Prolonged Exposure of T Cells to TNF Down-Regulates TCRζ and Expression of the TCR/CD3 Complex at the Cell Surface

Derivation, culture, and activation of T cell hybridomas

Mouse T cell hybridomas specific for human cartilage glycoprotein-39 (HCgp-39) were derived following immunization of HLA-DRαβ1*0401, human CD4 double-transgenic, MHC class II (I-Aβ-chain)-deficient mice with native Ag as previously described (31, 32). Peptide-specific T cell hybridomas were cloned by limiting dilution and propagated in complete medium (RPMI supplemented with 10% heat-inactivated FCS, 2 mM l-glutamine, 100 U/ml penicillin/streptomycin, 50 μM 2-ME, 1 mM sodium pyruvate, and 25 mM HEPES) as previously described (32). To study the effects of chronic cytokine exposure, T cells were cultured in the presence or the absence of recombinant mouse TNF (mTNF; B. Scallon, Centocor, Malvern, PA), human TNF (hTNF; Z. Kaymakcalan, BASF, Worcester, MA) or human IL-1α (hIL-1; U. Gubler, Hoffmann-La Roche, Nutley, NJ) for the indicated periods of time. Cytokines were added to cultures every second day, and the cells were washed extensively before stimulation. Where indicated, N-acetylcysteine (NAC; Sigma, Poole, U.K.) was added to cultures at a final concentration of 1 mM together with TNF. For T cell activation, 1 × 10⁵ T cells were stimulated for 24 h with peptides derived from HCgp-39 and 4 × 10⁵ EBV-transformed B cells homozygous for HLA-DRαβ1*0401 as APC in round-bottom 96-well plates or by plate-bound anti-CD3ε mAb. T cells were stimulated in the absence of cytokines, and all assays were performed in duplicate. IL-2 production in culture supernatants was determined by specific immunoassay using rat anti-mouse IL-2 mAb pairs (PharMingen, San Diego, CA). The assay was developed using streptavidin-europium detection system (Wallac Oy, Turku, Finland), and the results were analyzed on a Wallac fluorescence plate reader. The data are presented as IL-2 fluorescence units ± SD.

The HCQ6 mouse T cell hybridoma, specific for type II collagen and restricted to I-A^q, was transduced with a chimeric cell surface receptor by retroviral gene transfer using the pBabe retroviral vector as previously described (33). The chimeric receptor comprises a single-chain Fv constructed from the C2 mAb that is specific for native type II collagen coupled to the γ signaling subunit of FcεRI (C2/γ). This signaling subunit carries a single ITAM and is essential for transducing downstream signals necessary for IL-2 production in T cells. In addition, HCQ6 cells were transduced with a mutant receptor lacking the ITAM motif (C2/γ/IC⁻). C2 receptor expression was confirmed by flow cytometric analysis using a rabbit anti-C2 polyclonal Ab (33). The expression of C2/γ homodimers vs C2/γ-TCRζ heterodimers has been determined previously by immunoblotting of cell lysates with a rabbit polyclonal anti-C2 Ab under nonreducing conditions (33). Chronic culture in the presence of cytokines was performed as described above, and the chimeric receptor-mediated responses were tested by stimulating T cells with plate-bound native bovine collagen II (a gift from R. Williams, London, U.K.).

Abs and flow cytometry

The following mAbs were used for flow cytometric analysis: anti-mouse CD3ε-FITC, TCRβ-FITC, CD45-PE, and CD69-biotin (PharMingen) and anti-human CD4-FITC (Becton Dickinson, San Jose, CA). For flow cytometry, T cells were stained by standard methods and analyzed using a FACScan flow cytometer and CellQuest software (Becton Dickinson). To study the proportions of apoptotic and dead cells, annexin V-FITC and propidium iodide staining was performed using a commercial kit (Alexis, San Diego, CA). Glutathione levels were determined in whole cells by flow cytometry after staining with 50 nM Cell Tracker Green 5-chloromethylfluorescein diacetate according to the manufacturer’s instructions (Molecular Probes, Eugene, OR). For T cell activation, hamster anti-mouse CD3ε clones 145-2C11 and 500.A2 (PharMingen), mouse anti-human CD4 (OKT4) purified from a hybridoma supernatant, hamster Ig control (clone L2; R. Schreiber, Washington University School of Medicine, St. Louis, MO), and rabbit anti-C2 (33), followed by goat anti-hamster Ig (Cappel, via ICN Biomedicals, Eschwege, Germany) or goat anti-rabbit Ig (Pierce, Rockford, IL) were used for cross-linking. A blocking anti-HLA-DR mAb (L243) was provided by CellTech (Slough, U.K.). For immunoprecipitation and Western blotting experiments, polyclonal TCRζ antisera 98118 (J. Borst, The Netherlands Cancer Institute, Amsterdam, The Netherlands) and K2 (A. Kang, Veterans Administration Medical Center, Memphis, TN); monoclonal anti-TCRζ clones 6B10.2, which recognizes a transmembrane epitope (Santa Cruz Biotechnology, Santa Cruz, CA) and 8D3 that recognizes a C-terminal fragment (PharMingen); hamster anti-mouse CD3ε 145-2C11, polyclonal anti-mouse CD3γ, CD3δ, and CD3ε (E. Palmer, Basel Institute of Immunology, Basel, Switzerland); monoclonal anti-ZAP-70 (a gift from GlaxoWellcome, Stevenage, U.K.), monoclonal anti-Lck (3A5), and anti-Fyn (15) and polyclonal anti-ZAP-70 (LR; Santa Cruz Biotechnology); and polyclonal anti-linker for activation of T cells (anti-LAT) and monoclonal anti-phosphotyrosine (4G10) Abs (Upstate Biotechnology, Lake Placid, NY) were used.

