be induced inBBDR rats by immunological perturbation like Kilhamrat virus (KRV) infection, treatment with the TLR3-ligandpoly(I : C), depletion of Tregs using anti-ART2 monoclonalantibody or combinations of these agents [14]. Themechanism by which these agents trigger disease canbe explained by the so-called balance hypothesis, whichassumes that expression of T1D is dependent on thebalance between autoreactive T cells and Tregs [28].Like BBDP rats, BBDR rats also possess autoreactivecells. In BBDR rats, autoreactive cells are suppressedby Tregs, thereby preventing expression of diabetes.However, the delicate balance between autoreactive Tcells and Tregs can be disturbed by environmentalperturbation, as mentioned in the previous text. Thebalance hypothesis also explains why BBDP rats developdiabetes spontaneously. In BBDP rats, the mutation in theIan4 gene causes apoptosis of Tregs, causing an imbalancebetween autoreactive T cells and Tregs in favor of the first.This results in loss of peripheral tolerance and expressionof disease [25,26]. Transfer of Tregs (CD4+ART2+ T cellsfrom BBDR rats) can restore the balance in BBDP ratsand prevent diabetes development [28]. The regulatorycapacity within the CD4+ART2+ T cells has recently beenshown to reside not only within a subpopulation of Tcells expressing CD25+ T cells [29,30] but also in asubpopulation of CD45RC−CD25− T cells (Hillebrandset al., provisionally accepted).

The reason why BB rats harbor a population of β-cell-specific autoreactive T cells is still unknown, but isprobably related to defects in central tolerance and thepresentation of autoantigens in the context of RT1 B/Du

MHC class II molecules [31]. Indeed, negative selectionof T cells in the thymus is hampered in BB rats due toareas devoid of thymic epithelial cells and lack of MHCexpression [32,33].

Nonobese diabetic (NOD) mouse

The most extensively studied animal model that developsT1D spontaneously is the NOD mouse, described in

1980 by Makino et al. [34]. The NOD mouse modelhas several similarities with human T1D. These includethe presence of autoantibodies, T-cell–mediated insulitisand various susceptibility genes. Of the various genesinvolved, the MHC class II region is a major susceptibilitylocus. However, other characteristics of human T1D arenot shared by the NOD mouse. For instance, diabetesincidence is higher in female than in male NOD mice,whereas there is no such sex-bias in humans [14,35].Moreover, insulitis in NOD mice differs from that inhuman subjects with T1D. Whereas the human diabeticpancreas presents with massive leukocyte infiltrationwithin the pancreatic islets, insulitis in NOD mice ischaracterized by massive accumulation of leukocytesadjacent to and infiltrating the islets. Furthermore,peripheral lymphoid organs and the submandibularsalivary glands show leukocyte infiltration [36].

Defects in both central and peripheral tolerance leadto predisposition of autoimmunity in the NOD mouse[37]. However, full disease penetrance is only presentunder pathogen-free conditions. Since most interventions,including infection with different viruses, delay or evenprevent disease in NOD mice [36,38], this model seemsless valid to study the role and underlying mechanismsof viral pathogens in triggering or accelerating thedevelopment of human T1D.

Transgenic mouse models

The availability of mouse models in which a well-characterized viral protein is expressed by the β-cellsand/or thymus offers the possibility to follow theautoreactive immune response upon infection precisely.A well-characterized transgenic mouse model is therat insulin promoter-lymphocytic choriomeningitis virus(RIP-LCMV) mouse in which a nucleoprotein or glyco-protein of lymphocytic choriomeningitis virus (LCMV) isspecifically expressed in pancreatic β-cells under the con-trol of the rat insulin promoter (RIP) [39]. Because theviral transgene is inserted in the germline of the host, theviral protein is essentially a ‘self’ protein and does notprovoke an immune response and subsequent disease.However, infection with LCMV results in rapid onset T1Din more than 95% of RIP-LCMV mice. Since the LCMVviral protein is identical to the transgene expressed inβ-cells, 100% homology rather than molecular mimicry isresponsible for the high diabetes incidence [40]. Infectionof RIP-LCMV mice with Pichinde virus bearing an epitopethat shares homology with the LCMV-nucleoprotein epi-tope does not induce diabetes in RIP-LCMV nucleoproteinmice. However, infection with LCMV followed by infectionwith Pichinde virus at a later stage induces acceleratedonset of T1D [41]. In this study, the authors proposethat a cross-reactive immune response accelarates morelikely an already existing autoimmune process rather thaninitiates such a response [41].

Expression of the viral protein both in the pancreaticβ-cells and in the thymus results in deletion of the



majority of high-affinity T cells in the thymus by negativeselection. Therefore, usually only low-affinity T cells enterthe periphery. Despite the stringent thymic selectionprocess, some high-affinity T cells escape deletion, whichare activated upon viral infection. Activation of virus-specific T cells induces either rapid (<21 days) or slow(30–120 days) onset of T1D. Low-affinity CD8+ T cellsneed help from CD4+ T cells to produce disease, whereashigh-affinity CD8+ T cells are sufficient to induce rapidonset of T1D [40].

One of the most successful methods to regulate inflam-mation in vivo is blockade of particular receptor–ligandinteractions. Recently, Ejrnaes et al. showed that blockadeof CXCL10 using short-term administration of neutralizinganti-CXCL10 antibodies abrogated LCMV-induced TID inRIP-LCMV mice [42], suggesting a role for the chemokineCXCL10 in the induction of T1D.

Viruses in T1D

Potential mechanisms by which viralinfections induce T1D

Several viral infections have been associated withhuman T1D. These viruses include enterovirus [43,44],rubella [45,46], mumps [47,48], rotavirus [49] andcytomegalovirus (CMV) [50,51]. Mechanisms by whichthese viruses trigger T1D remain unclear, but variouspossibilities have been put forward, depending on the typeof virus involved. First of all, infection of pancreatic β-cells with virus may result in a cytolytic infection, leadingto β-cell lysis, that is, a non-immune-mediated formof diabetes. Second, viral infection of pancreatic β-cellsmay not result in cytolysis, but may break self-tolerancedue to the expression of viral antigens in the β-cells,exposure of the immune system to altered β-cell antigensand/or increased expression of MHC-antigens or cytokinesand chemokines [52]. Third, a viral infection mayinduce bystander activation of T lymphocytes includingautoreactive T cells directed against β-cell antigens [40].Fourth, some viruses may express proteins that act assuperantigens, leading to activation and proliferation ofspecific T cells expressing the appropriate VβTCR [53]. Ifautoreactive T cells also express the specific VβTCR, viralsuperantigens may trigger autoimmunity by activatingand expanding the pool of autoreactive T cells. Fifth,viruses may disturb the finely tuned balance betweenautoreactive T cells and Tregs in favor of the autoreactiveT cells. This may result from specific infection/destructionof Treg subsets and/or expansion of autoreactive Tcells [52]. The mechanism by which Tregs suppressautoreactive T cells is not fully understood. Direct cell–cellcontact between Tregs, Antigen Presenting Cell APCs andautoreactive effector T cells appear to be required forsuppression in vitro. In vivo, anti-inflammatory cytokineslike IL-10 and TGF-β are also required for Treg-mediatedsuppression [54]. Sixth, epitopes of viral proteinsmay have structural homology with β-cell autoantigens

(molecular mimicry) and activation of cross-reacting T orB cells may induce β-cell destruction [52]. However, theexistence of homology between viral- and self-epitopesdoes not necessarily imply that cross-activation leadsto similar immune responses. Homologues viral- andself-epitopes may have different effects on T cells, andcan either activate or abrogate effector functions [55].Furthermore, differences in avidity of self- versus viral-epitopes may influence the type of response induced uponrecognition [41]. Finally, successive infections with tworelated but independently encountered viral infectionsmay precipitate autoimmunity. A shared viral epitope maylead to cross-activation of existing virus-specific memory Tcells, modification of the T-cell repertoire and subsequentinduction of immunopathology [41,56,57].

In the next sections, an overview is given of the virusessupposedly associated with the induction of T1D and thepossible mechanisms by which these viruses may triggerβ-cell autoimmunity, leading to T1D. The major findingsof studies on the role of viral infections in precipitatingT1D are summarized in Table 1.

Enterovirus infection and T1D

Human enteroviruses are small nonenveloped RNAviruses belonging to the Picornavirus family, including thecoxsackie A and B viruses, echoviruses and polioviruses[91]. Accumulating evidence favors the causal role ofenterovirus infection in the pathogenesis of T1D [92].Evidence of the role of enteroviruses in the developmentof T1D was first obtained in studies showing increasedprevalence of enteroviral RNA in sera of patients withrecent-onset T1D compared with healthy controls [93,94].A prospective study of a cohort of children who wereat risk of developing T1D revealed that the presenceof enterovirus infection before the first appearance ofautoantibodies is more frequently found in children thatdeveloped autoantibodies against β-cell antigens than inchildren that remained free of autoantibodies during a6-month follow-up period [43]. Autoantibodies developduring the preclinical phase of diabetes developmentand may therefore function as a prognostic marker forthe development of T1D [95]. Recently, Williams et al.isolated an Echovirus 3 strain from a fecal sampleof an individual at the time of appearance of isletautoantibodies, also supporting a role for enteroviruses inthe pathogenesis of T1D [58]. However, in another cohortof children that were categorized as high- or as low-risksubjects, the relation between enterovirus infection andsubsequent development of autoantibodies could not beconfirmed [96], suggesting that enterovirus infection onlytriggers β-cell autoimmunity in some individuals at highrisk to develop T1D.

Congenital enteroviral infections may also play a rolein the pathogenesis of T1D. Dahlquist et al. showed anincreased prevalence of enteroviral RNA in neonatal bloodof subjects who developed T1D later in life comparedto subjects who did not develop the disease [97]. An



Table 1. Viral infections associated with the induction of T1D and the possible mechanisms involved

Virus SpeciesGeneticfactors Effect Reference

RNA viruses:Picornavirus family Men ND Associated with induction of islet autoantibodies 58Enterovirus Two case reports: isolated viruses cause diabetes in animals 59,60

Mice Yes Virus-induced β-cell lysis induces T1D 61,62Molecular mimicry 63Virus-induced bystander damage causes release of sequestered islet antigens andrestimulation of autoreactive T cells

64

Encephalomyocarditis virus Mice Yes Induction of diabetes 65High dose: virus-induced β-cell lysis, macrophages are involved 66–68Low dose: infection of β-cells, virus-activated macrophages play a central role andare recruited to the pancreatic islets where they contribute to the destruction

68,69

Togavirus family Men Yes Associated with induction of islet autoantibodies and T1D, especially congenitalrubella

46

Rubella virus T-cell cross-reactivity between rubella antigens and GAD 70Infection of β-cells, but without cytolytic effects 71–73

Hamsters ND Induction of diabetes 72Paramyxovirus family Men ND Associated with induction of islet autoantibodies and T1D 47,48Mumps virus Infection of β-cells, but without cytolytic effects 74,75

Increased expression of HLA class I and II on β-cells 75Reovirus family Men ND Associated with induction of islet autoantibodies 76Rotavirus Mice Yes Infection of β-cells 77

Molecular mimicry 78Reovirus Mice ND Infection of β-cells 79DNA viruses:Parvovirus family Men ND Associated with induction of autoimmune diseases incl. T1D (case report) 80Parvovirus

Rat Yes No infection of β-cells 81Virus-activated macrophages play a critical role in breakdown of tolerance 82,83Th1-type immune response is enhanced and Th2-type immune response (supposedTregs) is reduced

84,85

β-Herpes family Men ND Associated with induction of islet autoantibodies and T1D 51Cytomegalovirus Infection of β-cells, but without cytolytic effects 86

T-cell cross-reactivity between HCMV and GAD65 87Men/Rat Virus-induced clonal activation of T cells 88,89

Rat Yes Acceleration of diabetes onset 89,90Virus-induced recruitment of macrophages to the pancreas triggers the accelerateddevelopment of insulitis

ND, not determined.

explanation for the increased susceptibility of infantsfor persistent enterovirus infection is the fact that thefrequency of enterovirus infection in the population islow, resulting in an increased proportion of motherswho lack enterovirus antibodies. As a result, transfer ofantibodies (transplacental or via breast milk) is reducedin the infants, resulting in increased risk of persistententerovirus infection [98].

There is also direct evidence from case reportssupporting an association between enterovirus infectionand onset of T1D. For example, a 10-year-old boywho died of diabetic ketoacidosis showed lymphocyticinfiltration of the pancreatic islets and necrosis of β-cells. Inoculation of AJL/J male mice with homogenatesof the child’s pancreas led to isolation of a variant fromCoxsackievirus B4 (CVB4), which induced hyperglycemia,insulitis and β-cell necrosis [59]. Another case reportdescribed a girl who developed diabetic symptomsshortly after Coxsackie virus B5 (CVB5) infection. Aftera remission of 2.5 months, she developed definitiveT1D. The virus isolated from the girl’s feces inducedglucose intolerance in several mouse strains [60]. Thesetwo case reports clearly indicate that CVB variantsisolated from recent-onset human diabetic patients are

able to induce diabetes-like disease when injected intosusceptible mice.

The mechanism by which CVB may be involvedin the pathogenesis of T1D is, however, elusive, butseveral possibilities have been put forward based onobservations from experimental animal studies andin vitro data. As we discuss in the following text, thepossible mechanisms include non-T-cell–mediated β-cell destruction, induction of an enhanced autoantigen-specific T-cell response, molecular mimicry and bystander-activation of autoreactive T cells.

One of the mechanisms involved in CVB-induced T1D isdestruction of the pancreatic β-cells without involvementof an adaptive T-cell response. CVB has β-cell cytolyticactivity and directly destroys the β-cell mass, resulting inT1D in genetically susceptible mouse strains [61,62].Thecapacity of CVB to induce diabetes is dependent onthe genetic background of the host [71,99]. Flodstromet al. showed that CVB-induced β-cell lysis depends ona hampered virus-induced anti-viral IFN response of theβ-cells [100]. In this study, the authors show that SOCS-1transgenic mice are permissive for CVB3 infection. CVB3infection of wild-type mice induces destruction of theexocrine tissue, but spares the islets, resulting in survival



of most animals. In contrast, CVB3 infection of SOCS-1transgenic mice results in infection and destruction of bothislets and exocrine tissue, leading to an acute form of T1D.The observation that T-cell-deficient SOCS-1-Tg.scid micealso develop rapid diabetes upon CVB3 infection indicatethat in this model an adaptive T-cell response is notinvolved in the CVB3-induced development of T1D [100].Also, in children with T1D, a hampered anti-viral responsemay lead to CV-induced β-cell destruction. Skarsvik et al.showed that the TH1 immune response against CVB4in vitro is decreased (reduced IFN-γ ) in children withT1D compared with healthy controls and this could notbe explained by disease-associated HLA genotypes. Theauthors suggest that the decreased TH1 immune responseagainst CVB4 may lead to decreased clearance of the virusand consequently enhanced CVB4-related complications,such as β-cell damage [101].

Also, in T1D patients, enterovirus has a pancreastropism. Postmortem pancreatic specimens from severalT1D patients revealed the presence of enterovirus RNA-positive cells. Their location was exclusively confinedto the pancreatic islets [102]. The effect of enterovirusreplication on β-cell survival is strongly dependent on theenterovirus serotype [103].

In addition to direct cytolytic effects of CVB on β-cells, some data support a role for CVB in enhancing theautoreactive T-cell response by increased presentationof autoantigens to T cells [104]. BDC2.5 mice areTCR-transgenic NOD mice that harbor a large numberof resting, islet-specific T cells, which are not cross-reactive with CVB4 proteins. These mice do not developspontaneous diabetes. CVB4 infection, however, inducesT1D in BDC2.5 mice. CVB4 directly infects the β-cells,but without causing a lytic infection or necrosis. AfterCVB4 infection, insulin-containing macrophages/DCswere detected near the pancreatic islets, suggestingphagocytosis of β-cells. Furthermore, APCs from CVB4-infected NOD mice were able to increase proliferation ofautoreactive T cells in vitro and adoptive transfer of APCsfrom CVB4-infected NOD scid mice to na ıve BDC2.5 miceinduced diabetes in 33% of the recipients. These resultsindicate that APCs from CVB4-infected mice present β-cell autoantigens to the autoreactive T-cell population,resulting in the expression of T1D [104].

Alternative mechanisms by which CV might induceT1D are molecular mimicry and bystander damage. Themajority of patients with T1D develop humoral andcellular immune responses against islet-autoantigens suchas GAD65. Since GAD65 shows sequence homology withthe CV protein P2C, it is likely that upon CV infectiona P2C-specific immune response is induced, which iscross-reactive with GAD65. This possibility was analyzedby Schloot et al., who determined cross-reactivity ofGAD65-specific T-cell clones (generated from recent-onsetT1D patients) with P2C. However, no cross-reactivitysupportive of molecular mimicry was observed in thisstudy [105]. Harkonen et al. also analyzed molecularmimicry as the underlying mechanism and determined thepresence of cross-reactivity between tyrosine phosphatase

IA-2/IAR autoantibodies and enterovirus antigens [63]. Inthis study, IA-2/IAR autoantibodies from NOD mice, andof some humans (7%) were indeed cross-reactive with theenterovirus capsid protein [63]. These data indicate that,primarily at the B-cell level, molecular mimicry may playa role in enterovirus-induced T1D.

The final possible mechanism involved in CVB-inducedT1D is bystander damage, as proposed by Horwitzet al. [64]. In their study, mice bearing diabetes-susceptible MHC-alleles did not develop accelerateddiabetes following CVB4 infection, indicating that possiblecross-reactivity between the responses against the viralproteins P2C and GAD65 was not sufficient to induceT1D. However, BDC2.5 mice infected with CVB4 rapidlydeveloped T1D. BDC2.5 mice, as already mentioned, areTCR-transgenic and harbor diabetogenic T cells specificfor an islet granule antigen distinct from GAD65 whichare not cross-reactive with CVB4 proteins. The authorsproposed that in these mice CVB4 infection induces(bystander) tissue damage, leading to inflammation andrelease of sequestered islet antigens, resulting in therestimulation of the resting autoreactive T-cell populationand subsequent development of diabetes [64]. The roleof the cytokines IL-4 and IFN-γ in CVB4-induced T1Din NOD mice was addressed by Serreze et al. [106].NOD mice that already have a critical amount of insulitisdevelop accelerated diabetes following CVB4 infection.In contrast, the onset of diabetes is not accelerated afterCVB4 infection of age-matched IL-4 or IFN-γ -deficientNOD mice, suggesting that both IL-4 and IFN-γ play apivotal role in CVB4-accelerated diabetes at this stageof disease. However, the IL-4 and IFN-γ dependencyappears to be age and disease-stage dependent since, inolder NOD mice with more severe insulitis at the timeof infection, IL-4 and IFN-γ were no longer required toinduce accelerated diabetes onset [107].

Encephalomyocarditis virus and T1D

The encephalomyocarditis (EMC) virus belongs like theenteroviruses to the picornavirus family and is a small,single-stranded RNA virus with a genome of about 7.8 Kb.Although there is little evidence that the EMC virus isinvolved in the pathogenesis of T1D in humans, studiesperformed in mice provide clear evidence that variantsof EMC virus induce T1D in these animals (reviewed inref. 52). Since this experimental model for virus-inducedautoimmune diabetes is extensively studied, it providesa lot of information concerning the possible role of virusinfections in the induction of TID.

In genetically susceptible mouse strains, the M variantof EMC virus (EMC-M) induces a diabetes-like syndrome[108]. Plaque purification of the EMC-M virus resultedin the isolation of two stable variants: a diabetogenicEMC-D virus and a nondiabetogenic EMC-B virus [65].EMC-D virus induces diabetes in over 90% of infectedanimals, whereas EMC-B virus does not produce diabetesin infected mice [65]. Although the two variants



are antigenically indistinguishable, examination of thegenomes of both variants revealed a total of 14 nucleotidedifferences between them [109]. Further analysis ofseveral mutant viruses generated from EMC-B and EMC-D, revealed that only one amino acid, alanine (776thamino acid on polyprotein), is common to all diabetogenicvariants [110].

Only some mouse strains develop diabetes afterEMC-D viral infection [65]. Susceptibility to EMC-Dvirus–induced diabetes has been shown to be determinedby a single autosomal recessive gene that is inheritedin a Mendelian mode [111]. Inoculation of geneticallysusceptible SJL/J mice with a high titre of EMC-D virus(5 × 105 pfu/mouse) results in diabetes onset within4 days due to acute destruction of the pancreatic β-cellsby viral replication within these cells [66]. Treatmentof these mice with anti-T-cell antibodies did not affectdiabetes incidence, indicating that the destruction ofthe pancreatic β-cells by EMC-D virus in SJL/J miceis not T-cell–mediated [67,68]. However, treatment ofEMC-D-infected mice with anti-macrophage antibodiesresulted in a significantly lower diabetes incidence. Thisfinding indicates that replication of EMC-D in the β-cellsand virus-activated macrophages could act synergisticallyto destroy the pancreatic β-cells after infection with ahigh dose of EMC-D virus in SJL/J mice [67,68]. Sincesuch a high dose of EMC-D virus is unlikely to occurunder natural conditions, SJL/J mice were also infectedwith a low dose of EMC-D virus (50–100 pfu/mouse).Inoculation with a low dose of EMC-D virus results in theinitial replication of the virus in the pancreatic β-cells,followed by recruitment of macrophages to the infectedpancreatic islets [68]. In these mice, macrophages play acentral role, since activation of macrophages before viralinfection results in a significant increase in the diabetesincidence, and inactivation of macrophages before viralinfection completely prevents EMC-D virus–induceddiabetes [69]. Although the exact role of macrophagesin the pathogenesis of EMC-D virus–induced diabetesis not fully understood, the secretion of cytokines (IL-1,TNF-α, IFN-α) and oxygen-free radicals (NO) by activatedmacrophages may contribute to the destruction of β-cells[112]. Further investigation revealed that macrophagesare directly activated by the EMC-D virus throughtriggering of the tyrosine kinase signalling pathway [113].As a result, EMC-D virus–activated macrophages start toproduce TNF-α and NO, leading to the destruction of thepancreatic β-cells and eventually diabetes in geneticallysusceptible mice after infection with a low dose of theEMC-D virus.

Rubella virus and T1D

Rubella is an ssRNA-enveloped virus and member ofthe Togavirus family. Several reports have shown anincreased prevalence of T1D in patients with congenitalrubella [45,46,114]. In patients with congenital rubellasyndrome, the frequencies of the HLA antigens DR2

and DR3 are decreased in nondiabetic subjects, whereasthese frequencies are increased in rubella-positive patientsthat also developed T1D. These data indicate that ifrubella is indeed involved in the pathogenesis of T1D,this is restricted to those infected subjects that have thegenetic susceptibility to develop T1D [115]. It has beenshown that the percentage of subjects with antibodiesagainst islet cells is increased in patients with congenitalrubella syndrome compared with noninfected subjects[46]. However, a more recent study by Viskari et al.was not able to demonstrate such a correlation betweencongenital rubella syndrome and the development ofautoantibodies against islet antigens [116]. These dataled the authors to conclude that diabetes developmentin subjects with congenital rubella syndrome may not beimmune-mediated.

The role of rubella in the development of T1D hasalso been addressed in an experimental model for T1Din hamsters [72]. This study revealed that infection ofneonatal hamsters with rubella indeed induces diabetes,providing experimental evidence for a causal relationshipbetween rubella infection and the development of T1D[72]. Both direct infection of β-cells with rubella andmolecular mimicry have been put forward as mechanismsunderlying rubella-induced diabetes. Results from in vitroand in vivo studies indicate that rubella is able to infecthuman pancreatic β-cells [72,73] and reduce insulinsecretion [73]. It is, however, unlikely that rubellacauses direct destruction of β-cells since no signs ofcytopathological effects are observed after infection ofhuman islets [71]. An alternative mechanism is molecularmimicry. T cells of congenital rubella patients with T1Dwere found to be cross-reactive with rubella virus peptidesand GAD protein determinants [70].

Mumps virus and T1D

Mumps is an enveloped ssRNA virus and a memberof the Paramyxovirus family. Also, mumps has beenassociated with T1D, as a high frequency of childrenwith mumps appeared to have islet cell antibodies andoccasionally developed overt diabetes [47]. In addition,mumps outbreaks have been associated with an increasein the incidence of T1D 2–4 years later [48]. Themechanism through which mumps might be involvedin the development of T1D is elusive. In vitro studiesindicate that mumps is able to infect human β-cellsand other pancreatic cells, but without inducing a lyticinfection [74,75]. Cavallo et al. showed that the mumpsvirus induced the release of IL-1 and IL-6 in a humaninsulinoma cell line and increased the expression of HLAclass I and class II molecules [75]. These findings suggestthat the mechanism by which mumps induces diabetesinvolves loss of tolerance toward β-cells by making themmore vulnerable to immune-mediated destruction.

In the 1980s, vaccination against mumps, measlesand rubella was introduced in most Western-worldcountries. Since then, several studies have reported



on the relation between vaccination at childhood andthe development of T1D later in life [117–119].Hyoty et al. demonstrated that vaccination againstmumps/measles/rubella in Finland was followed by aplateau in the rising incidence of T1D 6–8 years later[117], suggesting a causal relation between these viralinfections and the development of T1D. However, thisplateau period was only temporary. Since prevention ofserious infection by vaccination at childhood did notstop the rising incidence of T1D, this study indicatesthat mumps, measles and rubella infections are not themajor environmental perturbants that trigger T1D. Otherstudies hypothesized that childhood vaccination wouldrather promote the development of T1D. Sofar, however,no evidence has been found for the triggering effect ofchildhood vaccination on the development of T1D later inlife [118,119].

Rotavirus infection and T1D

Rotavirus is a nonenveloped dsRNA virus and a memberof the Reovirus family. Rotavirus is the major causeof childhood gastroenteritis [120], but is also putforward as a candidate pathogen for the triggering orexacerbation of T1D [48,49]. Honeyman et al. reportedon the association between increased levels of rotavirus-specific IgG and IgA antibodies and the appearance orincrease in autoantibodies against islets in children at riskfor T1D [76]. In a similar study, however, Blomquist et al.were not able to show such an association [121]. One ofthe mechanisms involved in rotavirus-induced T1D mightbe direct infection of the pancreatic β-cells since it hasbeen shown in an in vitro study that rotavirus is able toinfect pancreatic islets of NOD mice [77]. Furthermore, itis reported that another member of the Reovirus family,reovirus, is also able to infect and destroy human β-cellsin vitro [79]. These data suggest that rotavirus infectionmay induce T1D by either direct cytolytic infection ofβ-cells or by initiating autoimmunity by infecting β-cells,resulting in release of β-cell (auto) antigens.

Since the rotavirus protein VP7 shows sequencehomology with the autoantigens tyrosine phosphatase IA-2 and GAD, molecular mimicry can also be put forwardas a mechanism for rotavirus-induced T1D [78]. So far, itis, however, unknown to what extent rotavirus-specific Tcells are actually able to trigger β-cell autoimmunity.

Endogenous retrovirus and T1D

A mouse mammary tumor virus (MMTV)-related humanendogenous retrovirus (IDDMK1,222) has been isolatedfrom pancreata of two recent-onset T1D patients, butnot from nondiabetic controls [53]. The virus, whichis incorporated in the human genome, encodes asuperantigen that is proposed to be involved in thesystemic activation of autoreactive T cells, leading topancreatic β-cell destruction [53]. However, data from

other studies indicate that the presence of IDDMK1,222does not predispose for the development of T1D,indicating that a major role for IDDMK1,222 in thepathogenesis of T1D is unlikely [122,123].

Parvovirus and T1D

Parvovirus belongs to the Parvovirus family. It has a rel-atively simple structure composed of three viral proteinsencoded by a linear ssDNA molecule. A link between par-vovirus infection and the induction of T1D was suggestedby Munakata et al., who described a case report of awoman who suffered from a persistent parvovirus infec-tion concurrently with the occurrence of T1D, Graves’disease and rheumatoid arthritis [80]. However, no dif-ferences were found in the presence of parvovirus B19 IgGbetween children with recent-onset T1D and healthy con-trols, arguing against an association between parvovirusB19 infection and development of T1D [124].

Although parvovirus infections have been suggested tobe involved in the induction of T1D in humans [80],one of the best-studied animal models for the role ofvirus infection in the pathogenesis of T1D is KRV-induceddiabetes in BBDR rats [125]. KRV selectively replicates individing cells, explaining its tropism for bone marrow, gut,spleen, thymus and lymph nodes [52]. KRV does not infectpancreatic β-cells [81]. Infection of BBDR rats with KRVinduces T1D in approximately one-third of the animals,whereas noninfected animals do not develop disease.Additional treatment with poly(inosinic : cytidylic) acid(poly(I : C) further increases the incidence to 80%[82,126]. KRV infection of prediabetic BBDP rats doesnot accelerate diabetes, but infection of protected BBDPrats (i.e. BBDP rats injected with BBDR splenocytes)accelerates diabetes just as in infected BBDR rats [125].Susceptibility of BBDR rats to KRV-induced diabetesrequires the presence of class I A(u) and class II B/D(u)gene products. However, mere expression of the RT1u

haplotype is not sufficient since RT1u WF rats do notdevelop diabetes following KRV infection [126].

To date, several studies have been reported onthe underlying mechanism of KRV-induced diabetes inBBDR rats. Since pancreatic β-cells are not permissivefor KRV [81], direct cytolysis of pancreatic β-cells isnot the underlying mechanism. Chung et al. showedthat selective depletion of macrophages prevents thedevelopment of KRV-induced diabetes in BBDR rats,suggesting that macrophages and their cytokines playa pivotal role in KRV-induced diabetes [82]. In linewith this, Mendez et al. studied the role of macrophage-derived NO in the activation of effector T cells in thepathogenesis of KRV-induced T1D in BBDR rats [83].Inhibition of NO production in KRV-infected BBDR ratsresulted in decreased expression of macrophage-derivedproinflammatory cytokines, a decreased percentage ofTh1-like CD4+CD45RC+ T cells and decreased T-cellactivation. Together, these data indicate that KRV-induced activation of macrophages plays a critical role in



enhancing the TH1-type immune response and activationof autoreactive effector cells, leading to T1D in BBDR rats[83].

Molecular mimicry is also one of the suggested mech-anisms of KRV-induced diabetes in BBDR rats. However,BBDR rats infected with recombinant Vaccinia virusexpressing KRV–derived structural and nonstructural pro-teins neither developed insulitis nor diabetes, despite theinduction of both a humoral and cellular response [84].Molecular mimicry is therefore unlikely, and KRV ratherseems to disturb the fine-tuned immune balance betweenautoreactive T cells and regulatory T cells. In line withthis, KRV infection was found to increase the percentage ofCD8+ T cells and Th1-like CD4+CD45RC+ T cells, whereasthe percentage of Th2-like CD4+CD45RC−ART2+ cells(supposedly Tregs) was decreased [84]. Adoptive trans-fer of activated CD8+ T cells and Th1-like CD4+CD45RC+T cells from KRV-infected BBDR rats to young BBDR ratsshowed that these two T-cell subsets are major effectorcells that work synergistically in β-cell destruction anddiabetes development [84]. In line with these findings,Zipris et al. demonstrated that KRV infection reducedthe percentage of splenic CD4+CD25+ Tregs in BBDRrats [85]. The decrease in CD4+CD25+ T cells wasfound to be virus-specific since infection with the KRVhomologue H1 neither reduced the frequency of splenicCD4+CD25+ Tregs nor induced diabetes. However, inthe peripancreatic lymph nodes of BBDR rats, but notWF rats, an increase in CD4+CD25+ Tregs was foundafter KRV, but not H1, infection. On the basis of theseresults, the authors concluded that KRV induces T1D inBBDR rats due to a failure to maintain function of Tregs[85].

In addition to its effects on cellular adaptive immunity(i.e. altered T cell subsets), KRV also appears to activateinnate immunity in both a virus- and rat strain–specificmanner [127]. In this study, Zipris et al. demonstratedthat infection of diabetes-susceptible BBDR rats withKRV, but not vaccinia virus or H1, resulted in increasedexpression of IL-12 p40, IP-10 and IFN-γ in particularlythe pancreatic lymph nodes. These observations areindicative of KRV-induced activation of innate immunity.The pivotal role of activation of innate immunity inKRV-induced diabetes was further supported by theobservation that treatment with synthetic and natural toll-like receptor (TLR) agonist synergized with KRV infectionand increased the frequency of diabetes. Whether KRVitself acts as a TLR ligand has not been reported, althoughit has been shown that various other viruses can directlyinteract with TLRs, resulting in cell activation [128]. Theobserved synergy of KRV and TLR-agonists in inducingdiabetes is also in line with the ‘fertile field hypothesis’proposed by von Herrath et al., stating that viral infectionsalone might not be able to induce disease and that otherinflammatory factors, like activation of innate immunityby TLR triggering, is required for increased diseasepenetrance [129,130].

