Reactive oxygen species differentially affect T cell receptor-signaling pathways.

Oxidative stress plays an important role in the induction of T lymphocyte hyporesponsiveness observed in several human pathologies including cancer, rheumatoid arthritis, leprosy, and AIDS. To investigate the molecular basis of oxidative stress-induced T cell hyporesponsiveness, we have developed an in vitro system in which T lymphocytes are rendered hyporesponsive by co-culture with oxygen radical-producing activated neutrophils. We have observed a direct correlation between the level of T cell hyporesponsiveness induced and the concentration of reactive oxygen species produced. Moreover, induction of T cell hyporesponsiveness is blocked by addition of N-acetyl cysteine, Mn(III)tetrakis(4-benzoic acid)porphyrin chloride, and catalase, confirming the critical role of oxidative stress in this system. The pattern of tyrosine-phosphorylated proteins was profoundly altered in hyporesponsive as compared with normal T cells. In hyporesponsive T cells, T cell receptor (TCR) ligation no longer induced phospholipase C-gamma1 activation and caused reduced Ca(2+) flux. In contrast, despite increased levels of ERK1/2 phosphorylation, TCR-dependent activation of mitogen-activated protein kinase ERK1/2 was unaltered in hyporesponsive T lymphocytes. A late TCR-signaling event such as caspase 3 activation was as well unaffected in hyporesponsive T lymphocytes. Our data indicate that TCR-signaling pathways are differentially affected by physiological levels of oxidative stress and would suggest that although "hyporesponsive" T cells have lost certain effector functions, they may have maintained or gained others.

T lymphocytes isolated from patients affected with human pathologies such as cancer, AIDS, rheumatoid arthritis (RA), 1 and leprosy display reduced proliferative responses upon TCR ligation ex vivo (1)(2)(3)(4). This observation appears to reflect an in vivo T cell hyporesponsiveness that in the case of cancer, lep-rosy, and AIDS may be expected to have deleterious effects but that in auto-immunity plays an as yet unidentified role. The responsible mechanisms depend on oxidative stress that can be generated by e.g. tumor macrophages (5,6).
Several TCR-signaling molecules are known to be affected by oxidative stress. In T lymphocytes from AIDS patients, p56 lck has a decreased activity that probably results from a conformational change due to an altered intracellular redox potential (7). T lymphocytes isolated from patients affected with certain cancers have decreased expression levels of TCR- (8 -12), and macrophages from tumor-bearing mice can induce such partial TCR-loss in normal T cells in vitro (6,13,14). T lymphocytes from RA synovial fluid express less TCR-and the linker for the activation of T cells (LAT) as well as lower levels or modified p56 lck (15)(16)(17)(18).
Although they are well known for their destructive effect on biomolecules, reactive oxygen species (ROS) are more and more accepted as necessary constituents in signaling pathways and modulators of responses in physiological and pathological conditions (19). It has been shown that ROS are produced in muscular cells upon binding of ligands such as angiotensin II (20). In addition, ROS production has been documented in a number of cells stimulated with cytokines such as tumor necrosis factor-␣, transforming growth factor-␤, and interleukin-1 (21)(22)(23) and growth factors such as bovine fibroblast growth factor, nerve growth factor, platelet-derived growth factor, and epidermal growth factor (24 -27). In T cells, it has been reported that the radical scavenger N-acetyl cysteine (NAC) inhibited the activation of NF-B by phorbol 12-myristate 13acetate, tumor necrosis factor-␣, and interleukin-1, strongly supporting the idea that oxygen radicals are implicated in physiological activation processes (28,29).
Reactive oxygen species trigger several proximal and distal signaling pathways in T lymphocytes, affect the activities of transcription factors, and lead to expression of specific genes (30). In Jurkat T cells ROS induce increases in protein tyrosine phosphorylation and activity of p56 lck , ZAP-70, and protein kinase C as well as elevations in intracellular Ca 2ϩ levels (31)(32)(33)(34). ROS are known to mediate the activation of NF-B (33,35,36), but chronic exposure to ROS inhibits NF-B phosphorylation and activation (37,38).