Studies of signaling pathways

T cells (5–10 × 10⁶ cells/time point) were harvested into serum-free RPMI, washed twice, and rested on ice for 1 h. Cells were then incubated with 10 μg/ml of anti-mouse CD3ε with or without anti-human CD4 mAbs or hamster Ig control for 20 min on ice. After washing with ice-cold RPMI, T cells were resuspended in serum-free medium at 37°C before stimulation with 10 μg/ml of goat anti-hamster Ig for the indicated times. For peptide stimulation, 5 × 10⁶ T cells were mixed with 20 × 10⁶ APCs pulsed with 50 μg/ml of specific or control peptide for 4 h at 37°C. Cells were centrifuged for 30 s at 3000 rpm and incubated for various times at 37°C. T cell activation was terminated by adding ice-cold PBS containing 0.5 mM Na₃VO₄. After centrifugation, cell pellets were resuspended in ice-cold lysis buffer (50 mM Tris (pH 8.0), 200 mM NaCl, 0.1 mM EDTA, 1% Nonidet P-40, 1 mM Na₃VO₄, 20 mM NaF, 1 mM PMSF, 1 μg/ml leupeptin, 2 μg/ml pepstatin, and 10 μg/ml aprotinin). For CD3ε immunoprecipitation, 0.5% Triton X-100 was used as a detergent to preserve the association between CD3 chains and TCRζ. Lysates were centrifuged for 10 min at 13,000 rpm to remove detergent-insoluble material. Postnuclear lysates were used for immunoprecipitation experiments or were resuspended into 2× SDS-PAGE sample buffer with or without 2-ME and studied by Western blotting.

Immunoprecipitation and in vitro kinase assays

After preclearing with protein A-agarose (Sigma), lysates were incubated with 5 μg of purified Ab or a 1/100 dilution of specific antiserum for 1–2 h on ice, followed by the addition of 35 μl of protein A-agarose in each tube. Samples were incubated for 2 h at 4°C with rotation, after which immunoprecipitates were washed three times with ice-cold lysis buffer and resuspended in 2× SDS-PAGE sample buffer with or without 2-ME. To immunoprecipitate cell surface CD3ε, the cells were stained with 10 μg/ml of the anti-CD3ε mAb for 20 min on ice before cell lysis. After three washes to remove the unbound mAb, the cells were lysed, and 35 μl of protein A-agarose was added directly. For Lck kinase assays, an additional wash in 20 mM Tris (pH 7.5) and 0.3 M LiCl, and two additional washes in kinase buffer (20 mM Tris (pH 7.5), 10 mM MnCl₂, 0.1% 2-ME, and 100 μM Na₃VO₄) were performed. Thereafter, immunoprecipitates were resuspended in 50 μl of kinase buffer containing 5 μg of purified substrate (a GST fusion protein including a C-terminal fragment corresponding to aa 331–443 of human SAM-68; provided by W. Kolanus, University of Munich, Munich, Germany) and 10 μM unlabeled ATP. Five to 10 μCi of [γ-³²P]ATP (3000 Ci/mmol; Amersham Pharmacia Biotech, Aylesbury, U.K.) was added to each tube, and the reaction was allowed to proceed for 15 min at room temperature with agitation. The reactions were terminated by adding 4× SDS-PAGE sample buffer. Boiled samples were resolved on SDS-PAGE before staining with Coomassie Brilliant Blue R-250 to verify equal substrate loading and equal amounts of immunoprecipitating Ab per sample. Radioactivity was quantified using a phosphorimager (FLA-2000; Fuji, Tokyo, Japan).

Cell surface biotinylation, Western blotting, and immunodetection

Biotinylation of cell surface proteins was performed as previously described (33). Whole cell lysates and immunoprecipitates were separated by SDS-PAGE using SeeBlue-prestained m.w. markers (Novex, San Diego, CA) as a reference, before transfer onto nitrocellulose membranes (Schleicher & Schuell, Keene, NH). Equal protein loading per lane was verified by staining with Ponceau S reagent (Sigma). After destaining, the membrane was blocked with Tris-buffered saline, 1 mM EDTA (pH 8), and 0.1% Tween 20 (TBS-Tween) containing 5% BSA or 5% skimmed milk for 1 h at room temperature. Immunoblotting with specific Abs was performed in TBS-Tween containing 1% BSA (anti-phosphotyrosine Ab) or 5% milk (other Abs) overnight at 4°C, followed by immunodetection using HRP-conjugated rabbit anti-mouse Ig or swine anti-rabbit Ig (DAKO, Glostrup, Denmark) and enhanced chemiluminescence (Amersham Pharmacia Biotech). For biotinylated proteins, blots were incubated with HRP-conjugated streptavidin (Amersham Pharmacia Biotech) before detection by enhanced chemiluminescence. The autoradiographs were scanned with an AGFAArcusII scanner (AGFA, Greenville, SC), and signal intensities were quantified using AIDA software (Raytest, Straubenhardt, Germany). Reprobing of membranes was undertaken after incubation in 74 mM Tris (pH 6.8), 2% SDS, and 0.7% 2-ME for 1 h at 60°C to remove the bound Abs, followed by three washes in TBS-Tween.