Cytomegalovirus and T1D

CMV belongs to the family of the β-herpesviridae. Thegenome of human CMV (HCMV) is large and consistsof 235 kb linear dsDNA. In 1979, a link between thepresence of HCMV and the development of T1D wassuggested for the first time [50]. In this study, a child wasdescribed that suffered from congenital CMV infectionand developed T1D at the age of 13 months [50]. Thisobservation was supported by Pak et al. who showed aclear correlation between the presence of CMV genomein PBMNCs and autoantibodies against islet antigens inpatients with recent-onset diabetes [51]. In addition,Nicoletti et al. showed that the presence of high titresof anti-CMV IgG antibodies positively correlates withthe presence of autoantibodies against islet antigens inhealthy siblings of diabetic children [131]. Together,these correlative studies suggest that HCMV may beinvolved in the pathogenesis of human T1D. In contrast,Banatvala et al. were not able to demonstrate CMV-specific IgM responses in recently diagnosed diabeticpatients [132]. Since CMV infection may have occurredearly in life, markers indicative of acute CMV infectionmay have been cleared at the time of expression ofdisease and consequently no correlation was detectedbetween the presence of CMV genome and/or anti-CMVantibodies and the presence of autoantibodies specificfor islet antigens. This might explain why Banatvalaet al. did not find a correlation between CMV-antibodiesand T1D [132]. For this reason, prospective studieswould be of more use than cross-sectional ones inproviding information about the possible relation betweenHCMV infection and development of T1D. To date,results of several prospective studies have been reported.A study from Sweden demonstrated that only 1.3%of children with congenital CMV infection developedT1D later in life, indicating that mere congenital CMVinfection is not a major cause of T1D in the generalpopulation [133]. Furthermore, in high-risk children(i.e. nondiabetic siblings of newly diagnosed diabeticsubjects), no evidence was found that supported anassociation between primary CMV infection and thesubsequent development of autoantibodies against isletantigens or overt diabetes. These data suggest that if CMVhas a role in the pathogenesis of T1D, it is limited to asubset of children [134].

Although prospective studies thus indicate that congen-ital CMV infection is not likely to be a main player in thepathogenesis of T1D, other studies do support a role forHCMV in the diabetogenic process. For example, charac-teristic CMV inclusion bodies were found in pancreaticislets of 44% of children who died of CMV infection, pro-viding evidence that CMV can infect and damage β-cellsin vivo [135]. However, no CMV genome was detected inpancreas from recent-onset T1D patients [136,137], indi-cating that, although CMV is able to infect pancreatic cellsin cases of severe CMV disease, this phenomenon is rarein subclinical CMV infection. In line with this is the obser-vation that islets cells are generally nonpermissive for



CMV infection in vitro. However, human fetal β-cells canbe productively infected with CMV in vitro, but withoutsignificant β-cell destruction or changes in insulin release[86]. Together, these findings indicate that, in general,congenital CMV infection does not induce direct cytolysisof the pancreatic β-cells and therefore direct β-cell lysis isnot the underlying mechanism of possible CMV-induceddevelopment of T1D.

Another mechanism by which CMV may be involvedin the pathogenesis of T1D is molecular mimicry, thatis, presence of anti-CMV antibodies that are cross-reactive with islet-autoantigens. Pak et al. showed thatCMV induces antibody formation that recognizes a 38-kDhuman pancreatic islet-specific protein [138]. It remainsto be determined whether CMV antigens actually havesequence homology with the 38-kD islet protein andwhether the antibodies are actually cross-reactive. Cross-reactivity may also exist at the T-cell level rather thanat the B-cell level since T cells are the main playersin the development of T1D, which can even develop inthe absence of B cells [139]. In line with this, Hiemstraet al. indeed obtained evidence for the existence of T-cellcross-reactivity between a HCMV-encoded protein and aβ-cell autoantigen [87]. In their study, the authors showthat clonal CD4+ T cells recognizing the major β-cellautoantigen GAD65 are cross-reactive with the HCMV-encoded major DNA-binding protein UL57. UL57 canbe efficiently processed for presentation by HLA-DR3,which predisposes to T1D [87]. This finding supportsa role for HCMV in the pathogenesis of T1D by amechanism of molecular mimicry between viral proteinsand autoantigens.

A third mechanism by which HCMV can modulatethe development of T1D is by bystander activation ofautoreactive T cells. CMV replication and latency islargely controlled by a cellular immune response. It hasbeen shown that ∼11% of individuals with subclinicalCMV infection and stable control of viral replicationhave high frequencies (10–43%) of CMV-specific CD4+T cells [88]. These CMV-specific CD4+ T cell wereshown to persist over several months, are oligoclonaland have an activated/memory phenotype. In line withthese observations, we showed that splenocytes derivedfrom BBDR rats infected with rat-specific CMV (RCMV)also display a strong proliferative response (∼50% ofall T cells proliferated) after in vitro restimulation withRCMV-infected syngeneic fibroblasts [89]. This responseis RCMV-specific as this response is only observed afterrestimulation of previously RCMV-primed splenocytes.Analysis of cell division patterns using carboxyfluoresceindiacetate ester CFSE-dilution revealed that in vitrorestimulation causes T-cell proliferation similar to thatobserved after stimulation with the polyclonal activatorConcanavalinA (N van der Werf, unpublished). These datasuggest RCMV induces polyclonal T-cell activation, whichmay include autoreactive T cells. Bystander activation ofT cells may also be the result of a superantigen-inducedmonoclonal expansion and Dobrescu et al. indeed showedthat HCMV encoded a superantigen that is able to induce

proliferation of a restricted subset of T cells expressingVβ12 [140]. We were, however, not able to induceproliferation of specific Vβ TCR families using RCMV[89]. Furthermore, primary infection in vivo is requiredto induce RCMV-specific T-cell proliferation in vitro,indicating that the high percentage of RCMV-specific T-cell proliferation is not induced by an RCMV-encodedsuperantigen.

As described above, so far no causal relation hasbeen demonstrated between HCMV infection and thedevelopment of human T1D; the role HCMV plays in thediabetogenic process is therefore still elusive. We were,however, able to demonstrate a causal relation betweenCMV infection and acceleration of the development of T1Dusing the well-established BB rat model for T1D (see alsosection on BioBreeding (BB) Rat Model for TID) [89,90].Infection of BBDP rats with RCMV at the age of 35 daysresulted in a significant acceleration of diabetes onset.In this model, RCMV appears to induce activation, andto some extent infection, of macrophages that selectivelymigrate into the pancreas. This macrophage influx appearsto be the trigger for activation and accelerated recruitmentof existing autoreactive T cells to the pancreatic islets,eventually resulting in destructive insulitis and T1D. Inthe model of RCMV-enhanced diabetes onset in BBDPrats, macrophages thus appear to be pivotal in theaccelerating effect of RCMV, suggesting the involvementof innate immunity. HCMV has been shown to bindto TLR2, resulting in the production of inflammatorycytokines [128,141]. In line with this, preliminary datafrom our laboratory suggest that RCMV is also able toactivate macrophages via TLR triggering, which may beinvolved in the induction of accelerated diabetes onset.The diabetogenic effect of RCMV has recently beenconfirmed in another rat model (LEW.1WR1 rats) ofspontaneous autoimmune diabetes [14]. LEW.1WR1 ratspossess a normal developed immune system and developspontaneous autoimmune diabetes at a low frequency(∼2%) [142].

It is unknown whether vaccination of children againstCMV is a feasible strategy to eliminate the possiblediabetogenic effect of CMV on the development of T1D.However, a recent preliminary report by Tirabassi et al.shows that maternal immunization of LEW.1WR1 ratswith RCMV protects their offspring from virus-induceddiabetes in a virus-specific fashion [143]. These datasuggest that preventing primary infection with CMV atyoung age by antibody-mediated mechanisms may indeedbe an effective way to eliminate the diabetogenic effectsof this virus.

Various possibilities by which CMV may be involvedin the pathogenesis of T1D are put forward, but clearmechanisms are still lacking. Since the prevalence ofHCMV seropositivity is very high (60–90%) [144] andthe incidence of T1D relatively low, the role of HCMVinfection in T1D must be a complementary and not aprimary one [145].



Discussion

Both data from correlative clinical studies as well asexperimental studies using rodent models for T1D haverevealed that various viruses are able to modulate theexpression of T1D via various mechanisms. These mech-anisms include the following: direct cytolysis of virus-infected pancreatic β-cells, induction of autoimmuneresponses to ‘altered self antigens’ as a result of increasedβ-cell antigenicity, bystander activation of (autoreactive)T cells that may be viral superantigen-driven, loss oftolerance due to virus-induced disturbance of the bal-ance between autoreactive T cells and Tregs, molecularmimicry resulting in the activation of autoreactive Tand/or B cells by viral antigens and the activation ofinnate immunity that may be mediated by triggering ofTLRs. It could well be that multiple mechanisms needto act in concert to result in precipitated disease onset.The different mechanisms are summarized in a schemedepicted in Figure 1.

We propose that whether or not a virus will modifydisease onset via one of the above-mentioned mechanismsin genetically diabetes-susceptible subjects depends on thepresence of precipitating factors such as the nature of thevirus, the viral load and the subjects immunological status.Data supporting the importance of these precipitatingfactors will be discussed in the following text.

The first precipitating factor may be the nature ofthe virus, that is, its structure, mode of transmission,cell tropism, entry mechanisms etc. The efficiency of avirus to evade recognition and activation of innate andadaptive immune responses determines its persistence.For instance, CMVs are highly sophisticated DNA viruseswith a large genome (235 kb for HCMV), encoding manyviral proteins that are involved in the promotion ofviral dissemination and evasion of immune surveillance[146]. Furthermore, viruses with a β-cell tropism suchas enteroviruses [61,102] may facilitate the induction ofT1D since β-cells may be directly (acute cytolysis) orindirectly (through recruitment of activated macrophagesor existing β-cell specific T cells) destroyed.

Figure 1. Influenced by genetic susceptibility to develop T1D and possibly other environmental events that modulate the immunestatus, viral infections may aggravate the development of T1D through various mechanisms that are not mutually exclusive.Viruses may infect the β-cells that reside in the pancreatic islets, resulting in direct cytolysis, and thereby elimination, of theinsulin-producing β-cells (1). Alternatively, viruses may infect β-cells without cytolysis, but with enhancement of the antigenicityof the β-cells, resulting in the induction of an immune response directed against ‘altered self antigens’ (2). Viruses may also induceactivation of macrophages through TLR triggering, resulting in the secretion of proinflammatory mediators (e.g. IL-1, IL-6, Il-8,MCP-1 and TNF-α). Proinflammatory mediators may enhance autoimmunity by increasing autoreactive T-cell responses by eitherbystander activation (including activation of autoreactive T cells) (3) or selective deletion/impairment of regulatory T cells (5).Moreover, proinflammatory mediators may directly affect β-cells, resulting in β-cell damage and increased immunogenicity (4).Molecular mimicry may also be involved through induction of virus-specific B and T cells by macrophages presenting viral peptides.These virus-specific B and T cells are also reactive with β-cell antigens due to presence of shared epitopes (6). The outcome ofthese various mechanisms is β-cell destruction, resulting in clinical manifestation of type 1 diabetes eventually. Abbreviations: BV:virus-specific B cell; IL: interleukin; M�: macrophage; MCP-1: monocyte chemoattractive protein-1, MHC: major histocompatibilitycomplex, T: T cell, TA: autoreactive T cell, TR: regulatory T cell, TV: virus-specific T cell, TLR: Toll-like receptor, TNF-α: tumornecrosis factor alpha



The second precipitating factor may be the viral load.Differences in viral load may influence the development ofT1D via different mechanisms. For instance, inoculationof diabetes-susceptible mice with a high dose of EMC-Dvirus resulted in the development of diabetes due to acutevirus-induced β-cell lysis [66]. However, inoculation ofthese mice with a lower, more natural dose of EMC-Dvirus resulted in virus-activated macrophages that weresupposed to play a major role in the β-cell destruction[69]. In addition, we demonstrated that infection with ratCMV-accelerated diabetes onset in BBDP rats, providedthat the inoculum consists of at least 5 × 105 plaqueforming units living virus (van der Werf et al., unpublisheddata). Infection with lower amounts of living virus ora similar amount of inactivated virus did not resultin acceleration of diabetes onset. Also, in some cases,protection from disease development was observed (vander Werf et al., unpublished results).

The third precipitating factor for virus-induced expres-sion of T1D may be the immunological status of anindividual. Viral infection may induce a shift in the deli-cate balance between autoreactive T cells and Tregs [28].Either enhancement of the autoreactive part or reductionof the regulatory part (or a combination of both) leadsto expression of disease. For instance, the rat parvovirusKRV induces downregulation of the regulatory part of thebalance [84,85], whereas rat CMV infection appears toinduce polyclonal expansion of the autoreactive part ofthe balance [89].

Although there is accumulating evidence describingassociations between viruses and the development ofT1D in genetically susceptible individuals, clear causativeviral agents have not been identified so far. A possibleexplanation for the difficulty in discerning viral triggersin human T1D is that, in addition to viral infection,other inflammatory factors are necessary to induceautoimmunity, supporting the ‘fertile field hypothesis’[129]. In addition, multiple viral infections (antigen-related or unrelated) encountered throughout life mayhave to act in concert to precipitate disease in diabetes-susceptible individuals [130]. Finally, the identificationof causative viral triggers in human T1D is furthercomplicated by the fact that certain viruses requirea certain genetic background in the host in order tobe diabetogenic. This is illustrated by the observationsreported by Blankenhorn et al., showing that, in additionto MHC-related genes, the diabetes-susceptibility lociIddm4 and Iddm20 are required for diabetes expressioninduced by viral infection [147].

Conclusions

In this review, we have recapitulated the currentknowledge about the role of different viruses, includingenteroviruses, EMC virus, rubella virus, mumps virus,rotavirus, parvovirus and CMV in the pathogenesisof T1D. Clinical and experimental studies revealed

that different viruses can modulate the expressionof diabetes via various mechanisms as described. Allclinical data reported so far are, however, descriptivein nature and direct evidence for causal relationshipsbetween viral infections and T1D are still lacking. Futurestudies should therefore aim at validating the dataobtained in experimental animal models and in vitrosystems in clinical T1D, thereby focussing on theclinically most relevant viruses such as enteroviruses andCMV. Currently, a major, multicenter prospective studysupported by the National Institutes of Health is beingperformed, which aims at identifying the genetic andenvironmental factors (including viral infections) that areinvolved in the induction of T1D in children. In this so-called TEDDY study (The Environmental Determinants ofDiabetes in the Young), neonates are recruited from fourdifferent countries (Finland, Germany, Sweden and NorthAmerica) and will be followed until the age of 15. Thisstudy and similar future studies will hopefully lead to theidentification of causative viral agents in human T1D. Thisknowledge will then provide further directions on how todeal with viral infections in diabetes-susceptible subjectsin order to delay or even prevent the diabetogenic process.

Acknowledgements

This study was supported by grants from the Dutch DiabetesFoundation (DFN 1999.028 and DFN 2001.05.003).

References

1. Onkamo P, Vaananen S, Karvonen M, et al. Worldwideincrease in incidence of Type I diabetes–the analysis of thedata on published incidence trends. Diabetologia 1999; 42:1395–1403.

2. Greiner DL, Rossini AA, Mordes JP. Translating data fromanimal models into methods for preventing humanautoimmune diabetes mellitus: caveat emptor and primumnon nocere. Clin Immunol 2001; 100: 134–143.

3. Notkins AL, Lernmark A. Autoimmune type 1 diabetes:resolved and unresolved issues. J Clin Invest 2001; 108:1247–1252.

4. Shapiro AMJ, Lakey JRT, Ryan EA, et al. Islet transplantationin seven patients with type 1 diabetes mellitus using aglucocorticoid-free immunosuppressive regimen. N Engl J Med2000; 343: 230–238.

5. Eiselein L, Schwartz HJ, Rutledge JC. The challenge of type 1diabetes mellitus. ILAR J 2004; 45: 231–236.

6. Hyttinen V, Kaprio J, Kinnunen L, et al. Genetic liability of type1 diabetes and the onset age among 22,650 young Finnishtwin pairs: a nationwide follow-up study. Diabetes 2003; 52:1052–1055.

7. Harjutsalo V, Podar T, Tuomilehto J. Cumulative incidence oftype 1 diabetes in 10,168 siblings of Finnish young-onset type1 diabetic patients. Diabetes 2005; 54: 563–569.

8. Spielman RS, Baker L, Zmijewski CM. Gene dosage andsusceptibility to insulin-dependent diabetes. Ann Hum Genet1980; 44: 135–150.

9. Field LL. Genetic linkage and association studies of TypeI diabetes: challenges and rewards. Diabetologia 2002; 45:21–35.

10. Anjos S, Polychronakos C. Mechanisms of genetic susceptibilityto type I diabetes: beyond HLA. Mol Genet Metab 2004; 81:187–195.

11. Kyvik KO, Green A, Beck-Nielsen H. Concordance rates ofinsulin dependent diabetes mellitus: a population based studyof young Danish twins. BMJ 1995; 311: 913–917.



12. Akerblom HK, Vaarala O, Hyoty H, et al. Environmental factorsin the etiology of type 1 diabetes. Am J Med Genet 2002; 115:18–29.

13. Eisenberg RA, Craven SY, Warren RW, et al. Stochastic controlof anti-Sm autoantibodies in MRL/Mp-lpr/lpr mice. J Clin Invest1987; 80: 691–697.

14. Mordes JP, Bortell R, Blankenhorn EP, et al. Rat models of type1 diabetes: genetics, environment, and autoimmunity. ILAR J2004; 45: 278–291.

15. Nakhooda AF, Like AA, Chappel CI, et al. The spontaneouslydiabetic Wistar rat. Metabolic and morphologic studies.Diabetes 1977; 26: 100–112.

16. Chappel CI, Chappel WR. The discovery and development ofthe BB rat colony: an animal model of spontaneous diabetesmellitus. Metabolism 1983; 32: 8–10.

17. Crisa L, Mordes JP, Rossini AA. Autoimmune diabetes mellitusin the BB rat. Diabetes Metab Rev 1992; 8: 4–37.

18. Rajatanavin R, Appel MC, Reinhardt W, et al. Variableprevalence of lymphocytic thyroiditis among diabetes-pronesublines of Bb/Wor rats. Endocrinology 1991; 128: 153–157.

19. Jacob HJ, Pettersson A, Wilson D, et al. Genetic dissection ofautoimmune type-I diabetes in the Bb rat. Nat Genet 1992; 2:56–60.

20. Jackson R, Rassi N, Crump T, et al. The Bb diabeticrat–profound T-cell lymphocytopenia. Diabetes 1981; 30:887–889.

21. Woda BA, Like AA, Padden C, et al. Deficiency of phenotypiccytotoxic-suppressor lymphocytes-T in the Bb/W rat. J Immunol1986; 136: 856–859.

22. Poussier P, Nakhooda AF, Falk JA, et al. Lymphopenia andabnormal lymphocyte subsets in the Bb rat–relationship tothe diabetic syndrome. Endocrinology 1982; 110: 1825–1827.

23. Jackson R, Kadison P, Buse J, et al. Lymphocyte abnormalitiesin the BB rat. Metabolism 1983; 32: 83–86.

24. Bortell R, Kanaitsuka T, Stevens LA, et al. The RT6 (Art2)family of ADP-ribosyltransferases in rat and mouse. Mol CellBiochem 1999; 193: 61–68.

25. Hornum L, Romer J, Markholst H. The diabetes-prone BB ratcarries a frameshift mutation in Ian4, a positional candidate ofIddm1. Diabetes 2002; 51: 1972–1979.

26. MacMurray AJ, Moralejo DH, Kwitek AE, et al. Lymphopeniain the BB rat model of type 1 diabetes is due to a mutationin a novel immune-associated nucleotide (lan)-related gene.Genome Res 2002; 12: 1029–1039.

27. Pandarpurkar M, Wilson-Fritch L, Corvera S, et al. lan4 isrequired for mitochondrial integrity and T cell survival. ProcNatl Acad Sci USA 2003; 100: 10382–10387.

28. Mordes JP, Bortell R, Doukas J, et al. The BB/Wor rat and thebalance hypothesis of autoimmunity. Diabetes Metab Rev 1996;12: 103–109.

29. Lundsgaard D, Holm TL, Hornum L, et al. In vivo control ofdiabetogenic T-cells by regulatory CD4(+)CD25(+) T-cellsexpressing Foxp3. Diabetes 2005; 54: 1040–1047.

30. Poussier P, Ning T, Murphy T, et al. Impaired post-thymicdevelopment of regulatory CD4(+) 25(+) T cells contributesto diabetes pathogenesis in BB rats. J Immunol 2005; 174:4081–4089.

31. Ellerman KE, Like AA. Susceptibility to diabetes is widelydistributed in normal class IIu haplotype rats. Diabetologia2000; 43: 890–898.

32. Doukas J, Mordes JP, Swymer C, et al. Thymic epithelialdefects and predisposition to autoimmune-disease in Bb rats.Am J Pathol 1994; 145: 1517–1525.

33. Rosmalen JGM, van Ewijk V, Leenen PJM. T-cell education inautoimmune diabetes: teachers and students. Trends Immunol2002; 23: 40–46.

34. Makino S, Kunimoto K, Muraoka Y, et al. Breeding of a non-obese, diabetic strain of mice. Jikken Dobutsu 1980; 29: 1–13.

35. Leiter EH, von Herrath M. Animal models have little to teachus about Type 1 diabetes: 2. In opposition to this proposal.Diabetologia 2004; 47: 1657–1660.

36. Atkinson MA, Leiter EH. The NOD mouse model of type 1diabetes: as good as it gets? Nat Med 1999; 5: 601–604.

37. Aoki CA, Borchers AT, Ridgway WM, et al. NOD mice andautoimmunity. Autoimmun Rev 2005; 4: 373–379.

38. Wetzel JD, Barton ES, Chappell JD, et al. Reovirus delaysdiabetes onset but does not prevent insulitis in nonobesediabetic mice. J Virol 2006; 80: 3078–3082.

39. von Herrath MG, Evans CF, Horwitz MS, et al. Using transgenicmouse models to dissect the pathogenesis of virus-inducedautoimmune disorders of the islets of Langerhans and thecentral nervous system. Immunol Rev 1996; 152: 111–143.

40. Oldstone MB. Molecular mimicry and immune-mediateddiseases. FASEB J 1998; 12: 1255–1265.

41. Christen U, von Herrath MG. Induction, acceleration orprevention of autoimmunity by molecular mimicry. MolImmunol 2004; 40: 1113–1120.

42. Ejrnaes M, von Herrath MG, Christen U. Cure of chronic viralinfection and virus-induced type 1 diabetes by neutralizingantibodies. Clin Dev Immunol 2006; 13: 67–77.

43. Lonnrot M, Korpela K, Knip M, et al. Enterovirus infection asa risk factor for beta-cell autoimmunity in a prospectivelyobserved birth cohort: the Finnish Diabetes Prediction andPrevention Study. Diabetes 2000; 49: 1314–1318.

44. Salminen K, Sadeharju K, Lonnrot M, et al. Enterovirusinfections are associated with the induction of beta-cellautoimmunity in a prospective birth Cohort study. J Med Virol2003; 69: 91–98.

45. Menser MA, Forrest JM, Bransby RD. Rubella infection anddiabetes-mellitus. Lancet 1978; 1: 57–60.

46. Ginsbergfellner F, Witt ME, Yagihashi S, et al. Congenital-rubella syndrome as a model for Type-1 (insulin-dependent)diabetes-mellitus–increased prevalence of islet cell-surfaceantibodies. Diabetologia 1984; 27: 87–89.

47. Helmke K, Otten A, Willems W. Islet cell antibodies in childrenwith mumps infection. Lancet 1980; 2: 211–212.

48. Hyoty H, Leinikki P, Reunanen A, et al. Mumps infections inthe etiology of type 1 (insulin-dependent) diabetes. DiabetesRes 1988; 9: 111–116.

49. Honeyman M. How robust is the evidence for viruses in theinduction of type 1 diabetes? Curr Opin Immunol 2005; 17:616–623.

50. Ward KP, Galloway WH, Auchterlonie IA. Congenital cytome-galo-virus infection and diabetes. Lancet 1979; 1: 497.

51. Pak CY, Mcarthur RG, Eun HM, et al. Association ofcytomegalo-virus infection with autoimmune Type-1 diabetes.Lancet 1988; 2: 1–4.

52. Jun HS, Yoon JW. A new took at viruses in type 1 diabetes.Diabetes Metab Res Rev 2003; 19: 8–31.

53. Conrad B, Weissmahr RN, Boni J, et al. A human endogenousretroviral superantigen as candidate autoimmune gene in typeI diabetes. Cell 1997; 90: 303–313.

54. Sakaguchi S. Naturally arising Foxp3-expressing CD25(+)CD4(+) regulatory T cells in immunological tolerance to selfand non-self. Nat Immunol 2005; 6: 345–352.

55. Kersh GJ, Allen PM. Structural basis for T cell recognition ofaltered peptide ligands: a single T cell receptor can productivelyrecognize a large continuum of related ligands. J Exp Med 1996;184: 1259–1268.

56. Welsh RM, Selin LK. No one is naive: the significance ofheterologous T-cell immunity. Nat Rev Immunol 2002; 2:417–426.

57. Merkler D, Horvath E, Bruck W, et al. ‘‘Viral deja vu’’ elicitsorgan-specific immune disease independent of reactivity toself. J Clin Invest 2006; 116: 1254–1263.

58. Williams CH, Oikarinen S, Tauriainen S, et al. Molecularanalysis of an echovirus 3 strain isolated from anindividual concurrently with appearance of islet cell and IA-2autoantibodies. J Clin Microbiol 2006; 44: 441–448.

59. Yoon JW, Austin M, Onodera T, et al. Isolation of a virus fromthe pancreas of a child with diabetic ketoacidosis. N Engl J Med1979; 300: 1173–1179.

60. Champsaur HF, Bottazzo GF, Bertrams J, et al. Virologic,immunologic, and genetic factors in insulin-dependent diabetesmellitus. J Pediatr 1982; 100: 15–20.

61. Yoon JW, Onodera T, Notkins AL. Virus-induced diabetesmellitus. XV. Beta cell damage and insulin-dependenthyperglycemia in mice infected with coxsackie virus B4. JExp Med 1978; 148: 1068–1080.

62. Toniolo A, Onodera T, Jordan G, et al. Virus-induced diabetesmellitus. Glucose abnormalities produced in mice by the sixmembers of the Coxsackie B virus group. Diabetes 1982; 31:496–499.

63. Harkonen T, Lankinen H, Davydova B, et al. Enterovirusinfection can induce immune responses that cross-react withbeta-cell autoantigen tyrosine phosphatase IA-2/IAR. J MedVirol 2002; 66: 340–350.



64. Horwitz MS, Bradley LM, Harbertson J, et al. Diabetes inducedby Coxsackie virus: initiation by bystander damage and notmolecular mimicry. Nat Med 1998; 4: 781–785.

65. Yoon JW, McClintock PR, Onodera T, et al. Virus-induceddiabetes mellitus. XVIII. Inhibition by a nondiabetogenicvariant of encephalomyocarditis virus. J Exp Med 1980; 152:878–892.

66. Yoon J, Onodera T, Notkins AL. Virus-induced diabetesmellitus: VIII. Passage of encephalomyocarditis virus andseverity of diabetes in susceptible and resistant strains of mice.J Gen Virol 1977; 37: 225–232.

67. Yoon JW, McClintock PR, Bachurski CJ, et al. Virus-induceddiabetes mellitus. No evidence for immune mechanismsin the destruction of beta-cells by the D-variant ofencephalomyocarditis virus. Diabetes 1985; 34: 922–925.

68. Baek HS, Yoon JW. Role of macrophages in the pathogenesisof encephalomyocarditis virus-induced diabetes in mice. J Virol1990; 64: 5708–5715.

69. Baek HS, Yoon JW. Direct involvement of macrophages indestruction of beta-cells leading to development of diabetesin virus-infected mice. Diabetes 1991; 40: 1586–1597.

70. Ou D, Mitchell LA, Metzger DL, et al. Cross-reactive rubellavirus and glutamic acid decarboxylase (65 and 67) proteindeterminants recognised by T cells of patients with type Idiabetes mellitus. Diabetologia 2000; 43: 750–762.

71. Numazaki K, Goldman H, Seemayer TA, et al. Infection byhuman cytomegalovirus and rubella virus of cultured humanfetal islets of Langerhans. In Vivo 1990; 4: 49–54.

72. Rayfield EJ, Kelly KJ, Yoon JW. Rubella virus-induced diabetesin the hamster. Diabetes 1986; 35: 1278–1281.

73. Numazaki K, Goldman H, Wong I, et al. Infection of culturedhuman fetal pancreatic islet cells by rubella virus. Am J ClinPathol 1989; 91: 446–451.

74. Vuorinen T, Nikolakaros G, Simell O, et al. Mumps andCoxsackie B3 virus infection of human fetal pancreatic islet-likecell clusters. Pancreas 1992; 7: 460–464.

75. Cavallo MG, Baroni MG, Toto A, et al. Viral infection inducescytokine release by beta islet cells. Immunology 1992; 75:664–668.

76. Honeyman MC, Coulson BS, Stone NL, et al. Associationbetween rotavirus infection and pancreatic islet autoimmunityin children at risk of developing type 1 diabetes. Diabetes 2000;49: 1319–1324.

77. Coulson BS, Witterick PD, Tan Y, et al. Growth of rotavirusesin primary pancreatic cells. J Virol 2002; 76: 9537–9544.

78. Honeyman MC, Stone NL, Harrison LC. T-cell epitopes in type1 diabetes autoantigen tyrosine phosphatase IA-2: potential formimicry with rotavirus and other environmental agents. MolMed 1998; 4: 231–239.

79. Yoon JW, Selvaggio S, Onodera T, et al. Infection of culturedhuman pancreatic B-cells with reovirus type-3. Diabetologia1981; 20: 462–467.

80. Munakata Y, Kodera T, Saito T, et al. Rheumatoid arthritis,type 1 diabetes, and Graves’ disease after acute parvovirus B19infection. Lancet 2005; 366: 780.

81. Brown DW, Welsh RM, Like AA. Infection of peripancreaticlymph-nodes but not islets precedes kilham rat virus-induceddiabetes in Bb/Wor rats. J Virol 1993; 67: 5873–5878.

82. Chung YH, Jun HS, Kang Y, et al. Role of macrophages andmacrophage-derived cytokines in the pathogenesis of Kilhamrat virus induced autoimmune diabetes in diabetes-resistantbiobreeding rats. J Immunol 1997; 159: 466–471.

83. Mendez II, Chung YH, Jun HS, et al. Immunoregulatory role ofnitric oxide in Kilham rat virus-induced autoimmune diabetesin DR-BB rats. J Immunol 2004; 173: 1327–1335.

84. Chung YH, Jun HS, Son M, et al. Cellular and molecularmechanism for Kilham rat virus-induced autoimmune diabetesin DR-BB rats. J Immunol 2000; 165: 2866–2876.

85. Zipris D, Hillebrands JL, Welsh RM, et al. Infections that induceautoimmune diabetes in BBDR rats modulate CD4(+)CD25(+)T cell populations. J Immunol 2003; 170: 3592–3602.

86. Numazaki K, Goldman H, Wong I, et al. Viral-infection ofhuman-fetal islets of langerhans–replication of humancytomegalo-virus in cultured human-fetal pancreatic-islets. AmJ Clin Pathol 1988; 90: 52–57.

87. Hiemstra HS, Schloot NC, van Veelen PA, et al. Cytomegalo-virus in autoimmunity: T cell crossreactivity to viral antigenand autoantigen glutamic acid decarboxylase. Proc Natl AcadSci U S A 2001; 98: 3988–3991.

88. Sester M, Sester U, Gartner B, et al. Sustained high frequenciesof specific CD4 T cells restricted to a single persistent virus. JVirol 2002; 76: 3748–3755.

89. van der Werf N, Hillebrands JL, Klatter FA, et al.Cytomegalovirus infection modulates cellular immunity inan experimental model for autoimmune diabetes. Clin DevImmunol 2003; 10: 153–160.

90. Hillebrands JL, van der Werf N, Klatter FA, et al. Roleof peritoneal macrophages in cytomegalovirus-inducedacceleration of autoimmune diabetes in BB-rats. Clin DevImmunol 2003; 10: 133–139.

91. Hyoty H, Taylor KW. The role of viruses in human diabetes.Diabetologia 2002; 45: 1353–1361.

92. Tauriainen S, Salminen K, Hyoty H. Can enteroviruses causetype 1 diabetes? Ann N Y Acad Sci 2003; 1005: 13–22.

93. Clements GB, Galbraith DN, Taylor KW. Coxsackie B virusinfection and onset of childhood diabetes. Lancet 1995; 346:221–223.

94. Andreoletti L, Hober D, Hober-Vandenberghe C, et al.Detection of coxsackie B virus RNA sequences in whole bloodsamples from adult patients at the onset of type I diabetesmellitus. J Med Virol 1997; 52: 121–127.

95. Kimpimaki T, Kulmala P, Savola K, et al. Disease-associatedautoantibodies as surrogate markers of type 1 diabetes inyoung children at increased genetic risk. Childhood Diabetesin Finland Study Group. J Clin Endocrinol Metab 2000; 85:1126–1132.

96. Graves PM, Rotbart HA, Nix WA, et al. Prospective studyof enteroviral infections and development of beta-cellautoimmunity. Diabetes Autoimmunity Study in the Young(DAISY). Diabetes Res Clin Pract 2003; 59: 51–61.

97. Dahlquist GG, Forsberg J, Hagenfeldt L, et al. Increasedprevalence of enteroviral RNA in blood spots from newbornchildren who later developed type 1 diabetes: a population-based case-control study. Diabetes Care 2004; 27: 285–286.

