loss of cd28 expression by liver-infiltrating t cells contributes to pathogenesis of primary...

Gastroenterology 2014;147:221–232

Loss of CD28 Expression by Liver-Infiltrating T Cells Contributes toPathogenesis of Primary Sclerosing CholangitisEvaggelia Liaskou,1 Louisa E. Jeffery,2 Palak J. Trivedi,1 Gary M. Reynolds,1 Shankar Suresh,1

Tony Bruns,1,3,4 David H. Adams,1 David M. Sansom,2,5 and Gideon M. Hirschfield1

1Centre for Liver Research and National Institute for Health Research Biomedical Research Unit in Liver Disease, 2MedicalResearch Council Centre for Immune Regulation, School of Immunity and Infection, Institute of Biomedical Research,University of Birmingham, Birmingham, United Kingdom; 3Department of Internal Medicine IV, Gastroenterology, Hepatologyand Infectious Disease, 4Center for Sepsis Control and Care, Jena University Hospital, Friedrich Schiller University of Jena,Jena, Germany; 5Institute for Immunity and Transplantation, University College London, Royal Free Campus, London, UnitedKingdom

Abbreviations used in this paper: BEC, biliary epithelial cell; CCR, C-C

BASICAN

DTR

ANSLAT

IONA

LLIVE

R

BACKGROUND & AIMS: T-cell–mediated biliary injury is afeature of primary sclerosing cholangitis (PSC). We studied theroles of CD28- T cells in PSC and their regulation by vitamin D.METHODS: Peripheral and liver-infiltrating mononuclear cellswere isolated from blood or fresh liver tissue. We analyzednumbers, phenotypes, functions, and localization patterns ofCD28- T cells, along with their ability to activate biliaryepithelial cells. We measured levels of tumor necrosis factor(TNF)a in liver tissues from patients with PSC and the effects ofexposure to active vitamin D (1,25[OH]2D3) on expression ofCD28. RESULTS: A significantly greater proportion of CD4þ andCD8þ T cells that infiltrated liver tissues of patients with PSCwere CD28-, compared with control liver tissue (CD4þ: 30.3%vs 2.5%, P < .0001; and CD8þ: 68.5% vs 31.9%, P < .05). Themean percentage of CD4þCD28- T cells in liver tissues frompatients with PSC was significantly higher than from patientswith primary biliary cirrhosis or nonalcoholic steatohepatitis(P < .05). CD28- T cells were activated CD69þCD45RA- C-Cchemokine receptor (CCR)7- effector memory and perforinþ

granzyme Bþ cytotoxic cells, which express CD11a, CX3CR1,C-X3-C motif receptor 6 (CXCR6), and CCR10—consistent withtheir infiltration of liver and localization around bile ducts.Compared with CD28þ T cells, activated CD28- T cells producedsignificantly higher levels of interferon g and TNFa (P < .05),and induced up-regulation of intercellular cell adhesionmolecule-1, HLA-DR, and CD40 by primary epithelial cells(3.6-fold, 1.5-fold, and 1.2-fold, respectively). Liver tissue frompatients with PSC contained high levels of TNFa; TNFa down-regulated the expression of CD28 by T cells in vitro(P < .01); this effect was prevented by administration of1,25(OH)2D3 (P < .05). CONCLUSIONS: Inflammatory CD28-

T cells accumulate in livers of patients with PSC and localizearound bile ducts. The TNFa-rich microenvironment of thistissue promotes inflammation; these effects are reversed byvitamin D in vitro.

chemokine receptor; CXCR, C-X3-C motif receptor; DC, disease control;ICAM1, intercellular cell adhesion molecule-1; IFNg, interferon g; IL,interleukin; LIMC, liver infiltrating mononuclear cell; NASH, nonalcoholicsteatohepatitis; PBC, primary biliary cirrhosis; PBMC, peripheral bloodmononuclear cell; PSC, primary sclerosing cholangitis; rTNFa, recombi-nant tumor necrosis factor a; TIM3, T-cell immunoglobulin domain andmucin domain 3; TNFa, tumor necrosis factor a; Treg, T regulatory cell.

© 2014 by the AGA Institute0016-5085/$36.00

Keywords: Interferon; Autoimmunity; Biliary Epithelial Cells;Immune Regulation.

rimary sclerosing cholangitis (PSC) is a poorly un-
http://dx.doi.org/10.1053/j.gastro.2014.04.003 Pderstood chronic immune-mediated biliary disease
lacking effective treatment, as well as validated animalmodels. Despite features supporting a classic autoimmuneetiology standard, immunosuppression is ineffective.1 As aresult of progress in genetic studies, several risk loci havebeen identified for PSC; numerous loci for immunologicallyrelevant proteins (eg, HLA, CD28, interleukin [IL]2RA, andmacrophage stimulating 1).2 These findings parallel histo-logic observations that include a mixed inflammatory cellinfiltrate consisting of lymphocytes, plasma cells, neutro-phils, natural killer cells, Kupffer cells, and perisinusoidalmacrophages.3 The majority of mononuclear cells in theportal infiltrates are T lymphocytes4 that produce highlevels of tumor necrosis factor a (TNFa), supporting PSC asa predominantly T-helper 1-mediated disease.5

The cell surface molecule CD28 is a co-stimulatorymolecule necessary for T-cell activation, survival, and pro-liferation,6 and the CD28 locus is a newly recognized riskfactor in PSC development. Prior work identified T cellslacking CD28 accumulating in various autoimmune dis-eases,7,8 and suggested that loss of CD28 occurs at chronicinflammatory sites as a consequence of continuous antigenicstimulation and TNFa exposure.9 CD28- T cells appear to bechronically activated immunopathogenic cells,10 less sus-ceptible to regulation by CD4þCD25þ T regulatory cells(Tregs), thus making them potentially important drivers ofpersistent chronic inflammation.11 Although immunogeneticprofiles underpin the risk of PSC, environmental factors areequally relevant. Vitamin D is an extrinsic factor repeatedlyassociated with autoimmunity, as well as cholestatic liverdiseases.12,13 The local activation of vitamin D by immunecells suppresses the development of proinflammatory

http://dx.doi.org/10.1053/j.gastro.2014.04.003

222 Liaskou et al Gastroenterology Vol. 147, No. 1

BASICAND

TRANSLATIONALLIVER

effector T cells while increasing the frequency and sup-pressive function of Tregs.14

We sought to elucidate the mechanisms of biliary injuryin PSC using patient-derived samples, to study T-cell infil-tration and CD28 expression, alongside intervention withvitamin D. Our data show expansion of CD28- T cells with anactivated phenotype in human PSC liver, localization close tobile ducts, release of proinflammatory cytokines, and in-duction of epithelial cell activation. A TNFa-rich PSCmicroenvironment was evident and TNFa down-regulatedT-cell CD28 expression in vitro, an effect prevented byvitamin D.

Materials and MethodsHuman Tissue and Blood

Fresh diseased liver tissue from our transplant programwas available, as was nondiseased liver from surgical re-sections. Whole blood was obtained from healthy volunteersand PSC patients. All samples were collected after localresearch ethics committee approval and patient consent.

Isolation of peripheral blood and liver-infiltratingmononuclear cells. Peripheral blood mononuclear cells(PBMCs) and liver-infiltrating mononuclear cells (LIMCs) wereisolated from peripheral blood and fresh human liver, respec-tively,15 as described in the Supplementary Materials andMethods section.

Isolation and culture of human primary biliaryepithelial cells. Human biliary epithelial cells (BECs) wereisolated from liver tissue, and cultured as previously reported16

and further described in the Supplementary Materials andMethods section.

Flow CytometryFlow cytometric analysis was performed on blood and liver-

infiltrating T cells using a Cyan flow cytometer (BeckmanCoulter, Bucks, United Kingdom), and analyzed using FlowJo(version 9; Treestar Inc, Ashland, OR) (see the SupplementaryMaterials and Methods section and Supplementary Table 1).