Northern blotting

Thirty micrograms of total RNA was separated on denaturing formaldehyde/1% agarose gels, transferred to Hybond XL membranes (Amersham Pharmacia Biotech) using capillary elution, and fixed by UV cross-linking. Radiolabeled RNA probes were generated by in vitro transcription using full-length mouse TCRζ cDNA (1.3-kb insert in pGEM-3Z vector, a gift from A. Weissman, National Cancer Institute, National Institutes of Health, Bethesda, MD) as a template (34). The murine GAPDH riboprobe (Ambion, Austin, TX) was used to confirm equivalent RNA loading. In vitro transcription was performed in a final volume of 20 μl containing 1× SP6/T7 transcription buffer; 500 μM each of ATP, GTP, and CTP; 50 or 500 μM UTP; 50 μCi of [α-³²P]UTP (∼800 Ci/mmol); 10 U of SP6 or T7 polymerase; 10 U of RNasin; and 1 μg of DNA template at 37°C for 1 h, after which DNA templates were removed by DNase treatment for 15 min. Reactions were then phenol/chloroform extracted, and unincorporated nucleotides were removed using a Microspin S-200 HR column (Amersham Pharmacia Biotech). Incorporation of radionucleotides was assessed by scintillation counting. Filters were prehybridized for 2 h and hybridized overnight at 68°C in ULTRAhyb solution (Ambion) containing 2.5 × 10⁶ cpm of radiolabeled RNA. Blots were then washed twice for 10 min each time in 2× SSC, 0.1% SDS, followed by two washes for 15 min each time in 0.1× SSC/0.1% SDS. Signals were quantified using a phosphorimager. For reprobing, the bound probe was removed by boiling the filter in 0.1% SDS, followed by two washes in 5× SSC.

Intracellular Ca²⁺ mobilization

Ca²⁺ flux in response to TCR or chimeric receptor ligation was investigated using variable wavelength Ca²⁺ indicators as described previously (22). T cells were loaded with 4 μM fura-2/AM (Molecular Probes) for 30 min at 37°C. Dye-loaded cells were washed and incubated on ice with anti-CD3ε and anti-CD4 mAbs or with anti-C2 polyclonal Abs (for C2 chimeric receptor-expressing cells) as described above. Fluorescence intensity was measured on a Perkin-Elmer luminescence spectrometer LS50 (Norwalk, CT) with excitation at 340 and 380 nm and emission fixed at 510 nm. After the baseline fluorescence was established, anti-hamster or anti-rabbit Ig was added to a final concentration of 10 μg/ml to cross-link surface receptors. Cells were maintained at 37°C with continuous gentle stirring throughout the analysis. Maximum fluorescence (F_max) was determined by lysing the cells with 0.1% Triton X-100, and minimum fluorescence (F_min) was determined by the addition of 40 mM Tris and 4 mM EGTA. Intracellular Ca²⁺ (Ca²⁺_i) was calculated using FL WinLab software (Perkin-Elmer, Norwalk, CT) according to the Grynkiewitz equation: Ca²⁺ = K_d × [(R − R_min)/(R_max − R)] × S_f/S_b where K_d = 135 nM for fura-2/AM and Ca²⁺, R is F at 340 nm/F at 380 nm, and S_f (or S_b) is F_min (or F_max) at 380 nm.

Chronic TNF exposure induces nondeletional and reversible T cell hyporesponsiveness

To investigate the mechanisms for T cell hyporesponsiveness induced by TNF, we used a model of prolonged TNF exposure with HCgp-39 specific, HLA-DR4-restricted mouse T cell hybridomas. T cell hybridomas were chosen to study the direct effects of TNF on T cells in the absence of accessory cells and to undertake a more detailed kinetic analysis of the effects observed. Hybridoma clones expressing p55 and p75 TNF-R (as well as other cell surface activation Ags) at levels similar to those observed on chronically activated T lymphocytes were selected for the study (data not shown) (35, 36).

Fig. 1⇓A shows that prolonged exposure of T cell hybridoma clones to mTNF or hTNF suppressed IL-2 production following stimulation with EBV-transformed human B cells pulsed with specific peptide. TNF could suppress T cell activation by up to 90% depending on the clone studied. Splenic APC derived from HLA-DR4 transgenic mice gave similar results, although the levels of IL-2 detected were somewhat lower (data not shown). Given that mTNF and hTNF attenuated T cell activation to a similar extent, and hTNF binds and signals through murine p55, but not p75 TNF-R, these results suggest that induction of T cell hyporesponsiveness is mediated by sustained signals transduced through the p55 TNF-R. IL-1 shares many signaling pathways with TNF, and yet culturing the cells in the presence of equimolar concentrations of hIL-1α had no suppressive effect on the activation of clones 11A2 and 20D11 (Fig. 1⇓A). However, a modest reduction in IL-2 production was noted for IL-1-treated 32A1 cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Chronic TNF exposure induces nondeletional and reversible T cell hyporesponsiveness. A, A panel of HC-gp39-specific T cell hybridoma clones were cultured in the presence or the absence of 5 ng/ml mTNF, hTNF or hIL-1 for 8 days before stimulation with APC pulsed with 10 μg/ml of HCgp-39 peptides 322–337 (clone 11A2), 334–349 (32A1), and 100–115 (20D11). Data represent IL-2 production (IL-2 fluorescence units ± SD) 24 h after stimulation. B, Clone 5G11 was cultured in the presence or the absence of increasing concentrations of hTNF for 4–12 days before stimulation with peptide-pulsed APC (10 μg/ml of peptide 262–277). IL-2 production was determined 24 h later. Data are presented as the percentage of IL-2 production by T cells cultured in the absence of TNF (control). C, Clone 11A2 was cultured in the presence or the absence of 5 ng/ml mTNF or hTNF for 8 days (baseline). T cells were washed extensively and cultured for a further 2 days in the absence of TNF (recovery period) before stimulation. D, Clone 11A2 was cultured in the presence or the absence of 5 ng/ml mTNF for 8 days. The proportion of apoptotic and dead cells (percentage) was determined following annexin V-FITC and propidium iodide (PI) staining by flow cytometric analysis.

