Role of Increased Guanosine Triphosphate Cyclohydrolase-1 Expression and Tetrahydrobiopterin Levels upon T Cell Activation

Abstract

Tetrahydrobiopterin (BH₄) is an essential co-factor for the nitric-oxide (NO) synthases, and in its absence these enzymes produce superoxide (O₂^˙̄) rather than NO. The rate-limiting enzyme for BH₄ production is guanosine triphosphate cyclohydrolase-1 (GTPCH-1). Because endogenously produced NO affects T cell function, we sought to determine whether antigen stimulation affected T cell GTPCH-1 expression and ultimately BH₄ levels. Resting T cells had minimal expression of inducible NOS (NOS2), endothelial NOS (NOS3), and GTPCH-1 protein and nearly undetectable levels of BH₄. Anti-CD3 stimulation of T cells robustly stimulated the coordinated expression of NOS2, NOS3, and GTPCH-1 and markedly increased both GTPCH-1 activity and T cell BH₄ levels. The newly expressed GTPCH-1 was phosphorylated on serine 72 and pharmacological inhibition of casein kinase II reduced GTPCH-1 phosphorylation and blunted the increase in T cell BH₄. Inhibition of GTPCH-1 with diaminohydroxypyrimidine (1 mmol/liter) prevented T cell BH₄ accumulation, reduced NO production, and increased T cell O₂^˙̄ production, due to both NOS2 and NOS3 uncoupling. GTPCH-1 inhibition also promoted TH₂ polarization in memory CD4 cells. Ovalbumin immunization of mice transgenic for an ovalbumin receptor (OT-II mice) confirmed a marked increase in T cell BH₄ in vivo. These studies identify a previously unidentified consequence of T cell activation, promoting BH₄ levels, NO production, and modulating T cell cytokine production.

Introduction

The nitric-oxide synthase (NOS)² enzymes were initially identified in neuronal cells (nNOS or NOS1), macrophages (iNOS or NOS2), and the endothelium (eNOS or NOS3). It is now recognized that NO has divergent signaling roles in various cells types where these enzymes are present (1). Of interest, T cells produce NO and express both NOS2 and NOS3 (2, 3). Endogenously produced NO has myriad effects on T cell function, including modulation of T cell memory, induction of apoptosis, regulation of cytokine production, and T cell polarization (2, 4). Recent studies have suggested that in CD4⁺ cells, ligation of the T cell receptor leads to immediate activation of NOS3, which translocates to the immune synapse and affects a variety of signaling events, increases IFN-γ and decreases IL-2 production (3, 5). Endogenously produced NO can either induce or suppress T regulatory cells depending on the nature of the external stimulus (6, 7).

All NOS enzymes require tetrahydrobiopterin (BH₄) as a co-factor. During NOS catalysis, electrons are transferred from the flavin-containing reductase domain to a heme group in the oxygenase domain. This reduces the heme iron to a ferrous state. The ferrous heme readily binds oxygen, forming a ferrous-dioxy (FeII-O₂) intermediate. BH₄, which is also bound in the oxygenase domain, donates an electron that causes scission of FeII-O₂ to an iron-oxy species (8), which in turn hydroxylates one of the guanidino nitrogens of l-arginine, initially to N-hydroxy-l-arginine and subsequently to citrulline and NO (9, 10). BH₄ also stabilizes dimers of some of NOS isoforms (11, 12). In the absence of BH₄, the NOS enzymes produce superoxide (O₂^˙̄) and other reactive oxygen species rather than NO. This condition is referred to as NOS uncoupling (13–15). Uncoupling of NOS3 contributes to the vascular pathology associated with common diseases such as hypertension, diabetes, and atherosclerosis (16, 17).

