pathogenic t cells in murine lupus exhibit spontaneous signaling activity through...

ARTHRITIS & RHEUMATISMVol. 48, No. 4, April 2003, pp 1071–1079DOI 10.1002/art.10900© 2003, American College of Rheumatology

Pathogenic T Cells in Murine Lupus Exhibit SpontaneousSignaling Activity Through Phosphatidylinositol 3-Kinase and

Mitogen-Activated Protein Kinase Pathways

Florin Niculescu,1 Phuong Nguyen,1 Teodora Niculescu,1 Horea Rus,1

Violeta Rus,1 and Charles S. Via2

Objective. To determine the activation status oftwo cytoplasmic signaling pathways, phosphatidylinosi-tol 3-kinase (PI 3-kinase) and the mitogen-activatedprotein kinase (MAPK) family.

Methods. We studied the pathogenic CD4� Tcells that drive disease in the parent-into-F1 mousemodel of lupus-like chronic graft-versus-host disease(GVHD). We determined immunoprecipitated kinaseactivity for PI 3-kinase and MAPK members (Raf-1,extracellular signal–regulated kinase 1 [ERK-1], c-JunN-terminal kinase 1 [JNK-1], and p38 MAPK) fromeither unfractionated splenocytes or purified donorCD4� T cells. Uninjected normal mice served as nega-tive controls, and acute GVHD mice served as positivecontrols.

Results. Compared with negative controls, un-fractionated splenocyte kinase activity from chronicGVHD mice was significantly increased for PI 3-kinaseand JNK-1, but not for Raf-1, p38 MAPK, or ERK-1.Increased PI 3-kinase and JNK-1 activity was also seenin acute GVHD splenocytes, as was increased Raf-1 andp38 MAPK activity. The pattern of increased PI3-kinase and JNK-1 activity seen in unfractionated

chronic GVHD splenocytes was also seen in isolateddonor, but not host, CD4� T cells from chronic GVHDmice, indicating that donor CD4� T cell signalingactivity accounted for at least a portion of the activityobserved in unfractionated splenocytes. IncreasedERK-1 activity was not seen in either donor or hostCD4� T cells. This pattern of cytoplasmic signalingpathway in donor CD4� T cells was associated withincreased T cell receptor membrane signaling activation(Lck and Fyn phosphorylation) and increased tran-scription activation (phosphorylation of inhibitor ofnuclear factor �B), confirming the biologic significanceof these observations.

Conclusion. The pathogenic T cells driving dis-ease in this murine model exhibit activation in the formof spontaneous cytoplasmic signaling pathway activitythat can be detected without in vitro restimulation andinvolves a T cell–specific (PI 3-kinase) and a nonspecificstress/cytokine pathway (JNK-1). These results raise thepossibility that a full characterization of the signalingpathways active in pathogenic lupus T cells might leadto new therapeutic targets.

T cells play a central role in the pathogenesis ofboth human and murine lupus (1–7). For example,pathogenic autoantibodies in lupus are typically IgG, a Tcell–dependent isotype, and exhibit features of anantigen-driven response (e.g., somatic mutation andclonal expansion) (8,9). Moreover, markers of Th cellactivation such as CD40 ligand are increased in peri-pheral blood lymphocytes from patients with systemiclupus erythematosus (SLE) (2,10). Despite evidence ofT cell activation in vivo, in vitro T cell function in lupusis widely reported as depressed, particularly with regardto proliferation, Th1 cytokine production, and cell-

Dr. Niculescu’s work was supported by an Engelicheff Fel-lowship Award from the Maryland Chapter of the Arthritis Founda-tion. Dr. H. Rus’s work was supported by NIH grant NS-42011. Dr.Via’s work was supported by NIH grant AI-47466 and by a Depart-ment of Veterans Affairs Merit Review award.

1Florin Niculescu, MD, PhD, Phuong Nguyen, BS, TeodoraNiculescu, MD, PhD, Horea Rus, MD, PhD, Violeta Rus, MD, PhD:University of Maryland School of Medicine, Baltimore; 2Charles S.Via, MD: University of Maryland School of Medicine, and BaltimoreVA Medical Center, Baltimore, Maryland.

Address correspondence and reprint requests to Charles S.Via, MD, Division of Rheumatology, MSTF 8-34, 10 South PineStreet, Baltimore, MD 21201. E-mail: [email protected].

Submitted for publication September 17, 2002; accepted inrevised form January 8, 2003.

1071

mediated immunity, including cytotoxic T lymphocyte(CTL) function (for review, see ref. 11). Additionally, awide variety of T cell signaling defects have beendescribed in lupus T cells (for review, see refs. 12 and13), including defects of T cell signaling through the Tcell receptor (TCR) and through the protein kinase Apathway (14,15). Typically, in vitro restimulation is re-quired to demonstrate defective T cell function in lupusT cells. Moreover, it has not been resolved whetherdefective in vitro T cell function, particularly defective Tcell signaling, reflects a primary, predisposing factor forlupus or instead represents a compensatory, down-regulatory mechanism in response to excessive in vivo Tcell activation (16).