To investigate the molecular basis of oxidative stress-induced T cell hyporesponsiveness, we developed an in vitro system in which T lymphocytes are rendered hyporesponsive by exposure to an oxidative environment generated by activated neutrophils. Using this system we analyzed the effects of ROS on several TCR-dependent signaling pathways. Here we report that oxidative stress does not cause a generalized inhibition of TCR-signaling pathways and suggest that hyporesponsive T lymphocytes may have lost certain effector func-tions but retained or gained others. This observation may have important implications for the physiopathology of cancer, RA, and other pathologies in which oxidative stress causes "T cell hyporesponsiveness."
Isolation of T Cells and Neutrophils-T lymphocytes and neutrophils were isolated from buffy coats of healthy donors or from peripheral blood and synovial fluid from RA patients. Briefly, mononuclear cells were collected upon centrifugation on Ficoll, washed three times, and resuspended in RPMI 1640 or Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 1 mM non essential amino acids, 1 mM sodium pyruvate, 1 mM HEPES, and antibiotics. Enriched T lymphocyte populations (Ͼ90%) were obtained after macrophage depletion by adherence to plastic for 1 h at 37°C and after B cell depletion with anti-CD19 mAb-coated magnetic beads. Neutrophils were separated from erythrocytes by dextran (T-500) sedimentation, and residual red blood cells were lysed with ice-cold 0.2% NaCl.
In Vitro Exposure to Activated Neutrophils-T lymphocytes (2 ϫ 10 6 cells/ml supplemented RPMI 1640 or Dulbecco's modified Eagle's medium) were cultured for 16 h with or without neutrophils at 1:1 ratio. Where stated, neutrophils were activated with 1 M N-formylmethionylleucylphenylalanine (fMLP) during the co-culture. Catalase (Sigma-Aldrich) and MnTBAP (Calbiochem) were added to cultures where indicated. T lymphocytes were subsequently isolated by neutrophil depletion with anti-CD66b mAb-coated magnetic beads. Where indicated T cells were subsequently cultured for 48 h in the presence of 5 mM N-acetyl cysteine (NAC). For caspase 3 activation, normal and hyporesponsive T cells were cultured for 38 h with or without plasticbound anti-CD3⑀ mAb OKT3 before lysis.
Flow Cytometric Analysis-For cell surface staining, T lymphocytes were washed once in PBS containing 2.5% fetal calf serum and 0.02% NaN 3 and incubated on ice for 20 min with saturating concentrations of the indicated antibodies. After three washes, the cells were incubated with appropriate secondary reagents for 20 min on ice. Stained cells were analyzed using a Coulter Epics XL cytometer (Coulter, Marseille, France), and the data were analyzed using WinMDI 2.8 software (facs.scripps.edu/software.html) or CellQuest (BD PharMingen). For intracellular staining, T lymphocytes were washed twice in PBS and fixed for 4 min with 2% paraformaldehyde. After 2 washes with PBS containing 2.5% fetal calf serum and 0.02% NaN 3 , cells were permeabilized in 1% saponin in PBS for 7 min at room temperature. Cells were subsequently incubated for 30 min with the indicated antibodies and washed three times with PBS, 2.5% fetal calf serum, 0.02% NaN 3 , and 0.1% saponin. Cells were subsequently incubated with the appropriate secondary reagents for 30 min, washed, and analyzed as described above. Lymphocytes were appropriately gated on forward and side scatter. For the detection of cell death, cells were stained with propidium iodide and fluorescein isothiocyanate-conjugated annexin V (Coulter-Immunotech, Marseille, France) according to the manufacture's instructions.