BH₄ is produced by a complex enzymatic pathway involving the serial actions of GTP cyclohydrolase-1 (GTPCH-1), pyruvoyl tetrahydropterin synthase, and sepiapterin reductase (18). The first of these enzymes, GTPCH-1, converts guanosine 5′-triphosphate to 7,8-dihydroneopterin triphosphate and is most often rate-limiting. This enzyme exists as a homodecamer, and its activity is regulated in a negative feedback fashion by levels of BH₄, which promotes binding of GTPCH-1 with its inhibitor protein (GFRP) (19). Thus, when cellular BH₄ levels are sufficient, GFRP binds to and inhibits GTPCH-1 function in a negative feedback fashion. Recently, we found that phosphorylation of human GTPCH-1 at serine 81 reduces its binding to GFRP and thus prevents inhibition by this negative feedback process (20).

Given the importance of GTPCH-1 and BH₄ in cellular NO production and the evolving understanding of NO in T cell function, the present study was performed to understand how T cell levels of BH₄ and T cell GTPCH-1 activity are regulated. Our studies show that resting T cells have virtually no NOS, GTPCH-1, or BH₄. Ligation of the T cell receptor leads to expression of GTPCH-1 and both NOS2 and NOS3. We further examined the importance of BH₄ production in modulating T cell NO and O₂^˙̄ production and CD4⁺ cytokine production.

EXPERIMENTAL PROCEDURES

T Cell Isolation and Culture

Total splenocytes were obtained from mouse spleens, and CD3⁺ T cells were isolated using the Pan T cell isolation kit (Milentyl Biotec, catalog no. 130-090-861). Respective T cell subtypes were isolated from CD3⁺ cells using CD4, CD8, and CD4⁺ CD62L T cell isolation kits obtained from Milentyl Biotec (catalog numbers 130-090-860, 130-090-859, and 130-049-701, respectively). Memory CD4⁺, CD62L CD44^high cells were isolated using a kit from R&D Systems (catalog number MAGM206). Cells were cultured in RPMI 1640 medium (Mediatech, Inc., catalog number 15-041-CV) supplemented with 1% (v/v) l-glutamine, 10% (v/v) FBS, 1% (v/v) penicillin-streptomycin, 50 μm β-mercaptoethanol. Cells were maintained at 37 °C in 5% CO₂. For T cell activation, cells were cultured for 48 h on 96-well plates coated with anti-CD3 antibody from BD Biosciences (catalog number 354720) in the above medium containing 20 units/ml IL-2. In some experiments, dasatinib (50 nmol/liter; Selleck Chemicals) was added to the tissue culture. In experiments involving prolonged T cell culture, medium without IL-2 was substituted after 48 h.

Mice Studied

C57BL/6, OT-II, NOS2^−/−, and NOS3^−/− mice were obtained from Jackson Laboratories. They were studied at 3 months of age.

Western Blotting and mRNA Quantification

Western blotting was performed as described previously (20). Antibodies for NOS1, NOS2, and NOS3 were from BD Biosciences (catalog numbers 610310, 610332, and 610298, respectively) and were used at a 1:3000 dilution. Polyclonal antibodies to GTPCH-1 amino acids 1–20 and a mouse serine 72 phosphoantibody were raised in rabbits and used at dilutions of 1:2000 and 1:1000, respectively. A secondary anti-rabbit antibody from Bio-Rad was employed at a dilution of 1:5000. Quantitative real-time PCRs were employed to quantify mRNA levels of GFRP using commercially available Taqman primers and an Applied Biosystems 7500 PCR system.

Quantification of Pterins, O₂^˙̄, and NO

Total biopterin, BH₄, and oxidized forms of BH₄ were quantified using differential oxidation in acid and alkaline conditions followed by HPLC as described previously (20, 21). Assays for GTPCH-1 activity were performed on homogenates of T cells as described recently for other tissues (20, 21). O₂^˙̄ was detected by monitoring the oxidation of dihydroethidium to the specific product 2-hydroxyethidium as described previously (22). NO was detected using the spin trap Fe[DETC]₂ and electron spin resonance as described previously (23).