Isolation of CD28þ/- T-Cell Subsets and Effect ofT-Cell–Conditioned Media on BECs

PBMCs from PSC patients were isolated and stained forCD3–fluorescein isothiocyanate, CD4-allophycocyanin, andCD28-phycoerythrin markers to allow isolation ofCD3þCD4þ/-CD28þ/- subsets by fluorescence-activated cellsorting. Isolated cells were activated overnight and their cell-free conditioned media was used to culture primary BECs for4 days (see the Supplementary Materials and Methods sectionand Supplementary Table 1).

Detection of Cytokine Expression Ex VivoPBMCs and LIMCs from PSC patients were isolated as

described earlier, and stimulated in 96-well, round-bottomedtissue culture plates for 6 hours with plate-bound anti-CD3(OKT3; 5 mg/mL) at 106 cells/well (see the SupplementaryMaterials and Methods section and Supplementary Table 1).

Long-term In Vitro Culture of CD4þ T CellsCD4þ T cells were enriched to 94%–98% purity from PSC

PBMCs using a negative selection antibody cocktail (StemCellTechnologies, Manchester, United Kingdom) and stimulatedwith aCD3/aCD28 Dynabeads (Life Technologies, Paisley,United Kingdom), IL2 (50 U/mL; Peprotech, London, UnitedKingdom), in the presence/absence of TNFa (10 ng/mL;Peprotech), with or without 1,25(OH)2D3 (10 nmol/L; ENZOLife Sciences, Exeter, United Kingdom). At 4 days, beads wereremoved using a magnet and cells were split to 0.5 � 106/mL.Cultures were assessed every 3–4 days and maintained at0.5–1 � 106/mL. Cytokines and 1,25(OH)2D3 also were re-supplemented at these times, the concentration of IL2 beingincreased to 100 U/mL after 2 weeks.

Immunohistochemical AnalysisDual-color immunohistochemistry was used to localize

CD4þ and CD8þ T-cell subsets, and lymphocytes that lackedCD28 expression in human liver tissue (see the SupplementaryMaterials and Methods section and Supplementary Tables 1and 2).

RNA AnalysisTotal RNA was extracted from normal, PSC, and primary

biliary cirrhosis (PBC) liver samples, using the RNeasy mini Kit(Qiagen, Manchester, United Kingdom), and analyzed asdescribed in the Supplementary Materials and Methods section,for expression of TNFa, interferon g (IFNg), and IL17A byquantitative polymerase chain reaction analysis.

Cytokine Secretion AssaysThe secretion of cytokines and chemokines from PSC LIMCs

was investigated using the cytokine array panel A kit (R&DSystems, Abingdon, United Kingdom) (see the SupplementaryMaterials and Methods section).

Vitamin D Supplementation in PSC PatientsPatients were classified as 25(OH)D sufficient (>50 nmol/L),

25(OH)D insufficient (30–49 nmol/L), or severely 25(OH)Ddeficient (<30 nmol/L). Routine supplementation is advised forindividuals with insufficient vitamin D levels (400 U/day cole-calciferol). Those with severe deficiency are treated withhigh-dose replacement therapy (ergocalciferol capsules 50,000U/wk for 5 weeks). Serum 25(OH)D levels were correlated withCD28- T-cell frequency in peripheral blood, before and aftervitamin D therapy.

Statistical AnalysesStatistical analyses were performed using GraphPad Prism

(GraphPad Software, Inc, La Jolla, CA) and SPSS software (SPSS,Inc, Chicago, IL). Data not normally distributed were evaluatedusing an unpaired Mann–Whitney, Wilcoxon matched-pairssigned rank, and Kruskal–Wallis tests. For correlations ofCD28- T-cell frequency and clinical characteristics, Kolmogor-ov–Smirnov testing and the Spearman rank correlation coeffi-cient were used. P values less than .05 were consideredsignificant.

July 2014 Loss of CD28 May Contribute to PSC 223

ResultsCD28- T Cells Are More Frequent in Human PSCLiver Tissue

In PSC liver a reversal of the CD4:CD8 ratio (1:1.4) wasdetected compared with peripheral blood (2.3:1) (data notshown). CD4þ T cells were found localized mainly to theportal tracts, and CD8þ T cells were found localized mainlyin the portal areas and in the parenchyma. Both CD4þ andCD8þ T cells were found in proximity to bile ducts(Figure 1A).

We studied the proportion of CD4þ and CD8þ T cells thatwere CD28- (Figure 1B). In the blood of PSC patients, 3.3%of CD4þ T cells were CD28-, whereas in PSC liver tissue thisfrequency was increased approximately 10-fold to 30.3% (P< .0001). By comparison, the frequencies of CD4þCD28- Tcells in normal blood and liver were only 0.96% and 2.5%,respectively. Higher frequencies of CD28- T cells were foundin the CD8þ compartment in both healthy controls and PSCpatients. In healthy controls the frequencies of CD8þCD28-

cells were not significantly different in blood and liver(27.05% and 31.9%, respectively), whereas in PSC,CD8þCD28- cells were increased in the blood (46.35%), andwere significantly higher in the PSC liver (68.5%) whencompared with uninflamed tissue (31.9%) (P < .05)(Figure 1C). We further studied the frequencies of CD28- Tcells in control groups consisting of individuals with PBCand nonalcoholic steatohepatitis (NASH) liver diseases(Figure 1D). The proportion of CD4þ T cells that were CD28-

was significantly higher in PSC livers (30.3%) comparedwith PBC (15.0%; P < .05) and NASH livers (14.2%; P <.05). No significant difference was detected for CD8þCD28-

T cells.

BASICAN

DTR

ANSLAT

IONA

LLIVE

R

CD28- T Cells Are Activated Memory Cells WithIntracellular Stores of Cytotoxic Molecules

CD45RA and CCR7 expression was evaluated to classifyCD28- T cells, in PSC and normal control livers, into naive(CD45RAþCCR7þ; T naive), central memory (CD45RA-

CCR7þ; T central memory), effector memory (CD45RA-

CCR7-; T effector memory), and terminally differentiatedeffector memory RA (CD45RAþCCR7-; terminally differen-tiated effector memory RA) populations (Figure 2A). Ourdata showed that CD4þCD28- T cells in both blood and liverof PSC patients were mainly effector memory cells. PSCliver-infiltrating CD8þCD28- T cells also were effectormemory cells; however, in peripheral circulation theyshowed a terminally differentiated effector memory RAphenotype (Figure 2A). The expression of activation andexhaustion markers was examined further on CD28- T cellsfrom PSC liver and blood, as shown in the representativecontour plots and histograms (Figure 2B). The expressionlevels between CD28- and CD28þ populations were notsignificantly different (Supplementary Figure 1A), but dif-ferences were seen between CD28- cells isolated from livervs blood (Figure 2C). A higher proportion of CD4þCD28- Tcells from liver expressed the activation marker CD69(42.8% vs 17.81%; P ¼ .05) (Figure 2C), and more

liver-infiltrating CD8þCD28- T cells compared with circu-lating CD8þCD28- T cells (49.38% vs 10.62%, respectively;P < .0001) (Figure 2C). Expression of the IL2-Ra chain(CD25) was generally low; only 11.7% and 4.8% of pe-ripheral blood CD4þCD28- and CD8þCD28- T cells, respec-tively, expressed CD25, and expression by liver-infiltratingCD28- cells was less than 2% (Figure 2C). CD8þCD28-

T cells from circulation and liver tissue expressed T-cellimmunoglobulin domain and mucin domain 3 (TIM3) atvery low to undetectable levels (<1%). However, PSC liver-infiltrating CD4þCD28- T cells showed 8.1% TIM3 expres-sion, which was 4 times higher compared with circulatingcells (2.0%; P < .05). Circulating CD4þ and CD8þCD28-

T cells (27.2% and 18.7%, respectively) expressed pro-grammed cell-death 1, whereas liver-infiltrating cellsexpressed slightly lower levels (24.6% and 13.7%, respec-tively) (Figure 2C). Freshly isolated peripheral blood CD4þ

and CD8þCD28- T cells contained intracellular stores ofgranzyme B and perforin, but these cytotoxic moleculeswere absent from their CD28þ counterparts (P < .01,P < .001) (Figure 2D and E). CD4þ and CD8þCD28- T cells inPSC liver and disease control (DC) (PBC and NASH) groupswere phenotypically similar. In the DC group, however,significantly higher CD25 expression was detected (P < .05)(Supplementary Figure 1B).