98. Viskari H, Ludvigsson J, Uibo R, et al. Relationship betweenthe incidence of type 1 diabetes and maternal enterovirusantibodies: time trends and geographical variation.Diabetologia 2005; 48: 1280–1287.

99. Karounos DG, Wolinsky JS, Thomas JW. Monoclonal antibodyto rubella virus capsid protein recognizes a beta-cell antigen. JImmunol 1993; 150: 3080–3085.

100. Flodstrom M, Tsai D, Fine C, et al. Diabetogenic potential ofhuman pathogens uncovered in experimentally permissivebeta-cells. Diabetes 2003; 52: 2025–2034.

101. Skarsvik S, Puranen J, Honkanen J, et al. Decreased in vitrotype 1 immune response against coxsackie virus B4 in childrenwith type 1 diabetes. Diabetes 2006; 55: 996–1003.

102. Roivainen M. Enteroviruses: new findings on the role ofenteroviruses in type 1 diabetes. Int J Biochem Cell Biol 2006;38: 721–725.

103. Roivainen M, Ylipaasto P, Savolainen C, et al. Functionalimpairment and killing of human beta cells by enteroviruses:the capacity is shared by a wide range of serotypes, but theextent is a characteristic of individual virus strains. Diabetologia2002; 45: 693–702.

104. Horwitz MS, Ilic A, Fine C, et al. Coxsackieviral-mediateddiabetes: induction requires antigen-presenting cells and isaccompanied by phagocytosis of beta cells. Clin Immunol 2004;110: 134–144.

105. Schloot NC, Willemen SJ, Duinkerken G, et al. Molecularmimicry in type 1 diabetes mellitus revisited: T-cell clonesto GAD65 peptides with sequence homology to Coxsackieor proinsulin peptides do not crossreact with homologouscounterpart. Hum Immunol 2001; 62: 299–309.

106. Serreze DV, Ottendorfer EW, Ellis TM, et al. Acceleration oftype 1 diabetes by a coxsackievirus infection requires apreexisting critical mass of autoreactive T-cells in pancreaticislets. Diabetes 2000; 49: 708–711.

107. Serreze DV, Wasserfall C, Ottendorfer EW, et al. Diabetesacceleration or prevention by a coxsackievirus B4 infection:critical requirements for both interleukin-4 and gammainterferon. J Virol 2005; 79: 1045–1052.

108. Craighead JE, McLane MF. Diabetes mellitus: induction in miceby encephalomyocarditis virus. Science 1968; 162: 913–914.

109. Bae YS, Eun HM, Yoon JW. Genomic differences betweenthe diabetogenic and nondiabetogenic variants ofencephalomyocarditis virus. Virology 1989; 170: 282–287.



110. Bae YS, Yoon JW. Determination of diabetogenicityattributable to a single amino acid, Ala776, on the polyproteinof encephalomyocarditis virus. Diabetes 1993; 42: 435–443.

111. Onodera T, Yoon JW, Brown KS, et al. Evidence for a singlelocus controlling susceptibility to virus-induced diabetesmellitus. Nature 1978; 274: 693–696.

112. Hirasawa K, Jun HS, Maeda K, et al. Possible role ofmacrophage-derived soluble mediators in the pathogenesis ofencephalomyocarditis virus-induced diabetes in mice. J Virol1997; 71: 4024–4031.

113. Hirasawa K, Jun HS, Han HS, et al. Prevention ofencephalomyocarditis virus-induced diabetes in mice byinhibition of the tyrosine kinase signalling pathway andsubsequent suppression of nitric oxide production inmacrophages. J Virol 1999; 73: 8541–8548.

114. Patterson K, Chandra RS, Jenson AB. Congenital rubella,insulitis, and diabetes mellitus in an infant. Lancet 1981; 1:1048–1049.

115. Rubinstein P, Walker ME, Fedun B, et al. The HLA system incongenital rubella patients with and without diabetes. Diabetes1982; 31: 1088–1091.

116. Viskari H, Paronen J, Keskinen P, et al. Humoral beta-cellautoimmunity is rare in patients with the congenital rubellasyndrome. Clin Exp Immunol 2003; 133: 378–383.

117. Hyoty H, Hiltunen M, Reunanen A, et al. Decline of mumpsantibodies in type 1 (insulin-dependent) diabetic children anda plateau in the rising incidence of type 1 diabetes afterintroduction of the mumps-measles-rubella vaccine in Finland.Childhood Diabetes in Finland Study Group. Diabetologia 1993;36: 1303–1308.

118. DeStefano F, Mullooly JP, Okoro CA, et al. Childhoodvaccinations, vaccination timing, and risk of type 1 diabetesmellitus. Pediatrics 2001; 108: E112.

119. Hviid A, Stellfeld M, Wohlfahrt J, et al. Childhood vaccinationand type 1 diabetes. N Engl J Med 2004; 350: 1398–1404.

120. Bishop RF, Unicomb LE, Barnes GL. Epidemiology of rotavirusserotypes in Melbourne, Australia, from 1973 to 1989. J ClinMicrobiol 1991; 29: 862–868.

121. Blomqvist M, Juhela S, Erkkila S, et al. Rotavirus infections anddevelopment of diabetes-associated autoantibodies during thefirst 2 years of life. Clin Exp Immunol 2002; 128: 511–515.

122. Kim A, Jun HS, Wong L, et al. Human endogenous retroviruswith a high genomic sequence homology with IDDMK(1,2)22is not specific for type I (insulin-dependent) diabetic patientsbut ubiquitous. Diabetologia 1999; 42: 413–418.

123. Jaeckel E, Heringlake S, Berger D, et al. No evidence forassociation between IDDMK(1,2)22, a novel isolatedretrovirus, and IDDM. Diabetes 1999; 48: 209–214.

124. O’Brayan TA, Beck MJ, Demers LM, et al. Human parvovirusB19 infection in children with new onset Type 1 diabetesmellitus. Diabet Med 2005; 22: 1778–1779.

125. Guberski DL, Thomas VA, Shek WR, et al. Induction of typeI diabetes by Kilham’s rat virus in diabetes-resistant BB/Worrats. Science 1991; 254: 1010–1013.

126. Ellerman KE, Richards CA, Guberski DL, et al. Kilham rat virustriggers T-cell-dependent autoimmune diabetes in multiplestrains of rat. Diabetes 1996; 45: 557–562.

127. Zipris D, Lien E, Xie JX, et al. TLR activation synergizes withKilham rat virus infection to induce diabetes in BBDR rats. JImmunol 2005; 174: 131–142.

128. Finberg RW, Kurt-Jones EA. Viruses and toll-like receptors.Microbes Infect 2004; 6: 1356–1360.

129. von Herrath MG, Fujinami RS, Whitton JL. Microorganismsand autoimmunity: making the barren field fertile? Nat RevMicrobiol 2003; 1: 151–157.

130. Filippi C, von Herrath M. How viral infections affect theautoimmune process leading to type I diabetes. Cell Immunol2005; 233: 125–132.

131. Nicoletti F, Scalia G, Lunetta M, et al. Correlation betweenislet cell antibodies and anticytomegalovirus Igm and Iggantibodies in healthy 1St-degree relatives of Type-1 (insulin-dependent) diabetic-patients. Clin Immunol Immunopathol1990; 55: 139–147.

132. Banatvala JE, Schernthaner G, Schober E, et al. Coxsackie-B,Mumps, Rubella, and Cytomegalo-Virus Specific Igm responsesin patients with Juvenile-Onset Insulin-dependent Diabetes-mellitus in Britain, Austria, and Australia. Lancet 1985; 1:1409–1412.

133. Ivarsson SA, Lindberg B, Nilsson KO, et al. The prevalenceof type-1 diabetes-mellitus at follow-up of Swedish infantscongenitally infected with cytomegalovirus. Diabet Med 1993;10: 521–523.

134. Hiltunen M, Hyoty H, Karjalainen J, et al. Serologicalevaluation of the role of cytomegalovirus in the pathogenesis ofIddm–a prospective-study. Diabetologia 1995; 38: 705–710.

135. Jenson AB, Rosenberg HS, Notkins AL. Pancreatic islet-celldamage in children with fatal viral-infections. Lancet 1980;2: 354–358.

136. Itoh N, Hanafusa T, Yamagata K, et al. No detectablecytomegalovirus and epstein-barr-virus genomes in thepancreas of recent-onset Iddm patients. Diabetologia 1995;38: 667–671.

137. Foulis AK, McGill M, Farquharson MA, et al. A search forevidence of viral infection in pancreases of newly diagnosedpatients with IDDM. Diabetologia 1997; 40: 53–61.

138. Pak CY, Cha CY, Rajotte RV, et al. Human pancreatic-islet cell specific 38 kilodalton autoantigen identified bycytomegalovirus-induced monoclonal islet cell autoantibody.Diabetologia 1990; 33: 569–572.

139. Martin S, Wolf-Eichbaum D, Duinkerken G, et al. Briefreport–development of type 1 diabetes despite severehereditary B-cell deficiency. N Engl J Med 2001; 345:1036–1040.

140. Dobrescu D, Ursea B, Pope M, et al. Enhanced Hiv-1 replicationin V-Beta-12 T-Cells due to human cytomegalovirus inMonocytes–evidence for a putative herpesvirus superantigen.Cell 1995; 82: 753–763.

141. Compton T, Kurt-Jones EA, Boehme KW, et al. Humancytomegalovirus activates inflammatory cytokine responses viaCD14 and toll-like receptor 2. J Virol 2003; 77: 4588–4596.

142. Mordes JP, Guberski DL, Leif JH, et al. LEW.1WR1 rats developautoimmune diabetes spontaneously and in response toenvironmental perturbation. Diabetes 2005; 54: 2727–2733.

143. Tirabassi RS, Guberski DL, Leif JH, et al. Maternal Immu-nization Protects Weanling LEW. 1WR1 Rats from Dia-betes Induced by Multiple Virus Infections. Diabetes (2006),http://scientificsessions.diabetes.org/Abstracts/index.cfm?fuseaction=Locator.SearchAbstracts (accessed October 27,2006).

144. Pass RF. Cytomegalovirus. In Fields Virology, Knipe DM,Howley PM (eds). Lippincott Williams and Wilkins:Philadelphia, PA, 2001; 2675–2705.

145. Numazaki K. Human cytomegalovirus infection of breast milk.FEMS Immunol Med Microbiol 1997; 18: 91–98.

146. Mocarski ES. Cytomegaloviruses and their replication. In FieldsVirology, Fields BN, Knipe DM, Howley PM (eds). LippincottWilliams and Wilkins: Philadelphia, PA, 2001; 2629–2673.

147. Blankenhorn EP, Rodemich L, Martin-Fernandez C, et al. Therat diabetes susceptibility locus Iddm4 and at least oneadditional gene are required for autoimmune diabetes inducedby viral infection. Diabetes 2005; 54: 1233–1237.


Virus Infections in Type 1 Diabetes

Ken T. Coppieters, Tobias Boettler, and Matthias von Herrath

Center for Type 1 Diabetes Research, La Jolla Institute for Allergy and Immunology, La Jolla,California 92037

Correspondence: [email protected]

The precise etiology of type 1 diabetes (T1D) is still unknown, but viruses have long beensuggested as a potential environmental trigger for the disease. However, despite decadesof research, the body of evidence supporting a relationship between viral infections andinitiation or acceleration of islet autoimmunity remains largely circumstantial. The mostrobust association with viruses and T1D involves enterovirus species, of which somestrains have the ability to induce or accelerate disease in animal models. Several hypotheseshave been formulated to mechanistically explain how viruses may affect islet autoimmunityand b-cell decay. The recent observation that certain viral infections, when encountered atthe right time and infectious dose, can prevent autoimmune diabetes illustrates that potentialrelationships may be more complex than previously thought. Here, we provide a concisesummary of data obtained in mouse models and humans, and identify future avenuestoward a better characterization of the association between viruses and T1D.

WHY AN ENVIRONMENTAL FACTORFOR T1D?

Although there is a well-documented geneticbasis for T1D, its rising incidence in devel-

oped countries has to be attributed to environ-mental changes (Bodansky et al. 1992; Gillespieet al. 2004). Population genetics simply do notchange enough between generations to accountfor the dramatic surge in T1D prevalence as ob-served in, for example, Finland (Harjutsalo et al.2008). Moreover, varying disease penetrancerates and the differing ages of disease onset inmonozygotic twins are suggestive of nongeneticvariables in T1D pathogenesis (Redondo et al.2008). It has been postulated that the largest in-crease in disease frequency is occurring in the

most developed countries, following a northto south gradient that is reminiscent of the dis-tribution seen in other autoimmune diseases(Borchers et al. 2010). Indeed, the “hygiene hy-pothesis” has frequently been referenced in thiscontext to explain high ratios of autoimmunedisease by the relative lack of exposure to infec-tious agents (Bach 2002). This theory, as well asthe relationship between hygiene standards andT1D incidence is, however, far from absolute asexemplified by the second-highest worldwide in-cidence in Sardinia, Italyor the substantially lowerincidences found in the countries surroundingFinland, which has the world’s highest rate. Thisarticle will provide an overview of the data thatsupport the implication ofviruses as environmen-tal triggers of islet autoimmunity in T1D.

Editors: Jeffrey A. Bluestone, Mark A. Atkinson, and Peter R. Arvan

Additional Perspectives on Type 1 Diabetes available at www.perspectivesinmedicine.org

Copyright # 2012 Cold Spring Harbor Laboratory Press; all rights reserved; doi: 10.1101/cshperspect.a007682

Cite this article as Cold Spring Harb Perspect Med 2012;2:a007732

1

ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

HISTORY OF VIRAL ASSOCIATIONSWITH T1D

Suspects Released without Charge

It has long been acknowledged in the medicalliterature that T1D onset within populationsfollows a seasonal pattern; a finding which ini-tially led to the formulation of a viral etiology(Adams 1926 ). Ever since, a variety of viruseshave come under scrutiny as potential induc-ers of T1D. In the majority of cases, however,associations are weak, irreproducible, or werein some instances convincingly disproved. Ex-amples of initially promising data include re-ports on T1D association with cytomegalovirus(CMV) (Pak et al. 1988), parvovirus (Guberskiet al. 1991; Kasuga et al. 1996), encephalomyo-carditis virus (Craighead and McLane 1968),and retroviruses (Conrad et al. 1997), all ofwhich were challenged or are awaiting scientificreplication.

Rotaviruses, responsible for a significantshare of childhood gastroenteritis, have alsobeen subjected to investigation for their rela-tionship to T1D. Interest was sparked by thedemonstration of sequence analogies betweenT-cell epitopes within the islet antigens GADand IA-2 and rotavirus protein, suggestingpotential cross-reactivity mechanisms (Honey-man et al. 1998). An association between rotavi-rus infection and islet autoantibody positivityin at-risk children was reported (Honeymanet al. 2000), but later challenged by studiesin the Finnish population (Blomqvist et al.2002; Makela et al. 2006). Therefore, it can beconcluded that the role of rotavirus in the etiol-ogy of T1D is unconfirmed.

Congenital rubella infection and subse-quent onset of diabetes after birth constitutesan interesting paradigm (Gale 2008). Congenitalrubella syndrome encompasses an array of phys-ical and behavioral abnormalities. Of note,diabetes development is associated with thepresence of the DR3-DQ2 T1D susceptibilityhaplotype (Menser et al. 1974). It is thoughtthat the virus causes diabetes by disturbingthe normal development of b-cell mass, ratherthan inducing islet autoimmunity (Pattersonet al. 1981). Since the introduction of an effi-

cient vaccine in 1969, the virus has been largelyeliminated in developed countries, and thusis not expected to be a factor in the currentincrease of T1D incidence. Beyond this, themere notion of whether this form of diabeteshas any relationship to T1D also remains indoubt (Gale 2008).

Mumps infection has also been implicatedin some instances, including a recent reportdocumenting a case of “fulminant” T1D (Gotoet al. 2008). In analogy to rubella, however, effi-cacious vaccination programs have not beenable to curb the rising T1D incidence levels(Honeyman 2005).

Enteroviruses and T1D: A Long History

The most robustly documented correlation be-tween a virus and T1D has been with enterovi-ruses, a viral single-stranded RNA (ssRNA) ge-nus belonging to the picornaviruses. A recentmeta-analysis by Yeung and coworkers estab-lished that there is a clinically significant as-sociation between enterovirus infection, detectedwith molecular methods, and T1D (Yeung et al.2011). Early reports suggesting a link betweencoxsackievirus, a member of the enterovirus ge-nus, and T1D showed higher neutralizing anti-body titers in serum from recent-onset patientsas compared to nondiabetic subjects (Gambleet al. 1969), and were later confirmed usingconventional polymerase chain reaction (PCR)testing (Clements et al. 1995). Some studies si-multaneously probed for antibodies againstother viruses and found that the most signifi-cant association was with coxsackievirus (Ba-natvala et al. 1985).

Cross-sectional studies have focused pre-dominantly on recent-onset individuals withT1D, although enterovirus was also identifiedas a risk factor in prediabetic children (Sade-harju et al. 2001) and pregnant women (Dahl-quist et al. 1995; Hyoty et al. 1995; Elfvinget al. 2008). There is still a lack of large prospec-tive studies that establish a clear temporal re-lation between enterovirus infection and thedevelopment of islet autoimmunity. That said,two recent studies offer support for a causativerole for enteroviruses in T1D. Studying blood

K.T. Coppieters et al.

2 Cite this article as Cold Spring Harb Perspect Med 2012;2:a007732

ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

samples collected by the Diabetes and Autoim-munity Study in the Young (DAISY) consor-tium, Stene et al. found that the rate of pro-gression from islet autoimmunity (detectionof islet autoantibodies) to T1D was significant-ly increased following detection of enteroviralRNA in serum (Stene et al. 2010). Oikarinenand colleagues further showed that detectionof enterovirus RNA is associated with increasedrisk for the primary development of islet auto-immunity, with a peak infection frequencyduring the 6 mo window that precedes theappearance of islet autoantibodies (Oikarinenet al. 2011). In combination, these reports sug-gest that enteroviruses may be involved in boththe initiation of islet autoimmunity as well asprogression to overt hyperglycemia and thus,act at multiple stages of disease development.

The timing of enteroviral infection as re-lated to T1D onset is, in a sense, an altogetherunresolved issue. Data from the nucleotideoligomerization domain (NOD) mouse seemto favor a scenario in which insulitis serves asa prerequisite for coxsackievirus to be diabeto-genic (see next section) (Horwitz et al. 2001;Drescher et al. 2004). Susceptible individualsmay thus suffer from subclinical insulitis foryears before a viral challenge eventually culmi-nates in hyperglycemia.

The notion that any potential association isnot absolute and depends, to a considerable de-gree, on genetic susceptibility, or perhaps actsin concert with other environmental factors, issupported by several studies noting no such cor-relation (Fuchtenbusch et al. 2001; Graves et al.2003). A 1971 study followed the diabetes inci-dence rate after a well-documented epidemic ofcoxsackievirus B4 (CVB4) infection in the re-mote Pribilof Islands (Alaska, USA). Five yearslater, the incidence of diabetes in the infectedversus noninfected persons was found to be un-affected (Dippe et al. 1975). CVB4 isolates do,reportedly, have the intrinsic capacity to infectb cells and cause insulitis and diabetes in sus-ceptible mouse strains (Yoon et al. 1978) follow-ing their direct isolation from a child after theironset of T1D (Yoon et al. 1979). Despite the vi-rus’ perceived islet tropism, host susceptibilityor additional environmental factors are thus

required for diabetogenicity. That said, someof these facets, including the identification ofmultiple strains (isolated from human pan-creata) having the ability to induce diabetes re-main rare in reports, and require more in theway of additional study.

ANIMAL MODELS AND MECHANISMS BYWHICH VIRUSES COULD IMPACT T1D

Animal Models for Studies of theRole for Viruses in T1D

Despite their shortcomings, animal models re-main indispensable tools to map pathologicalmechanisms. Although the use of rodents inT1D research is discussed in a distinct articleof this collection, we will summarize here whichviruses have shown the ability to induce or alterexperimental diabetogenic responses in vivo.

Mice

One of the oldest known and most unequivocalrelationships between viral infection and dia-betes development was revealed after inocula-tion of mice with encephalomyocarditis virus(EMC; picornavirus, ssRNA) (Craighead andMcLane 1968). Diabetes induction usuallyoccurs 3–4 d after infection and critically de-pends on the virus variant used (Onoderaet al. 1978b), dosing (Baek and Yoon 1991),and the genetic background of the host (Ono-dera et al. 1978b). The virus induces diabetesafter specifically infecting pancreatic b cells fol-lowed by direct cytolysis (high dose) or the re-cruitment of macrophages (low dose) (Baekand Yoon 1991; Yoon and Jun 2006). Parallelshave been drawn between this model and theacute subform of fulminant diabetes that isfound in the Japanese population based on itsrapid and aggressive onset, exocrine tissue dam-age, and lack of autoantibodies (Shimada andMaruyama 2004). A recent study exploited thismodel to show that the viral sensor moleculesMDA5 and TLR3, which are both involved inIFN-I responses to viral infection, are requiredto prevent diabetes in mice infected withEMCV (McCartney et al. 2011).

Virus Infections in Autoimmune Diabetes

Cite this article as Cold Spring Harb Perspect Med 2012;2:a007732 3

ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

A more ambiguous role is reserved for cox-sackie B viruses (CVB; picornavirus, ssRNA) inthe NOD mouse. Interest was sparked by theisolation of a virus resembling CVB4 from a10-yr-old boy, which famously triggered insu-litis and hyperglycemia in mice (Yoon et al.1979). Detailed analysis in the NOD mouse re-vealed that coxsackieviruses provoke diabetesonly when a preexisting mass of insulitis hasaccumulated (Serreze et al. 2000). When ad-ministered earlier, however, inoculation has astrong preventive outcome (Tracy et al. 2002;Filippi et al. 2009). This model is thus highlysuitable to study the combined effect of geneti-cally determined immunological abnormali-ties and an environmental factor (i.e., viral in-fection).

A final example of a widely studied viralmouse model differs from the previous two bythe fact that the host has been genetically alteredto express a viral antigen from the lymphocyt-ic choriomeningitis virus (LCMV; arenavirus;ssRNA) on its b cells (Ohashi et al. 1991; Old-stone et al. 1991). This is thus a pure mimicrymodel, in which antiviral T cells redirect tothe b cells after viral clearance and, dependingon expression of viral antigen in the thymus,cause rapid or slow onset of clinical hyperglyce-mia (von Herrath et al. 1994). Our laboratoryrecently developed an advanced two-photonimaging approach to allow for imaging of cyto-toxic T lymphocyte (CTL) effector kinetics invivo in the pancreas using this model (Fig. 1)(Coppieters et al. 2010). Although mechanismsof cross-reactivity at play in the rat insulin pro-moter (RIP)-LCMV model more than likely di-verge from the etiology of T1D, we propose thatthe behavior of activated CTL in the target tissuemay still serve as a suitable model of their ki-netics in T1D.

Rats

Although less commonly used as an animalmodel, valuable information on virus-induceddiabetes has been obtained from studies in rats.A well-documented diabetogenic virus is theKilham rat virus (KRV; parvovirus; ssDNA)that causes diabetes in the diabetes-resistant

BioBreeding (DR-BB) rat (Guberski et al. 1991).In contrast with EMCV, KRV does not infect bcells but rather promotes diabetes by inductionof autoimmunity against the b cells. Recruit-ment of macrophages and perturbation of regu-latory and autoreactive T-cell species have beenproposed as possible mechanisms that contrib-ute to b-cell decay in this model (Mordes et al.2004). More recently, infection with KRV, butalso rat cytomegalovirus (RCMV), was foundto result in autoimmune diabetes in LEW�1WR1rats, a strain that normally experiences spon-taneous onset in only 2% of cases (Tirabassiet al. 2010). Other viruses such as H-1, vaccinia,and CVB4, however, did not induce diabetes,indicating that diabetogenicity is virus-specific.

Table 1 lists other, less characterized, obser-vations in mice and rats that link viral infec-tionwith diabetes development. In the followingsubsections we will list some of the hypothesesthat have been put forward based primarily onresearch in the animal models outlined above(Fig. 2).

70 μm

0d00:02:00.000

Figure 1. Antiviral, diabetogenic CTL in the RIP-LCMV.GP diabetes model as visualized by intravitaltwo-photon imaging in the pancreas. The techniqueis described in Coppieters et al. (2010). Briefly, puri-fied and labeled P14 TCR-transgenic CTL is trans-ferred to RIP-LCMV.GP hosts, followed by infectionwith LCMV. As such, the kinetics of the diabetogen-ic CTL response can be dynamically visualized inthe pancreas. (Magenta) Insulin-producing b cells,(green) CTL, (red) vasculature after dye injection.



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

Evidence for Islet-Specific Infection

The possibility of a viral infection specificallyaffecting pancreatic endocrine cells constitutesa straightforward explanation for the selectivedemise of b cells, either through lysis inducedby cytopathic viruses or immune-mediated de-struction of infected b cells. The example ofEMCV-induced diabetes highlights the poten-tial of some viruses to specifically infect pan-creatic islets and cause b-cell decay. Coxsackie-virus, in contrast, displays pancreas tropismrather than a preference for b cells, and it hasbeen reported that it exclusively affects acinarcells while sparing the islet cells (Mena et al.2000). Although some evidence exists that CVBcan indeed infect islets in vivo (Drescher et al.2004), it is uncertain whether the virus can per-sist and whether islet-specific infection is re-quired for diabetogenicity. The detection of viralparticles in the human T1D pancreas will be dis-cussed separately.

Molecular Mimicry

The demonstration of remarkable sequencesimilarities between the 2C protein from cox-sackievirus and a GAD65 (glutamate decarbox-ylase) epitope, a major target antigen in T1D,gave rise to the idea of viral mimicry in T1D(Kaufman et al. 1992). Although molecularmimicry is undoubtedly the reason for rheu-matic heart disease following infection withStreptococcus pyogenes (Guilherme et al. 1995),in T1D, such type of connection proved difficultto establish. It was shown that CVB infection ofyoung NOD mice failed to activate or expandthe autoreactive precursors that are specific forthose b-cell epitopes that share structural simi-larities with viral epitopes (Horwitz et al. 1998).Analogously, autoreactive human T-cell clonesspecific for the GAD65 epitope did not pro-liferate following stimulation with the viralepitope (Schloot et al. 2001). Studies in themouse RIP-LCMV system suggest that complete

Table 1. Additional examples of viruses that can promote or prevent autoimmune diabetes in mice or rats

Virus Classification

Host species—

Diabetogenicity Reference

Ljungan virus Picornaviridae (ssRNA) Viral particles detected in

wild-trapped bank volesand BB rats.

Niklasson et al. 2003,

2007

MHV-68 Herperviridae (dsDNA) Delays diabetes in the NODin an age-dependentmanner.

Smith et al. 2007

Rotavirus Reoviridae (dsRNA) Delays or acceleratesdiabetes in the NOD in anage-dependent manner.

Graham et al. 2007,2008

Reovirus Reoviridae (dsRNA) Induces diabetes and infectsb cells. Delays diabetes in

the NOD mouse.

Onodera et al. 1978a;Wetzel et al. 2006

Retrovirus Retroviridae (ssRNA) Presence of endogenousretrovirus correlates withdiabetes susceptibility

and progression.

Suenaga and Yoon1988; Gaskinset al. 1992

Mengovirus Picornaviridae (ssRNA) Causes diabetes in additionto encephalitis owing toislet necrosis.

Yoon et al. 1984

Mouse hepatitis virus Coronaviridae (ssRNA) Reduces incidence in the

NOD mouse.

Wilberz et al. 1991

Lactate dehydrogenasevirus

Arteriviridae (ssRNA) Suppresses diabetesdevelopment in the NODmouse.

Takei et al. 1992



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

sequence identity is required to initiate diabetesas single amino acid changes in the viral epitopeon b cells protected these mice from diabetes(Sevilla et al. 2000). Collectively, the currentbody of evidence favoring a simple event ofcross-reactivity between viral and self-antigen isweak (Richter et al. 1994; Horwitz et al. 1998;Schloot et al. 2001). Nevertheless, molecularmimicry has been shown to accelerate diseaseprogression under conditions of virus-induced

pancreatic inflammation, suggesting that se-quence homologies may not be the initiatingtrigger, but are able to codetermine the pace bywhich disease develops (Christen et al. 2004).

The Concept of “Bystander” Activation

Bystander T-cell activation is defined as the ac-tivation of a T cell through a mechanism that isindependent of specific T-cell receptor (TCR)

Inflamed islets withimmune infiltrates

Uninflamed islets

Virus infection

Insulitis and highviral titers

Low viral titers, mildinsulitis and eventuallyviral clearance

Increased chances for harmful viralinfluences through…

Bystanderactivation

Immuno-regulation

Protectiveimmunity

Emergence of regulatory cell subsetsand activation of regulatory pathwaysresulting in deletion of effectorimmune cells to preventimmunopathology

Exposure to and successful eliminationof pancreatotropic viruses prior to theonset of insulitis results in immunitythat protects a disease-susceptiblehost from potentially more harmfulinfections at a later time point

Antigenicspreading

Molecularmimicry

Unspecific activation of unrelated T-cell specificities due to ongoingtissue inflammation initiated by aviral infection

Presentation of self-antigens in thecontext of inflammation can lead tothe priming or activation of otherautoactive immune cell clones

Structural similarities between a viralepitope targeted by the immunesystem and a self-protein present inpancreatic islets

Increased chances for protective viralinfluences through…

Figure 2. Potential immune mechanisms leading to islet inflammation or tolerance following viral infection.



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

stimulation. Alternative mechanisms encom-pass activation through soluble factors or mem-brane-bound molecules that bind to receptorsother than the TCR. In a strict sense, mecha-nisms such as “epitope spreading” or “molecu-lar mimicry” are not included, because these in-volve specific recognition of a presented peptideby the TCR. Thus, bystander T-cell activationcircumvents the requirement for specific TCRstimulation.

In spontaneous diabetes models, antigenspecificity appears to be required for migrationof autoreactive T-cell species to the pancreaticislet (Lennon et al. 2009). Recognition of isletantigens was shown to occur early at the vascu-lar level, ensuring that only antigen-specific Tcells can extravasate and access the islets (Savi-nov et al. 2003). In the context of CVB infection,circulating naive islet-specific T cells became ac-tivated, either through release of sequesteredantigen or by “true” non-TCR bystander activa-tion, and rapidly triggered diabetes develop-ment (Horwitz et al. 1998). However, the rolefor non-TCR-dependent activation of naıve Tcells during viral infection appears limited(Ehl et al. 1997; Zarozinski and Welsh 1997).

In contrast to “bystander activation” ofnaıve T cells, the induction of “bystander dam-age” via, e.g., cytokine release may be the morerelevant scenario during virus-induced diabetes(Seewaldt et al. 2000). Here, the b cells are thebystanders and are killed in a non-TCR-de-pendent fashion through the release of solublefactors by antiviral effectors.

The Fertile Field Hypothesis

The fertile field hypothesis combines the afore-mentioned concepts and postulates that virus-induced inflammation preconditions (“fertil-izes”) the pancreas milieu for autoimmunity,which ultimately results in T1D only in suscep-tible individuals (von Herrath et al. 2003). Fol-lowing the establishment of localized inflam-mation by the virus, autoreactive T cells aregenerated by molecular mimicry or bystanderactivation, or a combination of both. The subse-quent demise of b cells and consequent presen-tation of b-cell antigens in the draining lymph

nodes would then lead to epitope and antigenicspreading. This would explain the emergenceof a broad autoreactive T-cell repertoire overtime in most T1D individuals. Mouse modelsof virally induced T1D and multiple sclerosis(MS) also clearly show the occurrence of second-ary autoreactivity against molecules that are notinitially targeted by the antiviral effectors thatdrive the early stages (Miller et al. 1997; Coonet al. 1999; Holz et al. 2000). Contradicting thefertile field hypothesis are studies that empha-size preexisting insulitis as a crucial prerequisitefor a viral infection to cause T1D (see above).Thus, it remains to be conclusively determinedwhether viral infections are required to producedisease-promoting conditions in autoimmun-ity-prone individuals or, conversely, if preexist-ing autoimmunity determines the outcome of apancreatropic viral infection.

EVIDENCE FOR VIRAL INFECTIONOF THE PANCREAS

The target organ in T1D, the pancreas, is un-fortunately extremely difficult to study becauseof its inaccessible anatomical location (Coppi-eters and von Herrath 2009). Owing to greatlyimproved clinical management of T1D, pan-creas samples from recently diagnosed T1Dpatients today only rarely become available.Programs such as the Network for PancreaticOrgan Donors (nPOD) (www.jdrfnpod.org),aimed at the nationwide procurement of tissuerelevant to T1D research, respond to this unmetneed and will also offer samples from non-diabetic, islet-antibody positive individuals. Atpresent, the important question as to whetherviruses, and in particular, enteroviruses, can di-rectly affect pancreatic islets has been addressedby a relatively small set of studies.