Cytokine Measurements

T cells were isolated from mouse spleen and cultured on anti-CD3 plates with or without the GTPCH-1-specific inhibitor 2,4 diamino-6-hydroxypyrimadine (DAHP) at a final concentration of 1 mmol/liter for 7 days. Following this, levels of cytokines IFN-γ, IL-4, IL-5, and TNF-α in the media were measured using cytokine bead array (catalog number 551287; BD Biosciences). In some experiments, either the NO donor DETA-NO (1 μmol/liter; catalog number 82120, Cayman Chemical) or PEG-SOD (200 units/ml; catalog number S9549, Sigma) were added to the media. Media and drugs were changed every 48 h.

Statistical Analysis

Data in this paper are presented as means ± S.E. When comparing two or three variables, two-tailed unpaired t tests were employed with a Bonferroni correction when appropriate. When comparing multiple variables, analysis of variance (ANOVA) was employed using a Dunnett's post hoc, allowing comparison to control values. Two-way ANOVA was also employed when comparing the effect of an intervention on different cells.

RESULTS

Effect of T Cell Activation on NOS Expression and Pterin Content

In initial experiments, we found that resting T cells express minimal levels of either NOS2 or NOS3 protein; however, stimulation with anti-CD3 for 48 h robustly stimulated expression of both of these (Fig. 1, A and B). NOS1 was not detectable in either resting or stimulated T cells as determined by Western blotting and real-time PCR (data not shown). Because the NOS enzymes require BH₄ as an essential co-factor for production of NO, we also examined levels of BH₄ in resting and stimulated T cells. Levels of BH₄ and its oxidized products were nearly undetectable in resting T cells, but were markedly increased by exposure to anti-CD3-coated plates for 48 h (Fig. 1C). To determine whether this was specific for either CD4 or CD8 T cells, we isolated subsets of these cells using negative selection and exposed these cells to anti-CD3. Pterin synthesis was stimulated to a roughly equal extent in both CD4 and CD8 T cells (Fig. 1D). The ratio of oxidized to reduced BH₄ was greater in CD8⁺ compared with CD4⁺ T cells (Fig. 1E). These data show that activation of murine T cells leads to a striking induction of both NOS enzymes and BH₄ production in a coordinated fashion.

FIGURE 1.

Levels of NOS and pterins in resting and stimulated T cells. A and B, murine T cells were isolated from splenocytes and studied at base line or after stimulation on anti-CD3 plates for 48 h. Levels of NOS2 and NOS3 protein were detected by Western blotting (A) and quantified by densitometry (B, n = 4 in each group). Cells were also homogenized and subjected to differential oxidation with iodine in acid and alkaline conditions, and HPLC was employed for detection of total pterins and BH₄. C shows effect of T cell stimulation on total pterins. D shows relative levels of BH₄ and oxidized pterins in CD4 and CD8 T cells. E shows ratio of BH₄/BH₂ in respective T cell subtypes (n = 5–7). p values are from unpaired t tests except for D, where two-way ANOVA was used.

Effect of T Cell Activation on GTPCH-1 Activity and Phosphorylation

The enzyme GTPCH-1 plays a critical role as the rate-limiting step in BH₄ biosynthesis. In keeping with the marked increase in BH₄ caused by T cell receptor ligation, we also found that this stimulus markedly increased GTPCH-1 protein levels (Fig. 2, A and B). Likewise, GTPCH-1 activity was essentially undetectable in resting T cells, but was markedly increased upon exposure of cells to anti-CD3 for 48 h (Fig. 2C). T cell activation was not associated with a change in mRNA levels of GFRP (Fig. 2D).

FIGURE 2.

Effect of T cell activation on GTPCH-1 expression, GTPCH-1 activity, and GFRP mRNA. A–C, total GTPCH-1 expression was determined by Western blotting in homogenates of T cells. GTPCH-1 activity was measured in homogenates of resting and anti-CD3 stimulated T cells by examining production of dihydroneopterin triphosphate using HPLC (C, n = 3–4). D, mRNA level of the GTPCH-1 feedback regulatory protein was determined by real-time PCR (n = 6). Data in B, C, and D were analyzed using unpaired t tests.