Cell Lysis, Precipitation, and Immunoblot Analysis-Cells were resuspended at 10 7 cells/ml in supplemented RPMI 1640 or Dulbecco's modified Eagle's medium containing 0.05 mM Na 3 VO 4 , and where indicated, stimulated for 3 min with 10 g/ml soluble anti-CD3⑀ mAb OKT3. Subsequently, cells were lysed for 10 min on ice in 50 mM Tris, pH 7.6, 150 mM NaCl, 10 mM Na 3 VO 4 , 10 mM NaF, 10 g/ml aprotinin, 10 g/ml leupeptin, 2 M phenylmethylsulfonyl fluoride, and 1% Triton X-100. Lysates were centrifuged at 20,000 ϫ g for 15 min at 4°C. Upon centrifugation, postnuclear supernatants were immunoprecipitated using mAbs previously bound to protein A-Sepharose beads. The eluted samples were resolved on SDS-PAGE under reducing conditions, transferred to polyvinylidene fluoride membrane, and immunoblotted with the indicated antibodies. Blots were revealed with ECL Western blotting kit (Amersham Biosciences). For ERK detection, blots were stripped for 30 min at 50°C in stripping buffer (62.5 mM Tris, pH 6.8, 2% SDS, 10 mM ␤-mercaptoethanol), washed intensively in PBS, 0.05% Tween 20 (Sigma-Aldrich), and reprobed with anti-ERK2 mAb. The Western blot analysis of Fig. 3 was performed on detergent-soluble and -insoluble material resolved on SDS-PAGE.
Assessment of O 2 . Production-Neutrophils were resuspended at 10 6 cells/ml in RPMI 1640 and activated with 1 M fMLP for 20 min at 37°C. Superoxide anion generation was assessed spectrophotometrically by the superoxide dismutase-inhibitable reduction of cytochrome c, as previously described (39). (40) characterized by the presence of infiltrating leukocytes, of which activated neutrophils are the main constituents (41). We therefore reasoned that RA SF neutrophils could induce T cell hyporesponsiveness by producing ROS. To test this hypothesis T lymphocytes were isolated from the peripheral blood (PB) of RA patients and co-cultured for 16 h with autologous SF or PB neutrophils. Subsequently, neutrophils were eliminated, and T lymphocytes were stimulated with immobilized anti-CD3⑀ mAb OKT3. As shown in Table I, proliferative responses of T cells preincubated with SF neutrophils were lower than those of T lymphocytes alone. Proliferative responses of T cells cultured with PB neutrophils decreased as well, but to a lesser extent. T cell proliferation was partially recovered when catalase was added during the coculture, indicating that hydrogen peroxide played a critical role in T cell hyporesponsiveness.

In Vitro Induction of T Cell Hyporesponsiveness Using Synovial Fluid-derived Neutrophils-RA synovial fluid (SF) is an oxidative environment
Induction of T Cell Hyporesponsiveness Using in Vitro Activated Neutrophils-Analysis of the molecular mechanisms in- volved in oxidative stress-induced T cell hyporesponsiveness is severely hampered by the limited number of T cells that can be isolated from e.g. synovial fluid biopsies. We therefore developed a system in which hyporesponsive T lymphocytes are generated in vitro through co-culture of buffy coat-derived T cells and fMLP-activated neutrophils (Fig. 1). The level of induced T cell hyporesponsiveness directly correlated with the amount of O 2 . produced in the different experiments (p Ͻ 0,01; Fig. 1A). Moreover, treatment with three different anti-oxidants, NAC, MnTBAP, and catalase, before the anti-CD3⑀ mAb-mediated stimulation restored the proliferative response (Fig. 1B). The same degree of hyporesponsiveness was seen with all the concentration of anti-CD3⑀ mAb tested (Fig. 1C). These data show that fMLP-activated buffy coat-derived neutrophils can induce T cell hyporesponsiveness through the action of ROS. Because T lymphocytes exposed to high doses of H 2 O 2 undergo apoptosis (42), we wished to investigate if the hyporesponsiveness induced in our system was due to T cell apoptosis. Annexin V and propidium iodide staining of T cells indicated that activated neutrophils did not induce cell death of cocultured T lymphocytes (Fig. 2). Furthermore, upon anti-CD3⑀ mAb-mediated stimulation of oxidative stress-exposed T cells, only limited cell death induction was observed (Fig. 2). In contrast, T lymphocytes shortly treated high concentrations of H 2 O 2 were dead (Fig. 2). These data show that T cell hyporesponsiveness induced by co-culture with fMLP-activated neutrophils is not due to induction of cell death.