To directly address this question, it will be nec-essary to study the pathogenic T cells driving B cellhyperactivity in SLE separately from the remainder ofthe T cell pool. Because the putative antigen(s) recog-nized by the pathogenic T cells are not fully defined andmay differ among patients, it has been difficult toidentify and study pathogenic lupus T cells separatelyfrom nonspecifically activated T cells in humans.

In mice, however, this obstacle can be overcomeby using the parent-into-F1 model of lupus-like chronicgraft-versus-host disease (GVHD). In this model, a diseasestrongly resembling human lupus can be induced in normalF1 mice following the injection of normal homozygousparental T cells. Lupus-like GVHD is mediated by acti-vated donor CD4� T cells which recognize host allogeneicmajor histocompatibility complex (MHC) class II mole-cules and produce predominantly Th2 cytokines, resultingin polyclonal B cell hyperactivity, production of autoanti-bodies typical of SLE, and an immune complex glomeru-lonephritis (17–21). In contrast, acute GVHD is observedin parent-into-F1 combinations that involve activation ofboth donor CD4� and CD8� T cells upon recognition ofallogeneic host MHC class II and class I molecules, respec-tively. Disease is characterized by a strong anti-host Th1-mediated CTL response and elimination of host lympho-cytes (17–19).

We have chosen to use this model of lupus to testthe hypothesis that the pathogenic T cells driving diseaseare chronically activated and that intracellular signalingpathways important in T cell activation and maturationwill be active. The status of TCR signaling in lupus Tcells following in vitro restimulation has been wellstudied (12). In this study, we addressed the status ofspontaneous signaling activity in lupus T cells andexamined two cytoplasmic intracellular signaling path-

ways important in T cell activation, differentiation, andeffector function (i.e., the mitogen-activated proteinkinase [MAPK] and phosphatidylinositol 3-kinase [PI3-kinase] pathways). The PI 3-kinase pathway has beenshown to be important in T cell activation followingTCR engagement, costimulation through CD28, or cy-tokine stimulation (22–25). MAPK pathways can beinitiated by a wide variety of agonists, including cyto-kines, stress, growth factors, and antigen–receptor sig-naling. There are 3 major groups of MAP kinases inmammalian cells (for review, see ref. 26): extracellularsignal–regulated kinases (ERK), p38 MAP kinases, andc-Jun N-terminal kinases (JNK), all of which wereexamined in this study. Our results indicate that both PI3-kinase and MAPK signaling pathway activation can bedetected in the pathogenic CD4� T cells without exog-enous stimulation and support a role for these pathwaysin lupus pathogenesis.

MATERIALS AND METHODS

Mice. C57BL/6J (B6), B6 Thyla (Thy-1.1), DBA/2J(DBA), and B6D2F1 (BDF1) male mice ages 6–8 weeks werepurchased from The Jackson Laboratory (Bar Harbor, ME).

Induction of GVHD. Single-cell suspensions of spleno-cytes obtained from normal DBA and B6 mice were preparedin phosphate buffered saline (PBS), filtered through sterilenylon mesh, diluted to a concentration of 108 viable cells/ml(viability determined by trypan blue exclusion), and injectedinto the tail vein of F1 recipients at the following doses: 50 �106 B6 cells (acute GVHD) and 80 � 106 DBA cells (chronicGVHD). For studies of isolated CD4� T cells, chronic GVHDwas induced in BDF1 mice (Thy-1.2) with 12 � 106 CD4� Tcells from B6 Thy-1.1 splenocytes depleted of CD8� T cells bymagnetic bead separation as described below. Control miceconsisted of uninjected age- and sex-matched F1 animals.

Isolation of donor and host CD4� T cells. Splenocytesfrom normal F1 and GVHD mice were harvested at theindicated time points, and donor (Thy-1.1�) and host (Thy-1.2�) CD4� T cells were isolated by negative and positiveselection, respectively. Briefly, erythrocytes and dead cellswere removed from unfractionated splenocytes withLympholyte-M separation medium (centrifugation for 20 min-utes at 1,500 revolutions per minute and 21°C followed by 2washings in Ca/Mg-free PBS). Splenocytes were then depletedof B cells, monocyte/macrophages, and CD8� T cells by serial1-hour incubations with the respective monoclonal antibody–coated magnetic beads (anti-B220, anti-CD14, anti-CD8; Dy-nal, Lake Success, NY) according to the manufacturer’s in-structions. Donor CD4� T cells were then negatively selectedfollowing incubation with anti–Thy-1.2–coated beads and lysedfor immunoprecipitation with lysis buffer (see below). Thepositively selected Thy-1.2�,CD4� T cells of the host were also

1072 NICULESCU ET AL

lysed and the beads removed magnetically. Purity of selectedCD4� T cells was determined by flow cytometry to be �90%.