Indirect Evidence: The “Viral Signature”

Foulis and coworkers were the first to systemati-cally document the hyperexpression of HLAClass I and interferon-awithin islets of recentlydiagnosed diabetic children (Fig. 3) (Foulis et al.1987). Normal islets are completely devoid ofthese markers. The presence of these markers



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

is commonly referred to as a “viral signature,” asthe up-regulated expression of HLA/MHC classI molecules is typically driven by type I IFNs fol-lowing viral infection. The islet-specific MHCclass I expression would render the cells suitabletargets to CD8 T cells that are known to be anintegral part of insulitic lesions. Indeed, in amouse model of type T1D, only b cells thathave been unmasked by MHC class I expressionwere attacked by activated, autoreactive T cells,demonstrating the necessity of MHC class Iup-regulation for immune-mediated b-cell de-struction (von Herrath et al. 1994). In this con-text it is interesting to note that defined poly-morphisms in the IFIH1 gene may result inlower levels of type I interferons (IFNs) in re-sponse to viral infections and may confer pro-tection from autoimmune diabetes (Nejentsevet al. 2009).

CVB-induced hyperexpression of the inter-feron-inducible chemokine CXCL-10 by pan-creatic islet cells was also proposed as an earlymolecular marker of infection (Berg et al.2006). Its localized production was found to co-incide with islet-specific enteroviral infectionin fulminant T1D, but also conventional T1D(Tanaka et al. 2009; Roep et al. 2010). Somedata from animal models suggests that virus-in-duced CXCL-10 is essential in the recruitmentof CXCR-3þ autoreactive T cells to the islets,although recent studies in our laboratory foundthat CXCL10 is in fact dispensable (Christenet al. 2003) (KT Coppieters and MG von Her-rath, unpubl.). Collectively, although the indis-pensable role of CXCL10 has yet to be con-firmed, these data suggest that viral infections

have the potential to establish a molecular “sig-nature” that aids in the recruitment of diabeto-genic T cells to pancreatic islets.

Direct Evidence of Viral Infectionof the Pancreas in T1D

Whereas initial efforts to directly detect viral se-quences in the islets from patients with MHCclass I up-regulation were unsuccessful (Fouliset al. 1990, 1997), the same samples were re-cently revisited using a more modern method-ology (Richardson et al. 2009). Using immu-nohistochemical detection, enteroviral particleswere found in islets from 44 out of 72 recent-on-set patients versus three out of 50 controls. It isworth noticing that positive detection was alsoachieved in 10 out of 25 type 2 diabetes patients.Another group probed 65 pancreata from T1Dpatients for enteroviral RNA by in situ hybrid-ization and found enterovirus-positive islet cellsin four cases, compared to none in nondiabetics(Ylipaasto et al. 2004). Likewise, Dotta and co-workers found immunohistochemical traces ofenteroviral protein in three out of six pancreaticislets from recent-onset patients and corrobo-rated these data by sequencing (Dotta et al.2007). Of interest, islets from these patientsshowed an unusual, NK cell-dominated formof insulitis. In contrast to the above analysesperformed after diagnosis, a recent case reportsampled pancreatic tissue from an autoanti-body-positive, nondiabetic child (Oikarinenet al. 2008b). It was found that the pancreatic is-lets had no signs of insulitis, in line with datafrom In’t Veld et al. in a larger cohort (In’t

INSULIN MHC class I CD8

Figure 3. Aberrant MHC class I hyperexpression in islets from a 12-yr-old boy with one year of T1D duration(nPOD case 6052). Frozen pancreas section was stained for immunofluorescence and analyzed by confocal mi-croscopy. Note the abundance of CD8þ cellular infiltration. (For details see Gianani et al. 2010.)



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

Veld et al. 2007). However, enterovirus was de-tected by immunohistochemistry specifically inthe islets but not by in situ hybridization. Await-ing confirmation from larger studies, it is possi-ble that this isolated case exposes the very earlystages of T1D development, showing the estab-lishment of local viral infection even before on-set of subclinical insulitis.

Anatomical sites close to the pancreas mayserve as a reservoir for viral infections in T1D.Oikarinen et al. subjected intestinal biopsiesfrom 12 T1D patients to in situ hybridiza-tion and immunohistochemistry for enterovi-rus (Oikarinen et al. 2008a). Detection was ac-hieved in 50% of T1D patients as compared tonone in nondiabetic subjects. Based on thesefindings, the authors postulated that the gut—which showed normal histopathological com-position in all cases—could represent a chronicinfectious site with close anatomical ties to thepancreatic milieu. Alternatively, immune recog-nition could occur at the intestinal level fol-lowed by subsequent homing of activated lym-phocytes to the pancreas.

It seems rather unlikely that the majority ofT1D cases are attributable to direct lysis of in-fected b cells. Such a scenario would result ina more acute onset of disease, which is not thedisease phenotype that is commonly observedin patients, although such cases have been de-scribed in the literature (Imagawa et al. 2000).Although evidence of enteroviral infection hasbeen shown in these fulminant diabetes cases,the pathogenesis observed in such patients lacksthe development of islet autoreactivity (e.g., is-let autoantibodies), a fundamental feature ofconventional T1D (Tanaka et al. 2009).

DIFFICULTIES IN IDENTIFYING A ROLEFOR VIRUSES IN T1D AND THERAPEUTICIMPLICATIONS

Temporal Divergence of Viral Insultand Clinical Onset

Despite decades of investigation, no evidenceexists for the involvement of a particular viralstrain with T1D. The search for a correlationwith certain viral agents can be expected to be

complex for a variety of reasons. In both mousemodels and T1D patients, development of clin-ical hyperglycemia is thought to represent the fi-nal stage of the autoimmune process. Therefore,it can be assumed that the event that initiatesthe loss of tolerance against islet antigens likelyprecedes the onset of diabetes by several monthsor years. This temporal discrepancy poses con-siderable restraints in studying the role of viralinfection in T1D development, as the vast ma-jority of patients are traditionally sampled afterdiagnosis. Taking into account that inciting vi-ral agents may use a “hit-and-run” strategy, oract by repeated, sequential infection, analysesaround clinical onset may at least in some occa-sions miss out on the culprits.

Alternatively, viral infection may only serveas an accelerating factor that, superimposedonto advanced insulitis, leads to rapid culmi-nation into hyperglycemia. The latter scenariowould suggest that detection of viral particlesaround onset is an achievable goal in determin-ing a causal relationship.

Multiple Viruses May Provoke Diseasein Similar Fashion

The fact that no absolute association has beenidentified with certain viral strains or even viralgenuses or families indicates that, if T1D isindeed caused by viruses, multiple infectiousstrains may result in the same disease pheno-type. Historically, samples from T1D patientshave been probed by a “one test-one pathogen”approach that unavoidably introduces experi-mental bias. Whether the assay is based on de-tection of virus-specific antibodies or nucleicacid sequences, such strategies are costly, in-efficient, and time-consuming and generallymake poor usage of the limited sample volumesthat are available from T1D patients. To cast thenet more widely in the evaluation of a viraletiology, emerging nucleic acid technologiesto detect pathogens on a broad-spectrum basisshould be applied on blood and pancreassamples from T1D patients at various stagespre- and postdiagnosis. Indeed high-densitymicroarrays, such as the Virochip pan-viralmicroarray and deep sequencing, can test for



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

thousands of potential pathogens simultane-ously (Wang et al. 2002, 2003).

Alignment with the Hygiene Hypothesis

Based on our present knowledge, enteroviruseswould appear associated with at least a fractionof T1D cases. But if enteroviruses are indeed amajor contributor to T1D pathogenesis, howcan we explain the increase in T1D incidencein countries where exposure to enteroviruses hasbeen dropping (e.g., Finland) (Viskari et al.2005)? In other words, is the theory that T1Dcan be caused by a viral infection compatiblewith the hygiene hypothesis? Based on the find-ings in the NOD mouse, one could argue thatthe lack of exposure to enteroviruses in devel-oped countries results in a reduced frequency ofindividuals with protective immunity throughearly childhood infections. When geneticallydriven islet inflammation occurs in these un-protected individuals, they would be more sus-ceptible to an enteroviral infection that has thepotential to initiate overt autoreactivity and b-cell damage.

Immunization Strategies: Why Not Now?

So why don’t we initiate population-wide vacci-nation programs to more thoroughly and di-rectly evaluate the role of enterovirus in T1D?Theoretically, virologists deem the developmentof enterovirus vaccines relatively straightfor-ward and achievable (S Tracy, pers. comm.).The main limitation at present is that the enter-ovirus genus of the Picornavirida family con-sists of five virus species. These virus speciesin turn contain many different strains and sero-logically distinct viruses (Fauquet 2005). Anyone or a combination of these viruses could bethe virus detected by, for example, the anti-VP1antibody that is commonly used in immuno-histochemical analysis. Finnish groups are cur-rently attempting to delineate which enteroviralstrains are most prevalent in T1D patients toclarify serotypes that should be immunizedagainst (Roivainen 2006).

A disturbing observation related to the ideaof prophylactic vaccination is the finding that

CVB infection protects against diabetes de-velopment in the young NOD mouse. This pro-tection is orchestrated via at least two distinctsuppressive immune mechanisms, the up-regu-lation of the inhibitory PD-1/PD-L1 pathwayand increasing numbers of circulating T cellswith regulatory capacities (Filippi et al. 2009).Such data illustrate the dual role of viral in-fections in autoimmunity, and portray T1D de-velopment as a balancing act between immune“education” by viruses (see “hygiene” hypothe-sis) and the induction of aberrant immunity inresponse to these agents. Moreover, they suggestthat the protective effect of viral infections isa proactive mechanism that involves the emer-gence of regulatory mechanisms and thus,exceeds the achievement of sterile immunitywhich would be the ultimate goal in vaccinationprograms.

As a final note, the introduction of child-hood immunizations programs and the grow-ing prevalence of T1D in developed countrieshave also provided rationale for assessing a pos-sible correlation between the two entities. Mul-tiple large-scale studies found no support forany causal relation between childhood vaccina-tion and T1D (Blom et al. 1991; EURODIABSubstudy 2 Study Group 2000; DeStefano et al.2001; Hviid et al. 2004). As there appears tobe no significant association between vaccina-tion and T1D, the risk-benefit ratio as of todaybalances strongly in favor of continued protec-tion efforts by means of immunization.

CONCLUDING REMARKS

The available data set on the role of viral infec-tions in T1D development and progression al-lows us to conclude with a reasonable degreeof confidence, that at least a fraction of patientsat some point suffer some type of viral insult.Viruses belonging to the enterovirus genushave the capacity to initiate and/or accelerateislet autoimmunity, but cannot fully explainthe etiology as a sole environmental trigger.What is needed is a comprehensive, microar-ray-based approach using patient samples atvarious stages pre- and postdiagnosis of disease.Emerging PCR-based technologies should be



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

used to offer a more definitive and unbiasedevaluation of the potential role of viruses inT1D pathophysiology. Results from large longi-tudinal follow-up studies such as DAISY andThe Environmental Determinants of Diabetesin the Young (TEDDY) study are expected tocontribute significantly in the near future. In-depth knowledge on which viruses act when,and at which anatomical level, could enable usto design rational vaccination approaches insusceptible individuals for T1D prevention.

ACKNOWLEDGMENTS

The sample in Figure 1 was obtained with thesupport of the Network for Pancreatic OrganDonors with Diabetes (nPOD), a collaborativetype 1 diabetes research project sponsoredby the Juvenile Diabetes Research FoundationInternational (JDRF). Organ Procurement Or-ganizations (OPO) partnering with nPOD toprovide research resources are listed at www.jdrfnpod.org/our-partners.php.

REFERENCES

Adams SF. 1926. The seasonal variation in the incidence ofacute diabetes. Arch Intern Med 37: 861–864.

Bach JF. 2002. The effect of infections on susceptibility toautoimmune and allergic diseases. N Engl J Med 347:911–920.

Baek HS, Yoon JW. 1991. Direct involvement of macro-phages in destruction of b-cells leading to developmentof diabetes in virus-infected mice. Diabetes 40: 1586–1597.

Banatvala JE, Bryant J, Schernthaner G, Borkenstein M,Schober E, Brown D, De Silva LM, Menser MA, SilinkM. 1985. Coxsackie B, mumps, rubella, and cytomegalo-virus specific IgM responses in patients with juvenile-onset insulin-dependent diabetes mellitus in Britain,Austria, and Australia. Lancet 1: 1409–1412.

Berg AK, Korsgren O, Frisk G. 2006. Induction of the che-mokine interferon-g-inducible protein-10 in humanpancreatic islets during enterovirus infection. Diabetolo-gia 49: 2697–2703.

Blom L, Nystrom L, Dahlquist G. 1991. The Swedish child-hood diabetes study. Vaccinations and infections as riskdeterminants for diabetes in childhood. Diabetologia34: 176–181.

Blomqvist M, Juhela S, Erkkila S, Korhonen S, Simell T,Kupila A, Vaarala O, Simell O, Knip M, Ilonen J. 2002.Rotavirus infections and development of diabetes-associ-ated autoantibodies during the first 2 years of life. ClinExp Immunol 128: 511–515.

Bodansky HJ, Staines A, Stephenson C, Haigh D, CartwrightR. 1992. Evidence for an environmental effect in the aeti-ology of insulin dependent diabetes in a transmigratorypopulation. BMJ 304: 1020–1022.

Borchers AT, Uibo R, Gershwin ME. 2010. The geoepidemi-ology of type 1 diabetes. Autoimmun Rev 9: A355–A365.

Christen U, McGavern DB, Luster AD, von Herrath MG,Oldstone MB. 2003. Among CXCR3 chemokines,IFN-g-inducible protein of 10 kDa (CXC chemokineligand [CXCL] 10) but not monokine induced byIFN-g(CXCL9) imprints a pattern for the subsequentdevelopment of autoimmune disease. J Immunol 171:6838–6845.

Christen U, Edelmann KH, McGavern DB, Wolfe T, Coon B,Teague MK, Miller SD, Oldstone MB, von Herrath MG.2004. Aviral epitope that mimics a self antigen can accel-erate but not initiate autoimmune diabetes. J Clin Invest114: 1290–1298.

Clements GB, Galbraith DN, Taylor KW. 1995. Coxsackie Bvirus infection and onset of childhood diabetes. Lancet346: 221–223.

Conrad B, Weissmahr RN, Boni J, Arcari R, Schupbach J,Mach B. 1997. A human endogenous retroviral superan-tigen as candidate autoimmune gene in type I diabetes.Cell 90: 303–313.

Coon B, An LL, Whitton JL, von Herrath MG. 1999. DNAimmunization to prevent autoimmune diabetes. J ClinInvest 104: 189–194.

Coppieters KT, von Herrath MG. 2009. Histopathology oftype 1 diabetes: Old paradigms and new insights. RevDiabet Stud 6: 85–96.

Coppieters K, Martinic MM, Kiosses WB, Amirian N, vonHerrath M. 2010. A novel technique for the in vivo imag-ing of autoimmune diabetes development in the pancreasby two-photon microscopy. PLoS One 5: e15732.

Craighead JE, McLane MF. 1968. Diabetes mellitus: Induc-tion in mice by encephalomyocarditis virus. Science162: 913–914.

Dahlquist G, Frisk G, Ivarsson SA, Svanberg L, Forsgren M,Diderholm H. 1995. Indications that maternal coxsackieB virus infection during pregnancy is a risk factor forchildhood-onset IDDM. Diabetologia 38: 1371–1373.

DeStefano F, Mullooly JP, Okoro CA, Chen RT, Marcy SM,Ward JI, Vadheim CM, Black SB, Shinefield HR, DavisRL, et al. 2001. Childhood vaccinations, vaccination tim-ing, and risk of type 1 diabetes mellitus. Pediatrics 108:E112.

Dippe SE, Bennett PH, Miller M, Maynard JE, Berquist KR.1975. Lack of causal association between Coxsackie B4virus infection and diabetes. Lancet 1: 1314–1317.

Dotta F, Censini S, van Halteren AG, Marselli L, Masini M,Dionisi S, Mosca F, Boggi U, Muda AO, Prato SD, et al.2007. Coxsackie B4 virus infection of b cells and naturalkiller cell insulitis in recent-onset type 1 diabetic patients.Proc Natl Acad Sci 104: 5115–5120.

Drescher KM, Kono K, Bopegamage S, Carson SD, Tracy S.2004. Coxsackievirus B3 infection and type 1 diabetes de-velopment in NOD mice: Insulitis determines suscepti-bility of pancreatic islets to virus infection. Virology329: 381–394.



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

Ehl S, Hombach J, Aichele P, Hengartner H, ZinkernagelRM. 1997. Bystander activation of cytotoxic T cells: Stud-ies on the mechanism and evaluation of in vivo signifi-cance in a transgenic mouse model. J Exp Med 185:1241–1251.

Elfving M, Svensson J, Oikarinen S, Jonsson B, Olofsson P,Sundkvist G, Lindberg B, Lernmark A, Hyoty H, IvarssonSA. 2008. Maternal enterovirus infection during preg-nancy as a risk factor in offspring diagnosed with type1 diabetes between 15 and 30 years of age. Exp DiabetesRes 2008: 271958.

EURODIAB Substudy 2 Study Group. 2000. Infections andvaccinations as risk factors for childhood type I (insulin-dependent) diabetes mellitus: A multicentre case-controlinvestigation. Diabetologia 43: 47–53.

Fauquet CM, Mayo MA, Maniloff J, Desselberger U, Ball LA.2005. Virus taxonomy: Eighth report of the internationalcommittee on taxonomy of viruses. Academic Press, SanDiego.

Filippi CM, Estes EA, Oldham JE, von Herrath MG. 2009.Immunoregulatory mechanisms triggered by viral infec-tions protect from type 1 diabetes in mice. J Clin Invest119: 1515–1523.

Foulis AK, Farquharson MA, Meager A. 1987. Immunoreac-tivea-interferon in insulin-secretingb cells in type 1 dia-betes mellitus. Lancet 2: 1423–1427.

Foulis AK, Farquharson MA, Cameron SO, McGill M,Schonke H, Kandolf R. 1990. A search for the presenceof the enteroviral capsid protein VP1 in pancreases of pa-tients with type 1 (insulin-dependent) diabetes and pan-creases and hearts of infants who died of coxsackieviralmyocarditis. Diabetologia 33: 290–298.

Foulis AK, McGill M, Farquharson MA, Hilton DA. 1997. Asearch for evidence of viral infection in pancreases ofnewly diagnosed patients with IDDM. Diabetologia 40:53–61.

Fuchtenbusch M, Irnstetter A, Jager G, Ziegler AG. 2001. Noevidence for an association of coxsackie virus infectionsduring pregnancy and early childhood with developmentof islet autoantibodies in offspring of mothers or fatherswith type 1 diabetes. J Autoimmun 17: 333–340.

Gale EA. 2008. Congenital rubella: Citation virus or viralcause of type 1 diabetes? Diabetologia 51: 1559–1566.

Gamble DR, Kinsley ML, FitzGerald MG, Bolton R, TaylorKW. 1969. Viral antibodies in diabetes mellitus. Br MedJ 3: 627–630.

Gaskins HR, Prochazka M, Hamaguchi K, Serreze DV, LeiterEH. 1992. b cell expression of endogenous xenotropicretrovirus distinguishes diabetes-susceptible NOD/Ltfrom resistant NON/Lt mice. J Clin Invest 90: 2220–2227.

Gianani R, Campbell-Thompson M, Sarkar SA, WasserfallC, Pugliese A, Solis JM, Kent SC, Hering BJ, West E, SteckA, et al. 2010. Dimorphic histopathology of long-stand-ing childhood-onset diabetes. Diabetologia 53: 690–698.

Gillespie KM, Bain SC, Barnett AH, Bingley PJ, Christie MR,Gill GV, Gale EA. 2004. The rising incidence of childhoodtype 1 diabetes and reduced contribution of high-riskHLA haplotypes. Lancet 364: 1699–1700.

Goto A, Takahashi Y, Kishimoto M, Nakajima Y, NakanishiK, Kajio H, Noda M. 2008. A case of fulminant type 1 dia-

betes associated with significant elevation of mumpstiters. Endocr J 55: 561–564.

Graham KL, O’Donnell JA, Tan Y, Sanders N, CarringtonEM, Allison J, Coulson BS. 2007. Rotavirus infection ofinfant and young adult nonobese diabetic mice involvesextraintestinal spread and delays diabetes onset. J Virol81: 6446–6458.

Graham KL, Sanders N, Tan Y, Allison J, Kay TW, CoulsonBS. 2008. Rotavirus infection accelerates type 1 diabetesin mice with established insulitis. J Virol 82: 6139–6149.

Graves PM, Rotbart HA, Nix WA, Pallansch MA, Erlich HA,Norris JM, Hoffman M, Eisenbarth GS, Rewers M. 2003.Prospective study of enteroviral infections and develop-ment of b-cell autoimmunity. Diabetes autoimmunitystudy in the young (DAISY). Diabetes Res Clin Pract 59:51–61.

Guberski DL, Thomas VA, Shek WR, Like AA, Handler ES,Rossini AA, Wallace JE, Welsh RM. 1991. Induction oftype I diabetes by Kilham’s rat virus in diabetes-resistantBB/Wor rats. Science 254: 1010–1013.

Guilherme L, Cunha-Neto E, Coelho V, Snitcowsky R, Pom-erantzeff PM, Assis RV, Pedra F, Neumann J, GoldbergA, Patarroyo ME, et al. 1995. Human heart-infiltratingT-cell clones from rheumatic heart disease patients rec-ognize both streptococcal and cardiac proteins. Circula-tion 92: 415–420.

Harjutsalo V, Sjoberg L, Tuomilehto J. 2008. Time trends inthe incidence of type 1 diabetes in Finnish children: Acohort study. Lancet 371: 1777–1782.

Holz A, Dyrberg T, Hagopian W, Homann D, von HerrathM, Oldstone MB. 2000. Neither B lymphocytes norantibodies directed against self antigens of the islets ofLangerhans are required for development of virus-in-duced autoimmune diabetes. J Immunol 165: 5945–5953.

Honeyman M. 2005. How robust is the evidence for virusesin the induction of type 1 diabetes? Curr Opin Immunol17: 616–623.

Honeyman MC, Stone NL, Harrison LC. 1998. T-cell epito-pes in type 1 diabetes autoantigen tyrosine phosphataseIA-2: Potential for mimicry with rotavirus and other en-vironmental agents. Mol Med 4: 231–239.

Honeyman MC, Coulson BS, Stone NL, Gellert SA, Gold-water PN, Steele CE, Couper JJ, Tait BD, Colman PG,Harrison LC. 2000. Association between rotavirus infec-tion and pancreatic islet autoimmunity in children at riskof developing type 1 diabetes. Diabetes 49: 1319–1324.

Horwitz MS, Bradley LM, Harbertson J, Krahl T, Lee J, Sar-vetnick N. 1998. Diabetes induced by Coxsackie virus:Initiation by bystander damage and not molecular mimi-cry. Nat Med 4: 781–785.

Horwitz MS, Fine C, Ilic A, Sarvetnick N. 2001. Require-ments for viral-mediated autoimmune diabetes: b-celldamage and immune infiltration. J Autoimmun 16:211–217.

Hviid A, Stellfeld M, Wohlfahrt J, Melbye M. 2004. Child-hood vaccination and type 1 diabetes. N Engl J Med350: 1398–1404.

Hyoty H, Hiltunen M, Knip M, Laakkonen M, Vahasalo P,Karjalainen J, Koskela P, Roivainen M, Leinikki P, HoviT, et al. 1995. A prospective study of the role of coxsackieB and other enterovirus infections in the pathogenesis of



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

IDDM. Childhood Diabetes in Finland (DiMe) StudyGroup. Diabetes 44: 652–657.

Imagawa A, Hanafusa T, Miyagawa J, Matsuzawa Y. 2000. Anovel subtype of type 1 diabetes mellitus characterized bya rapid onset and an absence of diabetes-related antibod-ies. Osaka IDDM Study Group. N Engl J Med 342:301–307.

In’t Veld P, Lievens D, De Grijse J, Ling Z, Van der Auwera B,Pipeleers-Marichal M, Gorus F, Pipeleers D. 2007.Screening for insulitis in adult autoantibody-positive or-gan donors. Diabetes 56: 2400–2404.

Kasuga A, Harada R, Saruta T. 1996. Insulin-dependent dia-betes mellitus associated with parvovirus B19 infection.Ann Intern Med 125: 700–701.

Kaufman DL, Erlander MG, Clare-Salzler M, Atkinson MA,Maclaren NK, Tobin AJ. 1992. Autoimmunity to twoforms of glutamate decarboxylase in insulin-dependentdiabetes mellitus. J Clin Invest 89: 283–292.

Lennon GP, Bettini M, Burton AR, Vincent E, Arnold PY,Santamaria P, Vignali DA. 2009. T cell islet accumulationin type 1 diabetes is a tightly regulated, cell-autonomousevent. Immunity 31: 643–653.

Makela M, Oling V, Marttila J, Waris M, Knip M, Simell O,Ilonen J. 2006. Rotavirus-specific T cell responses andcytokine mRNA expression in children with diabetes-associated autoantibodies and type 1 diabetes. Clin ExpImmunol 145: 261–270.

McCartney SA, Vermi W, Lonardi S, Rossini C, Otero K, Cal-deron B, Gilfillan S, Diamond MS, Unanue ER, ColonnaM. 2011. RNA sensor-induced type I IFN prevents diabe-tes caused by a b cell-tropic virus in mice. J Clin Invest121: 1497–1507.

Mena I, Fischer C, Gebhard JR, Perry CM, Harkins S, Whit-ton JL. 2000. Coxsackievirus infection of the pancreas:Evaluation of receptor expression, pathogenesis, and im-munopathology. Virology 271: 276–288.

Menser MA, Forrest JM, Honeyman MC, Burgess JA. 1974.Letter: Diabetes, HL-A antigens, and congenital rubella.Lancet 2: 1508–1509.

Miller SD, Vanderlugt CL, Begolka WS, Pao W, Yauch RL,Neville KL, Katz-Levy Y, Carrizosa A, Kim BS. 1997. Per-sistent infection with Theiler’s virus leads to CNS auto-immunity via epitope spreading. Nat Med 3: 1133–1136.

Mordes JP, Bortell R, Blankenhorn EP, Rossini AA, GreinerDL. 2004. Rat models of type 1 diabetes: Genetics, envi-ronment, and autoimmunity. Ilar J 45: 278–291.

Nejentsev S, Walker N, Riches D, Egholm M, Todd JA. 2009.Rare variants of IFIH1, a gene implicated in antiviralresponses, protect against type 1 diabetes. Science 324:387–389.

Niklasson B, Hornfeldt B, Nyholm E, Niedrig M, Donoso-Mantke O, Gelderblom HR, Lernmark A. 2003. Type 1diabetes in Swedish bank voles (Clethrionomys glareolus):Signs of disease in both colonized and wild cyclic popu-lations at peak density. Ann NY Acad Sci 1005: 170–175.

Niklasson B, Hultman T, Kallies R, Niedrig M, Nilsson R,Berggren PO, Juntti-Berggren L, Efendic S, LernmarkA, Klitz W. 2007. The BioBreeding rat diabetes model isinfected with Ljungan virus. Diabetologia 50: 1559–1560.

Ohashi PS, Oehen S, Buerki K, Pircher H, Ohashi CT, Oder-matt B, Malissen B, Zinkernagel RM, Hengartner H.

1991. Ablation of “tolerance” and induction of diabetesby virus infection in viral antigen transgenic mice. Cell65: 305–317.

Oikarinen M, Tauriainen S, Honkanen T, Oikarinen S,Vuori K, Kaukinen K, Rantala I, Maki M, Hyoty H.2008a. Detection of enteroviruses in the intestine oftype 1 diabetic patients. Clin Exp Immunol 151: 71–75.

Oikarinen M, Tauriainen S, Honkanen T, Vuori K, Karhu-nen P, Vasama-Nolvi C, Oikarinen S, Verbeke C, BlairGE, Rantala I, et al. 2008b. Analysis of pancreas tissuein a child positive for islet cell antibodies. Diabetologia51: 1796–1802.

Oikarinen S, Martiskainen M, Tauriainen S, Huhtala H,Ilonen J, Veijola R, Simell O, Knip M, Hyoty H. 2011.Enterovirus RNA in blood is linked to the developmentof type 1 diabetes. Diabetes 60: 276–279.

Oldstone MB, Nerenberg M, Southern P, Price J, Lewicki H.1991. Virus infection triggers insulin-dependent diabetesmellitus in a transgenic model: Role of anti-self (virus)immune response. Cell 65: 319–331.

Onodera T, Jenson AB, Yoon JW, Notkins AL. 1978a.Virus-induced diabetes mellitus: Reovirus infection ofpancreatic b cells in mice. Science 201: 529–531.

Onodera T, Yoon JW, Brown KS, Notkina AL. 1978b. Evi-dence for a single locus controlling susceptibility tovirus-induced diabetes mellitus. Nature 274: 693–696.

Pak CY, Eun HM, McArthur RG, Yoon JW. 1988. Associationof cytomegalovirus infection with autoimmune type 1diabetes. Lancet 2: 1–4.

Patterson K, Chandra RS, Jenson AB. 1981. Congenital ru-bella, insulitis, and diabetes mellitus in an infant. Lancet1: 1048–1049.

Redondo MJ, Jeffrey J, Fain PR, Eisenbarth GS, Orban T.2008. Concordance for islet autoimmunityamong mono-zygotic twins. N Engl J Med 359: 2849–2850.

Richardson SJ, Willcox A, Bone AJ, Foulis AK, Morgan NG.2009. The prevalence of enteroviral capsid protein vp1immunostaining in pancreatic islets in human type 1 dia-betes. Diabetologia 52: 1143–1151.

Richter W, Mertens T, Schoel B, Muir P, Ritzkowsky A,Scherbaum WA, Boehm BO. 1994. Sequence homologyof the diabetes-associated autoantigen glutamate decar-boxylase with coxsackie B4-2C protein and heat shockprotein 60 mediates no molecular mimicry of autoanti-bodies. J Exp Med 180: 721–726.

Roep BO, Kleijwegt FS, van Halteren AG, Bonato V, Boggi U,Vendrame F, Marchetti P, Dotta F. 2010. Islet inflamma-tion and CXCL10 in recent-onset type 1 diabetes. ClinExp Immunol 159: 338–343.

Roivainen M. 2006. Enteroviruses: New findings on the roleof enteroviruses in type 1 diabetes. Int J Biochem Cell Biol38: 721–725.

Sadeharju K, Lonnrot M, Kimpimaki T, Savola K, Erkkila S,Kalliokoski T, Savolainen P, Koskela P, Ilonen J, Simell O,et al. 2001. Enterovirus antibody levels during the firsttwo years of life in prediabetic autoantibody-positivechildren. Diabetologia 44: 818–823.

Savinov AY, Wong FS, Stonebraker AC, Chervonsky AV. 2003.Presentation of antigen by endothelial cells and chemoat-traction are required for homing of insulin-specificCD8þ T cells. J Exp Med 197: 643–656.



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

Schloot NC, Willemen SJ, Duinkerken G, Drijfhout JW, deVries RR, Roep BO. 2001. Molecular mimicry in type 1diabetes mellitus revisited: T-cell clones to GAD65 pepti-des with sequence homology to Coxsackie or proinsulinpeptides do not crossreact with homologous counter-part. Hum Immunol 62: 299–309.

Seewaldt S, Thomas HE, Ejrnaes M, Christen U, Wolfe T,Rodrigo E, Coon B, Michelsen B, Kay TW, von HerrathMG. 2000. Virus-induced autoimmune diabetes: Mostb-cells die through inflammatory cytokines and not per-forin from autoreactive (anti-viral) cytotoxic T-lympho-cytes. Diabetes 49: 1801–1809.

Serreze DV, Ottendorfer EW, Ellis TM, Gauntt CJ, AtkinsonMA. 2000. Acceleration of type 1 diabetes by a coxsackie-virus infection requires a preexisting critical mass ofautoreactive T-cells in pancreatic islets. Diabetes 49:708–711.

Sevilla N, Homann D, von Herrath M, Rodriguez F, HarkinsS, Whitton JL, Oldstone MB. 2000. Virus-induced diabe-tes in a transgenic model: Role of cross-reacting virusesand quantitation of effector T cells needed to cause dis-ease. J Virol 74: 3284–3292.

Shimada A, Maruyama T. 2004. Encephalomyocarditis-vi-rus-induced diabetes model resembles “fulminant” type1 diabetes in humans. Diabetologia 47: 1854–1855.

Smith KA, Efstathiou S, Cooke A. 2007. Murine gammaher-pesvirus-68 infection alters self-antigen presentationand type 1 diabetes onset in NOD mice. J Immunol179: 7325–7333.

Stene LC, Oikarinen S, Hyoty H, Barriga KJ, Norris JM,Klingensmith G, Hutton JC, Erlich HA, Eisenbarth GS,Rewers M. 2010. Enterovirus infection and progressionfrom islet autoimmunity to type 1 diabetes: The diabetesand autoimmunity study in the young (DAISY). Diabetes59: 3174–3180.

Suenaga K, Yoon JW. 1988. Association of b-cell-specific ex-pression of endogenous retrovirus with development ofinsulitis and diabetes in NOD mouse. Diabetes 37:1722–1726.

Takei I, Asaba Y, Kasatani T, Maruyama T, Watanabe K, Ya-nagawa T, Saruta T, Ishii T. 1992. Suppression of develop-ment of diabetes in NOD mice by lactate dehydrogenasevirus infection. J Autoimmun 5: 665–673.

