Decreased Expression of the CD3{zeta} Chain in T Cells Infiltrating the Synovial Membrane of Patients with Osteoarthritis

Osteoarthritis (OA) is a heterogeneous disease which rheumatologistsconsider to be noninflammatory. However, recent studies suggestthat, at least in certain patients, OA is an inflammatory diseaseand that patients often exhibit inflammatory infiltrates inthe synovial membranes (SMs) of macrophages and activated Tcells expressing proinflammatory cytokines. We report here thatthe expression of CD3 ${zeta}$ is significantly decreased in T cellsinfiltrating the SMs of patients with OA. The CD3 ${zeta}$ chain is involvedin the T-cell signal transduction cascade, which is initiatedby the engagement of the T-cell antigen receptor and which culminatesin T-cell activation. Double immunofluorescence of single-cellsuspensions derived from the SMs from nine patients with OArevealed significantly increased proportions of CD3 ${varepsilon}$ -positive(CD3 ${varepsilon}$ ⁺) cells compared with the proportions of CD3 ${zeta}$ -positive (CD3 ${zeta}$ ⁺)T cells (means ± standard errors of the means, 80.48%± 3.92% and 69.02% ± 6.51%, respectively; P =0.0096), whereas there were no differences in the proportionsof these cells in peripheral blood mononuclear cells (PBMCs)from healthy donors (94.73% ± 1.39% and 93.79% ±1.08%, respectively; not significant). The CD3 ${zeta}$ ⁺ cell/CD3 ${varepsilon}$ ⁺ cellratio was also significantly decreased for T cells from theSMs of patients with OA compared with that for T cells fromthe PBMCs of healthy donors (0.84 ± 0.17 and 0.99 ±0.01, respectively; P = 0.0302). The proportions of CD3 ${varepsilon}$ ⁺ CD3 ${zeta}$ ⁺cells were lower in the SMs of patients with OA than in thePBMCs of healthy donors (65.04% ± 6.7% and 90.81% ±1.99%, respectively; P = 0.0047). Substantial proportions (about15%) of CD3 ${varepsilon}$ ⁺ CD3 ${zeta}$ -negative (CD3 ${zeta}$ ^-) and CD3 ${varepsilon}$ -negative (CD3 ${varepsilon}$ ^-) CD3 ${zeta}$ ^-cells were found in the SMs of patients with OA. Amplificationof the CD3 ${zeta}$ and CD3 ${delta}$ transcripts from the SMs of patients withOA by reverse transcriptase PCR consistently exhibited strongerbands for CD3 ${delta}$ cDNA than for CD3 ${zeta}$ cDNA The CD3 ${zeta}$ /CD3 ${delta}$ transcriptratio in the SMs of patients with OA was significantly lowerthan that in PBMCs from healthy controls (P < 0.0001). Theseresults were confirmed by competitive MIMIC PCR. Immunoreactivitiesfor the CD3 ${zeta}$ protein were detected in the SMs of 10 of 19 patientswith OA, and they were of various intensities, whereas SMs fromall patients were CD3 ${varepsilon}$ ⁺ (P = 0.0023). The decreased expressionof the CD3 ${zeta}$ transcript and protein in T cells from the SMs ofpatients with OA relative to that of the CD3 ${varepsilon}$ transcript is suggestiveof chronic T-cell stimulation and supports the concept of T-cellinvolvement in OA.

INTRODUCTION

Top

Introduction

Osteoarthritis (OA) is a heterogeneous disease (3). Rheumatologistsgenerally consider OA to be a noninflammatory disease (3), althoughpatients with OA often exhibit inflammatory infiltrates in thesynovial membrane (SM) (16, 19, 26, 37, 48, 51, 60). These infiltratesmostly consist of T cells and macrophages (16, 19, 26, 37, 48,51, 60). The infiltrating T cells, at least in advanced OA,have many of the features found in rheumatoid arthritis (RA),including the following: (i) they often exhibit a nodular pattern(51, 60); (ii) they express early, intermediate, and late cellsurface activation antigens (51); (iii) they express a TH1 cytokinepattern (18, 51, 52, 60); and (iv) they contain substantialproportions of oligoclonal T cells [S. R. Scanzello, L. I. Sakkas,N. Johanson, and C. D. Platsoucas, Arthritis Rheum. 42(Suppl.):S257,1999 (abstract); S. R. Scanzello, L. I. Sakkas, N. Johanson,and C. D. Platsoucas, Scand. J. Immunol. 54(Suppl. 1):59, 2001(abstract)], suggesting an antigen-driven immune response. Antigen-drivenactivation of T cells is thought to be important in the pathogenesisof RA (15, 45, 50).