Signaling studies. The following materials were pur-chased from Santa Cruz Biotechnology (Santa Cruz, CA):Syntide-2, N-79 c-Jun peptide, myelin basic protein 95–98peptide, and rabbit polyclonal IgG antibodies to Raf-1, ERK-1,JNK-1, p38 MAPK, and p85 PI 3-kinase, goat polyclonal IgGanti-mouse ZAP-70, CD3, rabbit polyclonal IgG anti-mouseLck, Fyn, and inhibitor of nuclear factor �B kinase (IKK), andmonoclonal antiphosphotyrosine antibody (PY20). The mono-clonal IgG anti–phospho–inhibitor of nuclear factor �B�(anti–phospho-I�B�; Ser32) antibody was purchased from CellSignaling Technology (Beverly, MA). For all kinase assays,splenocytes were lysed in lysis buffer as described (27), and 200�g of total protein from each lysate was immunoprecipitatedwith specific antibody.

Raf-1 kinase assay. Lysates were immunoprecipitatedby incubation at 4°C overnight with 10 �g of rabbit IgG anti-body to mouse/human Raf-1 and 10 �l of protein A/G–agaroseper sample. The agarose beads were washed 3 times in coldlysis buffer, then suspended in 50 �l of 20 mM PIPES reactionbuffer (pH 7.0) containing 10 mM MnCl2, 1 �l/sample of�32P-ATP (10 �Ci), and 2 �g/sample of Syntide-2 as a specificsubstrate of Raf-1 kinase. Following incubation for 30 minutesat 30°C, Syntide-2 phosphorylation was determined by loading10 �l of the reaction supernatant onto P81 phosphocellulosepaper and counting the radioactivity as described (27).

ERK-1 kinase assay. Cell lysates were immunoprecipi-tated with 10 �g of rabbit IgG antibody to mouse/humanERK-1 and 10 �l of protein A/G–agarose per sample. Theagarose beads were suspended in 50 �l reaction buffer con-taining 1 �l/sample of �32P-ATP (10 �Ci) and 10 �l/samplemyelin basic protein 95–98 peptide for 30 minutes at 30°C. Thepeptide phosphorylation was assessed by loading 10 �l of thereaction supernatant onto P81 phosphocellulose paper, fol-lowed by counting the radioactivity (27).

JNK-1 assay. Cell lysates were incubated at 4°C with 10�l each of anti–JNK-1 IgG antibody and protein A/G–agarose.After washing with lysis buffer, JNK-1 activity was determinedby incubation for 20 minutes at 30°C with 1 �g of N79 c-Junpeptide in the reaction buffer containing �32P-ATP (1 �Ci/sample). Radioactivity incorporated into the N-79 c-Jun pep-tide was then quantitated by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), transferredby blotting, and analyzed by autoradiography of the blot (28).

p38 MAPK assay. Cell lysates were immunoprecipitatedwith anti–p38 MAPK IgG antibody in 50 �l reaction buffer(12.5 mM MOPS [pH 7.5], 12.5 mM �-glycerophosphate, 7.5 mMMgCl2, 0.5 mM EGTA) containing 1 �g of activating transcrip-tion factor 2 as specific substrate for p38 MAPK and �32P-ATP (1�Ci/sample) for 30 minutes at 30°C. Samples were then quanti-tated by 12% SDS-PAGE and analyzed by autoradiography asdescribed (28).

PI 3-kinase assay. Cell lysates were immunoprecipi-tated overnight with 10 �l each of anti–p85 PI 3-kinase IgGantibody and protein A/G–agarose. The beads were washedwith lysis and reaction buffer (20 mM Tris HCl [pH 7.5], 100

mM EDTA, 0.5 mM EGTA, 1 �M ATP, 10 mM MgCl2), andthe kinase reaction was developed in reaction buffer contain-ing 40 �g/sample of PI and �32P-ATP (1 �Ci/sample) for 30minutes at 37°C. The reaction was stopped by the addition of4M HCl, and lipids were extracted with 1:1 chloroform:methanol mixture. Samples were spotted on thin-layer chro-matography plates and developed in an organic cocktail, andPI phosphorylation was calculated from the incorporatedradioactivity as described (28).

Membrane and transcription activation studies. Celllysates were immunoprecipitated overnight as described abovewith antibodies to either Lck or Fyn. Protein phosphorylationwas then assayed by Western blot using antiphosphotyrosineantibody and the enhanced chemiluminescence system (Pierce,Rockford, IL). Nuclear factor �B (NF-�B) activation wasdetermined by Western blot using the anti–phospho-I�B�(Ser32) monoclonal antibody, which identifies the serine 32phosphorylation of the I�B� molecule in the inhibitor ofNF-�B kinase � (IKK�) immunoprecipitate.

Statistical analysis. Data were examined for normalityand equal variance (Kolmogorov-Smirnov test). If these weresatisfactory, groups were compared by Student’s 2-tailed t-test; ifnot, groups were compared by the Mann-Whitney rank sum test.