Tanaka S, Nishida Y, Aida K, Maruyama T, Shimada A, Su-zuki M, Shimura H, Takizawa S, Takahashi M, AkiyamaD, et al. 2009. Enterovirus infection, CXC chemokine li-gand 10 (CXCL10), and CXCR3 circuit: A mechanism ofaccelerated b-cell failure in fulminant type 1 diabetes.Diabetes 58: 2285–2291.

Tirabassi RS, Guberski DL, Blankenhorn EP, Leif JH, WodaBA, Liu Z, Winans D, Greiner DL, Mordes JP. 2010. Infec-tion with viruses from several families triggers autoim-mune diabetes in LEW�1WR1 rats: Prevention of diabe-tes by maternal immunization. Diabetes 59: 110–118.

Tracy S, Drescher KM, Chapman NM, Kim KS, Carson SD,Pirruccello S, Lane PH, Romero JR, Leser JS. 2002. To-ward testing the hypothesis that group B coxsackieviruses(CVB) trigger insulin-dependent diabetes: Inoculating

nonobese diabetic mice with CVB markedly lowers dia-betes incidence. J Virol 76: 12097–12111.

Viskari H, Ludvigsson J, Uibo R, Salur L, Marciulionyte D,Hermann R, Soltesz G, Fuchtenbusch M, Ziegler AG,Kondrashova A, et al. 2005. Relationship between theincidence of type 1 diabetes and maternal enterovirusantibodies: Time trends and geographical variation. Dia-betologia 48: 1280–1287.

von Herrath MG, Dockter J, Oldstone MB. 1994. How virusinduces a rapid or slow onset insulin-dependent diabetesmellitus in a transgenic model. Immunity 1: 231–242.

von Herrath MG, Fujinami RS, Whitton JL. 2003. Microor-ganisms and autoimmunity: Making the barren field fer-tile? Nat Rev Microbiol 1: 151–157.

Wang D, Coscoy L, Zylberberg M, Avila PC, Boushey HA,Ganem D, DeRisi JL. 2002. Microarray-based detectionand genotyping of viral pathogens. Proc Natl Acad Sci99: 15687–15692.

Wang D, Urisman A, Liu YT, Springer M, Ksiazek TG, Erd-man DD, Mardis ER, Hickenbotham M, Magrini V, El-dred J, et al. 2003. Viral discovery and sequence recoveryusing DNA microarrays. PLoS Biol 1: E2.

Wetzel JD, Barton ES, Chappell JD, Baer GS, Mochow-Grundy M, Rodgers SE, Shyr Y, Powers AC, ThomasJW, Dermody TS. 2006. Reovirus delays diabetes onsetbut does not prevent insulitis in nonobese diabeticmice. J Virol 80: 3078–3082.

Wilberz S, Partke HJ, Dagnaes-Hansen F, Herberg L. 1991.Persistent MHV (mouse hepatitis virus) infection re-duces the incidence of diabetes mellitus in non-obese di-abetic mice. Diabetologia 34: 2–5.

Yeung WC, Rawlinson WD, Craig ME. 2011. Enterovirus in-fection and type 1 diabetes mellitus: Systematic reviewand meta-analysis of observational molecular studies.BMJ 342: d35.

Ylipaasto P, Klingel K, Lindberg AM, Otonkoski T, KandolfR, Hovi T, Roivainen M. 2004. Enterovirus infection inhuman pancreatic islet cells, islet tropism in vivo and re-ceptor involvement in cultured islet b cells. Diabetologia47: 225–239.

Yoon JW, Jun HS. 2006. Viruses cause type 1 diabetes in an-imals. Ann NY Acad Sci 1079: 138–146.

Yoon JW, Onodera T, Notkins AL. 1978. Virus-induced dia-betes mellitus. XV. b cell damage and insulin-dependenthyperglycemia in mice infected with coxsackie virus B4. JExp Med 148: 1068–1080.

Yoon JW, Austin M, Onodera T, Notkins AL. 1979. Isolationof a virus from the pancreas of a child with diabetic keto-acidosis. New Engl J Med 300: 1173–1179.

Yoon JW, Morishima T, McClintock PR, Austin M, NotkinsAL. 1984. Virus-induced diabetes mellitus: Mengovirusinfects pancreatic b cells in strains of mice resistant tothe diabetogenic effect of encephalomyocarditis virus. JVirol 50: 684–690.

Zarozinski CC, Welsh RM. 1997. Minimal bystander activa-tion of CD8 T cells during the virus-induced polyclonal Tcell response. J Exp Med 185: 1629–1639.



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

RESEARCH Open Access

The involvement of interleukin-22 in theexpression of pancreatic beta cell regenerativeReg genesThomas Hill1, Olga Krougly1, Enayat Nikoopour1, Stacey Bellemore1, Edwin Lee-Chan1, Lynette A Fouser2,David J Hill3,4 and Bhagirath Singh1,5,6*

Abstract

Background: In Type 1 diabetes, the insulin-producing β-cells within the pancreatic islets of Langerhans aredestroyed. We showed previously that immunotherapy with Bacillus Calmette-Guerin (BCG) or complete Freund’sadjuvant (CFA) of non-obese diabetic (NOD) mice can prevent disease process and pancreatic β-cell loss. This wasassociated with increased islet Regenerating (Reg) genes expression, and elevated IL-22-producing Th17 T-cells inthe pancreas.

Results: We hypothesized that IL-22 was responsible for the increased Reg gene expression in the pancreas. Wetherefore quantified the Reg1, Reg2, and Reg3δ (INGAP) mRNA expression in isolated pre-diabetic NOD islets treated withIL-22. We measured IL-22, and IL-22 receptor(R)-α mRNA expression in the pancreas and spleen of pre-diabetic anddiabetic NOD mice. Our results showed: 1) Reg1 and Reg2 mRNA abundance to be significantly increased in IL-22-treatedislets in vitro; 2) IL-22 mRNA expression in the pre-diabetic mouse pancreas increased with time following CFA treatment;3) a reduced expression of IL-22Rα following CFA treatment; 4) a down-regulation in Reg1 and Reg2 mRNA expression inthe pancreas of pre-diabetic mice injected with an IL-22 neutralizing antibody; and 5) an increased islet β-cell DNAsynthesis in vitro in the presence of IL-22.

Conclusions: We conclude that IL-22 may contribute to the regeneration of β-cells by up-regulating Regenerating Reg1and Reg2 genes in the islets.

Keywords: Adjuvant immunotherapy, Interleukin-22, Regenerating (Reg) genes, Beta-cell regeneration, Type 1 diabetes

BackgroundType 1 diabetes (T1D), also known as juvenile diabetes, isan autoimmune disease characterized by the destructionof insulin-producing β-cells within the pancreatic islets ofLangerhans. Type 1 diabetes is thought to be caused by acomplex interaction between environmental and geneticfactors that is still not fully understood [1]. This inter-action causes initial damage to the pancreatic islets leadingto a T-cell mediated autoimmune response that inducesdysregulation and destruction of β-cells via the release ofinflammatory cytokines [2]. Current T1D management

includes insulin therapy and pancreas or islet cell trans-plantation; however, none of these procedures will ensurethe complete removal of diabetic complications. There-fore, studying the endogenous regeneration of pancreaticβ-cells may suggest strategies for alternative and long-lasting approaches for T1D management [3].Endogenous β-cell regeneration is thought to occur ei-

ther by whole islet neogenesis (WIN) via the differentiationof progenitor cells within the adult pancreas, or by β-cellreplication (BCR) and the regeneration of new β-cells frompre-existing β-cells [3,4]. Using non-obese diabetic (NOD)or streptozotocin-injected C57/BL6 mice as a model forT1D, several groups have shown a possible role for tran-scription factors, growth factors, and regeneration genes instimulating β-cell regeneration. This could result in wholeislet neogenesis (WIN) and subsequent insulin production

* Correspondence: [email protected] of Microbiology and Immunology, University of WesternOntario, London, ON, Canada5Centre for Human Immunology, University of Western Ontario, London, ON,CanadaFull list of author information is available at the end of the article

© 2013 Hill et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly cited.

Hill et al. Cell Regeneration 2013, 2:2http://www.cellregenerationjournal.com/content/2/1/2

mailto:[email protected]

http://creativecommons.org/licenses/by/2.0

and diabetes reversal. Some of these candidate factorsinclude pancreatic and duodenal homeobox 1 (PDX-1),glucagon-like peptide-1 (GLP-1), islet neogenesis-associated protein (INGAP) and Regenerating protein 1and 2 (Reg1 and Reg2) [3,5-9]. Recently, platelet-derivedgrowth factor (PDGF) has also been shown to stimulate β-cell regeneration via activating cyclin D1 and inducing aG1 to S transition of the β-cell cycle [10].Previously, we have shown that a single injection of

Mycobacterium-containing Complete Freund’s Adjuvant(CFA) into NOD mice has a protective effect against T1Dby down- regulating autoimmunity and restoring normo-glycemia via the induction of various regulatory T (Treg)cells [7,11,12]. The role of CFA on endogenous β-cellreplenishment, identified by histological analysis, however,still remains controversial and the exact mechanism ispresently unknown [5,7].Several groups have identified the Regeneration gene

family (Reg) as being expressed during the process ofWIN and β-cell regeneration in the pancreas [3,5,7,9,12].There are seven types of Reg genes present in the mouse(located on chromosome six with the exception of Reg4)but only five have been shown to be present in the hu-man (located on chromosome 2) [13]. The Reg proteinsencoded by these genes are C-type lectins, and havebeen found to be also expressed in a variety of tissuessuch as the liver, kidney, brain, and gastrointestinal tis-sues [3,13]. Once secreted, these soluble proteins act inan autocrine and/or paracrine manner to exert their ef-fects on their cognate receptors, where they may stimu-late an anti-microbial, anti-inflammatory, anti-apoptotic,or regenerative response depending on the tissue type[3,7,9,13]. In the last decade, Reg1 and Reg3δ (INGAP),expressed by β-cells and acinar cells respectively, havebeen shown to be linked with pancreatic β-cell regener-ation in the mouse by activating cyclin D1 and promot-ing β-cell cycle progression [9,13,14]. More recently, wehave shown that Reg2 is substantially up-regulated in thepancreatic islets, particularly in the β-cells, following ad-juvant treatment in diabetic and non-diabetic NOD fe-male mice, as well as in C57BL/6 mice treated withstreptozotocin (STZ). This increased Reg2 expressionhas been shown to be associated with an increase in in-sulin production, a partial reversal of insulitis, and animproved glucose tolerance test in STZ-treated diabeticC57BL/6 mice [5]. This led to the conclusion that β-cellregeneration via up-regulation of the Reg2 gene mayhave the capacity to reverse T1D in diabetic mice immu-nized with CFA using a pathway that is similar to Reg1and Reg3δ [5,7].CFA immunization has been shown to induce CD4+

Th17 T cells to produce interleukin IL-17, IL-22, IL-10and IFN-γ in NOD mice [15]. This finding leads to theconcept that adjuvant-induced cytokines may have the

potential to activate transcription factors that stimulateReg proteins such as Reg1, Reg2, and Reg3β (PAP1) [7].Among these cytokines, IL-22 has been of specific inter-est because it is released by CD4+ Th17 T cells in dis-eases such as hepatitis and inflammatory bowel disease,where IL-22 expression levels have been shown to pro-mote cell regeneration and survival in hepatocytes andcolonal epithelial cells. Interestingly, IL-22 has also beenfound recently to induce Reg gene expression in the pan-creatic acinar cells surrounding pancreatic islets [16]. IL-22 is a member of the IL-10 family of cytokines and thegene is located on human chromosome 12. Like IL-17Aand IL-17F, IL-22 is a glycoprotein, which binds to itsreceptor complex (composed of IL-10Rβ and IL-22Rα) as a homodimer [17,18]. The IL-22 receptor com-plex is highly expressed in the pancreatic α and β cells[19]. Upon receptor binding, the tyrosine kinases, JAKand Tyk1 are phosphorylated which leads to the activa-tion of the STAT3 transcription factor and subsequentup-regulation of the Reg genes in the mouse and human[17,18]. However, the possible induction of the Reg genesvia IL-22 in the pancreatic islets or NOD mouse modelhas not yet been examined.This study sought to test the effects of IL-22 on the regu-

lation of the pancreatic Reg genes, with a focus on Reg2, indiabetic and pre-diabetic NOD female mice immunizedwith CFA. This was accomplished in a two-step process:Firstly, by studying the direct effect of IL-22 on Reg geneup-regulation in the NOD mouse pancreatic islets; andsecondly by examining whether CFA immunization couldup-regulate IL-22 expression in the whole pancreas. It washypothesized that CFA immunization in pre-diabetic micewould lead to the production of IL-22 via the induction ofTh17 cells, and that the resulting IL-22 cytokine would ac-tivate a JAK-STAT3 signal transduction pathway followingbinding to its receptor complex on pancreatic β-cells. IL-22 signaling would then result in the up-regulation of Reggene expression that may be linked to β-cell regenerationand the reversal of hyperglycemia in T1D. This, however,requires inhibition of autoimmunity to prevent the reversalof disease. Immunotherapies with mycobacterial adjuvantssuch as BCG vaccination have the potential to achieve boththese objectives [20].

ResultsRecombinant IL-22 up-regulates Reg2 and Reg1 mRNAexpression in the pancreatic isletsTo establish whether IL-22 induces Reg gene expression, 6-week-old pre-diabetic NOD mouse islets were isolated andincubated with recombinant IL-22 at 10 and 50 ng/mLconcentrations or supernatants from Th17 polarizedsplenocytes containing IL-17 and IL-22 [15] (Figure 1).Quantitative RT-PCR was performed on total extracted

Hill et al. Cell Regeneration 2013, 2:2 Page 2 of 11http://www.cellregenerationjournal.com/content/2/1/2

RNA from the treated pancreatic islets. Reg2 and Reg1mRNA levels were found to be significantly higher, by ap-proximately 3 and 4.2-fold respectively, in islets incubatedwith 10 ng/mL of IL-22 when compared to the media con-trol (P<0.05) (Figure 1A and 1B). Interestingly, the higher50 ng/mL dose of IL-22 significantly down-regulated Reg2and Reg1 expression (P<0.05). This was not due to the tox-icity of the cytokine as there was no change in the viabilityof the treated islets cells as compared to media control.The highest fold increase in Reg2 gene expression wasobserved in islets treated with the supernatants of 4 daysculture of BDC2.5 NOD splenocytes conditioned withIL-23 and IL-6 (Figure 1A). The supernatants alsocontained IL-22 and produced a significant 27-fold increasein Reg2 mRNA abundance when compared to the mediacontrol (P<0.05). Islets treated with the Th17 T cell super-natant also showed a significant up-regulation of Reg1 geneexpression by approximately 4.3-fold, when compared tothe control islets (P<0.05).However, unlike Reg2, this Reg1expression was not significantly different when comparedto islets treated with 10 ng/mL of recombinant IL-22. Nosignificant change in Reg3δ expression was observed in re-sponse to IL-22 or the Th17 T cell supernatant (Figure 1C).The effect of recombinant IL-22 and the Th17 cell super-

natant on IL-22 receptor expression was also tested on thesame islet samples using gene-specific primers for theIL-22 receptor chain, IL-22Rα. As shown in Figure 1Dthe mean mRNA levels for IL-22Rα did not increase in is-lets treated with 10 ng/ml of recombinant IL-22 and in factwas further reduced with 50 ng/ml of IL-22. Islets treatedwith the Th17 cell supernatant induced a significant 2.4-fold increase in IL-22Rα expression when compared tocontrol-treated islets (control [1.36±0.21], Th17 T cells[3.72±0.98], P<0.05).

CFA treatment increases the relative mRNA expression forIL-22 in pancreatic tissueWe have previously shown that CFA injection in non-diabetic and diabetic NOD mice results in a substantial in-crease in Reg gene expression [5]. To determine whetherthe development of diabetes was accompanied by similar

Figure 1 Reg genes and IL-22 receptor (IL-22Rα) expression inthe pancreatic islet cells after IL-22 treatment. QuantitativeRT-PCR analysis was performed using gene-specific primers (Table 1) forReg1, Reg2, Reg3δ and IL-22Rα on total RNA isolated from 6-week-oldpancreatic islets (A-D). Islets were harvested 48 hrs after incubation witheither DMEM medium (control); 10, or 50 ng/mL of recombinant IL-22; ora supernatant of Th17 cells polarized with IL-6 plus IL-23 (2 ml). Resultsshown represent the average fold-change in mRNA expression ± SEMwhen compared to control-treated islets. N = 3 mice (12 – 14 islets permouse). Means indicated by the asterisk (*) are significantlydifferent (P<0.05).


changes in IL-22 gene expression in vivo, qRT-PCR analysiswas performed using IL-22 gene-specific primers for cDNAusing the various NOD mouse treatment/age groups. Asshown in Figure 2A a significant effect of age on IL-22 ex-pression was detected (F= 6.374; df= 3; P=0.021). Post-hocanalysis revealed IL-22 mRNA abundance in the pancreasto be significantly higher in diabetic mice when comparedto non-diabetic (4-week-old) animals (P<0.05). The meanIL-22 expression tended to be higher in the pre-diabetic (8-week-old) mice and in diabetic mice following CFA treat-ment (by approximately 2 and 4-fold respectively).

CFA treatment down-regulates the pancreatic expressionof the IL-22 receptor sub-unit, IL-22RαIt has previously been reported by Shioya et al. [19] thatthe IL-22Rα chain, a member of the IL-22 receptor

complex, is specifically expressed in the α- and β-cellswithin the human pancreas. To determine whether CFAtreatment affected the relative mRNA expression levels ofIL-22Rα in the NOD mouse pancreas, qRT-PCR analysiswas performed using total RNA extracted from non-diabetic (4-week-old), pre-diabetic (8-week-old and 12-week-old) and diabetic mice (Figure 2B). A significantinteraction was identified between mouse age and treat-ment with CFA (F=6.146; df=3; P=0001). When analyzingthe variables separately, CFA treatment was found to sig-nificantly down-regulate IL-22Rα mRNA expression whencompared to the saline-treated controls for all mouse ages(F=82.411; df=3; P<0.001) (Figure 2B). IL-22Rα mRNA ex-pression was also found to significantly alter with age(F=24.212; df=3; P<0.001). Post-hoc analysis revealed IL-22Rα mRNA abundance to be significantly higher in dia-betic mice (P<0.05), when compared to non-diabetic andpre-diabetic mice. Pre-diabetic mice at 12-weeks of agewere found to have an IL-22Rα mRNA abundance thatwas significantly lower (P<0.05), when compared to non-diabetic (4-week-old), pre-diabetic (8-week-old), and dia-betic animals.

Reg2 and Reg1 expression is down-regulated in CFA-treated mice following treatment with an IL-22neutralizing antibodyTo determine whether IL-22 was directly involved in theup-regulation of Reg expression following CFA treatment,the abundance of Reg2, Reg1and Reg3δ mRNAs were mea-sured in the pancreas of 6-week-old NOD mice immu-nized with CFA followed by i.p injection of an IL-22neutralizing antibody 1 h later. Mice immunized with CFAand treated with the IL-22 neutralizing antibody [21] werefound to have a significant reduction (approximately 480-fold) in Reg2 gene expression when compared to mice im-munized with CFA and treated with an isotype controlantibody (P<0.001) (Figure 3A). Likewise, Reg1 gene ex-pression was significantly down-regulated in mice injectedwith the IL-22 neutralizing antibody (P<0.001) (Figure 3B).In contrast, however, no change in Reg3δ expression wasobserved for mice immunized with CFA in the presence ofIL-22 neutralizing antibody when compared to miceinjected with the isotype control antibody (Figure 3C).

IL-22 and Reg mRNA expression in the spleenSince the spleen is a major peripheral lymphoid tissue, weperformed quantitative RT-PCR analysis on splenocytesderived from 4-week-old, pre-diabetic NOD mice, with orwithout CFA treatment, to investigate the presence ofIL-22-producing Th17 cells during autoimmune insulitis(Figure 4). CFA treatment was found to significantly up-regulate IL-22 expression in the non-diabetic, 4 week-oldmice by approximately 12-fold (P<0.05). To determinewhether CFA-mediated Reg gene up-regulation occurred

Figure 2 Expression of IL-22 and IL-22 receptor (IL-22Rα) in thepancreatic islet cells after CFA immunization. The relative mRNAexpression levels of IL-22 and IL-22Rα (A and B) was carried out inthe pancreas of non-diabetic (4-week old), pre-diabetic (8-week oldand 12-week old), and older (>16 week-old) diabetic NOD micefollowing CFA immunization. Female NOD mice were injected i.p.with either 100 μl of CFA emulsified in saline or with saline aloneand sacrificed 48 hrs following immunization. Whole pancreatictissue was homogenized and total RNA extracted for reversetranscription quantitative real-time PCR analysis using gene-specificprimers. The results shown for each treatment and age group havebeen compared to 4-wk-old saline-treated mice and represent theaverage fold change of mRNA expression ± SEM. The relativeexpression of mRNA was taken from three pooled samples of RNAper group (with three mice per pooled sample). Age groupsindicated by a double asterisk (**) are significantly different (P<0.05);treatments indicated by the asterisk (*) are significantly differentfrom the saline control (P<0.05).


in the spleen as well as in pancreatic tissue, Reg1, Reg2and Reg3δ mRNA expression was also analyzed in wholesplenic tissue using non-diabetic mice with/without CFAinjection. Similar to Huszarik et al. [5], no effect of treat-ment was detected on Reg mRNA expression (datanot shown).

CFA treatment does not change Reg or IL-22 mRNAabundance in the immune-deficient NOD.Scid mousepancreasTo verify the finding by Huszarik et al. [5] that Reg up-regulation following CFA treatment is mediated by an in-filtrating T cell response, qRT-PCR analysis was performedon the pancreas of CFA treated non-diabetic (4-week-old)NOD.Scid mice. NOD-severe combined immune-deficient(Scid) mice have a deleterious single nucleotide poly-morphism (SNP) in the Prkdc gene that ultimately affectsT- and B- lymphocyte development. As reported byHuszarik et al. [5], we found no increase in Reg1, Reg2, orReg3δ mRNA abundance in the pancreas of NOD.Scidmice following CFA treatment (data not shown). This pro-cedure was repeated using IL-22 gene-specific primers toconfirm that IL-22 was a product of Th17 T lymphocytes,and was not expressed by the pancreatic cells. Asexpected, IL-22 mRNA expression did not change inCFA-treated NOD.Scid mice (data not shown).

IL-22 treatment induces islet β-cell DNA synthesis in vitroTo determine whether IL-22 can promote DNA synthesisin islet β-cells, we performed immunohistochemical ana-lysis on isolated islets from 5 to 6-week-old NOD micetreated with recombinant IL-22 (10 ng/ml) for 48 hoursand stained for nuclear EdU and cytoplasmic insulin(Figure 5A and B). The percentage of EdU positive β-cellswas found to be significantly higher in islets treated withrecombinant IL-22 when compared to control islets treatedwith media alone (P<0.001) (Figure 5C). To confirm results

Figure 3 Expression of Reg genes in the pancreatic islet cellsafter IL-22 neutralization. Reg1, Reg2, and Reg3δ gene expressionlevels (A-C) was determined in the pancreas of 6-week-old NODmice immunized with CFA and treated with an IL-22 neutralizing ora control antibody. Female NOD mice were injected i.p. with either100 μl of CFA emulsified in saline followed by 100 μl of IL-22neutralizing antibody (IL-22-01) 1 hr later, or 100 μl of CFAemulsified in saline followed by 100 μl of isotype control antibody.Mice were sacrificed 48 hrs following immunization. Wholepancreatic tissue was homogenized and total RNA extracted forreverse transcription quantitative real-time PCR analysis using gene-specific primers. The results shown for each treatment have beencompared to mice immunized with CFA and the IgG isotypeantibody, and represent the average fold change of mRNAexpression ± SEM. The relative expression of mRNA was taken fromthree pooled samples per group (with three mice per pooledsample). Treatments indicated by the asterisk (*) are significantlydifferent from the CFA control (P<0.001).

Figure 4 Expression of IL-22 in the spleen of 4-week-oldnon-diabetic NOD mice following CFA-treatment. The relativemRNA expression of IL-22 in the spleen of 4-week-old NOD micewas carried out following CFA-treatment. Quantitative RT-PCRanalysis was performed using gene-specific primers (Table 1) ontotal RNA isolated from whole splenic tissue from female NOD miceinjected i.p. with either 100 μl of CFA emulsified in saline or salinealone. Results shown represent the average fold-change in mRNAexpression ± SEM compared with saline-treated mice. Results weretaken from 3 mice per treatment. The asterisk (*) denotes asignificant difference from the saline control (P<0.05).


for DNA synthesis, slides containing IL-22 treated isletswere stained for nuclear Ki/67 protein, which is associ-ated with cell proliferation. The islets were stained withmouse anti-Ki/67 followed by incubation with second-ary antibody. The positive expression of Ki/67 in the is-lets validated the induction of islet β-cells by IL-22(data not shown). The ability of IL-22 to increase thepercentage of β-cells undergoing DNA synthesis, asdetected by nuclear labeling with EdU, was inversely re-lated to islet size (y = −29logx + 56, r2 = 0.31, p<0.001,n=46). Islet size was estimated from the total number ofinsulin-immuno-reactive β-cells present per islet withintissue sections. No such relationship existed betweenislet size and the percent nuclear labeling of β-cells withEdU for control incubations (y = 2.2logx + 4.3, r2 =0.01, non-significant, n=26). These findings suggest thatthe mitogenic actions of IL-22 on β-cells were greatestfor the smaller islets.

DiscussionWe have previously shown that immunizing young NODmice with CFA can cause regeneration of pancreatic β-cells by down-regulating autoimmunity and reducing insu-litis [5,11,12]. However, the cellular mechanisms by whichCFA is able to initiate a regenerative response are not wellunderstood [7,11]. Recently, CFA treatment in NOD micehas been shown to substantially up-regulate the regener-ation gene, Reg2, in pancreatic islets during insulitis, thusfunctioning as an important regenerative candidate for β-cells [5]. The objective of this study was to determinewhether the CFA-induced cytokine, IL-22, was in part re-sponsible for inducing Reg2 expression and other gene

family members, Reg1 and Reg3δ, leading to a subsequentincrease in β-cell mass and reversal of T1D development.The exact mechanism underlying CFA-induced Reg gene

upregulation has not been clearly defined. Experiments inthe past have shown IL-6 to be an intermediate for Reg2and Reg1 gene induction in the pancreas, since the IL-6upstream response element is conserved among these Reggenes [5,22]. Since IL-6 signaling leads to the activation ofStat3 transcription factors inside target cells [23], it wasbelieved that Reg2 gene induction was also Stat-3 medi-ated [7]. Our results confirm and extend the role of thecytokine, IL-22, in up-regulating the Reg2 gene as well asReg1, thus supporting our original hypothesis. We con-firmed a 3-fold and 4.2-fold increase in Reg2 and Reg1gene expression respectively when pre-diabetic (6-week-old) pancreatic islets were incubated with 10 ng/mL of re-combinant IL-22 in vitro for 48 hrs. We have previouslyshown that the PCR results for the expression of Reggenes in the pancreatic islets in our studies correlate withthe expression of Reg proteins by using Western blot assay[5]. This suggests that Reg2 gene expression, like Reg1, canbe induced via Stat3 signaling and that IL-22 may be animmune response-mediated agent for the induction ofthese Reg genes within the pancreatic islets during insulitis[7,16] (Figure 6). This finding is supported by a previousstudy by Aggarwal et al. [16], who had shown that thein vivo injection of IL-22 resulted in a rapid induction ofReg2 expression in the pancreas of C57/BL6 mice. Interest-ingly, in our in vitro study recombinant IL-22 resulted in anoticeable down-regulation of Reg2 and Reg1 expression inthe islets from 6-week-old NOD mice when incubated withhigh concentrations (50 ng/mL) of IL-22. This may suggestthat pancreatic β-cells exhibit a dose-dependent response

Figure 5 Detection of pancreatic islet regeneration after IL-22 treatment. Immunofluorescence localization of nuclear EdU and cytoplasmicinsulin within pancreatic islets was done in 5 to 6-week-old NOD mice. Islets were treated with either DMEM media alone (A) or 10 ng/mL ofrecombinant IL-22 for 48 hrs (B) (red = EdU; green = insulin; blue = DAPI). Examples of dual-labeled cells are shown with arrows. The size bar = 50 μm.(C) The mean percentage ± SEM of β-cells immunopositive for both EdU and insulin relative to insulin alone. Results were taken from 26 control isletsand 46 IL-22-treated islets derived from 4 animals for each group. The asterisk (*) denotes a significant difference from the control (P<0.001). Asdescribed in the Methods section, in parallel experiments we explored the nuclear expression of Ki/67 as a measure of islet cell proliferation (data notshown). This further confirmed the proliferation of islet cells by IL-22 treatment.


to IL-22, and that an excess of the cytokine may be inhibi-tory to islet Reg gene expression. In contrast to the otherReg genes, the expression of Reg3δ was not affected byIL-22 treatment in the pancreatic islets and thus confirmspast studies showing Reg3δ expression to be absent in theα- and β-cells. We also found Reg2 and Reg1 expression tobe drastically reduced in the pancreas of pre-diabetic miceinjected with a neutralizing IL-22 antibody, confirming thatIL-22 is an upstream activator for these Reg genes. Thesefindings support the study by Zhang et al. [24] who foundthe Reg1 and Reg2 genes to be induced by IL-22 in the co-lonic epithelial cells of Citrobacter rodentium-infected WTmice, but for expression to be completely abolished inIL-22 knock-out infected mice. As shown by our studieswith Th17 supernatant, cytokines other than IL-22 maycontribute to the upregulation of Reg genes. We are cur-rently exploring this possibility in our laboratory.Using the whole pancreas from NOD mice, Huszarik

et al. [5] found several members of the Reg gene familyto be up-regulated with age, and Reg2 to be furtherupregulated in diabetic mice following CFA treatment.Our results support the finding that Reg2 plays a dominantrole in endogenous β-cell regeneration following adjuvantimmunotherapy. Reg2 mRNA levels were found to grad-ually increase in pre-diabetic mice with age and were in-creased in diabetic mice following CFA immunization [5].Similar to Reg2, IL-22 gene expression gradually increased

with pre-diabetic age and was also increased followingCFA treatment in both pre-diabetes and following the on-set of diabetes. The similar pattern of expression betweenReg2 and IL-22 after CFA treatment further suggests Reg2to be a gene target in the IL-22 signaling pathway. How-ever, unlike IL-22, the expression of its receptor chain, IL-22Rα, unexpectedly dropped at all ages following CFAtreatment. In a study analyzing mantle cell lymphomagrowth caused by IL-22, Gelebart et al. [25] discoveredthat the gene promoter for IL-22Rα contained multiplebinding sites for NF-κB, a transcription factor that whenactivated contributes to the pathogenesis and progressiveloss of pancreatic β-cells by generating reactive oxygenspecies and promoting β-cell apoptosis. It is therefore pos-sible that the effects of CFA may lead to the disruption ordegradation of NF-κB that would in turn reduce/preventIL-22Rα. Whether CFA affects the integrity of NF-κB,however, would require further investigation. Alterna-tively, IL-22 binding protein (IL-22BP) may increase in cir-culation with age and inhibit the expression of IL-22receptor in pancreas.In the spleen, we found Reg mRNA levels to be absent,

thus supporting the finding by Huszarik et al. [5] thatthe Reg genes are pancreas-specific with/without CFAtreatment. In contrast to Reg2 expression, however, theIL-22 mRNA levels in splenocytes were found to be in-creased in non-diabetic (4-week-old) NOD mice

Figure 6 A model for cytokine-mediated up-regulation of the β-cell Reg genes leading to β-cell regeneration. Interleukin-22 (red) bindsto its receptor complex, IL-22Rα/IL-10Rβ, which activates the Stat3 transcription factor protein (1), STAT3 then migrates into the nucleus of theβ-cell and stimulates Reg gene transcription and translation (green). Secreted Reg proteins are then thought to activate Cyclin D1. This allows theβ-cell to enter the G1/S transition of the cell cycle leading to regeneration. IL-22 can also activate the MAP3 kinases leading to Cyclin D1 (2),which in turn inactivates the Retinoblastoma (Rb). Platelet-derived growth factor (PDGF) (blue) similarly leads the activation of Cyclin D1, which inturn inactivates Rb allowing the release of sequestered transcription factors (TF) essential for the G1–S progression of the cell cycle. Reg2 geneexpression by activating Stat3, believed to be caused by receptor-induced Src kinase activity. The known signaling pathways (black and blue) andthe proposed signaling pathways (green) are identified.


following CFA injection, which confirms the finding byNikoopour et al. [15] that IL-22 is a secretory productinduced by CFA.We have identified IL-22 as a Reg gene inducer in the

NOD mouse pancreatic islets. Reg proteins act as auto-crine/paracrine growth factor for pancreatic β-cell regen-eration. They increase ATF-2 that binds to the cyclin D1gene and leads to signaling pathway to induce β-cell re-generation [26]. It is also important to note, however, thatlike PDGF [10] and Reg1[9,13], IL-22 has been shown byRadaeva et al. [27] to be directly involved in promotingcell cycle progression and cell survival. Using in vitrotreatment of recombinant IL-22 at a concentration similarto our own study of 10 ng/mL, Radaeva et al. [27] wereable to promote cell growth and survival of human hepa-tocellular carcinoma HepG2 cells via the activation of aseries of anti-apoptotic and mitogenic proteins which in-cluded: Bcl-2, Bcl-xL, Mcl-1, c-myc, and cyclin D1. Due tothe ontological similarity between hepatic cells and pan-creatic cells, we also investigated the effect of IL-22 onpancreatic islet cell growth. We found increased DNAsynthesis in the IL-22 treated islets. The extent to whichthis increase was dependent on the production of Reg2 orReg1 by these islets, however, will require further investi-gation. The mitogenic effect of IL-22 was greater on thesmaller-sized islets, suggesting that this population of β-cells may have a greater regenerative capacity. Smukleret al. [28] reported a sub-population of insulin-expressingβ-cells located in small islets in both the mouse and hu-man pancreas that had a high mitogenic potential, but didnot express the glucose transporter-2. These cells wereshown to be lineage multipotent and capable of formingmultiple islet endocrine cell types, as well as neural cells,and may represent a β-cell progenitor population that cancontribute to islet regeneration.CFA-treated diabetic NOD mice have been previously

shown to partially reverse T1D by down-regulatingautoimmunity and stimulating endogenous β-cell regen-eration through a mechanism that is presently un-known. This regeneration process, however, is notenough to prevent the disease entirely, as newlyregenerated β-cells are still selectively destroyed by on-going autoimmunity, but is believed to involve the mito-genic Reg gene family. We have demonstrated that thecytokine, IL-22, can up-regulate Reg2 and Reg1 expres-sion when directly co-cultured with pre-diabetic pan-creatic islets in vitro. Thus, IL-22 represents a newlydiscovered mediator for up-regulating islet Reg expres-sion using a JAK/STAT3 signal transduction pathway.We have also shown that excess IL-22 may have an in-hibitory effect on islet Reg expression, that the mRNAlevels for IL-22 are up-regulated with increasing ageand development of disease in the NOD mouse, thatCFA treatment causes a down-regulation of the IL-22Rα

receptor chain, and that IL-22 can increase β-cell DNAsynthesis when directly applied to pancreatic isletsin vitro.In conclusion, this study has identified the cytokine,

IL-22, to be involved in up-regulating Reg2 and Reg1gene family members and increasing β-cell DNA synthe-sis within pancreatic islets. This could result in β-cell re-generation in islets and that could potentially beincorporated into future therapeutic strategies for dis-ease management in Type 1 diabetes.