Suppression of T cell activation by chronic TNF treatment was both time and dose dependent. A detailed kinetic analysis revealed detectable depression of peptide-specific IL-2 production within 4 days of culture in the presence of hTNF (Fig. 1⇑B). Moreover, a decrease in IL-2 production was observed at picomolar concentrations of hTNF (e.g., 0.6 ng/ml; <50 pM) after more prolonged culture, concentrations well within the range of TNF expression at sites of chronic inflammation (37). Finally, T cell hyporesponsiveness was reversible upon withdrawing TNF, because peptide-induced IL-2 production was restored to control levels after only 2 days of culture in the absence of TNF (Fig. 1⇑C). Consistent with this result, TNF did not significantly affect cell viability under these experimental conditions (Fig. 1⇑D).

Because TNF-treated T cell hybridomas reproduced many of the characteristics of T cells exposed to TNF in vivo reported previously (21, 22), this model system was used to explore in more detail the molecular basis for T cell hyporesponsiveness induced by TNF. For the purpose of these experiments, we selected experimental conditions that suppressed T cell activation by ∼70–80%; 2.5 ng/ml hTNF for 8–14 days was used for clone 11A2 unless otherwise stated.

TNF-treated T cells require stronger and more sustained TCR engagement for commitment to IL-2 production

We next studied T cell responses to a broad range of peptide and anti-CD3 mAb concentrations. T cells pretreated with mTNF or hTNF produced much lower levels of IL-2 regardless of the strength of the TCR stimulus (Fig. 2⇓A). Indeed, suppression of responses could not be overcome with anti-CD3 concentrations up to 8 μg/ml (Fig. 2⇓A) or with peptide concentrations up to 30 μg/ml (data not shown). In addition, these experiments demonstrated that stronger TCR engagement was required for TNF-treated T cells to produce detectable levels of IL-2 (Fig. 2⇓A).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

TNF-treated cells require stronger and more sustained TCR stimulation for commitment to IL-2 production. Clone 11A2 was chronically cultured in the presence or the absence of 2.5 ng/ml mTNF or hTNF. A, T cells were then stimulated by APC pulsed with the indicated concentrations of specific peptide or by immobilized anti-CD3ε mAb (500.A2), and IL-2 production was determined 24 h later. B, Cells were stimulated by peptide-pulsed APC (10 μg/ml), followed by the addition of blocking anti-HLA-DR mAb (50 μg/ml) to cultures at the indicated time points. Alternatively, T cells were stimulated with immobilized anti-CD3ε mAb (2 μg/ml) for the indicated times, and TCR engagement was interrupted by transferring the cells to uncoated wells. IL-2 production was determined after 24 h of incubation. Representative data from four (A) or six (B) independent experiments are shown.

We then determined the time period of TCR engagement required for induction of IL-2 production. Control and TNF-treated cells were stimulated with peptide-pulsed APC as before, and at predefined time points TCR ligation by MHC/peptide complexes was interrupted by adding a blocking anti-HLA-DR mAb to the cultures. This mAb completely blocked IL-2 production when added at the beginning of stimulation (see zero time point in Fig. 2⇑B). Using another approach, T cells were stimulated with plate-bound anti-CD3 mAb, and at predefined time points the cells were transferred to uncoated wells. All cultures were harvested after 24 h, and IL-2 production was determined. As shown in Fig. 2⇑B, TNF-treated T cells required longer periods of TCR engagement than control cells for commitment to IL-2 production.

Although a more precise quantitation of the parameters required for productive TCR stimulation in control and TNF-treated cells is difficult, multiple experiments indicated that TNF-treated T cells required approximately four times more peptide or anti-CD3 mAb for up to 3 h longer to produce detectable levels of IL-2 (using a stimulation index of 3 as the detection limit). Collectively, these results indicated that chronic TNF exposure increases the threshold required for T cell activation and commitment to IL-2 production.

Chronic TNF exposure down-regulates the expression of cell surface TCR/CD3 complex