Expression Level of Signaling Molecules in T Cells Rendered Hyporesponsive in Vitro-T cells isolated from patients affected by different diseases as RA, cancer, and patients affected with AIDS express lower levels of the TCR-chain and, in some cases, p56 lck (7-12, 18, 43). As shown in Fig. 3A, T cells rendered hyporesponsive in vitro appeared to express slightly lower levels of p56 lck , as detected using flow cytometry and Western blotting (Fig. 3A). The expression level of TCR-chain measured by flow cytometry with a monoclonal antibody against the cytoplasmic tail of the molecule was strongly downmodulated (Fig. 3B). In contrast, Western blot analysis using a monoclonal TCR--specific antibody specific for an epitope (amino acids 36 -54) containing a part of the transmembrane region of the molecule revealed similar expression levels in normal versus hyporesponsive T cells (Fig. 3B). These data suggest that in hyporesponsive T cells, TCR-could have undergone a conformational change. To analyze the role of ROS in the modulation of these signaling molecules, the expression level of p56 lck and TCR-was analyzed in T lymphocytes co- FIG. 1. Oxidative stress induces T cell hyporesponsiveness in  vitro. A, T lymphocytes were co-cultured with or without activated neutrophils. Upon co-culture, purified T cells were stimulated, and their proliferative response was measured. O 2 . production by neutrophils was measured. Depicted are the inhibition of T cell proliferation of oxidative stress-exposed T cells as compared with T cells cultured without neutrophils as a function of O 2 . produced by the neutrophils. cultured with neutrophils in the presence of catalase. As shown in Fig. 3C a normal expression level of p56 lck and TCR-is observed in the presence of this anti-oxidant. As detected by Western blot analysis, expression levels of p59 fyn , ZAP-70, LAT, and phosphatidylinositol 3-kinase were similar in the two T cell populations (Fig. 3D). Impaired Calcium Mobilization in Hyporesponsive T Lymphocytes upon TCR Engagement-To analyze signaling events associated with TCR engagement, we examined changes in [Ca 2ϩ ] i by flow cytometry. After stimulation with an anti-CD3⑀ mAb (OKT3), a rapid and sustained increase in [Ca 2ϩ ] i flux was observed in normal T lymphocytes, whereas a decreased mobilization of [Ca 2ϩ ] i was found in hyporesponsive T cells (Fig. 4). Thus, the TCR in hyporesponsive T cells is unable to efficiently couple to mechanisms responsible for increases in [Ca 2ϩ ] i , suggesting that ROS exposure results in an impaired signal transduction through the TCR.
To demonstrate that ROS were responsible for the alteration of TCR-dependent Ca 2ϩ mobilization, T lymphocytes were cocultured with neutrophils in the presence of catalase. As shown in Fig. 4, catalase was able to restore [Ca 2ϩ ] i mobilization, directly implicating ROS in this signaling defect.