T cells recognize peptides bound to self major histocompatibilitycomplex class I or class II through their alpha/beta ( ${alpha}$ ß)T-cell antigen receptor (TCR). In addition, other ${alpha}$ ßTCR-positive (TCR⁺) T cells that recognize nonpeptide antigenshave been identified (56, 72). The engagement of the TCR withantigen initiates a signal transduction cascade which culminatesin T-cell activation. The CD3 ${zeta}$ chain plays a very important rolein this pathway. Signal transduction involves a series of tyrosinephosphorylations and activation of protein tyrosine kinases(68). One of the earliest events upon TCR engagement with antigenis the phosphorylation of the CD3 ${zeta}$ chain, which leads to activationof the ZAP-70 protein tyrosine kinase (6, 68). ZAP-70 proteintyrosine kinase activates the phospholipase C ${gamma}$ 1, which in turnincreases intracellular Ca²⁺ and activates protein kinase C(49, 68).

Defective signal transduction with diminished tyrosine phosphorylationof the CD3 ${zeta}$ chain was found in synovial fluid (SF) T cells frompatients with RA, and this was associated with decreased CD3 ${zeta}$ chain protein expression (32). Decreased CD3 ${zeta}$ protein expressionwas also reported in peripheral blood T cells from patientswith RA (31) and systemic lupus erythematosus (SLE) (27) andin tumor-infiltrating lymphocytes (TILs) from patients withrenal (12), colorectal (36, 38), and ovarian [23; J. Pappas,A. D. Wolfson, C. W. Helm, D. P. Barton, A. D. Tsygankov, andC. D. Platsoucas, FASEB J. 10:A1472, 1996 (abstract); J. Pappas,A. D. Wolfson, W. Jung, C. W. Helm, A. D. Tsygankov, and C.D. Platsoucas, J. Allergy Clin. Immunol. 99:S447, 1997 (abstract)]carcinoma and Hodgkin's disease (47). In the study describedin this report, we investigated CD3 ${zeta}$ chain expression in theSMs from patients with OA at the transcript and protein levels.

MATERIALS AND METHODS

Top

Materials and Methods

Patients. Nineteen patients with OA (35) 9 females, 10 males; mean age,57 years; age range, 36 to 70 years) (35) and nine patientswith RA (7 females, 2 males; mean age, 64.8 years; age range,51 to 74 years) were included in this study. Patients with OAhad primary OA (17 patients) or secondary OA due to avascularnecrosis (2 patients). All patients with OA were treated withnonsteroidal anti-inflammatory drugs, whereas patients withRA were treated with nonsteroidal anti-inflammatory drugs andintramuscular gold (three patients) or methotrexate (one patient).Nine healthy controls (two females, seven males; mean age, 33.9years; age range, 25 to 49 years) were also included in thestudy. SM specimens were obtained at surgery during joint replacement.The use of these specimens has been approved by the InstitutionalReview Board of Temple University Hospital.

Preparation of single-cell suspensions from SMs. SM tissues were finely minced with scissors and digested at37°C for 2 h in RPMI 1640 culture medium supplemented with5% fetal calf serum, garamycin (10 µg/ml), collagenaseI (1.5 mg/ml; Sigma), collagenase IV (0.25 mg/ml; Sigma), andDNase I (0.15 mg/ml; Sigma). The digests were sequentially passedthrough a 200-µm-pore-size sieve (Fisher Scientific) anda 70-µm-pore-size sieve (Becton Dickinson). Single-cellsuspensions were collected by centrifugation on a Ficoll-Hypaquedensity cushion. Peripheral blood mononuclear cells (PBMCs)from healthy donors were used as a methodological control inthis study and were prepared by centrifugation on a Ficoll-Hypaquedensity cushion. These cells were immediately used for immunofluorescencestaining.

Antibodies. A phycoerythrin (PE)-conjugated anti-CD3 ${varepsilon}$ monoclonal antibody(MAb; mouse immunoglobulin G1 [IgG1]; clone UCHT1; PharMingen,San Diego, Calif.) and an anti-CD3 ${zeta}$ MAb (mouse IgGl; clone 2H2D9[Tia-2];Coulter/Immunotech) were used for flow cytometry. The anti-CD3 ${varepsilon}$ MAb (clone UCHT1; Novocastra, Newcastle upon Tyne, United Kingdom)and anti-CD3 ${zeta}$ MAb (Coulter) were used for immunohistochemistry.