MethodsAnimalsFemale NOD/Ltj mice at 4, 8 and 12 weeks of age, anddiabetic NOD mice, shortly after the diagnosis of diabetes(blood glucose greater than 11 mM and polyuria), as wellas NOD.SCID mice at 4 weeks or as adults, were obtainedfrom the University of Western Ontario, London, Ontario,Canada and housed in a pathogen-free environment. Micewere housed three per cage, with food and water providedad libitum, and maintained on a 12-hour light/dark cycle.Blood glucose was measured from a tail sample using aglucometer (Bayer, Elkhart, IN). All experiments followedthe Canadian Council for Animal Care guidelinesand were approved by the Animal Use Sub-Committee,University of Western Ontario.

TreatmentsComplete Freund’s adjuvant (CFA) was purchased fromSigma Aldrich (St. Louis, MO). To determine the effect ofCFA on Reg2, Reg1, Reg3δ, IL-22, and IL-22-receptor(R)-αexpression, mice were injected i.p. either with 100 μl ofCFA (50 μg/mL) emulsified in saline (1:1), or with 100 μl ofsaline alone for control in 4, 8, 12-week, and diabetic NODmice. To determine the effect of IL-22 neutralization onReg expression; three 5 to 6 week-old NOD mice wereinjected i.p. with 100 μl of CFA with saline (1:1) followed1 hr later with 100 μl (i.p.) of 3 mg/mL rat anti-mouseIL22-01 monoclonal antibody (21) or isotype matched con-trol antibody (Pfizer). Mice were sacrificed in a CO2 cham-ber 48 hrs later for tissue extraction.

Extraction of splenocytes and pancreasApproximately 50 mg of whole spleen or pancreatic tissuefrom NOD/Ltj and NOD.SCID mice was homogenized ina solution containing buffer RLT (Qiagen, Mississauga,ON) and 2-Mercaptoethanol (Sigma Aldrich, St. Louis,MO) using a PowerGen 700 homogenizer (Fisher Scien-tific, Pittsburgh, PL).

Isolation of pancreatic isletsNOD mice were sacrificed by cervical dislocation. Coldcollagenase XI (0.23 mg/mL, Sigma Aldrich, St. Louis,


MO) was directly injected into the pancreas through thecommon bile duct. Collagenase at 37°C was used to digestthe pancreatic tissue. Connective tissues and remainingcells were removed using Hank’s buffered salt solution(HBSS) (Invitrogen, Carlsbad, CA) containing 5% (v/v) fetalcalf serum (FCS) (HyClone Laboratories, Logan, UT). Isletseparation was accomplished by density gradient centrifu-gation using dextran (Sigma Aldrich, St. Louis, MO). Cellslocated at the interface of dextran gradient layers of 11%and 23% were harvested by pipette and washed in RoswellPark Memorial Institute (RPMI) 1640. Individual pancre-atic islets were handpicked and cultured in Dulbecco’smodified eagle medium (DMEM) (Invitrogen, Carlsbad,CA) supplemented with 100 U/mL streptomycin (GIBCO,Grand Island, NY), 5 μg/mL penicillin, and 10% (v/v) FCSat 37°C.

Culture of pancreatic islets cells with IL-22Pancreatic islets were isolated from young (4–6 wk old)NOD mice. They were equilibrated overnight at 37°C,5% CO2 in RPMI-1640 medium supplemented with 10%fetal calf serum (FCS), 100 U/mL streptomycin and5 μg/mL penicillin and islets were handpicked and dis-tributed into 24 well plate (~50 islets per well) andtreated with recombinant mouse IL-22 (10 ng/ml or50 ng/ml, R&D systems, Minneapolis, MN) or with 2 mlof Th17-cell supernatant from IL-6/IL-23 treatedsplenocytes [15]. Two days later, islets were harvestedfor RNA extraction.

RNA extractionTotal RNA from isolated splenocytes, pancreatic or isletcells was extracted from homogenates using an RNeasyMidi Kit (Qiagen, Mississauga, ON). A DNase 1 treatmentkit (Ambion, Austin, TX) was used to eliminate DNA con-tamination. The RNA content was quantified by measuringabsorbance at 260 nm using a Nanodrop 1000 spectropho-tometer (NanoDrop Products, Wilington, DL), and the in-tegrity of selected RNA samples was checked using agarosegel electrophoresis and ethidium bromide to identify thepresence of the 18S and 28S rRNA bands.

Real time polymerase chain reactionApproximately 1 to 5 μg of total RNA was taken fromeach sample and reverse transcribed into cDNA usingoligo dT12-18 primers from Superscript III first-strandSynthesis SuperMix for quantitative RT-PCR (Invitrogen,St. Louis, MO). The cDNA concentration was measuredusing a spectrophotometer and was diluted to 225 ng/μLin diethyl pyrocarbonate water and amplified byPCR using a Quantifast SYBR Green PCR Kit (Qiagen,Mississauga, ON) with specific primers for Reg1, Reg2,Reg3δ, IL-22 and IL-22Rα1.25 μL) (Table 1). DNA wasamplified by PCR in a two-step melting/annealing pro-gram for up to 40 cycles using a Corbett Rotor-Gene 6000thermocycler (Corbett Life Sciences, San Francisco, CA).This thermocycler program consisted of preheating thesamples at 95°C for 5 min, cycling at 95°C for 10 sec and60°C for 20 sec, and a melting phase of 60 to 95°C. β-actinwas used as the housekeeping gene for relative quantifica-tion. The efficiency of each set of primers was determinedby the standard curve method.

Fluorescence immunohistochemistry for detection ofβ-cell DNA synthesisApproximately 50 isolated islets from 5 to 6 week-oldNOD mice were added to four wells of a six well tissue cul-ture plate and treated either with 10 ng/mL of recombin-ant IL-22 as above in DMEM media, or in DMEM mediaalone as a control, for 48 hrs. Islets were fixed in 4% para-formaldehyde (PFA) for 1 hr at 4°C and re-suspended in2% low melting point (LMP) Agarose (Fisher Scientific)which was first liquefied by heating to 80°C. The islets em-bedded in Agarose were dehydrated in 70% ethanol andembedded in paraffin. Sections of 5 μm were cut andmounted on glass slides. EdU staining was performed usinga Click-it EdU cell proliferation kit (Invitrogen) accordingto the manufacturer’s instructions. Non-specific stainingwas blocked using Background Sniper (Biocare Medical,Concord, CA, U.S.A.) for 10 min, and the sections incu-bated overnight with the primary antibody, mouse anti-insulin (1:2000; Sigma Aldrich, St. Louis, MO) followed bya 1 h incubation with secondary antibody, Alexafluor 488goat anti-mouse (1:500; Invitrogen). Cells were co-stainedwith nuclear DAPI (1:500; Sigma Aldrich, St. Louis, MO).

Table 1 List of genes and primer sequences used for qRT-PCR

Gene symbol Forward primer (5’→3’) Reverse primer (5’→3’) Product size (bp)

Reg2 cactgccaaccgtggttat gacaaaggagtactgtgcctca 75

Reg1 catctgccaggatcagttgc aggtaccataggacag 549

Reg3δ (INGAP) ccatggtgtctcacaagacc tgatgcgtggagaagacagt 117

IL-22 tcagctcagctcctgtcacat tccccaatcgccttgatctct 117

IL-22Rα gctcgctgcagcacactacca tctgtgtcgggagtcaggcca 247

β-actin gcccagagcaagagaggtat cacacgcagctcattgtaga 116


Twenty-six control and forty-six treated islets were identi-fied and analyzed, and the percentage of EdU-positive β-cells was quantified. To confirm results for DNA synthesis,slides containing sections from 15 additional islets werestained for nuclear Ki/67 overnight using the primary anti-body, mouse anti-Ki/67 (1:2000; Sigma Aldrich), followedby a 1 h incubation with secondary antibody, Alexafluor555 goat anti-mouse (1:500; Invitrogen). Analysis of theislet sections was performed by fluorescence microscopyand the images were analyzed using Northern EclipseVersion 7.0 software (Empix Imaging, Inc., Mississauga,ON, Canada).

Statistical analysisNormally distributed data were analyzed with SPSS statis-tical software using a one-way or two-way ANOVA, andsignificant changes between variables were detected usinga Tukey’s test. The α value was set to 0.05.

Competing interestThe authors declare that they have no competing interests.

Authors’ contributionTH performed the experiments and wrote the initial draft of the manuscript.BS, DJH and LAF contributed to the design of the study. BS and DJHanalyzed the data and contributed to the writing of the manuscript. EN, SB,OK and EL-C contributed to the data collection. All authors have read andapprove the manuscript.

AcknowledgementsWe thank Ms. Brenda Strutt and Jessica Hill for technical help and Dr.Margery Ma from Pfizer, Cambridge, MA for anti-IL-22 antibody. We alsothank Dr. Kathleen Hill, Dr. Alexander Timoshenko and Morgan Kleiber forassistance in experimental design and interpretation of the data. This workwas supported by grants from the Canadian Institutes of Health Research(CIHR).

Author details1Department of Microbiology and Immunology, University of WesternOntario, London, ON, Canada. 2Inflammation and Immunology,Biotherapeutics Research and Development, Pfizer, Cambridge, MA 02140,USA. 3Department of Physiology and Pharmacology, University of WesternOntario, London, ON, Canada. 4Lawson Health Research Institute, London,ON, Canada. 5Centre for Human Immunology, University of Western Ontario,London, ON, Canada. 6Robarts Research Institute, University of WesternOntario, London, ON, Canada.

Received: 25 January 2013 Accepted: 2 April 2013Published: 4 April 2013

References1. Van Belle TL, Coppieters KT, von Herrath MG: Type 1 diabetes: etiology,

immunology, and therapeutic strategies. Physiol Rev 2011, 91:79–118.2. Mandrup-Poulsen T: Beta cell death and protection. Ann NY Acad Sci 2003,

1005:32–42.3. Levetan C: Distinctions between islet neogenesis and β-cell replication:

implications for reversal of type 1 and 2 diabetes. J Diabetes 2010,2:76–84.

4. Dor Y, Brown J, Martinez OI, Melton DA: Adult pancreatic β-cells areformed by self duplication rather than stem-cell differentiation.Nature 2004, 429:41–46.

5. Huszarik K, Wright B, Keller C, Nikoopour E, Krougly O, Lee-Chan E, Qin HY,Cameron MJ, Gurr WK, Hill DJ, Sherwin RS, Kelvin DJ, Singh B: Adjuvant

immunotherapy increases β cell regenerative factor Reg2 in thepancreas of diabetic mice. J Immunol 2010, 185:5120–5129.

6. Kaneto H, Matsuoka TA, Miyatsuka T, Kawamori D, Katakami N, Yamasaki Y,Matsuhisa M: PDX-1 functions as a master factor in the pancreas.Front Biosci 2008, 13:6406–6420.

7. Singh B, Nikoopour E, Huszarik K, Elliot JF, Jevnikar AM: Immunomodulationand regeneration of islet beta cells by cytokines in autoimmune type 1diabetes. J Interferon Cytokine Res 2011, 31:711–719.

8. Taylor-Fishwick DA, Rittman S, Kendall H, Roy L, Shi W, Cao Y, Pittenger GL,Vinik AI: Cloning genomic INGAP: a Reg-related family member withdistinct transcriptional regulation sites. Biochem. Biophys. Acta 2003,1638:83–89.

9. Parikh A, Stephan A, Tzanakakis ES: Regenerating proteins and theirexpression, regulation, and signaling. Biomol Concepts 2012, 3:57–70.

10. Chen H, Gu X, Liu Y, Wang J, Wirt SE, Bottino R, Schorle H, Sage J, Kim SK:PDGF signaling controls age-dependent proliferation in pancreaticβ-cells. Nature 2011, 478:349–355.

11. Sadelain MW, Qin HY, Lauzon J, Singh B: Prevention of type 1 diabetes inNOD mice by adjuvant immunotherapy. Diabetes 1990, 39:583–589.

12. Qin HY, Sadelain MW, Hitchon C, Lauzon J, Singh B: Complete freund’sadjuvant-induced T cells prevent the development and adoptivetransfer of diabetes in nonobese diabetic mice. J Immunol 1993,150:2072–2080.

13. Liu L, Cui W, Li B, Lu Y: Possible roles of Reg family proteins in pancreaticislet cell growth. Endocr Metab Immune Disord Drug Targets 2008, 8:1–10.

14. Unno M, Nata K, Noguchi N, Narushima Y, Akiyama T, Ikeda T, Nakagawa K,Takasawa S, Okamoto H: Production and characterization of Reg knockoutmice: reduced proliferation of pancreatic beta-cells in Reg knockoutmice. Diabetes 2002, 51(Suppl 3):S478–S483.

15. Nikoopour E, Schwartz JA, Huszarik K, Sandrock C, Krougly O, Lee-Chan E,Singh B: Th17 polarized cells from nonobese diabetic mice followingmycobacterial adjuvant immunotherapy delay type 1 diabetes.J Immunol 2010, 184:4779–4788.

16. Aggarwal S, Xie MH, Maruoka M, Foster J, Gurney AL: Acinar cells of thepancreas are a target of interleukin-22. J Interferon Cytokine Res 2001,2:1047–1053.

17. Eyerich S, Eyerich K, Cavani A, Schmidt-Weber C: IL-17 and IL-22: siblings,not twins. Trends Immunol 2010, 31:354–361.

18. Wolk K, Witte E, Warszawska K, Sabat R: Biology of interleukin-22.Semin Immunopathol 2010, 3:17–31.

19. Shioya M, Andoh A, Kakinoki S, Nishida A, Fujiyama Y: Interleukin 22receptor 1 expression in pancreas islets. Pancreas 2008, 36:197–199.

20. Faustman DL, Wang L, Okubo Y, Burger D, Ban L, Man G, Zheng H,Schoenfeld D, Pompei R, Avruch J, Nathan DM: Proof-of-concept,randomized, controlled clinical trial of Bacillus-Calmette-Guerin fortreatment of long-term type 1 diabetes. PLoS One 2012,7(8):e41756.

21. Ma HL, Liang S, Li J, Napierata L, Brown T, Benoit S, Senices M, Gill D,Dunussi-Joannopoulos K, Collins M, Nickerson-Nutter C, Fouser LA, YoungDA: IL-22 is required for Th17 cell-mediated pathology in a mousemodel of psoriasis-like skin inflammation. J Clin Invest 2008, 118:597–607.

22. Okamoto H, Takasawa S: Recent advances in the Okamoto model: theCD38-cyclic ADP-ribose signal system and the regenerating geneprotein (Reg)-Reg receptor system in beta-cells. Diabetes 2002,51(Suppl 3):S462–S473.

23. Oh YS, Lee YJ, Park EY, Jun HS: Interleukin-6 treatment induces beta-cellapoptosis via STAT-3-mediated nitric oxide production. Diabetes MetabRes Rev 2011, 27:813–819.

24. Zhang Y, Valdez PA, Danilenko DM, Hu Y, Sa SM, Gong Q, Abbas AR,Modrusan Z, Ghilardi N, de Sauvage FJ, Ouyang W: Interleukin-22 mediateshost defense against attacking and effacing bacterial pathogens.Nature Med 2008, 14:282–289.

25. Gelebart P, Zak Z, Dien-Bard J, Anand M, Lai R: Interleukin-22 signalingpromotes cell growth in mantle cell lymphoma. Transl Oncol 2011, 4:9–19.

26. Radaeva S, Sun R, Pan H-N, Hong F, Gao B: Interleukin 22 (IL-22) plays aprotective role in cell-mediated murine hepatitis: IL-22 is a survivalfactor for hepatocytes via STAT3 activation. Hepatology 2004,39:1332–1342.

27. Takasawa S, Ikeda T, Akiyama T, Nata K, Nakagawa K, Shervani NJ, NoguchiN, Murakami-Kawaguchi S, Yamauchi A, Takahashi I, Tomioka-Kumagai T,


Okamoto H: Cyclin D1 activation through ATF-2 in Reg-inducedpancreatic beta-cell regeneration. FEBS Lett 2006, 580:585–591.

28. Smukler SR, Arntfield ME, Razavi R, Bikopoulos G, Karpowicz P, Seaberg R,Dai F, Lee S, Ahrens R, Fraser PE, Wheeler MB, van der Kooy D: The adultmouse and human pancreas contain rare multipotent stem cells thatexpress insulin. Cell Stem Cell 2011, 8:281–293.

doi:10.1186/2045-9769-2-2Cite this article as: Hill et al.: The involvement of interleukin-22 in theexpression of pancreatic beta cell regenerative Reg genes. CellRegeneration 2013 2:2.

Submit your next manuscript to BioMed Centraland take full advantage of:

• Convenient online submission

• Thorough peer review

• No space constraints or color figure charges

• Immediate publication on acceptance

• Inclusion in PubMed, CAS, Scopus and Google Scholar

• Research which is freely available for redistribution

Submit your manuscript at www.biomedcentral.com/submit


© The American Society of Gene Therapyoriginal article

854� www.moleculartherapy.org vol.�16�no.�5,�854–861�may�2008

Remission of Diabetes by β-Cell Regeneration in Diabetic Mice Treated With a Recombinant Adenovirus Expressing BetacellulinSeungjin Shin1, Na Li1, Naoya Kobayashi2, Ji-Won Yoon1,* and Hee-Sook Jun1,3

1Rosalind Franklin Comprehensive Diabetes Center, Chicago Medical School, Rosalind Franklin University of Medicine and Science, North Chicago, Illinois, USA; 2Department of Surgery, Okayama University Graduate School of Medicine and Dentistry, Shikata-cho, Okayama, Japan; 3Laboratory of Beta-Cell Biology and Autoimmunity, Lee Gil Ya Cancer and Diabetes Institute, Gachon University of Medicine and Science, Songdo-dong, Yeongsu-ku, Incheon, Korea. *Deceased.

Type 1 diabetes results from destruction of the majority of the pancreatic β cells by β cell–specific autoimmune responses; therefore, expansion of the β-cell mass in vivo is a possible approach to its treatment. Betacellulin (BTC) is known to promote β-cell growth and differentiation. We investigated whether transient, constitutive expres-sion, and secretion of BTC would regenerate sufficient numbers of pancreatic β cells to restore normoglyce-mia in diabetic animals. We constructed a recombinant adenoviral vector (rAd-BTC) containing the cytomegalo-virus promoter/enhancer, β-globin chimeric intron, and albumin leader sequence to facilitate secretion, followed by BTC (1-80) complementary DNA (cDNA) encoding mature BTC. A single intravenous (IV) administration of rAd-BTC resulted in complete remission of streptozotocin (STZ)-induced diabetes within 2 weeks in mice. The mice remained normoglycemic for >100 days; glucose toler-ance tests showed kinetics similar to normal, nondiabetic mice. Pancreatic insulin content, β-cell mass, and serum insulin levels in rAd-BTC-treated mice were significantly higher than in the controls. Treatment of autoimmune diabetic mice with rAd-BTC in combination with an immune suppressor resulted in remission of diabetes. We conclude that transient expression of BTC by rAd-BTC administration results in prolonged remission of diabe-tes in mice, by the regeneration of sufficient numbers of β cells in the pancreas.

Received 29 June 2007; accepted 20 January 2008; published online 18 March 2008. doi:10.1038/mt.2008.22

INTRODUCTIONType 1 diabetes results from the destruction of the majority of insulin-producing pancreatic β cells by β cell–specific autoim-mune responses.1,2 A possible approach to the cure of diabetes is expansion of the β-cell mass in vivo by stimulating proliferation of existing β cells3 and, possibly, differentiation of β-cell stem

progenitors.4 However, naturally occurring β-cell regeneration is insufficient for the remission of diabetes, because of insuffi-cient numbers of β cells, and the re-attack that is mounted on newly formed β cells by existing β cell–specific autoreactive T cells.5 Therefore any therapy that utilizes β-cell regeneration for the control of type 1 diabetes must produce sufficient num-bers of β cells and also protect the newly formed β cells from autoimmune attack.

Considerable evidence has shown that intra-islet stem cells and pancreatic ductal or acinar cells can proliferate and differen-tiate into insulin-producing cells when exposed to appropriate environmental stimuli, such as morphogens or growth factors.6 Betacellulin (BTC), a member of the epidermal growth factor fam-ily, induces proliferation and differentiation of insulin-producing insulinoma cells7 and converts exocrine pancreatic cells (AR42J) into insulin-expressing cells when combined with activin A.8 Treatment with BTC induces the expression of the insulin gene in an α-cell line transfected with Pdx-1 (ref. 9) and in an intestinal cell line transfected with Pdx-1.10,11 Consistent with these in vitro studies, administration of recombinant human BTC improved glucose tolerance in alloxan-induced diabetic mice12 and in 90% of pancreatectomized rats, by promoting β-cell regeneration.13 It was also reported that combined treatment with activin A and BTC results in the regeneration of pancreatic β cells in neonatal streptozotocin (STZ)-treated rats.14 These reports suggest that BTC is a promising therapeutic tool for the potential cure of dia-betes by inducing the regeneration of insulin-producing cells. Although BTC promotes pancreatic β-cell growth and differen-tiation7,10,11,15 and improves glucose metabolism by stimulating β-cell regeneration in diabetic animals,12,13,16 even multiple sys-temic injections of recombinant BTC into diabetic animals fail to completely restore normoglycemia, probably because of the short half-life of the BTC protein.12 In this study, we determined whether continuous expression and secretion of BTC for a lim-ited time, by a gene therapy approach using adenoviral (Ad) vector, would regenerate sufficient numbers of pancreatic β cells to restore normoglycemia in diabetic mice.

Correspondence: Dr. Hee-Sook Jun, Laboratory of Beta-Cell Biology and Autoimmunity, Lee Gil Ya Cancer and Diabetes Institute, Gachon University of Medicine and Science, 7-45 Sondo-dong, Yeongsu-ku, Incheon 406-840, Korea. E-mail: [email protected]

mailto:[email protected]

http://www.nature.com/doifinder/10.1038/mt.2008.22

Molecular Therapy vol.�16�no.�5�may�2008� 855

© The American Society of Gene TherapyRemission of Diabetes by β-Cell Regeneration

RESULTSEfficient production and secretion of BTC by rAd-BTC both in vitro and in vivoWe constructed a recombinant Ad vector (rAd-BTC) containing the cytomegalovirus promoter/enhancer, β-globin chimeric intron, and albumin leader sequence to facilitate secretion, followed by BTC (1-80) complementary DNA (cDNA) encoding mature BTC (Figure 1a). In order to check whether rAd-BTC produces and secretes BTC efficiently, a human hepatocyte line, TTNT-16,17 was infected with rAd-BTC, and the expression of BTC protein was examined by immunohistochemical staining. BTC was clearly expressed in rAd-BTC-infected cells, but not in cells infected with rAd expressing β-galactosidase (rAd-βgal) (Figure 1b). Using enzyme-linked immunosorbent assay, we examined the secretion of BTC into the media by rAd-BTC-infected cells and found that secretion of BTC was detected when cells were infected with a multiplicity of infection of 1, and further increased when infected with a multiplicity of infection of 10 (Figure 1c).

In order to confirm the secretion and expression of BTC by rAd-BTC treatment in vivo, we injected rAd-BTC intravenously (IV) into nonobese diabetic/severe combined immunodeficiency (NOD.SCID) mice with STZ-induced diabetes, and determined the circulating BTC levels. The serum BTC level, which was 2.8 ± 0.6 ng/ml at day 3, fell to approximately one-third of that at week 1, further decreased at 2 weeks, and remained at a con-sistent low level between 2 weeks and 4 weeks after viral injec-tion. At 8 weeks after rAd-BTC injection, the serum BTC level was below the detection limit. No BTC was detected in sera from rAd-βgal-treated mice (Figure 2a). We examined the expression of BTC messenger RNA (mRNA) in various tissues by reverse transcriptase-polymerase chain reaction (PCR), using primers designed to detect only exogenous BTC mRNA. The expression

of BTC mRNA was detected in all the tissues that were tested (liver, lung, kidney, heart, spleen, and pancreas) at 4 weeks after viral injection, with the highest expression being found in the liver, whereas BTC mRNA was not detected in any of the tissues tested in rAd-βgal-treated mice (Figure 2b). Insulin mRNA was detected only in pancreatic tissues of rAd-BTC-treated mice, but not in other tissues including the liver at 4 weeks after viral injec-tion (Figure 2b), thereby suggesting that BTC-induced regenera-tion of insulin-producing cells is limited to the pancreas. Weak insulin mRNA expression was detected in the pancreas of rAd-βgal-treated mice (Figure 2b), probably because of some remain-ing β cells. These results indicate that rAd-BTC treatment results in efficient expression and secretion of BTC protein in vivo and may induce insulin-producing cells in the pancreas.

Remission of STZ-induced diabetes by rAd-BTC treatmentIn order to determine whether BTC expression in vivo results in remission of diabetes, we examined blood glucose levels in STZ-induced diabetic NOD.SCID mice after injecting rAd-BTC, and found that blood glucose levels gradually decreased and became normal within 2 weeks after rAd-BTC injection. Normoglycemia was maintained for >100 days, until the end of the experiment. In contrast, mice treated with rAd-βgal showed persistent hypergly-cemia and died (Figure 3a). In order to determine whether glucose is properly cleared in rAd-BTC-treated mice, intraperitoneal (IP) glucose tolerance tests were performed in the STZ-induced

CMV promoter/enhancer

a

b c

�-Globlin/IgGchimeric intron

Betacellulin (1–80)

Albumin leaderpeptide (1–25)

SV40 polyAsignal

rAd-�gal treated 14

12

10

8

Bet

acel

lulin

(ng

/106 c

ells

)

6

4

2

00 1 10 MOI

rAd-BTC treated

Figure 1 Efficient expression and secretion of betacellulin (BTC) in rAd-BTC-infected cells. (a) Schematic diagram of rAd-BTC. (b) TTNT-16 cells (1 × 104 cells) were infected with rAd-BTC or rAd-βgal at a multipli-city of infection (MOI) of 5, and the expression of BTC was determined by staining with anti-human BTC antibody at 24 hours after infection. (c) TTNT-16 cells (1 × 106 cells) were infected with rAd-BTC at various MOIs and incubated for 24 hours. BTC secretion into the supernatant was analyzed by enzyme-linked immunosorbent assay. Values are mean ± SD from three experiments. CMV, cytomegalovirus; IgG, immunoglobulin G; rAd, recombinant adenovirus; SV40, simian virus 40.

3

a

bBTC

L Lu K H S P L Lu K H S P

BTC/-RT

Insulin

Insulin/-RT

GAPDH

GAPDH-RT

rAd-�gal treated rAd-BTC treated

2

Ser

um B

TC

(ng

/ml)

1

00 10 20 30

Days after virus injection40 50 60

rAd-�gal

rAd-BTC

Figure 2 Efficient expression and secretion of betacellulin (BTC) after treatment with rAd-BTC in vivo. Streptozotocin-induced diabetic nonobese diabetic/severe combined immunodeficiency mice were injected with rAd-BTC. (a) Serum BTC levels were measured using enzyme-linked immunosorbent assay in rAd-BTC-treated mice (closed circles) and rAd-βgal-treated mice (open circles) at the indicated days after viral injection. Values are mean ± SD (n = 4/group). (b) Expression of BTC messenger RNA (mRNA) and insulin mRNA in various tissues was analyzed using reverse transcriptase (RT)-polymerase chain reaction at 4 weeks after viral injection. Glyceraldehyde-3-phosphate dehydroge-nase (GAPDH) mRNA was analyzed as an internal control. /-RT, without RT reaction; H, heart; K, kidney; L, liver; Lu, lung; P, pancreas; S, spleen; rAd, recombinant adenovirus.

856� www.moleculartherapy.org vol.�16�no.�5�may�2008


diabetic mice that achieved normoglycemia at 4 weeks after rAd-BTC treatment. rAd-BTC-treated mice showed the same kinetics of glucose clearance as normal mice, whereas rAd-βgal-treated control mice could not clear exogenous glucose (Figure 3b).

Regeneration of β cells in rAd-BTC-treated diabetic miceIn order to determine whether the remission of diabetes in rAd-BTC-treated mice was caused by the regeneration of insulin- producing cells, we stained liver and pancreatic sections with anti-insulin antibody. Insulin-positive cells were rarely seen in pancre-atic islets from diabetic mice at 4 weeks after rAd-βgal treatment. In contrast, insulin-positive cells were clearly seen in pancreatic islets from rAd-BTC-treated mice at 2 weeks after treatment, and more were seen at 4 weeks (Figure 4). The islets from rAd-BTC-treated mice were also C-peptide positive (data not shown). We did not find any insulin-positive cells in the liver (data not shown). Double-staining of the islets with anti-insulin and anti-glucagon antibodies at 2 weeks after rAd-BTC treatment showed that insulin-producing cells were interspersed with glucagon-producing α cells in the islets of rAd-BTC-treated mice, probably because of relocalization of glucagon-producing α cells after the destruction of β cells by STZ, and the subsequent appearance of newly formed β cells. However, at 4 weeks after rAd-BTC treatment, β cells were clustered centrally and surrounded by α cells, as found in normal mice (Figure 4).

We then measured the β-cell mass and found that the mass in rAd-BTC-treated mice was significantly higher than in rAd- βgal-treated mice and attained 32% (0.42 ± 0.12 mg/pancreas) and 47% (0.61 ± 0.2 mg/pancreas) of the mass in normal mice

700a600

500B

lood

glu

cose

leve

l(m

g/dl

)

400

300

200

100

00 20 40

Days after virus injection60 80 100

rAd-�gal (n = 9)rAd-BTC (n = 12)Normal (n = 8)

b

Blo

od g

luco

se le

vel

(mg/

dl)

00

100

200

300

400

500

600

50Time (minutes)

100 150

rAd-�gal (n = 6)rAd-BTC (n = 9)Normal (n = 8)

Figure 3 Remission of diabetes in streptozotocin (STZ)-induced dia-betic mice after systemic administration of rAd-BTC. (a) rAd-BTC was administrated into STZ-induced diabetic nonobese diabetic/severe com-bined immunodeficiency (NOD.SCID) mice (closed diamonds) and blood glucose levels were measured. Nondiabetic NOD.SCID mice (open circles) and STZ-induced diabetic NOD.SCID mice treated with rAd-βgal (closed squares) were used as controls. (b) rAd-BTC-treated, STZ-induced diabetic NOD.SCID mice in which blood glucose levels were normalized (closed diamonds) were made to fast for 4 hours and injected with glucose at 4 weeks after rAd-BTC treatment. Blood glucose levels were measured at the indicated times after glucose injection. Untreated NOD.SCID mice (open circles) and rAd-βgal-treated mice (closed squares) were used as controls. Values are mean ± SD. BTC, betacellulin; rAd, recombinant adenovirus.