Decreased Tyrosine Phosphorylation of PLC-␥1 but Normal TCR-Phosphorylation in Hyporesponsive T Lymphocytes-To examine upstream events of the TCR-signaling pathway, we analyzed the pattern of tyrosine-phosphorylated proteins induced by anti-CD3⑀ stimulation in normal and hyporesponsive T cells (Fig. 5). Anti-CD3⑀ stimulation induced an increase in tyrosine phosphorylation of a high molecular mass protein in normal T lymphocytes (e.g. ϳ150 kDa, Fig. 5), which was almost undetectable in oxidative stress-exposed T cells. Intriguingly, the TCR-dependent tyrosine phosphorylation of low molecular mass proteins (e.g. ϳ23 kDa, Fig. 5) was less affected. A protein of around 40 kDa was already substantially tyrosinephosphorylated in unstimulated hyporesponsive T cells, probably due to the known effects of ROS on kinase and phosphatase activities (44).
Tyrosine phosphorylation of TCR-chain (molecular mass, 21-23 kDa, phospho-) is one of the first biochemical events detectable in T cells upon TCR ligation. To assess whether the tyrosine-phosphorylated protein of ϳ23 kDa observed in our phosphotyrosine blot corresponded to phospho-, lysates from unstimulated and stimulated cells were immunoprecipitated with an mAb against the TCR-chain. The precipitates were resolved by SDS-PAGE and immunoblotted with either antiphosphotyrosine or anti-TCR-mAb. In normal and hyporesponsive T lymphocytes, TCR engagement led to a similar increase in TCR-phosphorylation, as demonstrated by the induction of the p21 and p23 forms of this subunit of the TCR complex. Blotting with anti-TCR-chain antibody showed that comparable amounts of the protein were immunoprecipitated in all the lanes (Fig. 6A). Therefore, ROS exposure does not profoundly alter the most proximal measurable TCR-signaling event in T lymphocytes.
Tyrosine phosphorylation of PLC-␥1 results in the hydrolysis of phosphatidylinositol-4 -5-biphosphate to inositol-1,4,5triphosphate and diacylglycerol. Inositol-1,4,5-triphosphate generation induces a sustained increase in intracellular cal-  A, B, and D). C, purified normal and hyporesponsive T lymphocytes (cultured in the presence or absence of 1000 units/ml of catalase) were fixed, permeabilized, and stained with anti-p56 lck or anti-TCR-antibodies and analyzed by flow cytometry. ctrl, control; U, units. cium, whereas diacylglycerol promotes the activation of protein kinase C. Because TCR ligation induced reduced Ca 2ϩ flux in hyporesponsive T lymphocytes, we wondered whether PLC-␥1 activation was defective in these cells. Consistent with this possibility, an inducible tyrosine-phosphorylated protein of ϳ150 kDa, seen in activated T cells, is not observed in hyporesponsive T lymphocytes (Fig. 5). Immunoprecipitation of lysates of normal and hyporesponsive T cells with an anti-PLC-␥1 antibody followed by phosphotyrosine blotting indicated that upon TCR stimulation PLC-␥1 does not get phosphorylated in hyporesponsive T lymphocytes (Fig. 6B). The same blot was reprobed with an anti-PLC-␥1 antibody to confirm that equal amounts of the protein were immunoprecipitated in all the lanes. To directly assess the role of ROS in the inhibition of TCR-mediated PLC-␥1 activation, T lymphocytes were co-cultured with neutrophils in the presence of catalase. As shown in Fig. 6B catalase restores TCR-dependent PLC-␥1 activation. The increase in steady state phosphorylation of PLC-␥1 observed in normal and hyporesponsive T cells treated with catalase does not preclude a subsequent TCRmediated activation of this enzyme. These results provide compelling evidence that ROS affect PLC-␥1 activation and, consequently,Ca 2ϩ mobilization in T lymphocytes.
Altered ERK Activity in Hyporesponsive T Lymphocytes-Because the most proximal signaling events upon TCR engagement in T lymphocytes are not affected by exposure to oxidative stress, we assessed if more distal events not dependent on PLC-␥1 phosphorylation, as ERK activation, were functional.