An increased activation threshold together with our previous results demonstrating suppressed intracellular Ca²⁺ mobilization by TNF (22) suggested that TNF might attenuate proximal TCR signal transduction pathways. To test this hypothesis, we first investigated cell surface TCR/CD3 expression after chronic TNF exposure. Indeed, flow cytometric analysis revealed a modest, yet reproducible, down-regulation of cell surface CD3ε expression in TNF-treated cells, whereas the expression of CD3ε was unaltered in cells chronically cultured in the presence of IL-1 (Fig. 3⇓A); similar results were observed using anti-TCRβ mAbs (data not shown). Down-regulation of cell surface molecules was not a generalized characteristic of TNF-treated cells, because no significant effect on the expression of hCD4 or CD45 was noted (Fig. 3⇓B), and the expression of the early activation Ag CD69 was consistently increased in TNF-treated cells. The effects of TNF on CD3ε expression were dose dependent (Fig. 3⇓C), and data from 10 experiments revealed that the median fluorescence intensities were 69.5, 77.8, and 82.4% of control in cells cultured in the presence of 2.5, 0.6, and 0.15 ng/ml of hTNF, respectively. It should be noted that at low TNF concentrations modest suppression of T cell IL-2 production (∼20–30%) was observed in the absence of any detectable decrease in TCR/CD3 expression.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Chronic TNF exposure down-regulates the expression of the cell surface TCR/CD3 complex. A, Flow cytometric analysis of CD3ε expression on control 11A2 cells (thick line) and on cells chronically cultured with 2.5 ng/ml mTNF or hTNF or with 5 ng/ml hIL-1 (thin line). Shaded histograms represent staining with negative control mAb. B, Flow cytometric analysis of hCD4, CD45, and CD69 expression on control cells (thick line) and cells chronically cultured in the presence of 2.5 ng/ml hTNF (thin line). C, Flow cytometric analysis of CD3ε expression on control cells (thick line) and on cells cultured with hTNF (thin line). D, CD3ε expression in whole cell lysates from control and TNF-treated cells was studied by Western blotting under reducing conditions. CD3ε was immunoprecipitated from whole cell lysates or from the cell surface of control cells and cells treated with 2.5 ng/ml hTNF, and CD3ε expression was studied by Western blotting under reducing conditions. Representative results from eight (A), three (B), ten (C), and four (D) separate experiments are shown.

The above results suggested that either TNF may affect the expression of CD3ε directly or the TCR/CD3 complex is not optimally assembled or transported to the cell surface. To investigate the first possibility, expression of CD3ε was studied by Western blotting. As shown in Fig. 3⇑D, chronic TNF exposure did not alter CD3ε levels in whole cell lysates. Consistent with these results, similar levels of CD3ε could be immunoprecipitated from whole cell lysates derived from control and TNF-treated cells (Fig. 3⇑D). In contrast, cell surface immunoprecipitates derived following incubation of whole cells with the immunoprecipitating mAbs before cell lysis revealed a reduction in cell surface CD3ε expression in TNF-treated cells compared with control cells (Fig. 3⇑D). These data suggested that chronic TNF exposure altered the subcellular localization of CD3ε and indicated to us that TNF perturbed TCR/CD3 complex assembly and/or transport to the cell surface.

Time- and dose-dependent decrease in TCRζ chain expression by chronic TNF

Because the association of TCRζ-ζ chain homodimers with the hexameric αβγεδε complex is a critical step in the formation of complete TCR/CD3 receptor complexes (5, 6), we explored the possibility that modulation of TCRζ chain levels by TNF could explain the differences in subcellular localization of CD3ε. Fig. 4⇓A shows that TNF profoundly down-regulated the expression of TCRζ-ζ homodimers (32 kDa) in whole cell lysates from clone 11A2. The extent of TCRζ down-regulation by mTNF and that by hTNF were similar, whereas hIL-1 had no effect. The decrease in TCR ζ-chain expression was selective, because no decrease in the expression of protein tyrosine kinase ZAP-70 was observed in the same lysates (Fig. 4⇓A), nor were the levels of Src kinases Lck and Fyn markedly altered by TNF (data not shown). Furthermore, streptavidin-HRP blotting of TCRζ immunoprecipitates derived from cell surface biotinylated T cells indicated that the expression of plasma membrane-associated TCRζ was substantially reduced in TNF-treated T cells (Fig. 4⇓B).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Down-regulation of TCRζ expression after chronic TNF exposure. Clone 11A2 was chronically cultured with 2.5 ng/ml mTNF or hTNF or with 5 ng/ml hIL-1. A, IL-2 production was investigated as described for Fig. 1⇑. TCRζ-ζ and ZAP-70 expression in whole cell lysates were studied by Western blotting under nonreducing or reducing conditions, respectively. B, Following cell surface biotinylation, TCRζ was immunoprecipitated from whole cell lysates. Biotinylated cell surface TCRζ-ζ was detected by Western blotting with streptavidin-HRP under nonreducing conditions. C, Clone 11A2 was cultured in the presence or the absence of hTNF for 12 days, the cells were washed extensively, and the culture was continued for an additional 4 days in the absence of TNF (recovery). TCRζ-ζ expression in whole cell lysates was studied by Western blotting under nonreducing conditions. D, The levels of TCRζ and GAPDH mRNA in T cells were studied by Northern blotting. The values indicate the ratios of TCRζ:GAPDH compared with those in control cells, determined by phosphorimaging. In T cells treated with 2.5 ng/ml TNF, TCRζ mRNA was reduced by 32% (mean percent reduction in three experiments). E, CD3ε was immunoprecipitated from whole cell lysates, and the expression of CD3ε and TCRζ-ζ was studied by Western blotting under reducing or nonreducing conditions, respectively. Representative results from six (A), three (B and D), two (C), and four (E) separate experiments are shown.

A kinetic analysis revealed that TCRζ down-regulation could be detected by 4 days of culture in the presence of 2.5 and 0.6 ng/ml of hTNF, but could also be detected after culturing T cells at very low TNF concentrations for longer periods (e.g., 0.15 ng/ml for 16–20 days; data not shown). Nevertheless, as in the case of TCR/CD3 expression, an ∼30% decrease in IL-2 production was observed without a detectable reduction in TCRζ expression, indicating that the correlation between TCRζ expression and IL-2 production may not be absolute, and that TNF may attenuate downstream signaling pathways. Consistent with the recovery of IL-2 production following TNF withdrawal (Fig. 1⇑D), TCRζ down-regulation was reversible, because levels were restored toward normal upon withdrawing TNF (Fig. 4⇑C).