CD28- T Cells Are Equipped With AdhesionMolecules and Chemokine Receptors ThatPromote Tissue Infiltration and Localization Closeto Bile Ducts

We investigated the mechanisms of CD28- T-cell infil-tration into liver tissue by studying the expression ofadhesion molecules and chemokine receptors previouslyassociated with migration into and retention within theliver. Higher frequencies of CX3CR1

þ CD4þ and CD8þCD28-

T cells were found in circulation than in liver (66.9% vs43.5% for CD4þ and 72.5% vs 20.8% for CD8þ in peripheralblood vs liver, respectively). Frequencies of CX3CR1

þ CD28þ

cells were significantly lower (P < .05) in the blood(Figure 3A). Of all liver-infiltrating CD4þCD28- and CD28þ Tcells in the PSC liver, 54.1% and 40.3% expressed CXCR6,respectively, levels much higher than their circulatingcounterparts. The expression of CXCR6 by CD8þCD28- andCD28þ cells was less than 38%, whereas more than 40% ofCD4þ T cells expressed CXCR6. PSC liver-infiltrating CD4þ

and CD8þCD28- T cells expressed higher levels of CCR10compared with their counterparts in the circulation. More-over, 26.6% and 22.7% of PSC liver-infiltrating CD4þCD28-

and CD28þ cells, respectively, expressed CCR9, at levelshigher than circulating cells (6.9% and 5.2%, respectively).CCR9 expression was observed in less than 6.5% of totalCD8þ T cells in both peripheral circulation and liver of PSCpatients. CD11a was expressed at high levels on both CD28-

and CD28þ T cells of both lineages in both blood and liver,whereas CD62L (L-selectin) was barely expressed (<6%) byliver-infiltrating CD28- and CD28þ T cells of both lineages,consistent with them being memory/effector T cells.Although CD62L expression was significantly higher in

Figure 1. CD28- T cells are more frequent in human PSC liver. (A) Single-color (magnification, 200�) and dual-color immu-nohistochemistry (CD4 [green] and CD8 [red]; 200� and 400�, respectively) showing localization of CD4þ and CD8þ T cells inhuman PSC liver tissue. BD, bile duct. (B) Representative flow cytometry dot plots showing the gating strategy defining CD28-

T cells. (C) The frequency of CD3þCD4þCD28- and CD3þCD8þCD28- T cells in human PSC blood (n ¼ 50 and n ¼ 20,respectively) and liver (n ¼ 11 and n ¼ 8, respectively) was analyzed by flow cytometry and compared with blood (n ¼ 6 andn ¼ 4, respectively) and liver of healthy controls (n ¼ 4, both). *P < .05, ****P < .0001. The ratio of CD4þCD28- T cells in PSCLIMCs:PBMCs was 9:1 and in normal LIMCs:PBMCs was 3:1. The ratio of CD8þCD28- T cells in PSC LIMCs:PBMCs was 1.4:1and in normal LIMCs:PBMCs was 1:1. (D) The frequency of CD3þCD4þCD28- and CD3þCD8þCD28- T cells was analyzed inhuman PSC liver (n ¼ 11 and n ¼ 8, respectively) and compared with PBC (n ¼ 5) and NASH (n ¼ 3) human livers. *P < .05. NB,normal blood; NL, normal liver; PSC B, PSC blood; PSC L, PSC liver.


BASICAND

TRANSLATIONALLIVER

CD4þ and CD8þ T cells in circulation, CD28- T cells fromboth subsets expressed significantly lower levels (44.8%and 40.6%, respectively) than their CD28þ counterparts(93.3% and 80.7%, respectively) (P < .01, P < .001)

(Figure 3A). We further observed that CD28- T cells localizearound bile ducts in human PSC liver tissue (Figure 3B).

The chemokine receptor expression was studied furtheron PBC patients. PSC liver-infiltrating CD4þCD28- T cells

Figure 2. Phenotypic characterization of CD28- T cells in blood and liver of PSC patients. (A) The expression of CD45RAand CCR7 on CD4þCD28- and CD8þCD28- T cells was analyzed by flow cytometry for normal blood (NB) PBMCs (n ¼ 4),PSC blood (PSC B) PBMCs (n ¼ 14), normal liver (NL) LIMCs (n ¼ 4), and PSC liver (PSC L) LIMCs (n ¼ 4). Cells wereclassified into naive (CD45RAþCCR7þ; T naive [Tn]), central memory (CD45RA-CCR7þ; T central memory [Tcm]), effectormemory (CD45RA–CCR7-; T effector memory [Tem]), and terminally differentiated effector memory RA (CD45RAþCCR7-;terminally differentiated effector memory RA [TEMRA]) populations. (B) CD3þCD4þ and CD3þCD8þ cells were selectedand CD28- T cells gated as shown in the representative contour plots were analyzed for CD69, programmed cell-death 1(PD-1), CD25, and TIM3 expression. Representative histograms for the marker (solid line) and its isotype control (shadedarea) are shown. (C) Data show the percentage (mean � SEM) of CD4þCD28- and CD8þCD28- T cells expressing CD69(n ¼ 23 [blood] and n ¼ 6 [liver]), CD25 (n ¼ 22 [blood] and n ¼ 7 [liver]), TIM3 (n ¼ 14 [blood] and n ¼ 5 [liver]), and PD-1(n ¼ 14 [blood] and n ¼ 6 [liver]). *P < .05, ***P < .001. (D and E) CD28- and CD28þ T lymphocytes from PSC PBMCs (n ¼9) and LIMCs (n ¼ 5) were analyzed by flow cytometry for the presence of intracellular deposits of granzyme B andperforin. **P < .01, ***P < .001.


BASICAN

DTR

ANSLAT

IONA

LLIVE

R

Figure 3. CD28- T cells areequipped with adhesionmolecules and chemokinereceptors that promote tis-sue infiltration and locali-zation close to the bileducts. (A) The expressionof chemokine receptorsCX3CR1 (n ¼ 7 [blood] andn ¼ 6 [liver]), CXCR6 (n ¼ 7[blood] and n ¼ 6 [liver]),CCR9 (n ¼ 4 [blood] andn ¼ 5 [liver]), and CCR10(n ¼ 9 [blood] and n ¼ 5[liver]), and adhesion mole-cules CD11a (n ¼ 4 [blood]and n ¼ 4 [liver]), andCD62L (n ¼ 7 [blood] andn ¼ 3 [liver]) on CD28- andCD28þ T cells of CD4þ andCD8þ T cells from blood(PSC B) and liver (PSC L) ofPSC patients was analyzedusing flow cytometry. Datashow the percentages ofCD28- and CD28þ T cellsthat express the chemo-kine receptors. *P< .05, **P< .01, ***P < .001. (B)Representative dual-colorimmunohistochemistry im-age showing the localiza-tion of CD4þCD28- andCD8þCD28- T cells in hu-man PSC liver tissue(magnification, 400�).Arrowheads point to CD28-ve T cells in red and arrowspoint to CD28+ T cells inblack and red.


BASICAND

TRANSLATIONALLIVER

expressed higher levels of CCR9 when compared with PBC(26.6% vs 12.5%) (Supplementary Figure 2B). Similarly, ahigher proportion of CD4þCD28þ T cells in PSC liverexpressed CCR9 when compared with PBC (22.7% vs 7.1%)(Supplementary Figure 2B). The absolute frequency ofCCR9þ T cells in peripheral blood of PSC patients was low,but remained higher compared with PBC patients(CD4þCD28-: 6.9% vs 4.3% and CD4þCD28þ: 5.2% vs 3.3%)(Supplementary Figure 2A). In PBC peripheral blood,significantly higher levels of CD8þCD28þ T cells expressedCX3CR1 (50%) when compared with PSC blood (12.1%; P <.05). Higher levels of PBC liver-infiltrating CD8þCD28þ/- Tcells also expressed CX3CR1 when compared with PSC. Ahigher proportion of PSC peripheral blood and liver-

infiltrating CD4þ and CD8þ CD28- T cells expressedCCR10 compared with PBC (Supplementary Figure 2).