The presence of activated ERK1/2 in T lymphocytes exposed or not to oxidative stress in vitro was analyzed by Western blot. In contrast to normal resting T cells, in non-stimulated hyporesponsive T cells, ERK1/2 phosphorylation was observed (Fig.7), showing that physiological concentrations of H 2 O 2 activated the MAP kinase signaling cascade, confirming and extending earlier reports using micromolar concentrations of H 2 O 2 (45,46). Surprisingly, this constitutive activation did not preclude further activation of the enzyme, as indicated by the increased phosphorylation of ERK1/2 induced by anti-CD3⑀ mAb OKT3 stimulation of hyporesponsive T cells (Fig 7).
Intact Caspase 3 Activation in Hyporesponsive T Lymphocytes-To determine whether other TCR-signaling pathways were functional in hyporesponsive T cells, we analyzed caspase 3 processing. It has been shown that TCR engagement leads to FIG. 4. Impaired calcium mobilization upon TCR engagement in hyporesponsive T lymphocytes. Purified normal and hyporesponsive T lymphocytes (cultured in the presence or absence of 1000 units/ml of catalase) were loaded with indo-1 for 45 min at 37°C, washed, and resuspended in Ca 2ϩ buffer. The cells were then stimulated with an anti-CD3⑀ mAb (OKT3, 10 g/ml, arrow), and changes in [Ca 2ϩ ] i were measured by flow cytometry. U , units.

FIG. 5. TCR-mediated protein phosphorylation substrates in hyporesponsive T lymphocytes.
Normal and hyporesponsive T cells stimulated or not with soluble anti-CD3⑀ mAb (OKT3, 10 g/ml) for 3 min were lysed, and tyrosine-phosphorylated proteins were immunoprecipitated (IP) with anti-phosphotyrosine mAb and analyzed by Western blot. caspase 3 activation and that this process is essential for T cell proliferation (47). Caspase 3 processing through its proteolytic cleavage generates protein species ranging from 17-to 24-kDa proliferation (47). To study the effect of oxidative stress on this newly described TCR-dependent signaling pathway, anti-CD3⑀ mAb OKT3-induced caspase 3 processing was analyzed in normal and hyporesponsive T lymphocytes. As shown in Fig. 8, cleavage of the p32 form of caspase 3 and concomitant appearance of a proteolytic fragment of around 20 -22 kDa was observed in normal as well as in hyporesponsive activated T cells. DISCUSSION Here we report that oxidative stress induces T cell hyporesponsiveness by targeting specific components of the TCR-signaling machinery. Oxidative stress inhibits TCR-dependent PLC-␥1 activation and, consequently, Ca 2ϩ mobilization. Importantly, the most proximal TCR signaling event analyzed, the phosphorylation of the TCRchain, is only marginally if at all affected in hyporesponsive T cells. This allows other downstream events such as activation of ERK1/2 and caspase 3 to take place upon TCR engagement.
The maintenance of both intra-and extracellular reducing conditions is a prerequisite for the proper functioning of T lymphocytes. Normally, T cells control the intracellular redox balance through various cytosolic anti-oxidant systems. However, the T cell microenvironment is always pushed to oxidation by various factors, one of which is the production of ROS by neutrophils and macrophages at the site of inflammation (48). It has been estimated that in the microenvironment of these cells the concentration of hydrogen peroxide can reach 10 -100 M (49 -51). This oxidative milieu can become chronic if inflammation persists and macrophages and neutrophils are continuously recruited. The generation of an oxidative environment has a strong influence on T lymphocytes. It has been shown that T cells isolated from patients affected with rheumatoid arthritis, cancer, leprosis, or AIDS show altered functional properties (1)(2)(3)(4)52), and an important role for oxidative stress has been suggested (6,7,11,53,54). Previous in vitro work on the effects of oxidative stress on T lymphocytes showed that ROS induce T cell hyporesponsiveness (55) and alter the expression levels of key T cell-signaling molecules such as TCRor p56 lck (6 -9, 11, 13, 14). Interpretation of these results is made difficult by the high doses of exogenously added hydrogen peroxide utilized (1-10 mM) that also lead to T cell apoptosis (42).