Immunofluorescence and flow cytometry. Cells (10⁶/ml) were at first fixed in 4% paraformaldehyde inphosphate-buffered saline (PBS) at room temperature for 10 minand permeabilized by incubation at room temperature for 7 minin 0.1% saponin-0.2% bovine serum albumin-PBS solution. Doublelabeling for CD3 ${varepsilon}$ -positive (CD3 ${varepsilon}$ ⁺) and CD3 ${zeta}$ -positive (CD3 ${zeta}$ ⁺) cellswas carried out by standard protocols. Briefly, the cells wereincubated with anti-CD3 ${zeta}$ MAb and then with fluorescein isothiocyanate(FITC)-conjugated anti-mouse IgG (indirect immunofluorescence).Next, the cells were incubated with PE-conjugated anti-CD3 ${varepsilon}$ MAb(direct immunofluorescence). The cells were incubated in a totalvolume of 100 µl on ice and were washed at 4°C in0.1% saponin-0.2% bovine serum albumin-PBS. Negative controlsincluded cells incubated alone without MAb or IgG antibodiesand cells incubated with FITC-conjugated anti-mouse IgG alone(omission of primary antibodies). Fluorescence-activated cellsorter analysis, after gating for lymphocytes only, was carriedout within 3 h of staining.

Immunohistochemistry. Six-micrometer-thick cryostat SM sections were air dried for1 h and fixed in 4% paraformaldehyde in PBS for 30 min. Endogenousperoxidase activity was blocked by incubation with 0.3% H₂O₂in ice-cold methanol, and the cells were permeabilized by incubationin 0.1% saponin-2% fetal calf serum-PBS. Then the sections werestained for CD3 ${zeta}$ by the avidin-biotin indirect immunoperoxidasemethod as described previously (51), with the exception thatincubations and washings were carried out in 0.1% saponin-PBS.Staining for CD3 ${zeta}$ was carried out as described previously (48).

RT-PCR. SM tissue samples were homogenized in tissue grinders (Kontes),and total RNA was obtained by using RNAzol B (Tel-Test). First-strandcDNA was synthesized from 5 µg of RNA by using SuperscriptII reverse transcriptase (RT) and 0.5 µg of oligo(dT)as a primer (GibcoBRL) and was kept at -30°C until it wasused. CD3 ${delta}$ and CD3 ${zeta}$ cDNAs were detected by 35-cycle PCR, witheach cycle consisting of the following incubation steps: 94°Cfor 45 s, 61°C for 45 s, and 72°C for 90 s, with a finalextension at 72°C for 7 min, as described previously (51).The primers (5' to 3') used for CD3 ${delta}$ amplification were CTGGACCTGGGAAAACGCATC(5' end) and GTACTGAGCATCATCTCGATC (3' end), and those usedfor CD3 ${zeta}$ amplification were ACAGAGCTTTGGCCTGCTGGATC (5' end)and TAGGTGTCCTTGGTGGCTGTACT (3' end) (Bio-Synthesis) (17). Theprimer sequences span at least one intron, so a PCR productof a larger size would be obtained if any contaminating DNAwas present. In addition, ß-actin was also amplifiedas a control (51). After gel electrophoresis the PCR productswere visualized under UV light, and the densities of the bandswere measured by densitometry with Sigma gel software. In someexperiments the PCR products were also quantitated by a competitiveMIMIC PCR (53)

Quantitative PCR. CD3 ${zeta}$ and CD3 ${delta}$ transcripts from SM specimens from five patientswith OA were quantitated by competitive MIMIC PCR, as describedpreviously (38, 51, 53). Internal nonhomologous competitivefragments (MIMIC DNA) for each transcript were constructed accordingto the instructions of the supplier (Clontech, Palo Alto, Calif.).Serial PCR mixtures containing a constant amount of sample cDNA(50 ng of RNA equivalents) were spiked with decreasing concentrationsof MIMIC DNA. The PCR products were separated by electrophoresisthrough an ethidium bromide-stained 1.6% agarose gel and quantitatedby comparison of the intensities of the transcript and the MIMICbands. When the intensities of the target DNA and the MIMICDNA product bands were equal, the amount of target DNA was equimolarto the amount of MIMIC DNA prior to amplification.