Down-regulation of TCRζ by TNF could not be explained by dissociation of dimers into monomeric TCR ζ-chains, because comparable decreases were observed in the expression of the 16-kDa TCRζ monomer studied under reducing conditions (data not shown), nor could it be explained by modification of a specific epitope, because similar results were obtained using two different TCRζ-specific mAbs, derived from clones 6B10.2 and 8D3, recognizing transmembrane and C-terminal epitopes, respectively (data not shown). By contrast, steady state TCRζ mRNA levels determined by Northern blotting were reduced by ∼30% in T cells cultured in the presence of 2.5 ng/ml TNF (Fig. 4⇑D). Nonetheless, TCRζ mRNA expression was not substantially decreased in T cells treated with lower concentrations of TNF despite consistent reductions of TCRζ protein expression (see Fig. 8⇓B below). These data indicate that in addition to its effects on TCRζ mRNA levels, TNF may down-regulate TCR ζ-chain expression through post-transcriptional and/or post-translational mechanisms.

We then investigated the association of TCRζ with the TCR/CD3 complex. Fig. 4⇑E shows that the amount of dimeric TCRζ associated with CD3ε was significantly reduced in TNF-treated cells. In contrast, similar levels of CD3γ and CD3δ were present in CD3ε immunoprecipitates from control and TNF-treated cells (data not shown). These findings imply that the decrease in cell surface TCR/CD3 expression after chronic TNF exposure could arise as a direct consequence of a selective reduction in the expression of TCRζ.

Chronic TNF exposure suppresses IL-2 production transduced through a chimeric receptor, which uses TCRζ for signaling, but not for assembly and expression

The results to date supported the idea that TNF suppressed T cell activation by down-regulating TCRζ expression, thereby perturbing the assembly and transport of stable TCR/CD3 complexes from the ER to the cell surface. To test the hypothesis that TNF targeted specifically TCRζ, as opposed to TCR/CD3 expression, we studied the responses of T cell hybridomas expressing a chimeric receptor that uses TCRζ, but not CD3, molecules for signaling. Briefly, a panel of receptors and mutants was generated based on a prototype receptor comprising a single-chain Fv of C2 mAb specific for type II collagen coupled to the γ signaling subunit of FcεRI (C2/γ) (33). In T cell hybridomas these chimeric receptors are expressed on the cell surface as either C2/γ homodimers or C2/γ/TCRζ heterodimers. In HCQ6 T cells, which express relatively high levels of TCRζ, the predominant form (∼80%) of the chimeric receptor is a C2/γ/TCRζ heterodimer (33). To examine responses transduced through a chimeric receptor complex with an absolute requirement for TCRζ to signal, we studied HCQ6 cells expressing mutant C2/γ receptors lacking the ITAM motif (C2/γ/IC⁻; see schematic in Fig. 5⇓A).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Chronic TNF exposure suppresses IL-2 responses transduced through a chimeric C2/γ/IC⁻ receptor. A, Schematic representation of the C2/γ/IC⁻ homodimer and the heterodimer with endogenous TCRζ chain. Only ligation of heterodimeric receptor by CII leads to IL-2 production. B, Flow cytometric analysis of C2/γ/IC⁻ expression levels on control HCQ6 cells (thick line) and on cells chronically cultured with 2.5 ng/ml hTNF (thin line). Shaded histograms represent staining with a negative control mAb. C, TCRζ-ζ and ZAP-70 expression in whole cell lysates from control HCQ6 cells and from cells chronically cultured with hTNF was studied by Western blotting under nonreducing or reducing conditions, respectively. D, Control and hTNF-treated HCQ6 cells were stimulated with immobilized anti-CD3ε mAb (0.5 μg/ml) or type II collagen (4 μg/ml), and IL-2 production was determined 24 h later. Representative results from seven separate experiments are shown.

In contrast to its effects on TCR/CD3 expression, chronic TNF exposure had no effect on cell surface expression of the chimeric receptor on transduced HCQ6 cells (Fig. 5⇑B). As expected, TNF reduced the expression of endogenous TCRζ chain in HCQ6 cells (Fig. 5⇑C). Furthermore, IL-2 responses to plate-bound collagen II were suppressed by TNF to the same extent as IL-2 responses to TCR ligation (Fig. 5⇑D). These results indicated that TNF-induced changes in the expression of TCRζ, but not the cell surface receptor complex, were sufficient to suppress IL-2 production.