RESULTS

Increase of PI 3-kinase activity in acute andchronic GVHD. The purpose of these studies was todetermine whether increased signaling activity could bedetected in pathogenic lupus T cells without secondarystimulation in vitro. To address this question, we choseto examine signaling activity in chronic lupus-likeGVHD in the parent-into-F1 model, in which disease ismediated by activated donor CD4� T cells which pro-duce Th2 cytokines. Because of the role played by PI3-kinase and MAPK pathways in both T and B cellactivation and maturation, we hypothesized that thesepathways would be activated in splenocytes from chronicGVHD mice, particularly donor T cells. For a positivecontrol, we also examined signaling activity in spleno-cytes from mice with acute GVHD, a disease character-ized by strong activation of both donor CD4� andCD8� T cells and the production of Th1 cytokines (17).As shown in Figures 1A and B, significantly elevated(�6.5-fold) PI 3-kinase activity was detected in wholesplenocytes from acute GVHD mice 10 days after pa-rental cell transfer (P � 0.001). We were unable todetect significant signaling activity over backgroundfrom chronic GVHD splenocytes on days 10 or 14 afterparental cell transfer (data not shown); however, by 21days, a significant increase (3.8-fold) in PI 3-kinase

SIGNALING ACTIVITY OF PATHOGENIC T CELLS IN MURINE LUPUS 1073

activity over background was observed from chronicGVHD splenocytes (P � 0.001). These results do notidentify the source of the increased activity, but they doattest to the feasibility of our approach.

MAPK pathway activation in GVHD mice: onlyJNK-1 activity increased in chronic GVHD splenocytes;JNK-1, Raf-1, and p38 MAPK activity increased inacute GVHD splenocytes. Using the same experimentalapproach as that shown in Figure 1, whole splenocytesfrom day 10 acute GVHD mice and day 21 chronicGVHD mice were examined for MAPK pathway activa-tion components, including Raf-1, ERK-1, JNK-1, andp38 MAPK. Significantly increased JNK-1 activity wasobserved for both acute (�3-fold; P � 0.05) and chronic(�2-fold; P � 0.005) GVHD mice compared withuntreated F1 mice (Figures 2A and B). No signifi-cant increase in kinase activity for Raf-1, ERK-1, or p38MAPK was seen in chronic GVHD mice (Figures 3A–

C). In contrast, acute GVHD mice exhibited significantelevations in Raf-1 and p38 MAPK activity (�2-fold;both P � 0.05), but no significant increase in ERK-1activity, compared with control mice (Figures 3A–C).

Localization to donor T cells of increased PI3-kinase and JNK-1 activity in chronic GVHD mice.Based on the above data demonstrating our ability todetect increased signaling activity from whole spleen celllysates, we sought to determine the source(s) of theincreased PI 3-kinase and JNK-1 activity in chronicGVHD mice. For these experiments, chronic GVHDwas induced using CD4� donor T cells from B6 Thy-1.1mice, and BDF1 Thy-1.2 mice were used as recipients.Purified donor and host CD4� T cells were thenisolated from chronic GVHD mice on day 21 as de-scribed in Materials and Methods and assayed forspontaneous PI 3-kinase, JNK-1, and ERK-1 activity.Controls consisted of purified CD4� T cells from nor-

Figure 1. Phosphatidylinositol 3-kinase (PI-3K) activity in lysatesof whole splenocytes. Splenocytes from uninjected (control) F1 mice,mice with acute graft-versus-host disease (acute GVHD mice; 10 daysafter parental cell transfer), and chronic GVHD mice (21 days afterparental cell transfer) were prepared, and PI-3K activity was deter-mined as described in Materials and Methods. Spleens from 5 mice pergroup were pooled and tested in triplicate. Results are shown as A,kinase activity as measured by PI phosphorylation in the third position(PIP3) on polyethyleneimine cellulose from a single immunoprecipi-tation and B, mean and SEM kinase activity of each group, from 5different immunoprecipitations and kinase assays from the same lysatepool. � � P � 0.001 versus controls.

Figure 2. Activity of c-Jun N-terminal kinase 1 (JNK1) in lysatesfrom whole spleens. Experimental and control groups are as describedin Figure 1. Spleens from 5 mice per group were tested in duplicate,and JNK1 activity is shown as A, phosphorylation of the N79 c-Junpeptide (the specific substrate of JNK1) tested in duplicate. In B,densitometric analysis was performed on 3 different immunoprecipi-tations and kinase assays from the same lysate pool. Results are shownas the mean and SEM of each group. � � P � 0.05; �� P � 0.005,versus controls.


mal B6 Thy-1.1) or normal F1 Thy-1.2 mice. As shown inFigures 4A and B, PI 3-kinase activity was strikinglyincreased (�6-fold; P � 0.001) in donor (Thy-1.1)CD4� T cells from chronic GVHD mice, whereas nosignificant increase in activity over control cells was seenin host (Thy-1.2) CD4� T cells from chronic GVHDmice. Similarly, an �4-fold increase in JNK-1 activitywas seen for donor (Thy-1.1) CD4� T cells (P �0.0001), with no increase in JNK-1 activity observed forhost (Thy-1.2) CD4� T cells (Figure 5A). Of note,signaling activity from positively selected Thy-1.2 cells(both control and GVHD) did not differ from baselineactivity of negatively selected control Thy-1.1 cells,demonstrating that neither negative nor positive selec-tion alone leads to significant cellular activation. Finally,

no significant increase in ERK-1 activity was observed ineither donor or host CD4� T cells (Figure 5B).