RESULTS

Top

Results

Double immunofluorescence staining with an anti-CD3 ${zeta}$ MAb andan anti-CD3 ${varepsilon}$ MAb, followed by flow cytometry analysis of single-cellsuspensions derived from the SMs from nine patients with OA,revealed significantly increased proportions of CD3 ${varepsilon}$ ⁺ T cellscompared with the proportions of CD3 ${zeta}$ ⁺ T cells (means ±standard error of the means [SEMs], 80.48% ± 3.92% and69.02% ± 6.51%, respectively; P = 0.0096) (Table 1).In contrast, there were no significant differences in the proportionsof CD3 ${varepsilon}$ ⁺ and CD3 ${zeta}$ ⁺ T cells (means ± SEMs, 94.73% ±1.39% and 93.79% ± 1.08%, respectively; not significant)present in PBMCs from healthy donors (Table 1). The proportionsof CD3 ${varepsilon}$ ⁺ and CD3 ${zeta}$ ⁺ T cells in the SMs of patients with OA weresignificantly lower than those found in PBMCs from healthy donors(means ± SEMs for CD3 ${varepsilon}$ ⁺ T cells, 80.48% ± 3.92%and 94.73% ± 1.39%, respectively; P = 0.0064; means ±SEMs for CD3 ${zeta}$ ⁺ T cells, 69.02% ± 6.51% and 93.79% ±1.08%, respectively; P = 0.0051) (Table 1). Significant differenceswere also found in the CD3 ${zeta}$ ⁺ cell/CD3 ${varepsilon}$ ⁺ cell ratio, which wasdecreased for T cells from the SMs of patients with OA comparedto that for T cells from the PBMCs of healthy donors (means± SEMs, 0.84 ± 0.06 and 0.99 ± 0.01; respectively,P = 0.0302) (Table 1).

View this table:
[in this window]
[in a new window]
TABLE 1. Proportions of CD3 ${varepsilon}$ ⁺ and CD3 ${zeta}$ ⁺ cells from the SMs of patients with OA and peripheral blood of healthy individuals (controls) analyzed by double immunofluorescence and flow cytometry

The proportions of CD3 ${varepsilon}$ ⁺ CD3 ${zeta}$ -negative (CD3 ${zeta}$ ^-) cells were significantlyhigher in the SMs of patients with OA than in the PBMCs of healthydonors (means ± SEMs, 15.39% ± 3.27% and 3.92%± 0.99%, respectively; P = 0.0079) (Table 2), suggestingthe presence of a sizable population of CD3 ${varepsilon}$ ⁺ CD3 ${zeta}$ ^- cells in theSMs of patients with OA. In contrast, the proportions of CD3 ${varepsilon}$ ⁺CD3 ${zeta}$ ⁺ cells were significantly lower in the SMs of patients withOA than in the PBMCs of healthy donors (means ± SEMs,65.04% ± 6.70% and 90.81% ± 1.99%, respectively;P = 0.0047) (Table 2). No significant differences in the proportionsof CD3 ${varepsilon}$ ^- CD3 ${zeta}$ ⁺ cells were found in the SMs of patients with OAand the PBMCs from healthy donors (means ± SEMs, 3.93%± 0.87% and 2.98% ± 1.01%, respectively; P = 0.4849)(Table 2). Significant differences were observed in the proportionsof CD3 ${varepsilon}$ ^- CD3 ${zeta}$ ^- cells found in the SMs of patients with OA andin PBMCs from healthy donors (means ± SEMs, 15.60% ±3.58% and 2.29% ± 0.57%, respectively; P = 0.0058) (Table2), also suggesting the presence of a sizable population ofCD3 ${varepsilon}$ ^- CD3 ${zeta}$ ^- cells in the SMs of patients with OA. Comparison ofthe proportions of the CD3 ${varepsilon}$ ⁺ CD3 ${zeta}$ ^- cells to those of the CD3 ${varepsilon}$ ^-CD3 ${zeta}$ ⁺ cells in the SMs of patients with OA revealed significantlyincreased proportions of CD3 ${varepsilon}$ ⁺ CD3 ${zeta}$ ^- cells (means ± SEMs,15.39% ± 3.27% and 3.93% ± 0.87%, respectively;P = 0.0096) (Table 2). In contrast, there were no differencesin the proportions of CD3 ${varepsilon}$ ⁺ CD3 ${zeta}$ ^- cells and CD3 ${varepsilon}$ ^- CD3 ${zeta}$ ⁺ cells inPBMCs from healthy donors (means ± SEMs, 3.92% ±0.99% and 2.98% ± 1.01%, respectively; P = 0.3317) (Table2). The results of representative experiments by double immunofluorescenceand flow cytometry analysis of CD3 ${varepsilon}$ ⁺ CD3 ${zeta}$ ^-, CD3 ${varepsilon}$ ⁺ CD3 ${zeta}$ ⁺, CD3 ${varepsilon}$ ^- CD3 ${zeta}$ ⁺,and CD3 ${varepsilon}$ ^- CD3 ${zeta}$ ^- populations of mononuclear cells infiltratingthe SMs of three patients with OA (patients 7, 3, and 4) areshown in Fig. 1. The results of representative experiments forthe same cell populations from PBMCs from three healthy controlsare also shown.