Chronic TNF exposure leads to quantitative, but not qualitative, changes in TCRζ phosphorylation

The TCR ζ-chain is not only involved in TCR/CD3 assembly, but also initiates signal transduction cascades originating from the TCR (10). To study how long term TNF exposure affected early phosphorylation events, TCRζ was immunoprecipitated from control and TNF-treated cells after activation by anti-mouse CD3ε either alone or in combination with anti-human CD4 mAb. TCR ligation induced the characteristic pattern of phosphorylated ζ isoforms pp21 and pp23; activation with anti-CD3 alone mimicked a partial agonist signal (pp21 > pp23), whereas combined anti-CD3 and anti-CD4 stimulation induced a full agonist signal (pp23 > pp21; Fig. 6⇓A). This pattern was mirrored by reductions in the amount of unphosphorylated TCRζ detected by immunoblotting with mAb 8D3, which does not recognize phospho-ζ (Fig. 6⇓A). As shown in Fig. 6⇓A, chronic exposure of T cells to TNF led to reductions in both pp21 and pp23 levels. Moreover, consistent with the reductions in cell surface CD3ε expression, levels of phosphorylated CD3ε were clearly decreased in TNF-treated T cells following TCR engagement (Fig. 6⇓B). In CD3ε immunoprecipitates, levels of pp21/pp23 were also reduced in TNF-treated T cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

The effects of chronic TNF exposure on TCRζ and CD3ε phosphorylation and on Lck kinase activity after TCR ligation. Clone 11A2 was chronically cultured in the presence or the absence of 2.5 ng/ml hTNF. The cells were activated by anti-CD3ε with or without anti-CD4 mAbs or by APC pulsed with specific peptide for the indicated times at 37°C. TCR ζ-chain was immunoprecipitated using a TCRζ-specific antiserum (A) or a ZAP-70-specific Ab (C), whereas CD3ε was immunoprecipitated using a hamster anti-CD3ε Ab (B). Tyrosine phosphorylation was detected by Western blotting under reducing conditions. The blots were reprobed with anti-TCRζ mAb 8D3, which recognizes the unphosphorylated form of TCRζ (A), or with rabbit polyclonal anti-CD3ε Ab (B). D, Lck kinase activity was determined by in vitro kinase assay on anti-Lck immunoprecipitates using GST-SAM68 C-terminal fragment as substrate. The values indicate the incorporation of radioactivity compared with that in nonstimulated control cells determined by phosphorimaging. Coomassie blue staining demonstrates comparable levels of the substrate and immunoprecipitating anti-Lck mAb heavy chain in each lane. Representative results from six (A) and two (B–D) separate experiments are shown.

We then studied TCRζ phosphorylation following activation by peptide-pulsed APCs by examining levels of phospho-ζ in ZAP-70 immunoprecipitates. A similar decrease in phosphorylated ζ was observed in TNF-treated cells following activation with the specific peptide (Fig. 6⇑C). Interestingly, peptide 322–337 induced a partial agonist signal, similar to that observed with anti-CD3ε alone. Nevertheless, the ratio of pp23:pp21 in TNF-treated cells did not differ significantly from that in control cells regardless of the nature of TCR ligation (Fig. 6⇑, A and C). Consistently, a comparable reduction in the level of unphosphorylated ζ was observed in control and TNF-treated cells following T cell activation (Fig. 6⇑A). These results suggest that the phosphorylation process per se may proceed normally in T cells after chronic TNF exposure, and that reduced levels of phospho-ζ after TCR ligation could be explained by the lower levels of TCRζ protein in TNF-treated cells before activation.

Because the ratio of pp23:pp21 appeared to be unaltered in TNF-treated cells, we predicted that Lck kinase activity should not differ markedly between control and TNF-treated cells upon TCR engagement. To test this directly, we compared in vitro kinase activity in Lck immunoprecipitates from control and TNF-treated T cells after activation with anti-CD3 with or without anti-CD4 mAbs. Modest reductions in Lck kinase activity were observed at intermediate time points in TNF-treated cells, suggesting that the amplitude of kinase activity is marginally reduced (Fig. 6⇑D). Although the reason for this small reduction is not understood, we envisage that loss of surface TCR/CD3 expression may lead to a decrease in the proportion of total cellular Lck recruited to the TCR/CD3 cluster upon activation, and thus a small reduction in kinase activity. Nevertheless, Lck kinase activity at early and later time points did not appear to be affected by TNF treatment despite the fact that levels of phospho-ζ were consistently depressed at all time points (Fig. 6⇑C).

A decrease in TCRζ expression uncouples ZAP-70 and LAT phosphorylation and attenuates Ca²⁺ responses after TCR ligation

The downstream consequences of depressed TCRζ expression were tested further by investigating the phosphorylation of ZAP-70, which is initiated by the recruitment of ZAP-70 to phosphorylated TCR ζ-chains (16). As shown in Fig. 7⇓A, chronic treatment of T cells with TNF resulted in a significant down-regulation of ZAP-70 tyrosine phosphorylation after T cell activation by anti-CD3ε plus anti-CD4 mAbs at all time points studied. A combination of anti-CD3ε and anti-CD4 mAbs was optimal, because ZAP-70 phosphorylation in these T cells was not as strong and was less consistent following activation with either anti-CD3ε mAb alone or specific peptide (data not shown). The extent of the decrease in ZAP-70 phosphorylation correlated closely with the duration of TNF exposure (8 vs 14 days), and with TCRζ expression in whole cell lysates before TCR ligation (data not shown). In addition, Fig. 6⇑C shows reduced levels of phospho-ζ in ZAP-70 immunoprecipitates following TCR ligation in TNF-treated cells, indicating that less ZAP-70 is recruited to phosphorylated TCRζ chains. Finally, additional experiments demonstrated that ZAP-70 phosphorylation was depressed only under circumstances where TCRζ was significantly down-regulated (data not shown), further suggesting that decreased TCR ζ-chain expression may account for attenuation of ZAP-70 phosphorylation.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