NormalHE

Insulin

Glucagon

Insulin/Glucagon

Diabetic BTC-2 weeks BTC-4 weeks

Figure 4 Histochemical and immunohistochemical analysis of pan-creatic islets in rAd-BTC-treated mice. Streptozotocin (STZ)-induced diabetic nonobese diabetic/severe combined immunodeficiency (NOD.SCID) mice were injected with rAd-BTC and killed 2 weeks (BTC-2 weeks) or 4 weeks (BTC-4 weeks) later. rAd-βgal-treated diabetic NOD.SCID mice (diabetic) and untreated NOD.SCID mice (normal) were used as controls. Pancreata were removed and serial sections were stained with hematoxylin and eosin (HE) or anti-insulin antibody (red) and anti-glucagon antibody (green). Similarly sized islets were chosen for the photographs. Bar = 20 μm. BTC, betacellulin; rAd, recombinant adenovirus.

Normal

a b

c

Diabetic

BTC-2 weeks

BTC-4 weeks

Normal

Diabetic

BTC-2 weeks

BTC-4 weeks

Normal

Diabetic

BTC-2 weeks

BTC-4 weeks

0.0 0.5 1.0

*

**

*

**

*

**

�-Cell mass (mg/pancreas)1.5 2.0 0 200 400

Pancreatic insulin (ng/mg pancreas)600 800

0 200 400Pancreatic insulin (pg/ml)

600 800 1,000 1,200 1,400

Figure 5 Increases in β-cell mass and insulin levels in rAd-BTC-treated mice. Streptozotocin (STZ)-induced diabetic nonobese dia-betic/severe combined immunodeficiency (NOD.SCID) mice were injected with rAd-BTC and killed 2 weeks (BTC-2w) or 4 weeks (BTC-4w) later. rAd-βgal-treated diabetic NOD.SCID mice (diabetic) and untreated NOD.SCID mice (normal) were used as controls. (a) The β-cell mass was calculated from the mean β-cell fraction as measured by immunohistochemical staining, and the weight of the pancreas. *P < 0.05, **P < 0.01 compared with diabetic control. (b) The insulin content of the pancreas was measured by enzyme-linked immunoassay. *P < 0.05, **P < 0.001 compared with diabetic control. (c) Before the mice were killed, serum insulin concentration was measured using at 30 minutes after glucose loading. *P < 0.01, **P < 0.001 compared with diabetic control. Values are mean ± SD (n = 3/group). BTC, betacellulin; rAd, recombinant adenovirus.



(1.3 ± 0.3 mg/pancreas) at 2 weeks and 4 weeks after treatment, respectively (Figure 5a). In order to confirm that rAd-BTC treat-ment results in the increase of insulin-producing cells in the pancreas, we measured the insulin content of pancreatic extracts from rAd-BTC-treated mice using enzyme-linked immunoas-say. The insulin levels of rAd-BTC-treated mice were significantly higher than those of rAd-βgal-treated mice at 2 weeks as well as 4 weeks after treatment, although lower than those of normal mice (Figure 5b). We also measured serum insulin levels using enzyme-linked immunoassay after glucose loading, and found that serum insulin levels in rAd-BTC-treated mice were significantly higher than in rAd-βgal-treated diabetic mice, attaining~70% of the level found in normal mice at 2 weeks and >80% at 4 weeks (Figure 5c). All these results clearly show that the number of insulin-producing cells increased in the pancreas of rAd-BTC-treated mice, thereby suggesting that the remission of diabetes by rAd-BTC treatment was mainly on account of the regeneration of β cells in the pancreas.

In order to determine the source of regenerated β cells in rAd-BTC-treated mice, we first examined whether rAd-BTC treatment results in the proliferation of the remaining β cells. We injected STZ- and rAd-BTC-treated NOD.SCID mice with 5-bromodeoxyuridine (BrdU; 100 mg/kg body weight, IP)

(Sigma, St. Louis, MO), which labels dividing cells, for 5 days beginning on day 3 after rAd-BTC injection, and stained pancre-atic sections with anti-BrdU and anti-Pdx-1 antibodies or anti- insulin antibodies. We found a threefold increase in BrdU/Pdx-1 (Figure 6a) and BrdU/insulin (Figure 6b) double-positive cells in rAd-BTC-treated mice during this short period of BrdU exposure in comparison with rAd-βgal-treated control mice. In order to examine whether BTC can induce the differentiation of β cells from ductal cells, we stained pancreatic sections with anti-BrdU antibody and anti-cytokeratin antibody, a marker for ductal cells. We found that BrdU/cytokeratin double-posi-tive cells increased fourfold in the pancreas of rAd-BTC-treated mice (Figure 6c), but we failed to find any Pdx-1/cytokeratin double-positive cells (data not shown) or insulin/cytokeratin double-positive cells (Figure 6d). In order to examine whether the newly generated β cells transdifferentiated from α or δ cells, we stained pancreatic sections of rAd-BTC-treated NOD.SCID mice with anti-insulin and anti-glucagon antibodies or anti-insulin and anti-somatostatin antibodies, but failed to find any double-positive cells (Figure 6d). We did find some BrdU/ glucagon double-positive cells, and very few BrdU/somatostatin double-positive cells (data not shown).

Remission of diabetes in rAd-BTC-treated autoimmune diabetic miceIn order to examine the efficacy of rAd-BTC gene therapy in auto-immune diabetic NOD mice, we injected diabetic NOD mice hav-ing newly developed diabetes (blood glucose levels > 500 mg/dl) with rAd-BTC (4 × 1011 particles, IV—the optimal dose for NOD mice as determined by a preliminary experiment) and examined blood glucose levels. The treatment of chronically diabetic mice with rAd-BTC did not result in remission of diabetes, even tempo-rarily (data not shown). However, in diabetic NOD mice with newly developed diabetes, blood glucose levels decreased to <300 mg/dl at days 4–5 after viral injection, but returned to levels >500 mg/dl at 2 weeks, probably because of regenerated insulin-producing cells being attacked by autoimmune responses (Figure 7a). We then injected complete Freund’s adjuvant (CFA; 100 μl, subcutaneously) into female diabetic NOD mice having newly developed diabetes, so as to prevent the autoimmune response, treated them with rAd-BTC 3 days later, and then examined blood glucose levels. We found that blood glucose levels were normalized at 2–3 weeks after viral injection, and normoglycemia was maintained for >100 days, until the end of the experiment. The treatment of diabetic NOD mice with CFA alone, or with CFA along with rAd-βgal, had no effect on blood glucose levels (Figure 7a). When we performed IP glucose tolerance tests in normoglycemic NOD mice at day 100 after rAd-BTC/CFA treatment, blood glucose levels peaked at 30 minutes after glucose loading and returned to normal at 120 minutes, similar to the pattern observed in nondiabetic NOD mice. In contrast, blood glucose levels of diabetic NOD control mice injected with CFA were significantly higher than those of rAd-BTC/CFA-treated mice at all time points of measurement (Figure 7b). When we examined pancreatic sections of rAd-BTC-treated NOD mice 3 months after rAd-BTC injection, we found that regenerated insulin-positive islets were surrounded by immunocytes, but were not infiltrated by them. In contrast, the islets from diabetic NOD mice were severely

0.0

rAd-BTC

rAd-BTC

rAd-�gal

rAd-BTC

rAd-�gal

rAd-BTC

rAd-�gal

pdx-

1/B

rdu

Cyt

oker

atin

/Brd

u

Insu

linIn

sulin

/Brd

u

BrdU labeling index (%)

BrdU labeling index (%)

rAd-�gal

rAd-BTCrAd-�gal

a

c d

b rAd-BTCrAd-�gal

0.5 1.0 1.5 2.0 2.5 3.0

0 2 4 6 8 10

0.0BrdU labeling index (%)

0.5 1.0 1.5 2.0 2.5 3.0

P < 0.05 P < 0.05

P < 0.01

Cytokeratin Glucagon Somatostatin

rAd-

�gal

rAd-

BT

C

Figure 6 Identification of proliferating cells in the islets of rAd-BTC-treated mice. Streptozotocin-induced diabetic nonobese diabetic/severe combined immunodeficiency (NOD.SCID) mice were treated with rAd-βgal or rAd-BTC, and then injected with 5-bromodeoxyuridine (BrdU) daily from day 3 to day 7 after virus treatment. (a) Regenerated cells from β-cell lineage. Pancreatic sections were double-stained with anti-Pdx-1 and anti-BrdU antibodies (top panel). Proliferation of Pdx-1-positive cells is shown in terms of the number of BrdU/Pdx-1 double-positive cells as a percent-age of the total number of Pdx-1-positive cells (bottom panel). (b) Insulin production in regenerated cells. Pancreatic sections were stained with anti-insulin and anti-BrdU antibodies (top panel). The proliferation of insu-lin-positive cells is shown (bottom panel). (c) Regenerated cells of ductal lineage. Pancreatic sections were double-stained with anti-cytokeratin and anti-BrdU antibodies (top panel). Proliferation of cytokeratin-positive cells is shown (bottom panel). (d) Absence of insulin production in non-β cells. Pancreatic sections from rAd-BTC- and rAd-βgal-treated mice, prepared at 1 week after virus injection, were stained with anti-cytokeratin and anti-insulin antibodies (left), anti-insulin and anti-glucagon antibodies (center), or anti-insulin and anti-somatostatin antibodies (right). Bars = 20 μm. Arrows indicate colocalization. Values are mean ± SD (n = 3/group). BTC, betacellulin; rAd, recombinant adenovirus.



infiltrated with immunocytes, and no insulin-positive cells were found (Figure 8a). The pancreatic insulin content of rAd-BTC/CFA-treated NOD mice was significantly higher than that of CFA only–treated NOD mice but lower than that of nondiabetic 6-week-old NOD mice (Figure 8b). In order to examine whether insulin secretion in response to glucose was normal, we examined serum insulin levels 30 minutes after glucose loading. Serum insulin levels of rAd-BTC/CFA-treated NOD mice were significantly higher than those of CFA only–treated NOD mice, and almost attained the levels found in normal mice (Figure 8c).

DISCUSSIONTransplantation of pancreatic islets is a promising method for the treatment of type 1 diabetes; however, this method is severely hampered by insufficient supplies of donor tissue. Possible meth-ods of producing β cells for transplantation have been proposed, such as establishment of human pancreatic β-cell lines18,19 or dif-ferentiation of embryonic or pancreatic stem cells into β cells.20,21 An alternative approach is the in vivo expansion of β cells with growth factors.22 In this study, we attempted to expand β cells in vivo by using an rAd vector that constitutively expresses a β-cell growth factor, BTC. We found that a single injection of rAd-BTC produced remission of diabetes for >100 days by inducing the regeneration of insulin-producing β cells in diabetic mice.

The results of our study and those of other studies utilizing BTC indicate that transient expression of sufficient BTC over a sufficient period of time appears to be important for successfully generat-ing β cells in adequate numbers to achieve remission of diabetes. Multiple systemic injections of recombinant BTC protein failed to completely restore normoglycemia, probably because of insuffi-cient amounts and the short half-life of the BTC protein, although glucose metabolism was improved.12,13,16 The administration of a helper-dependent Ad vector expressing BTC into STZ-induced diabetic mice also failed to remit diabetes,22 probably because of its low transduction efficiency as compared to the rAd vector used in this study. In addition, BTC gene therapy by ultrasound-targeted microbubble destruction in STZ-induced diabetic rats did not lower blood glucose levels, perhaps because of the low expression of the BTC gene, although a combination of Pdx-1 and BTC gene therapy did maintain blood glucose levels at <200 mg/dl for 10 days.23 IV administration of our rAd-BTC, which was designed to facilitate BTC secretion from the liver, resulted in high serum BTC levels within 3 days after viral injection in diabetic mice. Serum BTC lev-els were significantly higher than those in rAd-βgal-treated mice for 4 weeks after viral injection, and they gradually fell to undetectable levels by 8 weeks. Such a transient expression of BTC is preferable, considering that long-term expression of growth factors may have deleterious effects on growth regulation of other cells. Given that glucose levels and β-cell mass improved as early as 2 weeks after rAd-BTC treatment, with further improvement at 4 weeks, the major effect of BTC may occur within 1 month in mice, but can confer long-lasting effects on normalization of blood glucose levels and remission of diabetes. BTC mRNA expression was detected in various tissues including the liver, spleen, heart, lung, kidney, and pancreas, with the highest expression in the liver, as shown by anal-ysis using reverse transcriptase-PCR at 4 weeks after viral injection. However, insulin mRNA was detected only in the pancreas, thereby

700arAd-�gal/CFA-treated NOD (n = 5)CFA-treated NOD (n = 6)rAd-BTC-treated NOD (n = 6)rAd-BTC/CFA-treated NOD (n = 7)Nondiabetic (n = 5)

600

500

400

300

200

100

0

Blo

od

glu

cose

leve

l(m

g/d

l)

0 20 40 60Days after virus injection

80 100 120

b800

Blo

od

glu

cose

leve

l(m

g/d

l)

600

400

200

0

Time (minutes)0 906030 120 150 180

CFA-treated (n = 6)

rAd-BTC/CFA (n = 7)Nondiabetic (n = 8)

Figure 7 Remission of diabetes in autoimmune diabetic nonobese diabetic (NOD) mice by treatment with rAd-BTC. Autoimmune dia-betic NOD mice (blood glucose levels > 500 mg/dl) were injected with complete Freund’s adjuvant (CFA) subcutaneously 3 days before injec-tion of rAd-BTC (closed diamonds), or with CFA 3 days before injection of rAd-βgal (open circles), or with CFA alone (closed circles), or with rAd-BTC alone (open triangles). Age-matched, nondiabetic NOD.SCID mice (open squares) were also used as controls. (a) Blood glucose levels were measured. (b) Diabetic NOD mice in which blood glucose levels were restored to normalcy after rAd-BTC/CFA treatment (100 days after treat-ment; closed diamonds) were made to fast for 4 hours and then injected with glucose. Blood glucose levels were measured at the indicated times after glucose injection. CFA-treated diabetic NOD mice (closed circles) and nondiabetic NOD mice (open squares) were used as controls. Values are mean ± SD. BTC, betacellulin; SCID, severe combined immunodefi-ciency; rAd, recombinant adenovirus.

rAd-

BT

C/C

FAD

iabe

tic

0

Nondiabetic

CFA treated

P < 0.001

P < 0.001

rAd-BTC/CFAtreated

a b

c

Nondiabetic

CFA treated

rAd-BTC/CFAtreated

200 400Pancreatic insulin (ng/mg pancreas)

600 800 1,000

0 200 400

Serum insulin (pg/ml)

600 800 1,000 1,200

Figure 8 Increases in β-cell mass and insulin levels in rAd-BTC/com-plete Freund’s adjuvant (CFA)-treated mice. (a) At 3 months after viral injection, pancreatic sections from rAd-BTC/CFA-treated nonobese diabetic (NOD) mice were stained with anti-insulin antibody. The islets were strongly stained with anti-insulin antibody and surrounded by lymphocytes (top). In diabetic NOD control mice, insulin-positive cells were scarce and the islets were heavily infiltrated by lymphocytes (bottom). Bars = 40 μm. (b) The insulin content of the pancreata from rAd-BTC/CFA-treated NOD mice was measured by enzyme-linked immunoassay (EIA). (c) Before the mice were killed, serum insulin concentrations were measured using EIA at 30 minutes after glucose loading. CFA-treated diabetic NOD and nondiabetic NOD (6 weeks old) were used as controls. Values are mean ± SD (n = 3/group). BTC, betacellulin; rAd, recombinant adenovirus.



suggesting that the effect of BTC on the regeneration of insulin- producing cells is limited to the pancreas. Therefore BTC, in the circulation and expressed locally in the pancreas, may contribute to the regeneration of β cells in the pancreas.

In order to examine the source of regenerated β cells, we first determined whether the increase in the number of β cells in rAd-BTC-treated mice was the result of the proliferation of the remaining β cells (a few β cells always remain in STZ-treated dia-betic mice) (e.g., Figure 4, diabetic). The number of BrdU/Pdx-1 double-positive cells in the pancreas was significantly higher (over threefold) than in the rAd-βgal-treated diabetic controls. We con-firmed the proliferation of β cells by showing that BrdU/insulin double-positive cells were also increased in rAd-BTC-treated mice. We observed some BrdU-positive islet cells that were nega-tive for Pdx-1, insulin, glucagon, and somatostatin. We speculate that these cells might be islet progenitor cells, but further studies are needed for their identification. BTC is known to be a potent mitogen for a variety of cells including the endocrine cells of the human fetal pancreas,24 adult rat islets,25 and insulinoma cells.26 Our results therefore suggest that BTC might directly induce the replication of existing β cells, thereby confirming the report that adult pancreatic β cells form by self-duplication.3

BTC was shown to convert glucagonoma cells and intestinal epitheloid cells into insulin-producing cells when Pdx-1 was also expressed in these cells.9,11 In addition, other studies suggest that BTC may induce transdifferentiation of δ cells into β cells.12,14 On the basis of this evidence, we used immunohistochemical markers to examine the possibility of transdifferentiation from α or δ cells, but failed to detect any insulin/glucagon or insulin/somatostatin double-positive cells in the pancreas of rAd-BTC-treated mice. It has also been reported that islet cells can differentiate from progen-itor cells associated with the ductal epithelium.26,27 More recently, BTC transduction through retrograde pancreatic duct injection of an Ad vector was shown to induce β-cell neogenesis from ductal cells as well as proliferation of existing β cells in mice.28 We exam-ined this possibility by double-staining with anti-cytokeratin antibody, a marker for ductal cells, and anti-Pdx-1 or anti-insulin antibodies, but failed to find any cytokeratin/Pdx-1 or cytokeratin/insulin double-positive cells. The difference between our results and those of the earlier report might be because our experimental design resulted in expression of BTC primarily by the liver, with secretion into the circulation, whereas the design of the previous report resulted in BTC overexpression in pancreatic ductal cells.

Taken together, our results indicate that β-cell replication is likely to be the major source of new β cells generated by BTC. However, the existence of a mechanism of transdifferentiation from ductal cells or other pancreatic endocrine cells cannot be completely excluded. It is possible that dedifferentiation and subsequent redifferentiation might occur from non-β endocrine cells, such as α and δ cells, into β cells; if this were so, markers for terminally differentiated cells, such as glucagon and somatostatin, would not be seen simultaneously with insulin. We also cannot exclude the possibility that BTC induces dif-ferentiation of pancreatic progenitor/stem cells into β cells; further studies, including lineage-tracing studies, will be necessary for clari-fying the role of BTC in increasing the β-cell mass.

Finally, we examined whether our rAd-BTC gene therapy shows similar effects on the remission of diabetes in autoimmune diabetic

NOD mice, an animal model of human type 1 diabetes. In contrast with the results seen in chemically induced diabetic mice, rAd-BTC treatment of autoimmune diabetic mice resulted in lowering of blood glucose only transiently, probably as a result of autoimmune attack against newly generated β cells. CFA is known to prevent dia-betes in NOD mice by downregulating autoreactive T cells through stimulation of natural killer cells29 and induction of regulatory T cells.30,31 We therefore attempted to prevent autoimmune attack in diabetic NOD mice by injecting CFA before treating with rAd-BTC. Pretreatment with CFA resulted in long-term remission of diabetes (>100 days) after rAd-BTC treatment, even though the BTC gene was expressed for a relatively short time. This result suggests that once β cells are regenerated and protected from autoimmune attack, further expression of BTC is unnecessary. However, blood glucose levels in diabetic NOD mice treated with rAd-BTC/CFA were not completely normalized, in contrast to STZ-induced diabetic NOD.SCID mice treated with rAd-BTC. This might be because of different degrees of β-cell destruction in the two models. STZ treatment gen-erally does not destroy all the β cells; however, in the spontaneously diabetic NOD mice model, the β cells are extensively destroyed by the time symptoms appear, and are almost completely destroyed at the chronic stage of the disease. In fact, when we treated chronically diabetic NOD mice with rAd-BTC, there was no remission of the disease. This suggests that some β cells are required to remain in order for the rAd-BTC therapy to be successful.

Using two animal models of diabetes, chemically induced diabetic mice and spontaneous autoimmune diabetic mice, we have shown that a single administration of rAd-BTC results in the complete remission of diabetes as a result of BTC-induced regen-eration of β cells. The long-term remission of diabetes by BTC-induced regeneration of β cells, achieved despite the relatively short period during which the BTC gene was active, suggests that there is therapeutic potential for this method of regenerating the β-cell mass from remaining β cells as a possible treatment for type 1 diabetes in humans.

MATERIALS AND METHODSProduction of rAds expressing human BTC cDNA (rAd-BTC). Human BTC cDNA encoding the mature 80-amino acid protein was purchased from American Type Culture Collection (ATCC #1887012; Rockville, MD). The cDNA was cloned into pCR259 (Qbiogene, Montreal, Canada) Ad transfer vector at SmaI and NotI sites. The albumin leader pep-tide sequence was then inserted at SalI and SmaI sites, and the 6-base pair sequence, which was additionally inserted by the SmaI recognition sequence, was removed by site-directed mutagenesis. The expression cas-sette contained a cytomegalovirus promoter, β-globin/immunoglobulin G (IgG) chimeric intron, and simian virus 40 poly A signal. An Ad vec-tor carrying these genes was constructed using the Transpose-Ad method (Qbiogene, Montreal, Canada) in accordance with the manufacturer’s protocol. Viruses were produced in HEK-293 cells and purified using CsCl2-gradient ultracentrifugation.32 As a control, an rAd expressing β-galactosidase (rAd-βgal) was produced by replacing BTC cDNA with β-galactosidase cDNA. The viral titer was determined by 50% tissue culture infection dose assay on 293 cells.

Enzyme-linked immunosorbent assay for BTC. Ninety-six-well microtiter plates were coated with 100 μl of anti-human BTC antibody (1 μg/ml in phosphate-buffered saline; AB-261-NA, R&D Systems, Minneapolis, MN) overnight at 4°C. After blocking with 5% horse serum, 1% bovine serum albumin, and 0.05% Tween-20 in phosphate-buffered saline, serum samples



were added and incubated at room temperature for 2 hours. The plates were washed and incubated with biotinylated anti-human BTC antibody (100 ng/ml; BAF261, R&D Systems, Minneapolis, MN) for 1 hour at 37 °C, then incubated with biotinylated horseradish peroxidase–conjugated streptavidin for 1 hour at room temperature. The reaction was visualized by addition of the substrate 3,3′,5,5′-tetramethylbenzidine, and the absor-bance at λ = 450 nm was measured by an enzyme-linked immunosorbent assay plate reader. The concentration was calculated from a standard curve of recombinant human BTC (R&D Systems, Minneapolis, MN).

Reverse transcriptase-PCR analysis of BTC and insulin mRNA. Various tis-sues were removed from STZ-induced diabetic NOD.SCID mice treated with rAd-BTC at 4 weeks after viral injection, and the expression of insulin mRNA was analyzed by conventional two-step reverse transcriptase-PCR followed by agarose gel electrophoresis. PCR was run for 35 cycles using primers specific for transgenic BTC that recognize the albumin leader sequence and the BTC-encoding region: 5′-AGTGGGTAACCTTTATTTCC-3′ and 5′-GTAAAACAAGTCAACTCTCTC-3′ and primers for mouse insulin: 5′-AGGCTTTTGTCAAGCAG-3′ and 5′-CTGATCTACAATGCCACG-3′. As a control, mouse glyceraldehyde-3-phosphate dehydrogenase mRNA was amplified for 25 cycles using primers 5′-AACGACCCCTTCATT GACCTC-3′ and 5′-CCTTGACTGTGCCGTTGAATT-3′.

In vivo treatment of diabetic mice with rAd-BTC. Six-week-old NOD.SCID male mice (Jackson Labs) were made hyperglycemic by two injec-tions of STZ given 5 days apart (100 mg/kg body weight in citrate buf-fer, pH 4.5, IP). STZ-induced diabetic mice or autoimmune diabetic NOD mice (blood glucose > 500 mg/dl for 2 consecutive days) were injected IV with rAd-BTC or rAd-βgal (2 × 1011 particles for NOD.SCID mice and 4 × 1011 particles for NOD mice) through the tail vein under methoxyflurane anesthesia. Blood glucose levels were measured on alternate days using a One-Touch Ultra Glucometer. In order to prevent immune attack of the newly generated β cells in diabetic NOD mice, CFA (100 μl/mouse, subcutaneously) was injected 3 days before the viral injection. All experiments using mice were approved by the Institutional Animal Care and Use Committee at the Rosalind Franklin University of Medicine and Science.

IP glucose tolerance tests. The mice were made to fast for 4 hours and a glu-cose solution (2 g/kg body weight) was injected IP. Blood glucose levels were measured at 0, 30, 60, 90, 120, 150, and 180 minutes after glucose injection.

Measurement of the β-cell area and mass. Quantitative evaluation of the β-cell area was performed on insulin-stained sections using the UTHSCSA Image Tool program and an Olympus BH-1 microscope connected to a Nikon 4500 camera system (magnification ×100). More than 400 serial sections (6-μm thick) were prepared from each pancreas, and every twentieth section was stained with anti-insulin antibody. The insulin-positive area of all islets, regardless of size, and interspersed insulin-positive cells was measured. The ratio of the β-cell area was calculated by dividing the area of all insulin-positive cells from at least 20 sections/mouse by the total area of the pancreas. The β-cell mass was calculated by multiplying the weight of the pancreas by the ratio of the β-cell area.

Measurement of pancreatic and serum insulin. Insulin was extracted from the pancreas as described earlier.33 At 30 minutes after glucose load-ing (2 g/kg body weight, IP), sera were collected from the mice. Pancreatic and serum insulin were measured using insulin (mouse) Ultrasensitive EIA Kit (Alpco Diagnostics, Windham, NH).

Immunohistochemical analyses. Pancreata and livers were fixed in either methacarn or 10% buffered formalin and embedded in paraffin. The tissue sections were placed in an oven (95 °C for 15 minutes, 10 mmol/l citrate, pH 6.0) for antigen retrieval, and blocked with block-ing solution (5% goat or horse serum, 1% bovine serum albumin, and

0.05% Tween-20 in phosphate-buffered saline). The tissues were then incubated with primary antibody solutions: goat anti-human BTC (R&D Systems, 1:50); guinea-pig anti-insulin (DAKO, Carpinteria, CA; 1:500); rabbit anti-glucagon, rabbit anti-somatostatin, or rabbit anti-cytokeratin (DAKO, Carpinteria, CA; 1:200); mouse anti-BrdU (DAKO, Carpinteria, CA; 1:100); or rabbit anti-pancreatic and duodenal homeobox factor 1 (Pdx-1; provided by Dr. C. Wright, Vanderbilt University, 1:500). Cy3-conjugated goat anti-guinea-pig IgG (1:200), Cy2-conjugated goat anti-rabbit IgG (1:200), or Cy3-conjugated goat anti-mouse IgG (1:200) (Jackson ImmunoRes, West Grove, PA) or horseradish peroxidase– conjugated goat anti-rabbit IgG (1:500) or horseradish peroxidase– conjugated horse anti-goat IgG (1:500) (Chemicon) were used as second-ary antibodies. Hematoxylin (Sigma, St. Louis, MO) was used as a nuclear counterstain in light microscopy. Fluorescence was imaged using a laser scanning confocal fluorescent microscope (Olympus Fluoview 300), and peroxidase staining was performed with VIP as a chromogen (purple color) (VIP kit; Vector Laboratories, Burlingame, CA).

Statistical analysis. The statistical significance of the differences between groups was analyzed by two-tailed Student’s t-test for comparisons of two groups or analysis of variance followed by Tukey’s post hoc test for multiple comparisons. A level of P < 0.05 was accepted as significant.

REFERENCES1. Rabinovitch, A (2004). Roles of cell-mediated immunity and cytokines in the

pathogenesis of type 1 diabetes mellitus. In: LeRoith, D, Olefsky, JM and Taylor, SI (eds.). Diabetes Mellitus: A Fundamental and Clinical Text, Lippincott Williams & Wilkins: Philadelphia, PA. pp. 519–538.

2. Yoon, JW and Jun, HS (1998). Insulin-dependent diabetes mellitus. In: Roitt, IM and Delves, PJ (eds.) Encyclopedia of Immunology, Academic Press: London. pp. 1390–1398.

3. Dor, Y, Brown, J, Martinez, OI and Melton, DA (2004). Adult pancreatic β-cells are formed by self-duplication rather than stem-cell differentiation. Nature 429: 41–46.

4. Trucco, M (2005). Regeneration of the pancreatic β cell. J Clin Invest 115: 5–12.5. Yoon, JW and Jun, HS (2005). Autoimmune destruction of pancreatic β cells. Am J Ther

12: 580–591.6. Bonner-Weir, S and Sharma, A (2002). Pancreatic stem cells. J Pathol 197: 519–526.7. Huotari, MA, Palgi, J and Otonkoski, T (1998). Growth factor-mediated proliferation

and differentiation of insulin-producing INS-1 and RINm5F cells: identification of betacellulin as a novel β-cell mitogen. Endocrinology 139: 1494–1499.

8. Mashima, H, Ohnishi, H, Wakabayashi, K, Mine, T, Miyagawa, J, Hanafusa, T et al. (1996). Betacellulin and activin A coordinately convert amylase-secreting pancreatic AR42J cells into insulin-secreting cells. J Clin Invest 97: 1647–1654.

9. Watada, H, Kajimoto, Y, Miyagawa, J, Hanafusa, T, Hamaguchi, K, Matsuoka, T et al. (1996). PDX-1 induces insulin and glucokinase gene expressions in αTC1 clone 6 cells in the presence of betacellulin. Diabetes 45: 1826–1831.

10. Kojima, H, Nakamura, T, Fujita, Y, Kishi, A, Fujimiya, M, Yamada, S et al. (2002). Combined expression of pancreatic duodenal homeobox 1 and islet factor 1 induces immature enterocytes to produce insulin. Diabetes 51: 1398–1408.

11. Yoshida, S, Kajimoto, Y, Yasuda, T, Watada, H, Fujitani, Y, Kosaka, H et al. (2002). PDX-1 induces differentiation of intestinal epithelioid IEC-6 into insulin-producing cells. Diabetes 51: 2505–2513.

12. Yamamoto, K, Miyagawa, J, Waguri, M, Sasada, R, Igarashi, K, Li, M et al. (2000). Recombinant human betacellulin promotes the neogenesis of β-cells and ameliorates glucose intolerance in mice with diabetes induced by selective alloxan perfusion. Diabetes 49: 2021–2027.

13. Li, L, Seno, M, Yamada, H and Kojima, I (2001). Promotion of β-cell regeneration by betacellulin in ninety percent-pancreatectomized rats. Endocrinology 142: 5379–5385.

14. Li, L, Yi, Z, Seno, M and Kojima, I (2004). Activin A and betacellulin: effect on regeneration of pancreatic β-cells in neonatal streptozotocin-treated rats. Diabetes 53: 608–615.

15. Ouziel-Yahalom, L, Zalzman, M, Anker-Kitai, L, Knoller, S, Bar, Y, Glandt, M et al. (2006). Expansion and redifferentiation of adult human pancreatic islet cells. Biochem Biophys Res Commun 341: 291–298.

16. Li, L, Seno, M, Yamada, H and Kojima, I (2003). Betacellulin improves glucose metabolism by promoting conversion of intraislet precursor cells to β-cells in streptozotocin-treated mice. Am J Physiol Endocrinol Metab 285: E577–E583.

17. Okitsu, T, Kobayashi, N, Jun, HS, Shin, S, Kim, SJ, Han, J et al. (2004). Transplantation of reversibly immortalized insulin-secreting human hepatocytes controls diabetes in pancreatectomized pigs. Diabetes 53: 105–112.

18. de la Tour, D, Halvorsen, T, Demeterco, C, Tyrberg, B, Itkin-Ansari, P, Loy, M et al. (2001). β-Cell differentiation from a human pancreatic cell line in vitro and in vivo. Mol Endocrinol 15: 476–483.

19. Narushima, M, Kobayashi, N, Okitsu, T, Tanaka, Y, Li, SA, Chen, Y et al. (2005). A human β-cell line for transplantation therapy to control type 1 diabetes. Nat Biotechnol 23: 1274–1282.

20. Assady, S, Maor, G, Amit, M, Itskovitz-Eldor, J, Skorecki, KL and Tzukerman, M (2001). Insulin production by human embryonic stem cells. Diabetes 50: 1691–1697.

21. Noguchi, H (2007). Stem cells for the treatment of diabetes. Endocr J 54: 7–16.



22. Kojima, H, Fujimiya, M, Matsumura, K, Younan, P, Imaeda, H, Maeda, M et al. (2003). NeuroD-betacellulin gene therapy induces islet neogenesis in the liver and reverses diabetes in mice. Nat Med 9: 596–603.

23. Chen, S, Ding, J, Yu, C, Yang, B, Wood, DR and Grayburn, PA (2007). Reversal of streptozotocin-induced diabetes in rats by gene therapy with betacellulin and pancreatic duodenal homeobox-1. Gene Ther 14: 1102–1110.

24. Demeterco, C, Beattie, GM, Dib, SA, Lopez, AD and Hayek, A (2000). A role for activin A and betacellulin in human fetal pancreatic cell differentiation and growth. J Clin Endocrinol Metab 85: 3892–3897.

25. Brun, T, Franklin, I, St-Onge, L, Biason-Lauber, A, Schoenle, EJ, Wollheim, CB et al. (2004). The diabetes-linked transcription factor PAX4 promotes β-cell proliferation and survival in rat and human islets. J Cell Biol 167: 1123–1135.