CD28- T Cells Release ProinflammatoryCytokines and Induce Activation of BECs

To further examine the role of CD28- T cells in PSC pa-thology, we studied their ability to produce proinflammatorycytokines immediately after isolation. More than 40% of pe-ripheral blood CD4þCD28- T cells produced IFNg after 6hours of stimulation with anti-CD3, which was significantlyhigher compared with their CD28þ counterparts (P < .05).Peripheral blood CD8þCD28- T cells produced significantlymore IFNg (28%) when compared with CD8þCD28þ T cells

Figure 4. CD28- T cellsrelease proinflammatorycytokines and their super-natants are able to activatehuman primary BECsin vitro. Data from (A) 6PSC peripheral bloodsamples and (B) 6 PSCliver samples showingTNFa and IFNg productionby CD4þCD28þ/- andCD8þCD28þ/-. *P < .05.(C) Representative flowcytometry plots showingthe gating strategy fordefining T-regulatory cells.(D) Data show the per-centage of CD3þCD4þ Tcells that are CD25hi

CD127low, in blood andliver of normal and PSCpatients (normal blood[NB], n ¼ 6; PSC blood[PSC B], n ¼ 8; normal liver[NL], n ¼ 5; PSC liver [PSCL], n ¼ 6). Each shaperepresents the value fromeach individual, and theline represents the meanvalue � SEM.


BASICAN

DTR

ANSLAT

IONA

LLIVE

R

(P < .05) (Figure 4A). Both CD4þ and CD8þCD28- T cellssecreted more TNFa (14.5% and 7.2%, respectively) thanCD28þ T cells (Figure 4A). PSC liver-infiltrating CD28- T cellsfrom both CD4þ and CD8þ subsets produced IFNg (16.6%and 11.0%, respectively), but levels were not significantlydifferent from their CD28þ counterparts (Figure 4B). Liver-infiltrating cells also produced TNFa but at a lower levelthan IFNg. In contrast to blood, liver-infiltrating CD28þCD4þ

and CD8þ T cells produced more TNFa when compared withCD28- T cells (Figure 4B).

Studies have reported that Tregs can inhibit cytokineproduction by CD28- T cells.11 In PSC liver tissue wedetected low frequencies of CD3þCD4þCD25hiCD127low

Tregs at levels similar to those seen in normal tissue(Figure 4C and D).

Because CD28- and CD28þ T cells showed differentproinflammatory cytokine profiles, we were interested toobserve the effect of their secretomes on BEC activation andsurvival. Conditioned media from activated CD4þCD28- andCD8þCD28- T cells induced the expression of intercellularcell adhesion molecule-1 (ICAM1), HLA-DR, and CD40(Figure 5A–C) on primary BECs. No significant differences,

however, were detected on the effects between CD28- andCD28þ T cells. Neither CD28- nor CD28þ cells inducedleukocyte function-associated antigen 3 expression (datanot shown). Only 8% of BECs were viable after 4 days ofculture with activated CD4þCD28- T-cell–conditioned media,whereas 29%–41% of BECs were alive when cultured withthe media of all other subsets (Figure 5D).

TNFa Induces the Differentiation of CD28- T Cellsand Vitamin D Overcomes This Effect

We detected significantly increased levels of TNFa andIFNg messenger RNA in the PSC liver microenvironmentcompared with normal tissue (4.8 vs 1.7, P < .05; and 17.5vs 3.2, P < .01, respectively) (Figure 6A) and increasedlevels of IL17A messenger RNA relative to PBC(Supplementary Figure 3A). PSC liver-infiltrating mono-nuclear cells were able to release TNFa and IFNg after 24hours of culture even without stimulation. IL2, IL4, and IL10were low to undetectable (Figure 6B). PSC LIMCs also wereproducers of IL17, IL23, IL6, IL8, CCL2, CCL3, and CCL4(Supplementary Figure 3B).

Figure 5. Supernatantsfrom CD28- T cells are ableto activate human primaryBECs in vitro. (A–C) Datashow the indirect effects ofuntreated (UT) and aCD3-/aCD28-treated, cell-sortedCD28- and CD28þ T cellson the percentageexpression of ICAM1,HLA-DR, and CD40 (n ¼ 5)on BECs. (D) Data showthe percentage of liveBECs after 4 days of co-culture with the T-cell–conditioned media (n ¼ 3).


BASICAND

TRANSLATIONALLIVER

We cultured PSC CD4þ T cells with recombinant (r)TNFain vitro and found a significant increase in CD4þCD28- Tcells after 7–21 days. After 21 days, 22.9% of CD4þ T cellsspontaneously lost CD28 expression whereas this expres-sion increased to 37.6% in the presence of exogenousrTNFa. Given the immunomodulatory effects of vitamin Dwe tested further whether the active form of vitamin D1,25(OH)2D3 had an effect on CD28 expression in thepresence and/or absence of TNFa. Our data showed that1,25(OH)2D3 markedly suppressed the development ofCD28- T cells (2%; P < .05) in vitro, even in cultures inwhich exogenous rTNFa was present (3.7%; P < .05)(Figure 6C and D).

Frequencies of CD4þCD28- T Cells in PSCPatients Supplemented With Vitamin D

To examine whether vitamin D status influences CD28-

T-cell frequencies in PSC patients we measured theirperipheral blood CD28- T-cell frequencies and serumvitamin D (25[OH]D) concentration. A trend toward anegative correlation with serum vitamin D was seen(nonparametric Spearman rank correlation test: P ¼.1699; and Spearman r(r): �0.1451) (Figure 7A). Nosignificant associations were found between peripheralCD28- T-cell frequency and patient age, ethnicity,severity/activity of liver disease, or inflammatory boweldisease status (data not shown). In 14 PSC patients withsevere vitamin D deficiency who returned for a routineclinical review, we repeated CD4þCD28- T-cell frequencyand vitamin D measurements. Our data showed that

supplementation of vitamin D increased levels sufficiently(Figure 7B) and in 7 of these patients the CD4þCD28- T-cell frequency was reduced by 5.8-fold (Figure 7C).

DiscussionWe provide evidence that a significant proportion of PSC

liver-infiltrating T cells lack CD28, an important immuno-regulatory protein implicated in PSC pathogenesis by ge-netic association studies. We showed that CD28- T cells arerecently activated memory/effector cells, equipped with acombination of adhesion molecules and chemokine re-ceptors that allow infiltration into liver tissue and specificlocalization around bile ducts. Upon re-activation, CD28-

memory/effector T cells release high levels of TNFa andIFNg proinflammatory cytokines, which can activate neigh-boring BECs to express adhesion and costimulatory mole-cules. The important role of local TNFa is suggested by thefinding of high intrahepatic levels of TNFa in PSC and itsability to induce CD28 loss in cultured T cells. The ability ofvitamin D to prevent this effect suggests a novel therapeuticaction of this vitamin in PSC. However, more extensivestudies are required to evaluate this effect, particularly inlarger cohorts of PSC patients, with vitamin D supplemen-tation remaining a speculative approach to intervention atthis time.