Down-regulation of ZAP-70 and LAT tyrosine phosphorylation and Ca²⁺ mobilization following TCR or C2 ligation in TNF-treated T cells. Clones 11A2 (A–C) and HCQ6 (D) were chronically cultured in the presence or absence of 2.5 ng/ml hTNF. A and B, T cells were activated by anti-CD3 with or without anti-CD4 mAbs for the indicated times at 37°C. ZAP-70 (A) or LAT (B) was immunoprecipitated using specific Abs, and tyrosine phosphorylation was detected by Western blotting under reducing conditions. The blots were reprobed with anti-ZAP-70 (A) or anti-LAT Abs (B) to verify equal loading in each lane. C and D, Cells loaded with fura-2 AM were stained with anti-CD3ε and anti-CD4 mAbs (C) or anti-C2 Abs (D) on ice. After establishing baseline fluorescence, cross-linking anti-hamster or anti-rabbit Ig was added, and fluorescence intensity was measured with a luminescence spectrometer. Representative results from six (A), two (B and D), and three (C) separate experiments are shown.

LAT is a naturally occurring ZAP-70 substrate and links the phosphorylation of ZAP-70 to the activation of downstream Ras and Ca²⁺ signaling pathways (18, 19). Consistent with the reductions observed in ZAP-70 phosphorylation, tyrosine phosphorylation of p36LAT was depressed in cells chronically treated with TNF after TCR ligation by anti-CD3 mAb alone and in combination with anti-CD4 mAb (Fig. 7⇑B). Furthermore, intracellular Ca²⁺ mobilization following TCR ligation was significantly attenuated in cells after chronic TNF exposure (Fig. 7⇑C), to an extent remarkably similar to that previously reported in TNF-treated TCR transgenic T cells (22) as well as in T cells derived from TCRζ-deficient mice (38).

Finally, we studied Ca²⁺ responses in HCQ6 cells expressing the chimeric receptor C2/γ/IC⁻, because this receptor requires endogenous TCRζ chains to signal (see Fig. 5⇑A). As shown in Fig. 7⇑C, we detected a clear reduction in the Ca²⁺ mobilization after cross-linking of the C2/γ/IC⁻ receptor in TNF-treated cells (Fig. 7⇑D). Because these chimeric receptor complexes do not appear to associate with other CD3 invariant chains (33), these data suggest a more direct link between loss of TCRζ expression and attenuation of downstream signaling pathways. Nonetheless, the data cannot rule out the possibility that TNF uncouples additional downstream signaling pathways common to TCR and the C2 receptor independently of its effects on TCRζ.

Inhibition of TNF-induced TCRζ down-regulation by NAC is accompanied by partial, but not complete, reversal of T cell hyporesponsiveness

Alterations in intracellular redox balance following depletion of the abundant intracellular anti-oxidant glutathione have been shown to suppress T cell proliferation and TCR signaling pathways (39, 40, 41). To explore the possibility that TNF down-regulated TCRζ expression and IL-2 production by reducing glutathione levels, we evaluated the effects of NAC, a biosynthetic precursor of glutathione that scavenges reactive oxygen species. Control or TNF-treated cells were cultured in the presence or the absence of 1 mM NAC for up to 20 days before stimulation. Under these conditions, NAC completely inhibited the reduction in glutathione levels observed following culture with TNF, as determined by flow cytometry (Fig. 8⇓A). NAC had no consistent effects on TCRζ expression or IL-2 production by untreated T cells (data not shown). In contrast, data from multiple experiments with TNF-treated cells indicated that the effects of NAC on TCRζ expression varied according to the duration of the culture and TNF concentration. Specifically, inhibition of TNF-induced TCRζ down-regulation by NAC was observed after prolonged culture at intermediate and low doses of TNF, whereas the expression of ZAP-70 was not altered by NAC in the same cells (Fig. 8⇓B; 20 days of culture). These results suggested that down-regulation of glutathione levels is not the only mechanism by which TNF influences TCRζ expression. Nevertheless, coculture of TNF-treated cells with NAC allowed us to investigate the relationship between TCRζ expression and T cell IL-2 production. Several important observations were noted from these experiments. Firstly, the effects of NAC on IL-2 production by TNF-treated T cells correlated to changes in TCRζ expression, because T cell hyporesponsiveness was reversed only when down-regulation of TCRζ expression was inhibited by NAC (Fig. 8⇓, B and C, and data not shown). Secondly, NAC only partially inhibited TCRζ down-regulation and suppression of IL-2 production at the highest concentration of TNF. Thirdly, ∼40% suppression of IL-2 production was observed in T cells cultured in the presence of NAC and intermediate concentrations of TNF (0.6 ng/ml) despite TCRζ levels comparable to those in control cells. Although NAC may have a wide variety of cellular targets besides TCRζ, these data provide further evidence that prolonged exposure to TNF may uncouple additional TCR signaling pathways, perhaps downstream of TCRζ.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

The effect of NAC on intracellular glutathione levels, TCRζ expression, and peptide-induced IL-2 production. Clone 11A2 was cultured in the presence or the absence of increasing concentrations of hTNF and 1 mM NAC for 20 days. A, Flow cytometric analysis of glutathione levels in control cells (thick line) and in cells cultured with 2.5 ng/ml hTNF (thin line), determined by Cell Tracker Green 5-chloromethylfluorescein diacetate fluorescence. Shaded histograms represent unstained cells. B, TCRζ-ζ and ZAP-70 expression in whole cell lysates was studied by Western blotting under nonreducing or reducing conditions, respectively. C, IL-2 production was studied as described for Fig. 1⇑. Representative results from six separate experiments are shown.