In other cells, cytokines induce expression of GTPCH-1 (24). Thus, the increase in this protein and T cell levels of BH₄ could have been due to IL-2, which was used in culture of the T cells, rather than T cell receptor ligation per se. We found, however, that GTPCH-1 expression was increased by anti-CD3 even when IL-2 was excluded from the media, indicating that this cytokine does not significantly alter GTPCH-1 expression (data not shown). An important downstream signaling event of T cell receptor ligation is activation of the tyrosine kinase Lck, which is inhibited by dasatinib (25). Addition of dasatinib during exposure to anti-CD3 completely abrogated the increase in GTPCH-1 protein (Fig. 3, A and B).

FIGURE 3.

Effect of T cell signaling and GTPCH-1 phosphorylation on GTPCH-1 and BH₄ levels. Resting T cells were obtained from splenocytes by negative selection. Cells were cultured in the presence and absence of the Lck inhibitor dasatinib (50 nmol/liter) or the casein kinase II inhibitor TBB (30 μmol/liter). A shows an example of Western blotting, and B shows mean data, illustrating the effect of Lck inhibition on GTPCH-1 expression in T cells. C shows an example Western blotting (upper) and mean data (lower) for effect of TBB on phosphorylation of GTPCH-1 on serine 72 (n = 4–6). D, levels of BH₄ were detected by HPLC after exposure to anti-CD3 in the presence and absence of TBB (n = 9 for each). The p values in B and C were obtained using ANOVA with Dunnett's post hoc test (n = 5–6). Values in D were compared using unpaired t tests.

Previously, we found that human endothelial GTPCH-1 is phosphorylated by casein kinase II at serine 81 in response to mechanical shear stress and that this increases its activity by 30-fold (20). In the present study, we found that the newly synthesized GTPCH-1 in T cells was also phosphorylated at serine 72 (the mouse equivalent of human serine 81; Fig. 2C). In keeping with prior findings in human endothelial cells, the casein kinase II inhibitor 4,5,6,7-tetrabromobenzotriazole (TBB) (26), reduced GTPCH-1 phosphorylation (Fig. 3C) and decreased GTPCH-1 activity (Fig. 3D).

Role of BH₄ Stimulation in Modulation of T Cell NO and Superoxide Production

We next sought to determine whether GTPCH-1 activation is necessary for function of NOS within T cells. To accomplish this, we exposed cells to anti-CD3 in the presence and absence of the GTPCH-1 inhibitor DAHP (1 mmol/liter) (27). As evident in Fig. 4A, DAHP prevented the increase in both reduced and oxidized BH₄ in T cells exposed to anti-CD3. In keeping with a critical role of BH₄ in regulating NO production, exposure of T cells to anti-CD3 stimulated NO production as detected by the spin trap Fe[DETC]₂ and electron spin resonance, and this increase was abolished by concomitant treatment of cells with DAHP (Fig. 4B). In the absence of BH₄, the NOS enzymes produce O₂^˙̄ rather than NO (14, 15). We sought to determine whether this was true in T cells and to examine the relative contributions of NOS2 and NOS3 to this process. We therefore examined production of O₂^˙̄ by T cells isolated from spleens of wild-type, NOS2^−/−, and NOS3^−/− mice. At base line, these produced similar levels of O₂^˙̄ (Fig. 4C). Exposure of wild-type T cells to DAHP increased T cell O₂^˙̄ production by almost 2-fold (Fig. 4C). Both NOS2 and NOS3 seemed to be sources of this radical because DAHP did not increase O₂^˙̄ in cells lacking these compared with the wild-type cells. To examine further a role of NOS2, we treated NOS3-deficient cells with the specific NOS2 inhibitor 1400W, which dissociates the heme from NOS2 and prevents the ability of this enzyme to produce O₂^˙̄ (28, 29). Addition of 1400W decreased NOS3^−/− cell production of O₂^˙̄ both at base line and following DAHP treatment. Thus, when murine T cells are stimulated by anti-CD3, the increase in BH₄ is essential to allow adequate function of the newly synthesized NOS isoforms.