We show that CD28- T cells express high levels of CD69and no CD45RA, suggesting an active effector/memoryphenotype. Both CD4þ and CD8þCD28- T cells expressedprogrammed cell-death 1, a marker associated with bothactivated and exhausted T cells.17,18 PSC liver-infiltrating

Figure 6. TNFa enhancesthe emergence of CD28- Tcells and 1,25(OH)2D3overcomes this effect. (A)TNFa and IFNgmessengerRNA (mRNA) expression in6 normal liver (NL) and 9PSC liver tissues wasmeasured by quantitativepolymerase chain reaction.Scatter dot plots showrelative mRNA levels indiseased livers withrespect to 1 NL tissue(mean � SEM). *P < .05,**P < .01. (B) Cytokinesand chemokines releasedfrom 3 PSC LIMCs asanalyzed with the HumanCytokine Array kit (R&DSystems, Abingdon,United Kingdom). Expres-sion levels are reported asmean pixel density inarbitrary units. (C) CD4þ Tcells from blood of PSCpatients were stimulatedwith aCD3/aCD28 beadsand cultured for 21 days inthe presence or absenceof TNFa, with or without1,25(OH)2D3. The fre-quency of CD28- cells wasmeasured at 0, 7, 14, and21 days by flow cytometryas shown in the represen-tative contour plots. (D)Data from 7 donors.


BASICAN

DTR

ANSLAT

IONA

LLIVE

R

CD4þCD28- T cells expressed low levels of TIM3, whichgenerally is considered a marker of T-cell exhaustion,suggesting that CD28- cells in PSC liver are activatedeffector cells.19 TIM3 levels were higher on CD28- thanCD28þ cells, further suggesting that CD28- cells have un-dergone many cycles of activation. The higher CD25expression on CD28- cells of the DC group agrees with ourdata showing the presence of more CD28- cells in PSCbecause CD25 expression is dependent on CD28-mediatedsignals. Both CD4þ and CD8þ CD28- cells readily producedproinflammatory cytokines upon activation and containedperforin and granzyme B.

Our data suggest mechanisms to explain the accumu-lation of CD28- T cells in human PSC liver tissue bymigration from peripheral blood. We found that CD28- Tcells are equipped with CD11a and chemokine receptorsCX3CR1, CXCR6, and CCR10, which allow their infiltrationinto tissue and migration toward BECs, which stronglyexpress the chemokine ligands CX3CL1, CXCL16, andCCL28.20–22 This recruitment may be amplified in aparacrine fashion because CD28- T cells could releaseTNFa and IFNg proinflammatory cytokines, which in-crease adhesion molecule expression and chemokinesecretion from BECs. The ability of CD28- T cells to up-

Figure 7. CD4þCD28- T cells in PSC patients supplementedwith vitamin D. (A) Serum 25(OH)D levels of 92 PSC patientswere correlated with the frequency of CD4þCD28- T cells incirculation. Each dot represents the value from each indi-vidual. (B) PSC patients who had insufficient serum vitamin Dlevels were supplemented with vitamin D as part of theirmedical treatment. The levels of serum 25(OH) vitamin Dbefore and after supplementation are shown. (C) Data showthe percentage of CD4þCD28- T cells after several weekswhen serum vitamin D levels reached sufficiency. Data wereanalyzed using the Wilcoxon matched-pairs signed rank test(P ¼ .669).


BASICAND

TRANSLATIONALLIVER
regulate ICAM1, HLA-DR, and CD40 on BECs could pro-
mote BECs as T-cell targets in PSC. Increased ICAM1expression on bile ducts is a characteristic feature of liverdiseases in which bile ducts are the major targets ofimmune-mediated destruction, such as PSC and PBC.23

Increased binding of effector cells to cholangiocyteICAM1 not only promotes accumulation but also allowsrecognition of major histocompatibility complex antigensand the activation of cytolytic mechanisms that inducecholangiocyte apoptosis.24

Previous studies showed the ability of Tregs to inhibitcytokine production by CD28- T cells but not their pro-liferation.11 In PSC liver tissue we detected low fre-quencies of CD4þCD25hiCD127low Tregs at levels similarto those seen in normal tissue, and lower than the fre-quencies previously reported for other autoimmune andchronic inflammatory conditions.25 A reduction in thefrequency of peripheral blood Tregs and in Foxp3þ T cellsin PSC liver, with an apparent impaired suppressive ca-pacity, recently was reported.26 Reduced numbers ofTregs in PSC liver therefore may lead to uncontrolledcytokine production by CD28- T cells. Other investigators

have shown that TNFa can modulate the expression ofCD28 at a transcriptional level.9 Our data showed highlevels of TNFa in human PSC liver, suggesting that thelocal cytokine microenvironment may promote the dif-ferentiation and accumulation of CD28- T cells. We alsoshowed that activated liver-infiltrating CD28þ T cellsproduce high levels of TNFa, which may act in an auto-crine manner to down-regulate their own CD28 expres-sion. CD28 signaling is required for IL2 production, whichin turn is needed for induction of Tregs. Therefore, in PSCliver, in which 30.3% of CD4þ T cells are CD28- T cells, IL2production likely is defective and this might be acontributing factor for the reduced frequency of Tregs inPSC liver tissue. Because both CD28- and CD28þ T cellslocalize close to bile ducts and produce TNFa, we suggestthat such local cytokine production can induce theemergence of CD28- T cells and the amplification of theinflammatory response as indicated by their ability toactivate BECs. TNFa also can activate BECs to expressreceptors such as tumor necrosis factor receptor super-family members involved in apoptosis and cholangiocytedeath,6 further implicating TNFa in the pathogenesis ofbiliary destruction in PSC. Correspondingly, in our co-culture system, we observed that peripheral bloodCD4þCD28-, which produced the highest levels of TNFa,promoted the greatest death of BECs. CD28- T cells alsoare cytotoxic cells, expressing perforin and granzyme B,and other investigators have shown the synergistic effectof them in cholangiocyte injury.27

We also show that 1,25(OH)2D3 can increase the ab-solute expression of CD28 on in vitro stimulated cells andprevent emergence of the CD28- population during long-term culture, even in the presence of exogenous rTNFa.This finding is consistent with recent data showing thatvitamin D could increase the median fluorescence of CD28in CD4þ T cells from healthy controls and multiple scle-rosis patients during in vitro stimulation.28 We observedthat in a subset of PSC patients with profound deficiency invitamin D, the frequency of CD28- T cells in the circulationwas reduced after supplementation and correction ofhypovitaminosis. However, in as many patients an increaseor stability of CD28- T-cell frequency was observed. Nosignificant clinical differences were detected between thecohorts experiencing a reduction in CD28- frequency vsthose in whom CD28- frequency remained stable/increased (Supplementary Table 3). We suggest that,in vivo, a more complex set of immunoregulatory in-teractions is likely to participate and other factors in thedisease background of the patients also might affect re-sults. Moreover, we speculate that in patients in whomsupplementation of vitamin D had no obvious effect, itactually may have prevented the further expansion ofCD28- proinflammatory cells. We therefore highlight apotential therapeutic use of vitamin D in PSC that meritsfurther in-depth evaluation.

Our findings are important in the context of a diseasethat lacks therapy and has poorly refined animal modelsin which to test new therapies. We recognize that theincreased frequency of CD28- T cells is a feature of


BASICAN

DRA

NSLA

TION

ALLIVE

R

chronic inflammatory and autoimmune disease. Notably,our data show a significant difference in the frequency ofCD4þCD28- T cells in PSC liver, compared with PBC andNASH, with PSC liver-infiltrating T cells also showinghigh CCR9 expression, further suggesting the specificityof these cells for PSC. It is important, however, to high-light the inherent difficulty of comparing and contrastingliver diseases that are fundamentally distinct because ofthe different patterns of liver injury and response toinjury over time. Although CD28- T cells were detected inall 3 chronic liver diseases studied, we believe specificitylies in the antigen specificity of these cells. Genetic as-sociations at the CD28 locus have been identified inseveral autoimmune diseases (eg, PSC, celiac disease,type 1 diabetes, rheumatoid arthritis, alopecia areata,and autoimmune thyroid disease), but notably not inwell-powered assessments in inflammatory bowel dis-ease or PBC. The basis of the profound pleiotropyobserved for risk genes in immune-mediated conditionsis poorly understood, and our findings provide thecontextual data for a PSC-centered evaluation of CD28risk variants.