Evidence of membrane activation and transcrip-tion activation in donor CD4� cells. Additional studieswere performed to confirm that the cytoplasmic signal-ing pathways identified above were associated withevidence of membrane and nuclear transcription activa-tion. Using phosphorylation of Lck and Fyn as markersof membrane (TCR) activation, we observed that bothLck and Fyn phosphorylation were increased in donorGVHD Thy-1.1 cells compared with host GVHD Thy-1.2 cells (Figure 6), confirming the involvement of theTCR (Lck phosphorylation 138,704 scan units in GVHDdonor cells, 109,179 scan units in GVHD host cells; Fynphosphorylation 97,364 scan units in GVHD donor cells,

Figure 3. Raf-1, extracellular signal–regulated kinase 1 (ERK1), andp38 mitogen-activated protein kinase (p38 MAPK) activity from wholespleen cell lysates. Experimental groups are as described in Figure 1.For all kinase activity, spleens from 5 mice per group were pooled andtested in quadruplicate, and 3 individual runs were performed from thesame lysate. Kinase activity is as described in Materials and Methodsand shown as the group mean and SEM for A, Raf-1 activity,determined as phosphorylation of Syntide-2, B, ERK1 activity, deter-mined as phosphorylation of myelin basic protein 95–98 peptide, andC, p38 MAPK activity, determined as phosphorylation of activatingtranscription factor 2. � � P � 0.05 versus controls.


50,908 scan units in GVHD host cells; phosphorylationwas �50,000 scan units in all Lck and Fyn controls).Both host and donor T cells were activated comparedwith control CD4� T cells, consistent with previousreports (29,30).

NF-�B is a major nuclear transcription activationfactor in activated T cells, and, following its release fromthe cytoplasmic complex of NF-�B–I�B, it translocatesto the nucleus to induce transcription activation (forreview, see ref. 31). The first step in NF-�B release isI�B� phosphorylation at serine 32 by IKK�. Coprecipi-tation of IKK� and the I�B� substrate of this kinase areshown in Figure 7 for donor and host CD4� T cells fromchronic GVHD mice. The upper panel shows the immu-noprecipitated IKK� (confirming the physical presence

of the kinase). The lower panel shows the phosphoryla-tion at the serine 32 residue of I�B� in the immunopre-cipitated IKK� enzyme (demonstrating its activity).Donor CD4� T cells showed an increase in I�B�phosphorylation (184,574 scan units) compared withhost CD4� T cells (142,917 scan units) or donor andhost control cells (127,392 and 131,293 scan units, re-spectively). Taken together, these findings confirm that

Figure 4. Increased PI-3K activity in donor CD4� T cells fromchronic GVHD mice. Chronic GVHD was induced as described inMaterials and Methods, using C57BL/6J (B6) Thy-1.1 donor CD4� Tcells and B6D2F1 (BDF1) Thy-1.2 hosts. Twenty-one days after celltransfer, donor and host CD4� T cells were isolated from pooledsplenocytes from normal BDF1 Thy-1.2 mice, normal B6 Thy-1.1 mice,and 3 groups of chronic GVHD mice (5 mice pooled per group). PI-3Kactivity was determined as described in Materials and Methods. A,Representative individual results are shown as PI-3K activity in donorand host CD4� T cells from control BDF1 mice and 2 pools of chronicGVHD mice. B, Results are shown as mean and SEM PI-3K activityfor all groups tested in 2 separate immunoprecipitation and kinaseassays. � � P � 0.001 versus controls. See Figure 1 for otherdefinitions.

Figure 5. Extracellular signal–regulated kinase 1 (ERK1) and c-JunN-terminal kinase 1 (JNK1) activity in donor and host CD4� T cellsfrom mice with chronic graft-versus-host disease (GVHD). Experi-mental groups and CD4� T cell isolation are as described in Figure 4.Kinase activity was determined as described in Materials and Methodsand in Figures 2 and 3. Results are shown as the mean and SEM ofspecific substrate phosphorylation for A, JNK1 and B, ERK1. � � P �0.0001 versus controls.


the cytoplasmic signaling mechanisms shown in Figures4 and 5 are associated with membrane signaling andtranscription activation.

DISCUSSION

The present study used a well-defined murinemodel of lupus to address the status of several signalingpathways important in T cell activation and differentia-tion. An important advantage of the GVHD model usedin these studies is that the T cells responsible forinitiating and perpetuating B cell hyperactivity are ofdonor origin and can be studied separately from non-specifically activated T cells of host origin. It has beenpreviously shown that chronic GVHD and lupus-likedisease in this model result from the selective activationof donor CD4� T cells and the production of predom-inantly Th2 cytokines, whereas acute GVHD resultsfrom the activation of both CD4� and CD8� donor Tcells and predominantly Th1 cytokines (17). We initiallyexamined T cell signaling in both acute and chronicGVHD splenocytes to determine whether detection ofsignaling activity without additional restimulation invitro was possible in either disease. Ten days afterGVHD induction, strong spontaneous signaling activitywas readily detectable only in acute GVHD splenocytesand was manifested by increased PI 3-kinase, JNK-1, p38MAPK, and Raf-1 activity. No activity was seen in

chronic GVHD splenocytes until day 21 of disease, whenthere were increases in both PI 3-kinase and JNK-1activity but no increases in p38 MAPK or Raf-1 activityas were seen in acute GVHD.