View this table:
[in this window]
[in a new window]
TABLE 2. CD3 ${varepsilon}$ ⁺ CD3 ${zeta}$ ^-, CD3 ${varepsilon}$ ⁺ CD3 ${zeta}$ ⁺, CD3 ${varepsilon}$ ^- CD3 ${zeta}$ ⁺, and CD3 ${varepsilon}$ ^- CD3 ${zeta}$ ^- cells from the SMs of patients with OA and peripheral blood of healthy individuals (controls) analyzed by double immunofluorescence and flow cytometry

View larger version (72K):
[in this window]
[in a new window]
FIG. 1. Results of representative experiments by double immunofluorescence and flow cytometry analysis of CD3 ${varepsilon}$ ⁺ CD3 ${zeta}$ ^-, CD3 ${varepsilon}$ ⁺ CD3 ${zeta}$ ⁺, CD3 ${varepsilon}$ ^- CD3 ${zeta}$ ⁺, and CD3 ${varepsilon}$ ^- CD3 ${zeta}$ ^- populations of mononuclear cells infiltrating the SMs of three patients with OA (patients 7, 3, and 4, from left to right, respectively) (OA-SM) and of peripheral blood mononuclear cells from three healthy controls (C-PB). The y axis shows CD3 ${varepsilon}$ staining, and the x axis shows CD3 ${zeta}$ staining. Immunofluorescence staining was carried out as described in Materials and Methods.

Amplification of the CD3 ${zeta}$ and CD3 ${delta}$ transcripts by RT-PCR showedinvariably stronger bands for the CD3 ${delta}$ cDNA than for the CD3 ${zeta}$ cDNA. Representative results for 11 patients with OA are shownin Fig. 2. The CD3 ${zeta}$ /CD3 ${delta}$ transcript ratio in SMs from patientswith OA (n = 11) was significantly lower than the CD3 ${zeta}$ /CD3 ${delta}$ transcriptratio in PBMCs from healthy controls (means ± SEMs, 0.063± 0.028 and 0.787 ± 0.023, respectively; P <0.0001) (Fig. 3A). These results for five individuals were confirmedby semiquantitative competitive MIMIC PCR, as described previously(51) (data not shown). The CD3 ${zeta}$ /CD3 ${delta}$ transcript ratio in SMs frompatients with RA (n = 9) was also significantly lower than theCD3 ${zeta}$ /CD3 ${delta}$ transcript ratio in peripheral blood from healthy controls(means ± SEMs, 0.327 ± 0.068 and 0.787 ±0.023, respectively; P < 0.0001) (Fig. 3B).

View larger version (28K):
[in this window]
[in a new window]
FIG. 2. Amplification of reverse-transcribed cDNA of CD3 ${zeta}$ and CD3 ${delta}$ chains from the SMs of 11 patients with OA (lanes 1 to 11) and from the peripheral blood of a healthy donor (lane 12). Lane M, 100-bp ladder.

View larger version (45K):
[in this window]
[in a new window]
FIG. 3. (A) CD3 ${zeta}$ /CD3 ${delta}$ transcript ratio in the SMs of patients with OA (OA SM) and peripheral blood (PB) from healthy controls (P < 0.0001). (B) CD3 ${zeta}$ /CD3 ${delta}$ transcript ratio in the SMs of patients with RA (RA SM) and peripheral blood from healthy controls (P < 0.0001).

The expression of CD3 ${zeta}$ and CD3 ${varepsilon}$ proteins was investigated by immunohistochemicalstaining of frozen sections of the SMs from 19 patients withOA and 11 patients with RA by using the anti-CD3 ${zeta}$ MAb or theanti-CD3 ${varepsilon}$ MAb, respectively, as described in Materials and Methods,and the ABC indirect immunoperoxidase method. Immunoreactivityfor CD3 ${zeta}$ was detected in the SMs of 10 of 19 patients with OAand the SMs of 9 of 11 patients with RA. The intensity of theimmunoreactivity for CD3 ${zeta}$ , when present, was variable. As anticipated,CD3 ${zeta}$ protein expression was not observed in those specimens thatlacked expression of CD3 ${zeta}$ transcripts. However, certain specimensthat expressed CD3 ${zeta}$ transcripts (often at low levels) did notexpress the CD3 ${zeta}$ protein, suggesting the involvement, among othermechanisms, of a transcriptional mechanism in the downregulationof CD3 ${zeta}$ protein expression. In contrast, immunoreactivity forCD3 ${varepsilon}$ was detected in all patients with OA and RA. These resultsare summarized in Table 3. The results of a representative immunostainingexperiment are shown in Fig. 4.