26. Bonner-Weir, S, Baxter, LA, Schuppin, GT and Smith, FE (1993). A second pathway for regeneration of adult exocrine and endocrine pancreas. A possible recapitulation of embryonic development. Diabetes 42: 1715–1720.

27. Bernard, C, Berthault, MF, Saulnier, C and Ktorza, A (1999). Neogenesis vs. apoptosis as main components of pancreatic β cell mass changes in glucose-infused normal and mildly diabetic adult rats. FASEB J 13: 1195–1205.

28. Tokui, Y, Kozawa, J, Yamagata, K, Zhang, J, Ohmoto, H, Tochino, Y et al. (2006). Neogenesis and proliferation of β-cells induced by human betacellulin gene transduction via retrograde pancreatic duct injection of an adenovirus vector. Biochem Biophys Res Commun 350: 987–993.

29. Lee, IF, Qin, H, Trudeau, J, Dutz, J and Tan, R (2004). Regulation of autoimmune diabetes by complete Freund’s adjuvant is mediated by NK cells. J Immunol 172: 937–942.

30. Qin, HY, Sadelain, MW, Hitchon, C, Lauzon, J and Singh, B (1993). Complete Freund’s adjuvant-induced T cells prevent the development and adoptive transfer of diabetes in nonobese diabetic mice. J Immunol 150: 2072–2080.

31. Qin, HY, Mukherjee, R, Lee-Chan, E, Ewen, C, Bleackley, RC and Singh, B (2006). A novel mechanism of regulatory T cell-mediated down-regulation of autoimmunity. Int Immunol 18: 1001–1015.

32. Becker, TC, Noel, RJ, Coats, WS, Gómez-Foix, AM, Alam, T, Gerard, RD et al. (1994). Use of recombinant adenovirus for metabolic engineering of mammalian cells. Methods Cell Biol 43 Pt A: 161–189.

33. Yoon, JW, Lesniak, MA, Fussganger, R and Notkins, AL (1976). Genetic differences in susceptibility of pancreatic β cells to virus-induced diabetes mellitus. Nature 264: 178–180.

Vitamin D and type 1 diabetes mellitus:state of the artChantal Mathieu1 and Klaus Badenhoop2

1Laboratory for Experimental Medicine and Endocrinology, Katholieke Universiteit Leuven, Leuven, B-3000, Belgium2Division of Endocrinology, Diabetes and Metabolism, Department of Medicine I, University Hospital Frankfurt am Main, Frankfurt,

D-60590, Germany

Recent evidence suggests a role for vitamin D in

pathogenesis and prevention of diabetes mellitus.

Active vitamin D, 1a,25(OH)2D3, prevents type 1 diabetes

in animal models, modifies T-cell differentiation, modu-

lates dendritic cell action and induces cytokine

secretion, shifting the balance to regulatory T cells.

High-dose vitamin D supplementation early in life

protects against type 1 diabetes. 1a,25(OH)2D3 activity

is mediated through its receptor, and targets include

transcriptional regulators; therefore, 1a,25(OH)2D3 influ-

ences gene transcription. 1a,25(OH)2D3 also affects

pancreatic b-cell function. Genomic variations of vitamin

Dmetabolism and target cell action predispose to type 1

diabetes. Vitamin D deficiency in pregnancy probably

increases the incidence of autoimmune diseases, such

as type 1 diabetes, in genetically predisposed individ-

uals. Pharmacotherapy with 1a,25(OH)2D3 analogues

might help prevent and treat diabetes.

Introduction

Vitamin D can be taken up from food (e.g. fatty fishes andtheir oils), but most people achieve their vitamin D needsthrough direct ultraviolet B (UVB)-mediated synthesis inthe skin. Two hydroxylation steps are needed to activatevitamin D, one in the liver (by the enzyme 25-hydroxylase,or CYP2R1, leading to 25-hydroxyvitamin D3-25-OHD3)and a second in the kidney, leading to the activesecosteroid hormone, 1a,25-dihydroxyvitamin D3

[1a,25(OH)2D3] [1]. This final hydroxylation (carried outby the enzyme 1-a hydroxylase or CYB27B1) can alsooccur in several cells outside the kidney, allowing theparacrine secretion of 1a,25(OH)2D3 in other tissues, suchas sites of inflammation, where activatedmacrophages areimportant sources of the active vitamin D metabolite [2](Figure 1). Vitamin D and its metabolites are transportedin the circulation by vitamin D-binding protein and thecomplex enters the cell together with megalin and cubilin,recently characterized carrier proteins [3]. Vitamin Dexerts its actions in a variety of cell types by binding to thenuclear vitamin D receptor (VDR), which shares itsstructure with many other nuclear steroid hormonereceptors, such as the glucocorticoid, thyroid hormoneand estrogen receptors.

This review presents the currently available data on apotential link between vitamin D, its metabolites andreceptor system with type 1 diabetes. The effects ofvitamin D on the immune system and on pancreaticb cells are reviewed and the implications for type 2diabetes are discussed.

Environmental sources of vitamin D

Vitamin D concentrations in the blood depend on sunexposure and alimentary intake. Levels of 1a,25(OH)2D3

are also determined by the activity of the enzymesresponsible for its final hydroxylation (CYP27B1) andcatabolism (mainly CYP24). The vitamin D status isusually assessed by measuring 25-OHD3 levels in theblood. Vitamin D serum concentrations are low or overtlydeficient in a sizeable proportion of studied populationsand decrease with age. Therefore, routine supplemen-tation is advised, particularly for those at risk of orsuffering from osteoporosis, and also in other conditions,such as pregnancy, lactation and early childhood [4].Optimal dosing is still a matter of debate, but it appearsthat the current recommendation of 200–400 U dK1

(5–10 mg dK1) is not sufficient, and doses below1000 U dK1 will not raise 25-OHD3 levels to15–80 ng mlK1 [5], currently thought to be normal. More-over, in recent years new data suggest that the physio-logical range of 25-OHD3 should be raised to even higherlevels, bringing the target zone to between 30 and80 ng mlK1 [6,7], although the official bodies have notyet accepted these new ranges. To reach these highervitamin D concentrations [5], even higher doses of vitaminD would be required (3000–5000 U dK1) [8,9].

Genetics of the vitamin D system

Vitamin D deficiency often runs in families, suggestingthat genetic variation might account for differences invitamin D concentrations; however, the genes regulatingvitamin D concentrations remain to be identified.

Genetic variation occurs in nearly all genes of thevitamin D system, but most investigations have beenperformed for the VDR. The VDR gene spans nearly100 kb on chromosome 12q12–14. The CYB27B1 gene isalso found on chromosome 12, at 12q.13.1–13.3, 10 Mbcentromeric of the VDR gene. Whereas mutations inCYP27B1 cause vitamin D-dependent rickets, polymorph-isms of the gene are associated with type 1 diabetes,

Corresponding author: Mathieu, C. ([email protected]).

DTD 5 ARTICLE IN PRESS TEM 267

Review TRENDS in Endocrinology and Metabolism Vol.xx No.xx Monthxxxx

www.sciencedirect.com 1043-2760/$ - see front matter Q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.tem.2005.06.004

http://www.sciencedirect.com

Addison’s and Graves’ disease, and Hashimoto’s thyroid-itis [10,11]. The CYP27B1 promoter (K1260) variant C ismore often transmitted to offspring with type 1 diabetes.This genotype also occurred significantly more frequentlyin case–control studies of patients with autoimmuneendocrinopathies, where promoter variant C was consist-ently associated with all four disorders [10].

Polymorphisms of the VDR: from association studies to

function

Common polymorphisms of the VDR gene have beenreported to affect the risk of breast, colon [12] and prostatecancer, in addition to bone mineral density and immunedisorders, including type 1 diabetes [13,14]. The

association of VDR variants with protection from coloncancer might be related to the recent finding that the VDRacts as an intestinal sensor for toxic bile acids: the VDRhas a greater sensitivity for binding lithocholic acid thanother nuclear receptors [15].

Steroid hormone nuclear receptors are polymorphic:glucocorticoid receptors affect the neuroendocrine controlof glucocorticoid function [16], the androgen receptor isinvolved in testosterone action [17,18] and hormonereplacement therapy (via the estrogen receptor) affectslipid metabolism and E-selectin [19]. In addition to theireffects on osteoporosis,VDRpolymorphisms correlatedwithmuscle strength, fat mass and body weight in healthypremenopausal Swedish women [20] in a population-based

TRENDS in Endocrinology & Metabolism

7-DHCpre-vitamin D3

DBPDBP

DBP

DBP

vitamin D3

CYP2R1CYP27B1

Blood circulation

Skin

Intestine

Targettissue

Liver

Macrophage

Kidney

25(OH)D3

1α,25(OH)2D3

CH2

CH2

CH2

HO

HO

HO

HO

OH

OH

OH

OH

1α,25(OH)2D3

UVB290–315 nm

CH2

OH

CH2

HO

CH3

HO

HO

Figure 1. The complex vitamin D system including the regulation of vitamin D serum concentrations, transcellular transport and intracellular metabolism, in addition to its

target action. Vitamin D is taken up from food (vitamin D2 and D3) or synthesized in the skin [sun light with the photon energy wavelength of 290–315 nm causes photolysis of

7-dehydrocholesterol (7-DHC; provitamin D3) to pre-vitamin D3, the immediate precursor in the biosynthetic pathway to vitamin D3]. In the circulation, all vitamin D

metabolites are bound to vitamin D-binding protein (DBP). To become active, vitamin D3 is hydroxylated first in the liver (25-hydroxylases) and successively in the kidney

(1a-hydroxylase) to its final form, 1a,25(OH)2D3. Note that macrophages and other cells are also capable of hydroxylating vitamin D to 1a,25(OH)2D3, thereby facilitating

locally regulated tissue concentrations.

Review TRENDS in Endocrinology and Metabolism Vol.xx No.xx Monthxxxx2

www.sciencedirect.com



study. A similar correlation with muscle strength hadpreviously been shown for non-obese elderly women [21].One VDR variant has been characterized for its transcrip-tional activity [22]: the polymorphic FokI site in the VDRaffects its trans-activating potential, with the shorter FokIvariant of 424 amino acids (absence of the FokI restrictionsite ‘F’ abolishes the start codon, and translation startsfurther downstream) displaying a stronger degree oftrans-activation than the less active ‘f ’ variant with 427amino acids. This results in a difference in the modulationof the transcription factor IIB. It is conceivable that,depending on the target tissue, the differences in localconcentrations of coactivators and abundance of localpromoters could lead to augmented or attenuated trans-activation. In addition, the vitamin D-mediated responsedepends not only on VDR structure, with its polymorphicN-terminus, but also on its DNA binding and ligandaffinity. VDR functions as a heterodimer with the retinoicX receptor, which undergoes conformational changes afterboth DNA and ligand interaction. This process ultimatelydetermines the diversified response to 1a,25(OH)2D3.VDR polymorphisms affect the function of peripheralblood mononuclear cells. Phytohaemagglutinin-stimu-lated growth inhibition by 1a,25(OH)2D3 was significantlydifferent depending on the FokI status (F: absence of therestriction enzyme site; f: presence): FF homozygotes hadthe lowest ED50 (i.e. the strongest growth inhibition ofstimulated mononuclear cells) [23]. This would affectmacrophage function in the presence of vitamin Ddeficiency.

VDR polymorphisms are associated with type 1diabetes mellitus in Caucasians – with two independentstudies from Germany showing an association both in acase–control study and in family-based probands [13,24],in Bangladeshi Indians [14], and in Japanese [25],although such an association was not seen in the Finnishpopulation [26] and in a combined large scale analysisfrom the UK, Romania and Finland [27]. Whereas initialstudies used only the polymorphic FokI, BsmI, ApaI andTaqI variants, recently more polymorphisms were ident-ified with the use of a high resolution single nucleotidepolymorphism (SNP) map [28]. Haplotypes rather thanindividual SNPs have recently been associated withasthma in a Quebec-based family study [29].

Vitamin D and the physiology of b-cell secretion and

insulin action

The VDR can be viewed as a master regulator oftranscription. VDRs are present in pancreatic b-cells andvitamin D is essential for normal insulin secretion [30].Islet cell insulin secretion is reduced in vitaminD-deficient animals and can be corrected by vitamin Dsupplementation [30–32]. Interestingly, an animal modelwith a mutated VDR has been reported to have impairedinsulin secretion [33]. These mice with functionallyinactive VDRs have a severely disrupted vitamin Dsignalling system. They show greatly impaired oralglucose tolerance, expression of the gene encoding insulinand insulin secretion. However, in this model, the back-ground of the mice appears to be crucial, becauseVdr-knockout mice with a different genetic background

had normal b-cell function [34]. When NOD mice, ananimal model for human type 1 diabetes, are renderedvitamin D deficient in early life, impaired glucosetolerance is seen by 100 days of age, with a doubling ofdiabetes incidence at 200 days [35].

The impact of vitamin D deficiency on b-cell functionseen in vitro and in vivo in animal models has beenmatched by vitamin D studies in human volunteersundergoing hyperglycaemic clamps [36]. In this study ofindividuals with different racial origins, 26% of theCaucasians and 54% of the African Americans werevitamin D deficient [defined as 25-OHD3 levels!20 ng mlK1]. There was a significant negative corre-lation between plasma glucose values at 60, 90 and120 min and serum 25-OHD3 levels in those who under-went an oral glucose challenge. Furthermore, b-cellfunction, as measured by the first and second insulinresponses, also correlated negatively with 25-OHD3 levels.This correlation between glucose-induced insulinsecretion and vitamin D status was also seen in a studyof men aged 70–88, 39% of whomwere vitamin D depleted.Their 1-h glucose values and area under a glucose curvenegatively correlated with vitamin D concentrations [37].Therefore, vitamin D insufficiency might contribute to therelative insulin deficiency seen in type 2 diabetes. VitaminD insufficiency has also been associated with an increasedrisk for type 2 diabetes, with lower 25-OHD3 concen-trations in patients with type 2 diabetes than in controls[38]. In a large cross-sectional survey of Americansperformed between 1988 and 1994 there was an inverseassociation between vitamin D status and diabetes in non-Hispanic whites and in Mexican Americans [39].

Vitamin D and vitamin D analogs and their effects on the

immune system in type 1 diabetes

Type 1 diabetes is an autoimmune disease, and the selfimmune system plays a central role in the destruction ofthe b cell. The detection of the VDR in almost all cells ofthe immune system, especially antigen-presenting cells(macrophages and dendritic cells) and activated T cells,led to the investigation of a potential role for1a,25(OH)2D3 as an immunomodulator [40,41]. Not onlyare VDRs present in the immune system, but immunecells themselves, in particular activated macrophages anddendritic cells, are able to synthesize and secrete1a,25(OH)2D3 [42]. These cells possess the tools necessaryfor the final activating step in the synthesis of1a,25(OH)2D3 (the enzyme 1a-hydroxylase). The effectsof 1a,25(OH)2D3 on the immune system are multiple, butall lead to the generation of tolerance and anergy ratherthan immune activation [40]. In the presence of1a,25(OH)2D3, dendritic cells mature in the direction ofa tolerogenic cell [43], with less expression of majorhistocompatibility complex (MHC) class II molecules andadhesion molecules necessary for full T-cell stimulation[44]. In addition, cytokines crucial to the recruitment andactivation of T cells are suppressed by1a,25(OH)2D3. Themajor cytokine driving the immune system towardsT helper type 1 (Th1) development, interleukin 12(IL-12), is almost completely inhibited in the presence of1a,25(OH)2D3 through interference with the nuclear

Review TRENDS in Endocrinology and Metabolism Vol.xx No.xx Monthxxxx 3




factor kB pathway [45]. Several T-cell cytokines are alsodirect targets for 1a,25(OH)2D3, leading to inhibition ofTh1 cytokines, such as IL-2 and interferon g (IFN-g), andstimulation of Th2 cytokines, such as IL-4 [46–48](Figure 2).

Chronic administration of pharmacological doses of1a,25(OH) reduces the incidence of both insulitis anddiabetes in NOD mice [49,50], and reduces diabetes in amodel of multiple low dose streptozotocin diabetes [51].The basis for this protection appears to be mainly arestoration of suppressor cell function in NOD mice [52].

A major obstacle to human application of 1a,25(OH)2D3

is its possible side effect of hypercalcaemia as a result ofincreased Ca2C resorption from intestine and bone.However, in view of the therapeutic potential, structuralanalogues have been designed and synthesized, with theaim of achieving a dissociation between calcaemic andimmune effects. Several of the most promising analogueshave been successfully tested in the NOD mouse [53]. Themechanism of protection against insulitis and diabetesappears to be similar to that of 1a,25(OH)2D3. Effects ofthe analogues on dendritic cell phenotype, regulatory cellinduction and b-cell protection have been described[40,54]. The strongest protection against b-cell destructionby 1a,25(OH)2D3 or its analogues can be seen in situations

of chronic administration in primary prevention, withtreatment being initiated before autoimmune destructionhas started.When administered inNODmice in situationssimilar to the human situation, where people at high riskof type 1 diabetes are identified on the basis of alreadyhaving circulating autoantibodies against the b cell,1a,25(OH)2D3 or its analogues can only achieve diseaseprevention when combined with a short course of an anti-T-cell immunosuppressant (e.g. cyclosporine A) [55]. Thisobservation again confirms the in vitro observations thatthe main target cell for 1a,25(OH)2D3 action is theantigen-presenting dendritic cell, whereas its anti-T-celleffects are relatively weak.

In humans, several epidemiological studies provideevidence that vitamin D intake can prevent type 1diabetes (Box 1).

Conclusion and outlook

Vitamin D, the essential vitamin in Ca2C and bonemetabolism, has beneficial effects on b-cell function andnormal immunity. Its activated form, 1a,25(OH)2D3,prevents diabetes in NOD mice. Vitamin D insufficiencyis a risk factor for autoimmune disease and otherdisorders. Although optimal supplement dosing withregard to immune and b-cell function is not known,

TRENDS in Endocrinology & Metabolism

DC

MCH II

Ag

IL-12

Th1

Th1

Th1

Th1

Th1

IFN-γ

IL-2

Th2

Th2

Th2

IL-4IL-5IL-10

CD80/86

CD28/CTLA4Treg IL-10

TGF-β

Tc

Mf

IL-1TNF-αFree radicals

β cell

Th

TCR

CD40

CD40L

1α,25(OH)2D3

Figure 2. The immunomodulatory effects of 1a,25(OH)2D3. At the level of the antigen-presenting cell (such as dendritic cells; DCs), 1a,25(OH)2D3 inhibits the surface

expression of MHC class II-complexed antigen and of co-stimulatory molecules, in addition to production of the cytokine IL-12, thereby indirectly shifting the polarization of

T cells from a Th1 towards a Th2 phenotype. In addition, 1a,25(OH)2D3 has immunomodulatory effects directly at the level of the T cell, by inhibiting the production of the Th1

cytokines IL-2 and IFN-g and stimulating the production of Th2 cytokines. Moreover, 1a,25(OH)2D3 favours the induction of regulatory T cells. Both Th2 and Tregs can inhibit

Th1 cells through the production of counteracting or inhibitory cytokines. Together, these immunomodulatory effects of 1a,25(OH)2D3 can lead to the protection of target

tissues, such as b cells, in autoimmune diseases and transplantation. CD40L, CD40 ligand; Mf, macrophage; Tc, cytotoxic T cell; TGF-b, transforming growth factor b; Th1,

T helper type 1; TNF-a, tumour necrosis factor a; Treg, regulatory T cell.





substantially higher doses than those currently rec-ommended, possibly as high as 50 mg (2000 U) dK, mightbe required. A major clinical lesson that can be drawn atthis moment is that avoidance of vitamin D deficiency isessential for b-cell function and might contribute toprotection against type 1 diabetes in later life. Epidemio-logical studies have shown that vitamin D deficiencyshould be avoided in pregnancy, not only because of itseffects on bone development, but also because it mightincrease the incidence of autoimmune diseases, such astype 1 diabetes in genetically at-risk individuals. Exploit-ing the immunomodulatory effects of 1a,25(OH)2D3 inhumans will require the development of safe structuralanalogues with a dissociation between calcaemic andimmune effects. Novel analogues have been developedthat are more potent in T cell and dendritic cellmodulation and less calcaemic, thus allowing higherdoses to target the immune system. These analogues arecurrently being analysed for their therapeutic potential.

Acknowledgements

C.M. has a grant from the Flemish research council (FWO) and K.B. has agrant from the European Foundation for the study of Diabetes (EFSD).We thank Evelyne van Etten, Conny Gysemans and Heinrich Kahles forfigure preparation.

References

1 Holick, M.F. (2003) Vitamin D: a millenium perspective. J. Cell.Biochem. 88, 296–307

2 Overbergh, L. et al. (2000) Identification and immune regulation of 25-hydroxyvitamin D–1a-hydroxylase in murinemacrophages.Clin. Exp.Immunol. 120, 139–146

3 Christensen, E.I. et al. (2003) Loss of chloride channel ClC-5 impairsendocytosis by defective trafficking of megalin and cubilin in kidneyproximal tubules. Proc. Natl. Acad. Sci. U. S. A. 100, 8472–8477

4 Weisberg, P. et al. (2004) Nutritional rickets among children in theUnited States: review of cases reported between 1986 and 2003. Am.J. Clin. Nutr. 80, 1697S–1705S

5 Hollis, B.W. and Wagner, C.L. (2004) Assessment of dietary vitamin Drequirements during pregnancy and lactation. Am. J. Clin. Nutr. 79,717–726

6 Hollis, B.W. (2005) Circulating 25-hydroxyvitamin D levels indicativeof vitamin D sufficiency: implications for establishing a new effectivedietary intake recommendation for vitamin D. J. Nutr. 135, 317–322

7 Hollis, B.W. and Wagner, C.L. (2005) Normal serum vitamin D levels.N. Engl. J. Med. 352, 515–516

8 Vieth, R. (1999) Vitamin D supplementation, 25-hydroxyvitamin Dconcentrations, and safety. Am. J. Clin. Nutr. 69, 842–856

9 Heaney, R.P. et al. (2003) Human serum 25-hydroxycholecalciferolresponse to extended oral dosing with cholecalciferol. Am. J. Clin.Nutr. 77, 204–210

10 Lopez, E.R. et al. (2004) A promoter polymorphism of the CYP27B1gene is associated with Addison’s disease, Hashimoto’s thyroiditis,Graves’ disease and type 1 diabetes mellitus in Germans. Eur. J.Endocrinol. 151, 193–197

11 Lopez, E.R. et al. (2004) CYP27B1 polymorphism variants areassociated with type 1 diabetes mellitus in Germans. J. SteroidBiochem. Mol. Biol. 89–90, 155–157

12 Slattery, M.L. et al. (2004) Dietary calcium, vitamin D, VDR genotypesand colorectal cancer. Int. J. Cancer 111, 750–756

13 Pani, M.A. et al. (2000) Vitamin D receptor allele combinationsinfluence genetic susceptibility to IDDM in Germans. Diabetes 49,504–507

14 McDermott, M.F. et al. (1997) Allelic variation in the vitamin Dreceptor influences susceptibility to IDDM in Indian Asians.Diabetologia 40, 971–975

15 Makishima, M. et al. (2002) Vitamin D receptor as an intestinal bileacid sensor. Science 296, 1313–1316

16 Wust, S. et al. (2004) Common polymorphisms in the glucocorticoidreceptor gene are associated with adrenocortical responses topsychosocial stress. J. Clin. Endocrinol. Metab. 89, 565–573

17 Zitzmann, M. et al. (2004) X-chromosome inactivation patterns andandrogen receptor functionality influence phenotype and socialcharacteristics as well as pharmacogenetics of testosterone therapyin klinefelter patients. J. Clin. Endocrinol. Metab. 89, 6208–6217

18 Zitzmann, M. et al. (2003) Prostate volume and growth in testoster-one-substituted hypogonadal men are dependent on the CAG repeatpolymorphism of the androgen receptor gene: a longitudinal pharma-cogenetic study. J. Clin. Endocrinol. Metab. 88, 2049–2054

19 Herrington, D.M. et al. (2002) Common estrogen receptor polymorph-ism augments effects of hormone replacement therapy on E-selectinbut not C-reactive protein. Circulation 105, 1879–1882

20 Grundberg, E. et al. (2004) Genetic variation in the human vitamin Dreceptor is associated with muscle strength, fat mass and body weightin Swedish women. Eur. J. Endocrinol. 150, 323–328

21 Vandevyver, C. et al. (1997) Influence of the vitamin D receptor genealleles on bone mineral density in postmenopausal and osteoporoticwomen. J. Bone Miner. Res. 12, 241–247

22 Jurutka, P.W. et al. (2000) The polymorphic N terminus in humanvitamin D receptor isoforms influences transcriptional activity bymodulating interaction with transcription factor IIB.Mol. Endocrinol.14, 401–420

23 Colin, E.M. et al. (2000) Consequences of vitamin D receptor genepolymorphisms for growth inhibition of cultured human peripheralblood mononuclear cells by 1, 25-dihydroxyvitamin D3. Clin.Endocrinol. (Oxf.) 52, 211–216

24 Fassbender, W.J. et al. (2002) VDR gene polymorphisms are over-represented in German patients with type 1 diabetes compared tohealthy controls without effect on biochemical parameters of bonemetabolism. Horm. Metab. Res. 34, 330–337

25 Ban, Y. et al. (2001) Vitamin D receptor initiation codon polymorphisminfluences genetic susceptibility to type 1 diabetes mellitus in theJapanese population. BMC Med. Genet. 2, 7

26 Turpeinen, H. et al. (2003) Vitamin D receptor polymorphisms: noassociation with type 1 diabetes in the Finnish population. Eur. J.Endocrinol. 149, 591–596

27 Nejentsev, S. et al. (2004) Analysis of the vitamin D receptor genesequence variants in type 1 diabetes. Diabetes 53, 2709–2712

28 Nejentsev, S. et al. (2004) Comparative high-resolution analysis oflinkage disequilibrium and tag single nucleotide polymorphismsbetween populations in the vitamin D receptor gene. Hum. Mol.Genet. 13, 1633–1639

29 Poon, A.H. et al. (2004) Association of vitamin D receptor geneticvariants with susceptibility to asthma and atopy. Am. J. Respir. Crit.Care Med. 170, 967–973

30 Norman, A.W. et al. (1980) Vitamin D deficiency inhibits pancreaticsecretion of insulin. Science 209, 823–825

Box 1. Vitamin D intake and type 1 diabetes prevention

The intake of vitamin D, either as a supplement or via food, has been

the subject of recent studies examining populations with a high risk

for type 1 diabetes. Hypponen et al. found a significantly reduced risk

of 0.22 for type 1 diabetes in a birth-cohort study when high-dose

vitamin D supplementation (O50 mg dK1, 2000 U dK1) was given

regularly or irregularly [56]. By contrast, those children with

suspected rickets during the first year of life had a threefold

increased risk of developing type 1 diabetes during later life.

Similarly, increased vitamin D intake during pregnancy significantly

reduced b-cell autoimmunity in offspring as detected by islet

autoantibodies [57]. However, this effect was restricted to vitamin

D intake from food. In a Norwegian study, the use of cod liver oil

either during pregnancy or in the first year of life was associated with

a lower incidence of type 1 diabetes [58]. Whether this was the result

of the content of vitamin D or long-chain n-3 fatty acids (or a

combination) merits further investigation [59]. Furthermore, a

EURODIAB (European Community Concerted Action Programme in

Diabetes) subgroup multicentre study of cases and controls found

that the risk for type 1 diabetes was significantly reduced in countries

with vitamin D supplementation during childhood [60].

Review TRENDS in Endocrinology and Metabolism Vol.xx No.xx Monthxxxx 5




31 Chertow, B.S. et al. (1983) Cellular mechanisms of insulin release: theeffects of vitamin D deficiency and repletion on rat insulin secretion.Endocrinology 113, 1511–1518

32 Nyomba, B.L. et al. (1986) Pancreatic secretion in man withsubclinical vitamin D deficiency. Diabetologia 29, 34–38

33 Zeitz, U. et al. (2003) Impaired insulin secretory capacity in micelacking a functional vitamin D receptor. FASEB J. 17, 509–511

34 Mathieu, C. et al. (2001) In vitro and in vivo analysis of the immunesystem of vitamin D receptor knockout mice. J. Bone Miner. Res. 16,2057–2065

35 Giulietti, A. et al. (2004) Vitamin D deficiency in early life acceleratestype 1 diabetes in non-obese diabetic mice. Diabetologia 47, 451–462

36 Chiu, K.C. et al. (2004) Hypovitaminosis D is associated with insulinresistance and b cell dysfunction. Am. J. Clin. Nutr. 79, 820–825

37 Baynes, K.C. et al. (1997) Vitamin D, glucose tolerance andinsulinaemia in elderly men. Diabetologia 40, 344–347

38 Isaia, G. et al. (2001) High prevalence of hypovitaminosis D in femaletype 2 diabetic population. Diabetes Care 24, 1496

39 Scragg, R. et al. (2004) Serum 25-hydroxyvitamin D, diabetes, andethnicity in the Third National Health and Nutrition ExaminationSurvey. Diabetes Care 27, 2813–2818

40 Mathieu, C. and Adorini, L. (2002) The coming of age of 1,25-dihydroxyvitamin D(3) analogs as immunomodulatory agents. TrendsMol. Med. 8, 174–179

41 Mathieu, C. et al. (2004) Vitamin D and 1,25-dihydroxyvitamin D3 asmodulators in the immune system. J. Steroid Biochem. Mol. Biol.89–90, 449–452

42 Hewison, M. et al. (2003) Differential regulation of vitamin D receptorand its ligand in humanmonocyte-derived dendritic cells. J. Immunol.170, 5382–5390

43 vanHalteren, A.G. et al. (2004) 1a,25-DihydroxyvitaminD3 or analoguetreated dendritic cells modulate human autoreactive T cells via theselective induction of apoptosis. J. Autoimmun. 23, 233–239

44 van Halteren, A.G. et al. (2002) Redirection of human autoreactiveT-cells upon interaction with dendritic cells modulated by TX527, ananalog of 1,25 dihydroxyvitamin D(3). Diabetes 51, 2119–2125

45 D’Ambrosio, D. et al. (1998) Inhibition of IL-12 production by 1,25-dihydroxyvitamin D3. Involvement of NF-kB downregulation intranscriptional repression of the p40 gene. J. Clin. Invest. 101, 252–262

46 Takeuchi, A. et al. (1998) Nuclear factor of activated T cells (NFAT) asa molecular target for 1a,25-dihydroxyvitamin D3-mediated effects.J. Immunol. 160, 209–218

47 Overbergh, L. et al. (2000) 1a,25-Dihydroxyvitamin D3 induces anautoantigen-specific T-helper 1/T-helper 2 immune shift in NOD miceimmunized with GAD65 (p524-543). Diabetes 49, 1301–1307

48 Boonstra, A. et al. (2001) 1a,25-Dihydroxyvitamin D3 has a directeffect on naive CD4(C) Tcells to enhance the development of Th2 cells.J. Immunol. 167, 4974–4980

49 Mathieu, C. et al. (1992) 1,25-Dihydroxyvitamin D3 prevents insulitisin NOD mice. Diabetes 41, 1491–1495

50 Mathieu, C. et al. (1994) Prevention of autoimmune diabetes in NODmice by 1,25 dihydroxyvitamin D3. Diabetologia 37, 552–558

51 Inaba, M. et al. (1992) Partial protection of 1 a-hydroxyvitamin D3against the development of diabetes induced by multiple low-dosestreptozotocin injection in CD-1 mice. Metabolism 41, 631–635

52 Gregori, S. et al. (2002) A 1a,25-dihydroxyvitamin D(3) analogenhances regulatory T-cells and arrests autoimmune diabetes inNOD mice. Diabetes 51, 1367–1374

53 Mathieu, C. et al. (1995) Prevention of type I diabetes in NODmice bynonhypercalcemic doses of a new structural analog of 1,25-dihydrox-yvitamin D3, KH1060. Endocrinology 136, 866–872

54 Gysemans, C.A. et al. (2005) 1,25-dihydroxyvitamin D3 modulatesexpression of chemokines and cytokines in pancreatic islets: impli-cations for prevention of diabetes in NOD mice. Endocrinology 146,1956–1964

55 Casteels, K.M. et al. (1998) Prevention of type I diabetes in nonobesediabetic mice by late intervention with nonhypercalcemic analogs of1,25-dihydroxyvitamin D3 in combination with a short inductioncourse of cyclosporin A. Endocrinology 139, 95–102

56 Hypponen, E. et al. (2001) Intake of vitamin D and risk of type 1diabetes: a birth-cohort study. Lancet 358, 1500–1503

57 Fronczak, C.M. et al. (2003) In utero dietary exposures and risk of isletautoimmunity in children. Diabetes Care 26, 3237–3242

58 Stene, L.C. and Joner, G. (2003) Use of cod liver oil during the firstyear of life is associated with lower risk of childhood-onset type 1diabetes: a large, population-based, case–control study. Am. J. Clin.Nutr. 78, 1128–1134

59 Harris, S.S. (2004) Vitamin D and type 1 diabetes. Am. J. Clin. Nutr.79, 889–890

60 The EURODIAB substudy 2 study group. (1999) Vitamin Dsupplement in early childhood and risk for type I (insulin-dependent)diabetes mellitus. The EURODIAB Substudy 2 Study Group.Diabetologia 42, 51–54





collected articles on type 1 diabetes.pdf

Documents