It is possible that the common activation of PI3-kinase and JNK-1 in both acute and chronic GVHDrepresents the activation of CD4� T cells (seen in bothacute and chronic GVHD), whereas the additional twosignaling pathways activated in acute GVHD but not inchronic GVHD (p38 MAPK and Raf-1) reflect theadditional activation of donor CD8� T cells and Th1cytokine production (seen only in acute GVHD). Fur-ther studies will be needed to fully address this possibil-ity. The results from acute GVHD mice also serve as apositive control and indicate that our inability to detectactivity in unfractionated splenocytes from chronicGVHD mice until day 21 is probably an indicator of thesensitivity of this approach and reflects the reducedpercentage of donor T cells in chronic GVHD mice(2%) compared with that in acute GVHD mice (20%) at10–14 days of disease (32).

Using highly purified donor and host T cells, wewere able to confirm that the increased PI 3-kinase andJNK-1 activity seen in unfractionated splenocytes fromchronic GVHD mice could be localized to the donorCD4� T cells and was not observed in the host T cells.Importantly, signaling pathways that did not exhibitincreased activity in unfractionated chronic GVHDsplenocytes (i.e., ERK-1) did not show increased activityin isolated T cells of either host or donor origin. These

Figure 6. Increased membrane signaling activity in T cells from micewith graft-versus-host disease (GVHD). Experimental groups andCD4� T cell isolation are as described in Figure 4. Cell lysates wereimmunoprecipitated with anti-Lck or anti-Fyn antibodies as describedin Materials and Methods, and phosphorylation was assessed byWestern blotting with the PY20 monoclonal antiphosphotyrosineantibody (P-Tyr) and the enhanced chemiluminescence system. IgGlight chains from the immunoprecipitate (IP) are shown as a controlfor gel loading.

Figure 7. Increased phosphorylation of inhibitor of nuclear factor�B� (I�B�) in donor CD4� T cells from chronic GVHD mice.Experimental groups and T cell isolation are as described in Figure 4.Using cell lysates immunoprecipitated with polyclonal antibody toinhibitor of nuclear factor �B kinase � (IKK�), phosphorylation at theserine 32 residue (P-Ser[32]) of I�B� was assessed by Western blottingwith either anti-IKK� antibody (upper panel) or anti–phospho-I�B�(Ser32) (lower panel), as described in Materials and Methods. SeeFigure 6 for other definitions.


results do not exclude a non–T cell source (especially Bcells) as an additional potential contributor to the in-creased PI 3-kinase and JNK-1 signaling activity ob-served in unfractionated splenocytes, but they do under-score the idea that in this model of lupus in which anoligoclonal population of activated T cells drives disease,T cell signaling through PI 3-kinase and JNK-1 pathwaysis involved.

The full implications of increased PI 3-kinase andJNK-1 activity in the pathogenic T cells driving lupus arenot entirely clear. PI 3-kinase is known to play animportant role in T cell activation following TCR en-gagement, costimulation through CD28, or cytokinestimulation (22–25). The MAPK family of signalingmolecules is important in regulating the activation,expansion, and differentiation of naive CD4� T cellsinto effector Th1 and Th2 cells. Specifically, ERK-1 hasbeen shown to be required for Th2 differentiation (33),whereas p38 MAPK and JNK-1 have important roles inTh1 generation (26,34–36). Our results demonstratingincreased PI 3-kinase activity in donor CD4� T cellsfrom chronic GVHD mice are consistent with a role forpersistent T cell activation in lupus-like disease. Thisconclusion is supported by our data showing increasedactivity for TCR-associated Src-family kinases (Lck andFyn) (Figure 6) and I�B� phosphorylation at serine 32(Figure 7), the latter being a precursor of transcriptionactivation by NF-�B. However, if Th2 cytokines are thecritical effectors in lupus-like GVHD, our results aresomewhat surprising in that donor CD4� T cells exhib-ited an increase in JNK-1, which is reported to promoteTh1 effector cell maturation, but donor CD4� T cellsdid not exhibit significantly increased ERK-1, which isreported to promote Th2 differentiation.

Although interferon-� (IFN�) has been shown toplay an important role in disease pathogenesis in othermodels of murine lupus (37–39), the exact role of IFN�,if any, in the chronic GVHD model is still unclear. Wehave observed striking elevations of serum IFN� in acuteGVHD mice; however, serum IFN� levels in chronicGVHD mice are modestly but significantly increased(40). Studies are currently under way to determinewhether these small elevations in IFN� play a patho-genic role in disease severity in chronic GVHD mice.However, the increased JNK-1 on day 21 may be anindicator that IFN� and/or other Th1 cytokines areinvolved in the pathogenesis of established disease inchronic GVHD. Delineation of the exact role of PI3-kinase and MAPK pathway family members in thepathogenesis of lupus in this model will require a moreextensive kinetic analysis of purified CD4� T cells.