View this table:
[in this window]
[in a new window]
TABLE 3. Immunoreactivities of SMs from patients with OA or RA after staining with either anti-CD3 ${zeta}$ MAb or anti-CD3 ${varepsilon}$ MAb

View larger version (142K):
[in this window]
[in a new window]
FIG. 4. Immunoreactivities for CD3 ${zeta}$ (A) and CD3 ${varepsilon}$ (B) in the SMs of patients with OA. ABC immunoperoxidase staining was performed as described in Materials and Methods.

DISCUSSION

Top

Discussion

We report here that the expression of CD3 ${zeta}$ chain transcriptsand protein in the SMs from patients with OA is decreased. Decreasedexpression of the CD3 ${zeta}$ chain in patients with OA may be of functionalsignificance because the CD3 ${zeta}$ chain plays a significant rolein antigen-specific T-cell activation (40, 62) This decreasedexpression of the CD3 ${zeta}$ chain is associated with anergy and T-cellhyporesponsiveness and may reflect chronic antigenic stimulationof T cells [12, 22, 23, 27, 31, 32, 36, 38, 43, 47, 61, 73;Pappas et al., FASEB J. 10:A1472, 1996 (abstract)]. Althoughsuch antigens have not been identified in patients with OA,evidence strongly suggesting that such an antigen-specific chronicT-cell activation may occur in patients with OA and that T cellsmay play an important role in the pathogenesis of the diseasehas been accumulating.

Mononuclear cell infiltrates consisting primarily of CD3⁺ Tcells and monocytes (16, 19, 26, 37, 48, 51, 60) are often observedin the SMs of patients with OA (16, 19, 26, 37, 49). The T cellsinfiltrating the SMs of patients with OA often exhibit a nodularpattern (51, 60) and express certain early (CD69), intermediate(CD25), and late (CD45RO, HLA-DR) activation antigens. Theseinfiltrating T cells are often angiocentric [L. I. Sakkas, C.R. Scanzello, C. D. Katsetos, N. A. Johanson, and C. D. Platsoucas,Rheumatology 39(Suppl.):117, 2000 (abstract)], and their presenceis associated with the activation of vascular endothelial cells,as determined by the expression of E-selectin (5, 20, 63).

To determine whether a specific antigen-driven clonal immuneresponse is taking place in the SMs of patients with OA, ß-chainT-cell receptor (TCR) transcripts were amplified from the SMsof patients with OA by the nonpalindromic adaptor PCR methodor by Vß-specific PCR (54, 59). The amplified transcriptswere cloned and sequenced. Sequence analysis revealed substantialproportions of identical copies of ß-chain TCR transcripts,suggesting the presence of oligoclonal expansions of T cellsin the SMs from five of five patients with advanced OA [Scanzelloet al., Arthritis Rheum. 42(Suppl.):S257, 1999 (abstract); Scanzelloet al., Scand. J. Immunol. 54(Suppl. 1):59, 2001 (abstract)],strongly supporting a specific antigen-driven immune response(s)(C. Scanzello, L. I. Sakkas, B. Xu, C. N. Johanson, and C. D.Platsoucas, submitted for publication). The antigen(s) thatelicits these T-cell responses is not known. However, an autoimmuneresponse to chondrocyte SM components has been reported in patientswith OA (2). Inflammation in patients with OA may not be restrictedto the joints. Perivascular lymphocytic infiltrates have beenreported in muscle biopsy specimens from 18% of the patientswith OA examined (67).

Incidental infections with common herpesviruses may enrich thepool of activated T cells in the SMs of patients with OA. CD8⁺T cells specific for Epstein-Barr virus (EBV) and cytomegalovirusare present at an increased frequency in the SF of patientswith RA (57, 65) and certain patients with OA (57, 65) as wellas in the affected organs of patients with various chronic inflammatorydiseases (57). However, EBV gene expression is not altered inthe SMs of patients with RA or OA, despite the presence of EBVantigen-specific T-cell clones (9). The viral peptide-specificCD8⁺ T cells comprised approximately 4% of the total CD8⁺ Tcells in the SF of patients with RA (11) and were oligoclonal(11). However, the clonality of the remaining T cells that didnot react with these viral epitopes was not examined in thatstudy (11). In addition, the method used to determine the proportionsof these viral peptide-specific CD8⁺ T cells, HLA class I peptidetetramer staining, provided estimates of CD8⁺ T-cell frequenciesthat were, on average, 4.4-fold higher than the frequenciesobtained by enzyme-linked immunospot assays (64). These estimateswere, in turn, on average 5.3-fold higher than the frequenciesobtained by limiting dilution analysis (64). Although the viralpeptide-specific CD8⁺ T cells that are present in the SMs ofthese patients may exhibit decreased expression of the CD3 ${zeta}$ antigen,their presence in relatively low proportions (see above) alonecannot account for the decreases in the proportions of CD3 ${zeta}$ ⁺cells found in the SMs of patients with OA in this study.