The frequent, and often profound, deficiency of vitaminD associated with liver disease may make the mechanismparticularly applicable to cholestatic liver disease; indeed,increased numbers of CD4þCD28- T cells have been re-ported in association with bile duct damage in PBC.29 Ourvitamin D data are provocative and interesting, albeitpresently still are a hypothesis-generating concept at thistime.

In summary, in PSC patients we show the significance ofCD28- cells, which release proinflammatory cytokines andcytotoxic mediators and express the necessary receptors toallow their localization close to the bile ducts and inducetheir activation. We provide evidence that TNFa, which isenriched in the PSC liver microenvironment, can enhancethe accumulation of CD28- T cells and show that this effectcan be overcome with vitamin D.

T

Supplementary MaterialNote: To access the supplementary material accompanyingthis article, visit the online version of Gastroenterology atwww.gastrojournal.org, and at http://dx.doi.org/10.1053/j.gastro.2014.04.003.

References
1. Hirschfield GM, Karlsen TH, Lindor KD, et al. Primary
sclerosing cholangitis. Lancet 2013;382:1587–1599.2. Liu JZ, Hov JR, Folseraas T, et al. Dense genotyping of

immune-related disease regions identifies nine new riskloci for primary sclerosing cholangitis. Nat Genet 2013;45:670–675.

3. Hashimoto E, Lindor KD, Homburger HA, et al. Immu-nohistochemical characterization of hepatic lympho-cytes in primary biliary cirrhosis in comparison withprimary sclerosing cholangitis and autoimmunechronic active hepatitis. Mayo Clin Proc 1993;68:1049–1055.

4. Ponsioen CY, Kuiper H, Ten Kate FJ, et al. Immunohis-tochemical analysis of inflammation in primary sclerosingcholangitis. Eur J Gastroenterol Hepatol 1999;11:769–774.

5. Bo X, Broome U, Remberger M, et al. Tumour necrosisfactor alpha impairs function of liver derived T lympho-cytes and natural killer cells in patients with primarysclerosing cholangitis. Gut 2001;49:131–141.

6. Noel PJ, Boise LH, Thompson CB. Regulation of T cellactivation by CD28 and CTLA4. Adv Exp Med Biol 1996;406:209–217.

7. Fasth AE, Cao D, van Vollenhoven R, et al.CD28nullCD4þ T cells–characterization of an effectormemory T-cell population in patients with rheumatoidarthritis. Scand J Immunol 2004;60:199–208.

8. Markovic-Plese S, Cortese I, Wandinger KP, et al.CD4þCD28- costimulation-independent T cells inmultiple sclerosis. J Clin Invest 2001;108:1185–1194.

9. Bryl E, Vallejo AN, Matteson EL, et al. Modulation ofCD28 expression with anti-tumor necrosis factor alphatherapy in rheumatoid arthritis. Arthritis Rheum 2005;52:2996–3003.

10. Nakajima T, Schulte S, Warrington KJ, et al. T-cell-mediated lysis of endothelial cells in acute coronarysyndromes. Circulation 2002;105:570–575.

11. Thewissen M, Somers V, Hellings N, et al.CD4þCD28null T cells in autoimmune disease: patho-genic features and decreased susceptibility to immuno-regulation. J Immunol 2007;179:6514–6523.

12. L Ng KV, Nguyen LT. The role of vitamin D in primarybiliary cirrhosis: possible genetic and cell signalingmechanisms. Gastroenterol Res Pract 2013;2013:602321.

13. Mowry EM. Vitamin D: evidence for its role as a prog-nostic factor in multiple sclerosis. J Neurol Sci 2011;311:19–22.

14. Jeffery LE, Burke F, Mura M, et al. 1,25-DihydroxyvitaminD3 and IL-2 combine to inhibit T cell production of in-flammatory cytokines and promote development of reg-ulatory T cells expressing CTLA-4 and FoxP3. J Immunol2009;183:5458–5467.

15. Liaskou E, Zimmermann HW, Li KK, et al. Monocytesubsets in human liver disease show distinct phenotypicand functional characteristics. Hepatology 2013;57:385–398.

16. Humphreys EH, Williams KT, Adams DH, et al. Primaryand malignant cholangiocytes undergo CD40 mediatedFas dependent apoptosis, but are insensitive to directactivation with exogenous Fas ligand. PLoS One 2010;5:e14037.

17. Wei F, Zhong S, Ma Z, et al. Strength of PD-1 signalingdifferentially affects T-cell effector functions. Proc NatlAcad Sci U S A 2013;110:E2480–E2489.

18. Agata Y, Kawasaki A, Nishimura H, et al. Expression ofthe PD-1 antigen on the surface of stimulated mouse Tand B lymphocytes. Int Immunol 1996;8:765–772.

19. Sakhdari A, Mujib S, Vali B, et al. Tim-3 negatively reg-ulates cytotoxicity in exhausted CD8þ T cells in HIVinfection. PLoS One 2012;7:e40146.

http://www.gastrojournal.org



http://refhub.elsevier.com/S0016-5085(14)00452-1/sref1






























































































BASICAND

TRANSLATIONALLIVER

20. Heydtmann M, Lalor PF, Eksteen JA, et al. CXC che-mokine ligand 16 promotes integrin-mediated adhesionof liver-infiltrating lymphocytes to cholangiocytes andhepatocytes within the inflamed human liver. J Immunol2005;174:1055–1062.

21. Isse K, Harada K, Zen Y, et al. Fractalkine and CX3CR1are involved in the recruitment of intraepithelial lympho-cytes of intrahepatic bile ducts. Hepatology 2005;41:506–516.

22. Eksteen B, Miles A, Curbishley SM, et al. Epithelialinflammation is associated with CCL28 production andthe recruitment of regulatory T cells expressing CCR10.J Immunol 2006;177:593–603.

23. Volpes R, van den Oord JJ, Desmet VJ. Immunohisto-chemical study of adhesion molecules in liver inflam-mation. Hepatology 1990;12:59–65.

24. Ayres RC, Neuberger JM, Shaw J, et al. Intercellularadhesion molecule-1 and MHC antigens on humanintrahepatic bile duct cells: effect of pro-inflammatorycytokines. Gut 1993;34:1245–1249.

25. Oo YH, Weston CJ, Lalor PF, et al. Distinct roles forCCR4 and CXCR3 in the recruitment and positioning ofregulatory T cells in the inflamed human liver. J Immunol2010;184:2886–2898.

26. Sebode M, Peiseler M, Franke B, et al. Reduced FOXP3þregulatory T cells in patients with primary sclerosingcholangitis are associated with IL-2RA gene poly-morphisms. J Hepatol 2014;60:1010–1016.

27. Shivakumar P, Mourya R, Bezerra JA. Perforin andgranzymes work in synergy to mediate cholangiocyte

injury in experimental biliary atresia. J Hepatol 2014;60:370–376.

28. Kickler K, Ni Choileain S, Williams A, et al. Calcitriolmodulates the CD46 pathway in T cells. PLoS One 2012;7:e48486.

29. Isse K, Harada K, Sato Y, et al. Characterization of biliaryintra-epithelial lymphocytes at different anatomical levelsof intrahepatic bile ducts under normal and pathologicalconditions: numbers of CD4þCD28- intra-epitheliallymphocytes are increased in primary biliary cirrhosis.Pathol Int 2006;56:17–24.

Author names in bold designate shared co-first authors.

Received January 6, 2014. Accepted April 4, 2014.

Reprint requestsAddress requests for reprints to: Gideon M. Hirschfield, MB BChir, PhD, FRCP,Centre for Liver Research, University of Birmingham, Wolfson Drive,Birmingham B15 2TT, United Kingdom. e-mail: [email protected];fax: þ44121 415 8701.

Conflicts of interestThe authors disclose no conflicts.

FundingSupported in part by the National Institute for Health Research, Wellcome Trust(RCHX1704), and Primary Sclerosing Cholangitis Partners Seeking a Cure; andby the German Research Foundation (BR182/1-1 to T.B.).