In summary, our results support the concept thatchronic activation of an oligoclonal population ofantigen-specific CD4� T cells is sufficient to drive B cellhyperactivity resulting in lupus. This mechanism may beoperative in at least a subset of human lupus patients.Moreover, the antigen-specific T cells driving disease inthis murine model exhibit a characteristic pattern ofsignaling pathway activity that can be detected withoutin vitro restimulation and involves a T cell–specificpathway (PI 3-kinase) and a nonspecific stress/cytokinepathway (JNK-1). We have recently observed that un-fractionated T cells from human SLE patients alsoexhibit increased spontaneous JNK and PI 3-kinaseactivity (Rus V, Niculescu F, Via C: unpublished obser-vations). These results raise the possibility that a fullcharacterization of the signaling pathways active inpathogenic lupus T cells might lead to new therapeutictargets.

REFERENCES

1. Diamond B, Katz JB, Paul E, Aranow C, Lustgarten D, ScharffMD. The role of somatic mutation in the pathogenic anti-DNAresponse. Annu Rev Immunol 1992;10:731–57.

2. Koshy M, Berger D, Crow MK. Increased expression of CD40ligand on systemic lupus erythematosus lymphocytes. J Clin Invest1996;98:826–37.

3. Burlingame RW, Rubin RL, Balderas RS, Theofilopoulos AN.Genesis and evolution of antichromatin autoantibodies in murinelupus implicates T-dependent immunization with self antigen.J Clin Invest 1993;91:1687–96.

4. Desai-Mehta A, Mao C, Rajagopalan S, Robinson T, Datta SK.Structure and specificity of T cell receptors expressed by poten-tially pathogenic anti-DNA autoantibody-inducing T cells in hu-man lupus. J Clin Invest 1995;95:531–41.

5. Wofsy D, Ledbetter JA, Hendler PL, Seaman WE. Treatment ofmurine lupus with monoclonal anti-T cell antibody. J Immunol1985;134:852–7.

6. Santoro TJ, Portanova JP, Kotzin BL. The contribution of L3T4�T cells to lymphoproliferation and autoantibody production inMRL-lpr/lpr mice. J Exp Med 1988;167:1713–8.

7. Cohen PL, Litvin DA, Winfield JB. Association between endog-enously activated T cells and immunoglobulin-secreting B cells inpatients with active systemic lupus erythematosus. ArthritisRheum 1982;25:168–73.

8. Radic MZ, Weigert M. Genetic and structural evidence forantigen selection of anti-DNA antibodies. Annu Rev Immunol1994;12:487–520.

9. Pisetsky DS. Autoantibodies and their significance. Curr OpinRheumatol 1993;5:549–56.

10. Desai-Mehta A, Lu L, Ramsey-Goldman R, Datta SK. Hyperex-pression of CD40 ligand by B and T cells in human lupus and itsrole in pathogenic autoantibody production. J Clin Invest 1996;97:2063–73.

11. Horwitz DA, Stohl W, Gray JD. T lymphocytes, natural killer cells,and immune regulation. In: Wallace DJ, Hahn BH, editors.Dubois’ lupus erythematosus. Philadelphia: Lippincott Williams &Wilkins; 2002. p. 157–85.

12. Kammer GM, Perl A, Richardson BC, Tsokos GC. Abnormal T


cell signal transduction in systemic lupus erythematosus. ArthritisRheum 2002;46:1139–54.

13. Tsokos GC, Liossis SN. Immune cell signaling defects in lupus:activation, anergy and death. Immunol Today 1999;20:119–24.

14. Vassilopoulos D, Kovacs B, Tsokos GC. TCR/CD3 complex-mediated signal transduction pathway in T cells and T cell linesfrom patients with systemic lupus erythematosus. J Immunol1995;155:2269–81.

15. Kammer GM, Khan IU, Malemud CJ. Deficient type I proteinkinase A isozyme activity in systemic lupus erythematosus Tlymphocytes. J Clin Invest 1994;94:422–30.

16. Georgescu L, Vakkalanka RK, Elkon KB, Crow MK. Interleu-kin-10 promotes activation-induced cell death of SLE lymphocytesmediated by Fas ligand. J Clin Invest 1997;100:2622–33.

17. Rus V, Svetic A, Nguyen P, Gause WC, Via CS. Kinetics of Th1and Th2 cytokine production during the early course of acute andchronic murine graft-versus-host disease: regulatory role of donorCD8� T cells. J Immunol 1995;155:2396–406.

18. Gleichmann E, van Elven EH, van der Veen JP. A systemic lupuserythematosus (SLE)-like disease in mice induced by abnormalT-B cell cooperation: preferential formation of autoantibodiescharacteristic of SLE. Eur J Immunol 1982;12:152–9.