It appears that the recruitment of virus-specific CD8⁺ T cellsin inflammatory joints of patients with OA is chemokine driven.T cells infiltrating the SMs of patients with OA are of theTH1 type (51) and express high levels of the chemokine receptorCCR5 on their surfaces (29). The ligand of CCR5, the chemokinemacrophage inflammatory protein 1ß, is upregulatedin the SF of patients with OA (21). In the SF of patients withRA, the vast majority of virus-specific CD8⁺ T cells expressthe chemokine receptor CCR5 (11).

It is also possible that in the SMs of these patients the virus-specificT cells recognize self-antigenic epitopes as being of viralorigin by molecular mimicry. Molecular mimicry is defined asthe presence of epitopes with substantial structural homologybetween host proteins and the proteins of microorganisms suchas viruses, bacteria, and other organisms (13, 41, 42, 69).Therefore, a host immune response against a viral epitope mayrecognize as non-self a host epitope with substantial structuralhomology, even after the microorganism that elicited the immuneresponse is not present. These molecular mimicry responses maylead to the development of autoimmune disease. Along these lines,it has been reported that T cells specific for an EBV antigenalso respond to a self-peptide from a serine-threonine kinase(34). Intermolecular or intramolecular epitope spreading isanother mechanism that can lead to the development of T-cellresponses in the SMs of patients with OA (7, 24, 25, 58, 71).Epitope spreading is defined as the generation of de novo immuneresponses to new self-antigenic epitopes which are differentand which do not share structural homology (cross-reactivity)to those that elicited the initial immune response. Epitopespreading is generated in the environment of chronic inflammation.Epitope spreading was first identified in autoimmune demyelinatingdiseases of the central nervous system (1, 24, 25, 42, 58, 71)and is now identified in antitumor immunity (10) and the immuneresponse against organ grafts (7).

The SMs of patients with OA exhibit a TH1 cytokine profile (51,52, 60). It could be argued that this TH1 response may be drivenprimarily by nonspecific functions of macrophages instead ofT-cell responses to a specific antigen(s). Macrophages and synoviocytesproduce interleukin-12 (52), a potent cytokine that drives theTH1 immune response by inducing the production of proinflammatorycytokines. Increased concentrations of macrophage inflammatoryprotein-1ß (MIP-1ß) have been reported inthe SF of patients with OA (21). MIP-1ß is a ligandfor the chemokine receptor CCR5, which is expressed on TH1 cells(29). In a recent review, Pelletier et al. (46) proposed thatOA is an inflammatory disease and cited macrophages as the exclusivesource of inflammation in the pathogenesis of OA, ignoring therole of T cells (46, 55). Although the mononuclear cell infiltrationand the TH1 response may be explained by a nonspecific activationof T cells, such as cytokine-induced activation, our findings(see above) of T-cell oligoclonality in the SMs of five of fivepatients with advanced OA [Scanzello et al., Arthritis Rheum.42(Suppl.):S257, 1999 (abstract); Scanzello et al., Scand. J.Immunol. 54(Suppl. 1):59, 2001 (abstract)] strongly suggestthat T cells have undergone antigen-driven proliferation andclonal expansion in situ in the SMs of patients with OA in responseto an as yet unidentified antigen(s). Although the nature andcharacteristics of the antigen(s) remain to be determined, itis clear that the view traditionally held by many rheumatologistsand scientists alike that OA is a noninflammatory disease shouldbe abandoned and that the disease should be reclassified.