The views expressed are those of the authors(s) and not necessarily those ofthe National Health System, the National Institute for Health Research, or theDepartment of Health.




















































mailto:[email protected]

Supplementary Materials and MethodsIsolation of LIMCs

LIMCs were isolated from fresh human liver tissue.Tissue was washed in ice-cold phosphate-buffered saline toremove residuals of blood, cut into small pieces, mechani-cally digested in RPMI1640 using a Stomacher400 Circu-lator (Seward Stomacher400 Circulator [Cole-ParmerInstrument Co Ltd, London, United Kingdom]) for 330 sec-onds at 260 speed, filtered, and purified further by densitygradient centrifugation using Lympholyte (Cedarlane Labo-ratories, Burlington, Canada) for 20 minutes at 550 � g. TheLIMC population was extracted and further studied for theirphenotype and function.

Isolation of Human Primary BECsHuman BECs were isolated from approximately 150 g

of liver tissue. Tissue was digested enzymatically withcollagenase type 1A (Sigma, Dorset, United Kingdom),filtered, and purified further via density gradientcentrifugation over 33%/77% Percoll (Amersham Bio-sciences UK Ltd, Buckinghamshire, United Kingdom).BECs were extracted from the mixed nonparenchymalpopulation via magnetic selection using antibodiesagainst the cholangiocyte-specific receptor HEA-125 (50mg/mL; Progen Biotechnik, Heidelberg, Germany). HEA-125–positive BECs were plated in rat-tail collagen-coatedflasks and kept in culture and used at passage 4.

Flow CytometryBlood and liver-derived mononuclear cells were incu-

bated with mouse anti-human or rat anti-human mono-clonal fluorochrome-conjugated antibodies for 30 minutesin the dark before flow cytometry for surface markerphenotypic evaluation. Cells were washed and resuspendedin phosphate-buffered saline supplemented with 1% fetalcalf serum. Cells were fixed and permeabilized for immu-nofluorescent staining of intracytoplasmic cytokines andcytotoxic molecules.

Isolation of CD28þ/- T-Cell Subsets and Effect ofT-Cell–Conditioned Media on BECs

Fluorescence-activated, cell-sorted CD3þCD4þ/-

CD28þ/- subsets were cultured overnight using 96-well,round-bottomed plates coated with anti-CD3 (OKT3; 5mg/mL) and anti-CD28 (5 mg/mL) in media consisting ofRPMI1640, Dulbecco’s modified Eagle medium, and 10%heat-inactivated human AB serum (Life Technologies).Cells were cultured at a density of 50,000 cells/100 mL.Cell-free conditioned media then was transferred to pri-mary BECs cultured (10,000 cells/well) as a monolayer in96-well, flat-bottomed plates. After 4 days, the effect ofconditioned media on BEC activation and death wasmonitored by staining with a live/dead cell discriminationdye (Biolegend, London, United Kingdom), according tothe manufacturer’s instructions, followed by antibodysurface labeling for ICAM1, HLA-DR, CD40, and leukocyte

function-associated antigen 3 markers of BEC activation,and analysis by flow cytometry.

Detection of Cytokine Expression Ex VivoBrefeldin A (1 mL/mL) was added 2 hours after initi-

ation of stimulation of PSC PBMCs and LIMCs with plate-bound anti-CD3 (OKT3; 5 mg/mL) (at 106 cells/well) toallow intracellular cytokine accumulation. Cells werewashed and stained with CD3, CD4, CD8, and CD28 anti-bodies before fixation, permeabilization, intracellularcytokine staining for TNFa and IFNg, and detection withflow cytometry.

Immunohistochemical AnalysisLocalization of CD4þ and CD8þ T cells as well as local-

ization of T cells that had lost CD28 expression was testedusing immunohistochemistry. Alcohol-fixed frozen tissuesections (7- to 10-mm thick) initially were blocked withBLOXALL endogenous peroxidase (Vector Laboratories Ltd,Peterborough, United Kingdom) and alkaline phosphatasesolution for 10 minutes at room temperature. Sections thenwere washed with Tris-buffered saline–0.1% Tween20 andblocked with 2� casein block (Vector Labs) for 10 minutesbefore incubation with mouse anti-human CD8 (1:50; VectorLabs) for 30 minutes at room temperature. After washes,sections were incubated with the secondary antibody,ImmPRESS universal (Vector Labs), for 30 minutes, fol-lowed by developing using the Vector NovaRED PeroxidaseSubstrate kit (Vector Labs) for 5 minutes. Sections wereblocked further with 2� casein for 5 minutes before incu-bation with mouse anti-human CD4 (1:25; Vector Labs) for1 hour and the ImmPRESS-AP anti-mouse Ig (alkalinephosphatase) polymer detection kit (Vector Labs) for 30minutes. The StayGreen/AP substrate chromogen (Abcam,Cambridge, United Kingdom) then was used for 20 minutesfor green color development. Isotype-matched negativecontrol antibodies also were used (IgG1, 1:25; Dako UK Ltd,Cambridgeshire, United Kingdom; and IgG2b, 1:50; eBio-science, Ltd, Hatfield, United Kingdom).

For localization of CD4þ and CD8þ T cells that hadlost CD28 expression, alcohol-fixed frozen tissue sections(7- to 10-mm thick) also were used. Endogenous peroxi-dase was blocked using peroxidase blocking solution(Dako) for 10 minutes at room temperature. After washesin Tris-buffered saline–0.1% Tween20, sections wereblocked with 2� casein block (Vector Labs) for 10 mi-nutes followed by incubation with mouse anti-humanCD28 (20 mg/mL) for 30 minutes at room temperature.After washes, ImmPRESS anti-mouse kit (Vector Labs)was used for 30 minutes, followed by washes and devel-oping using the Vector DAB/nickel substrate (VectorLabs) for 5 minutes. The mouse on mouse elite peroxidasekit (Vector Labs) then was used according to the manu-facturer’s instructions, to continue the staining withmouse anti-human CD4 (1:15; Vector Labs) or mouseanti-human CD8 (1:40; Vector Labs) for 1 hour at roomtemperature. The ImmPRESS anti-mouse kit then was

July 2014 Loss of CD28 May Contribute to PSC 232.e1

used for 30 minutes followed by developing using the 3-amino-9-ethylcarbazole peroxidase substrate kit for 30minutes, counterstained with hematoxylin for 30 seconds,and mounted in aqueous mounting media. Isotype-matched negative control antibodies also were used(IgG1, 20 mg/mL for CD28 and at 1:15 for CD14 andIgG2b, 1:50; eBioscience).

RNA Extraction, Complementary DNA Synthesis,and Real-Time Quantitative Polymerase ChainReaction

Total RNA was extracted from normal (n ¼ 6) anddiseased PBC (n ¼ 5), and PSC (n ¼ 9) human liver tis-sues to test TNFa, IFNg, and IL17A messenger RNAexpression. The eluted RNA concentration was measuredusing a NanoDrop Spectrophotometer (Thermo Fisher

Scientific, Waltham, MA). Extracted RNA (50 mg) wastranscribed into complementary DNA using the iScriptcomplementary DNA synthesis kit (Bio-Rad, Hercules,CA). Quantitative analyses were performed using TaqManFluorogenic 5’ nuclease assays using gene-specific5’FAM-labeled probes (Life Technologies) run on an ABIPrism 7700 sequencer (Life Technologies) with b-actinused as an internal control. Differential expression levelswere calculated according to the 2-DDCt method.

Cytokine Secretion AssaysPSC liver-infiltrating mononuclear cells were cultured at

106 cells/mL for 24 hours in RPMI1640 supplemented with10% fetal calf serum. Cell-free supernatant was tested forcytokine and chemokine presence using the cytokine arraypanel A kit (R&D Systems).