To circumvent this problem, we established an in vitro system in which freshly isolated T cells are rendered hyporesponsive by exposing them to oxidative stress generated by activated syngenic neutrophils. We showed that RA SF neutrophils can induce T cell hyporesponsiveness ex vivo by secreting H 2 O 2 . Important also, fMLP-activated neutrophils (producing comparable concentrations of H 2 O 2 as SF RA neutrophils (i.e. ϳ0.01-0.02 nmol/10 6 neutrophils/20 min) induced T cell hyporesponsiveness. Although we can formally not exclude that fMLPstimulated neutrophils release other agents that contribute to the induction of hyporesponsiveness, the observation that Constitutive and TCR-mediated increased ERK1/2 activation in hyporesponsive T lymphocytes. Normal and hyporesponsive T cells were stimulated or not with soluble anti-CD3⑀ mAb (OKT3, 10 g/ml) for 3 min, and the induced ERK1/2 phosphorylation was analyzed by Western blot of total lysates. Upon stripping, blots were reprobed with an anti-ERK2 antibody to measure total ERK protein levels in the two cell types. NAC, MnTBAP, and catalase restored T cell responsiveness indicates the critical role of oxidative stress.
As compared with normal T cells, in hyporesponsive T lymphocytes TCR-dependent tyrosine phosphorylation of PLC-␥1 is defective, leading to a strongly decreased Ca 2ϩ flux in these cells. Importantly, these signaling defects are restored by the addition of catalase, indicating that they are ROS-dependent. Our results are in apparent contrast with the reported H 2 O 2induced PLC-␥1 activation in mouse embryonic fibroblasts. The discrepancy is probably due to differences in experimental systems, in particular to the high doses of H 2 O 2 utilized by Wang et al. (56). Moreover, it has previously been shown that sulfhydryl oxidation down-regulates PLC-␥1 activation in T cells (57). Taken together, these observations show that PLC-␥1 is a target of ROS and its inhibition causes a block of downstream TCR-signaling pathways.
In contrast to PLC-␥1, TCR signaling led to increased ERK-2 activation in normal and hyporesponsive T cells. However, ERK1/2 appeared already activated in non-TCR-stimulated hyporesponsive T cells. It is unknown whether ERK1/2 can be a direct target of ROS, but it has been previously shown that Ras is activated by H 2 O 2 , thereby probably leading to ERK1/2 activation (58). It has recently been shown that mild oxidative stress activates other MAP kinase-signaling pathways other than ERK1/2 (59). The difference between the latter and our data may be explained by differences in the nature of the oxidative stimulus; although Hehner and Droge (59) used agents altering intracellular glutathione levels, in our system ROS are generated by activated neutrophils. Potential inhibitory or activating effects of ROS-induced ERK1/2 activation on downstream effectors are unknown.
It has recently been reported that caspase-3 activation is a necessary event for normal T cell activation (47). TCR-mediated activation of this critical signaling cascade was normal in oxidative stress-exposed T cells. Although the ultimate effects of caspase-3 activation during T cell responses are unknown, this result suggests that certain effector functions may be maintained in hyporesponsive T cells. It has been shown that in T cells, H 2 O 2 -triggered cell death leads to the induced cleavage and activation of caspase 3 (60). The fact that caspase 3 cleavage is not observed in unstimulated hyporesponsive T cells further supports the validity of our in vitro system and its relevance for human pathology.
The results reported here uncover the divergent effects of ROS on T cell-signaling pathways. Using a source of ROS that mimic in vivo conditions, we have observed that TCRdependent PLC-␥1 activation is inhibited, whereas activation of other proteins such as TCR-, ERK, and caspase 3 is still functional. Our data indicate that TCR-signaling pathways are differentially affected by oxidative stress and urge a thorough investigation of T cell effector functions remaining operational in T lymphocytes exposed to oxidative stress in vivo or in vitro.