In a manner similar to that for other conditions characterizedby chronic T-cell activation (22), such as RA (31, 32), SLE(27), lepromatous leprosy (73), human immunodeficiency virusinfection (66), renal carcinoma (12), colorectal carcinoma (36,38), ovarian carcinoma [23; Pappas et al., FASEB J. 10:A1472,1996 (abstract); Pappas et al., J. Allergy Clin. Immunol. 99:S447,1997 (abstract)] and Hodgkin's disease (47), T cells from theSMs of patients with OA exhibit decreased expression of CD3 ${zeta}$ chain transcripts and protein. Although studies with T cellsfrom the SMs of patients with OA have not been reported, thisdecreased expression of the CD3 ${zeta}$ chain in these other diseasesis associated with anergy and T-cell hyporesponsiveness. Inpatients with RA, T cells from the peripheral blood and theSMs are functionally deficient or "frustrated," although theseT cells express early, intermediate, and late activation antigens(19, 51). Additionally, these T cells from patients with RAexhibit decreased in vitro proliferation and impaired increasesin intracellular Ca²⁺ levels after stimulation with mitogensand recall antigens (for a review, see reference 1). Similarly,in TILs [23, 28; Pappas et al., FASEB J. 10:A1472, 1996 (abstract)],peripheral blood T cells of patients infected with human immunodeficiencyvirus (66), and tumor-bearing mice (40), the decreased expressionof CD3 ${zeta}$ was associated with reduced T-cell proliferative responsesand decreased phosphorylation of the CD3 ${zeta}$ chain [23, 28; Pappaset al., FASEB J. 10:A1472, 1996 (abstract)]. Decreased CD3 ${zeta}$ chainexpression in T cells from the SF of patients with RA was accompaniedby decreased phosphorylation of the CD3 ${zeta}$ chain (32). Decreasedphosphorylation of the CD3 ${zeta}$ chain was also observed in antigen-inducedT-cell death (43) and in T cells that are in anergic state (30,33).

The hyporesponsiveness of T cells in patients with RA couldbe also attributed to chronic exposure to tumor necrosis factoralpha (TNF- ${alpha}$ ), since it was reversed in vivo by anti-TNF- ${alpha}$ antibodytreatment (8). The effect of anti-TNF- ${alpha}$ antibody treatment onCD3 ${zeta}$ chain expression was not investigated in that study (8).Apart from a chronic antigenic stimulation, other factors mayalso contribute to decreased CD3 ${zeta}$ chain expression. For example,activated macrophages can induce decreased CD3 ${zeta}$ chain expressionin T cells by cell contact-dependent interaction (4) or oxidativestress (44).

The defect in CD3 ${zeta}$ chain expression appears to be reversible,and it was restored in TILs after stimulation with interleukin-2[70; Pappas et al., J. Allergy Clin. Immunol. 99:S447, 1997(abstract)]. Similarly, CD3 ${zeta}$ chain expression was restored inperipheral blood T lymphocytes in patients with Hodgkin's diseaseafter stimulation with anti-CD3 and anti-CD28 MAbs (47). Inpatients with SLE, the loss of CD3 ${zeta}$ chain expression was foundto be associated with an increased frequency of alternativesplicing of the CD3 ${zeta}$ chain RNA (39). However, one study reportedthat the CD3 ${zeta}$ chain can serve as a substrate for caspase 3, andthe decreased expression of the CD3 ${zeta}$ chain was associated withapoptosis of TILs (14). Furthermore, two populations of T cells,CD3 ${varepsilon}$ ⁺ CD3 ${zeta}$ ^- cells and CD3 ${varepsilon}$ ^- CD3 ${zeta}$ ^- cells, were found to be presentin substantial proportions (about 15%) in the SMs of patientswith OA. These two populations of T cells are virtually absentfrom the peripheral blood of healthy donors. Their functionis not known at present. Also, it is not known whether thesepopulations can differentiate to CD3 ${varepsilon}$ ⁺ CD3 ${zeta}$ ⁺ T cells.

In conclusion, our finding of decreased expression of the CD3 ${zeta}$ chain by T lymphocytes from the SMs of patients with OA is suggestiveof a chronic antigenic T-cell stimulation and reinforces theconcept of T-cell involvement in OA.

ACKNOWLEDGMENTS

We thank Norman Johanson (formerly of the Department of Orthopedics,Temple University School of Medicine, and presently of the Departmentof Orthopedics, Hannemann University, Philadelphia, Pa.) forproviding specimens for this study.

This work was supported in part by grant T32 AI07101 from theNational Institutes of Health.

FOOTNOTES

* Corresponding author. Mailing address: Department of Microbiology and Immunology, Temple University School of Medicine, Philadelphia, PA 19140. Phone: (215) 707-7929. Fax: (215) 707-7788. E-mail: chris.platsoucas{at}temple.edu.

Present address: University of Thessaly, School of Medicine,Larisa 41222, Greece.

REFERENCES

Top