232.e2 Liaskou et al Gastroenterology Vol. 147, No. 1

Supplementary Figure 1. Phenotypic characterization of CD28þ T cells in blood and liver of PSC patients. The expression ofCD69, CD25, TIM3, and programmed cell-death 1 (PD-1) on CD28þ T cells isolated straight ex vivo also was analyzed usingflow cytometry. (A) Data show the percentage of CD4þCD28þ and CD8þCD28þ T cells expressing CD69 (n ¼ 23 [blood] andn ¼ 6 [liver]), CD25 (n ¼ 22 [blood] and n ¼ 7 [liver]), TIM-3 (n ¼ 14 [blood] and n ¼ 5 [liver]), and PD-1 (n ¼ 14 [blood] and n ¼ 6[liver]). Bars indicate the mean � SEM. *P < .05, **P < .01, ****P < .001 as compared with expression levels in PSC liversamples. (B) Phenotypic characteristics of CD28- T cells in the PSC liver and disease control group consisted of PBC andNASH livers. Data show the percentage of CD4þCD28- and CD8þCD28- T cells expressing CD69 (n ¼ 6 [PSC liver] and n ¼ 3[disease control]), TIM3 (n ¼ 5 [PSC liver] and n ¼ 2 [disease control]), and PD-1 (n ¼ 6 [PSC liver] and n ¼ 1 [disease control]).Bars indicate the mean � SEM. *P < .05.


Supplementary Figure 2. CD28- T cells show a differential expression of chemokine receptors between PSC and PBC. Theexpression of chemokine receptors on CD28- and CD28þ T cells of CD4 and CD8 T cells from PSC and PBC patients wasanalyzed using flow cytometry. (A, B) Data show the percentage of CD28- and CD28þ T cells that express the chemokinereceptors CX3CR1 (n ¼ 7 [PSC blood] vs n ¼ 3 [PBC blood], and n ¼ 6 [PSC livers] vs n ¼ 2 [PBC livers]), CCR9 (n ¼ 4 [PSCblood] vs n ¼ 4 [PBC blood], and n ¼ 5 [PSC livers] vs n ¼ 2 [PBC livers]), and CCR10 (n ¼ 6 [PSC blood] vs n ¼ 3 [PBC blood],and n ¼ 6 [PSC livers] vs n ¼ 2 [PBC livers]). *P < .05.


Supplementary Figure 3. Cytokine profile of PSC livermicroenvironment. (A) IL17A messenger RNA (mRNA)expression in 4 PSC and 4 PBC livers. Scatter dot plots showrelative mRNA levels in PSC livers with respect to PBC livers(mean � SEM). *P < .05. (B) Liver-infiltrating mononuclearcells from 3 PSC liver samples were cultured for 24 hoursbefore collection of their cell-free conditioned media. Cyto-kines and chemokines in the media were measured using aHuman Cytokine Array kit. Expression levels are reported asthe mean pixel density in arbitrary units.


Supplementary Table 1.Antibodies Used for Flow Cytometry

Antibody Clone Final concentration Source

CD3-Pacific Blue UCHT1 2 mL/106 cells BD PharmingenMouse IgG1, k–Pacific Blue X40 2 mL/106 cells BD PharmingenCD3-FITC UCHT1 2 mL/106 cells BD PharmingenCD4-V500 RPA-T4 2 mL/106 cells BD PharmingenMouse IgG1, k–V500 X40 2 mL/106 cells BD PharmingenCD4-APC RPA-T4 2 mL/106 cells BD PharmingenMouse IgG1, k–APC MOPC-21 2 mL/106 cells BD PharmingenCD8-PeCy7 RPA-T8 2 mL/106 cells BD PharmingenMouse IgG1, k–PE/CY7 MOPC-21 2 mL/106 cells BD PharmingenCD8-FITC RPA-T8 2 mL/106 cells BD PharmingenCD28-PE CD28.2 2 mL/106 cells BD PharmingenMouse IgG1 k–PE MOPC-21 2 mL/106 cells BD PharmingenCD28-APC CD28.2 2 mL/106 cells BD PharmingenCD69-FITC FN50 2 mL/106 cells BD PharmingenCD25-APC M-A251 5 mL/106 cells BD PharmingenPD-1-APC MIH4 5 mL/106 cells BD PharmingenMouse IgG1, k–APC MOPC-21 5 mL/106 cells BD PharmingenTIM3-PeCy7 F38-2E2 5 mL/106 cells eBioscienceCD45RA-V450 HI100 5 mL/106 cells BD PharmingenMouse IgG2b, k–V450 27-35 5 mL/106 cells BD PharmingenCD45RO-PE UCHL-1 5 mL/106 cells BD PharmingenMouse IgG2a, k–PE G155-178 5 mL/106 cells BD PharmingenCCR7-PE/Cy7 3D12 5 mL/106 cells BD PharmingenCD62L-APC DREG-56 5 mL/106 cells BD PharmingenCX3CR1- PE/Cy7 2A9-1 2 mL/106 cells BiolegendRat IgG2b, k-PE/Cy7 RTK4530 2 mg/mL BiolegendCXCR6-APC K041E5 4 mL/106 cells BiolegendMouse IgG2a-APC X39 4 mL/106 cells BD PharmingenCCR10-PE 314305 2 mg/mL R&D SystemsRat IgG2a-PE 54447 2 mg/mL R&D SystemsCD11a-PE HI111 2 mL/106 cells BD PharmingenGranzyme B- FITC GB11 5 mL/106 cells BD PharmingenPerforin -PE dG9 5 mL/106 cells BD PharmingenMouse IgG2b, k-PE 27-35 5 mL/106 cells BD PharmingenIFNg-APC B27 0.5 mL/106 cells BD PharmingenTNFa-FITC MAb11 3 mL/106 cells BD PharmingenICAM1-FITC RR1/1 3 mL/106 cells eBioscienceMouse IgG1-FITC MOPC-21 3 mL/106 cells BD PharmingenCD40-APC HI40a 5 mL/106 cells ImmunostepLFA-3-PE L300.4 5 mL/106 cells Becton DickinsonHLA-DR-PE/Cy7 L243 (G46-6) 2.5 mg/mL BD PharmingenMouse IgG2a-PE/Cy7 MOPC-173 2.5 mg/mL BD Pharmingen

APC, allophycocyanin; FITC, fluorescein isothiocyanate; LFA-3, leukocyte function-associated antigen 3; PD-1, programmedcell-death 1; PE, phycoerythrin.


Supplementary Table 2.Antibodies Used for Immunohistochemical Analysis

Antibody Clone Final concentration Source

CD4 1F6 1:25 Vector LaboratoriesMouse IgG1 1:25 DakoCD8 4B11 1:50 Vector LaboratoriesMouse IgG2b 1:40 eBioscienceCD28 CD28.2 20 mg/mL eBioscience

Supplementary Table 3.Clinical Characteristics of PSC Patients Supplemented With Vitamin D

Responders Nonresponders

Interval between sampling, da 92 (78–105) 91 (42–189)Age at time of sample, y 42 (20–61) 55 (35–69)UDCA exposure 5/7 6/7With IBD 5/7 6/7Active IBD, %b 0/5 2/6On immunosuppression 0/7 1/7Liver disease status

Precirrhotic 2/7 3/7Cirrhotic: compensated 5/7 3/7Cirrhotic: decompensated 0/7 1/7

Dominant strictures 0/7 1/7Ascending cholangitisb 1/7 3/7Increase in vitamin D level, nmol/La 57 (31–80) 43 (30–67)Laboratory parameters at baselinea

Bilirubin level, mmol/L 22 (12–39) 43 (13–75)ALT level, IU/L 61 (24–123) 74 (35–87)AST level, IU/L 71.5 (33–90) 72 (45–115)ALP level, IU/L 346 (115–393) 296 (172–430)Albumin level, g/dL 46 (42–50) 43 (33–44)Platelets, �103/mm3 204 (130–298) 244 (204–327)

IBD, inflammatory bowel disease; UDCA, ursodeoxycholic acid.aValues for continuous variables represent median (interquartile range).bRepresents clinical events within the preceding 2 months.


loss of cd28 expression by liver-infiltrating t cells contributes to pathogenesis of primary...

Documents