19. Gleichmann E, Pals ST, Rolink AG, Radaszkiewicz T, Gleich-mann H. Graft-versus-host reactions: clues to the etiopathology ofa spectrum of immunological diseases. Immunol Today 1984;5:324–32.

20. Morris SC, Cheek RL, Cohen PL, Eisenberg RA. Autoantibodiesin chronic graft versus host result from cognate T-B interactions. JExp Med 1990;171:503–17.

21. Eisenberg RA, Cohen PL. Class II major histocompatibility anti-gens and the etiology of systemic lupus erythematosus. ClinImmunol Immunopathol 1983;29:1–6.

22. Ward SG, Ley SC, MacPhee C, Cantrell DA. Regulation of D-3phosphoinositides during T cell activation via the T cell antigenreceptor/CD3 complex and CD2 antigens. Eur J Immunol 1992;22:45–9.

23. Ward SG, June CH, Olive D. PI 3-kinase: a pivotal pathway inT-cell activation? Immunol Today 1996;17:187–97.

24. Reif K, Burgering BM, Cantrell DA. Phosphatidylinositol 3-kinaselinks the interleukin-2 receptor to protein kinase B and p70 S6kinase. J Biol Chem 1997;272:14426–33.

25. Brennan P, Babbage JW, Burgering BM, Groner B, Reif K,Cantrell DA. Phosphatidylinositol 3-kinase couples the interleu-kin-2 receptor to the cell cycle regulator E2F. Immunity 1997;7:679–89.

26. Dong C, Davis RJ, Flavell RA. MAP kinases in the immuneresponse. Annu Rev Immunol 2002;20:55–72.

27. Niculescu FI, Rus H, van Biesen T, Shin ML. Activation of Ras

and mitogen-activated protein kinase pathway by terminal com-plement complexes is G protein dependent. J Immunol 1997;158:4405–12.

28. Niculescu F, Badea T, Rus H. Sublytic C5b-9 induces proliferationof human aortic smooth muscle cells: role of mitogen activatedprotein kinase and phosphatidylinositol 3-kinase. Atherosclerosis1999;142:47–56.

29. Via CS, Rus V, Nguyen P, Linsley P, Gause WC. Differentialeffect of CTLA4Ig on murine graft-versus-host disease (GVHD)development: CTLA4Ig prevents both acute and chronic GVHDdevelopment but reverses only chronic GVHD. J Immunol 1996;157:4258–67.

30. Chen F, Maldonado MA, Madaio M, Eisenberg RA. The role ofhost (endogenous) T cells in chronic graft-versus-host auto-immune disease. J Immunol 1998;161:5880–5.

31. Ghosh S, Karin M. Missing pieces in the NF-kappaB puzzle. Cell2002;109 Suppl:S81–96.

32. Via CS, Sharrow SO, Shearer GM. Role of cytotoxic T lympho-cytes in the prevention of lupus-like disease occurring in a murinemodel of graft-vs-host disease. J Immunol 1987;139:1840–9.

33. Yamashita M, Kimura M, Kubo M, Shimizu C, Tada T, PerlmutterRM, et al. T cell antigen receptor-mediated activation of theRas/mitogen-activated protein kinase pathway controls interleukin4 receptor function and type-2 helper T cell differentiation. ProcNatl Acad Sci U S A 1999;96:1024–9.

34. Rincon M, Enslen H, Raingeaud J, Recht M, Zapton T, Su MS, etal. Interferon-gamma expression by Th1 effector T cells mediatedby the p38 MAP kinase signaling pathway. EMBO J 1998;17:2817–29.

35. Constant SL, Dong C, Yang DD, Wysk M, Davis RJ, Flavell RA.JNK1 is required for T cell-mediated immunity against Leishma-nia major infection. J Immunol 2000;165:2671–6.

36. Dong C, Yang DD, Tournier C, Whitmarsh AJ, Xu J, Davis RJ, etal. JNK is required for effector T-cell function but not for T-cellactivation. Nature 2000;405:91–4.

37. Peng SL, Moslehi J, Craft J. Roles of interferon-gamma andinterleukin-4 in murine lupus. J Clin Invest 1997;99:1936–46.

38. Lawson BR, Prud’homme GJ, Chang Y, Gardner HA, Kuan J,Kono DH, et al. Treatment of murine lupus with cDNA encodingIFN-gammaR/Fc. J Clin Invest 2000;106:207–15.

39. Theofilopoulos AN, Koundouris S, Kono DH, Lawson BR. Therole of IFN-gamma in systemic lupus erythematosus: a challengeto the Th1/Th2 paradigm in autoimmunity. Arthritis Res 2001;3:136–41.

40. Via CS, Shustov A, Rus V, Lang T, Nguyen P, Finkelman FD. Invivo neutralization of TNF-alpha promotes humoral autoimmu-nity by preventing the induction of CTL. J Immunol 2001;167:6821–6.


pathogenic t cells in murine lupus exhibit spontaneous signaling activity through...

Documents