FIGURE 4.

Role of BH₄ in modulating T cell production of NO and O₂^˙̄. A, inhibition of GTPCH-1 with DAHP (1 mmol/liter) during anti-CD3 stimulation markedly reduced T cell levels of both reduced (BH₄) and oxidized pterins. B, values with and without DAHP were compared using unpaired t tests (n = 5). NO production was measured using the spin trap Fe[DETC]₂ and electron spin resonance (n = 4). C, superoxide was measured by monitoring oxidation of dihydroethidium to 2-hydroxyethidium in T cells from wild-type, NOS2^−/−, and NOS3^−/− mice at base line and following exposure to DAHP (1 mmol/liter) for 48 h (n = 3–4). The NOS2 inhibitor 1400W (100 nmol/liter) was added to NOS3^−/− cells during exposure to anti-CD3 in some experiments. In C, two-way ANOVA was also used to compare levels of O₂^˙̄ production by T cells from various strains of mice in the absence and presence of DAHP.

Role of NOS Uncoupling on T Cell Cytokine Production and Relative Effects of NO and O₂^˙̄

Both O₂^˙̄ and NO have been reported to affect T cell polarization (30–33). It is therefore likely that uncoupling of the NOS enzymes could modulate T cell cytokine production. To examine this, we isolated memory murine CD4⁺ T cells and exposed them to anti-CD3-coated plates for 7 days in the absence and presence of DAHP, to prevent BH₄ synthesis. Incubation of T cells with DAHP decreased IFN-γ production by one-half and doubled IL-4 production (Fig. 5, A and B). In contrast, DAHP co-incubation had no effect on either IL-5 or TNF-α production (Fig. 5, C and D). These changes in IL-4 and IFN-γ production could have been due to an increased O₂^˙̄ or decreased NO production. To determine which of these might have contributed, we added either PEG-SOD (200 units/ml) or the NO donor DETA-NO (1 μmol/liter) every 48 h throughout the incubation. Addition of the NO donor had no effect on production of any of the cytokines studied; however, PEG-SOD restored production of IFN-γ and IL-4 to levels observed in the absence of DAHP (Fig. 5, A and B). PEG-SOD also decreased IL-5 production (Fig. 5C). These studies indicate that NOS uncoupling leads to polarization of T cells toward a TH₂ phenotype and that this is largely due to an increase in cellular O₂^˙̄ production.

FIGURE 5.

Role of BH₄, O₂^˙̄, and NO in modulating T cell cytokine production. T cells were cultured on anti-CD3 plates as described under “Experimental Procedures” for 7 days with and without DAHP (1 mmol/liter). In some experiments, cells were supplemented with either DETA-NO (1 μmol/liter) or PEG-SOD (200 units/ml) every 48 h. Cytokines were measured by cytokine bead array. Values were compared using ANOVA with Dunnett's post hoc test (n = 5–6 for each).

Effect of T Cell Activation in Vivo on Pterin Levels

To determine whether T cell activation in vivo also stimulates levels of BH₄, we performed additional studies in OT-II mice. These animals are transgenic for a T cell receptor that induces production of CD4⁺-activated T cells when immunized using the ovalbumin peptide 323–339 (34). OT-II mice were injected intraperitoneally with this peptide (0.5 μg/μl) or vehicle (Al(OH)₃, 2.5 μg/μl) on two occasions 1 week apart. Five days following the final ovalbumin injection, CD62L (effector) T cells were isolated and pterin levels measured. As evident in Fig. 6, CD4 effector T cells from OT-II mice immunized with ovalbumin peptide contained ∼4-fold more total pterins and BH₄ than did cells from animals injected with the Al(OH)₃ vehicle. These findings indicate that T cell activation in vivo stimulates BH₄ synthesis in these cells.

FIGURE 6.

Effect of in vivo activation on T cell pterin levels. OT-II mice were injected intraperitoneally with ovalbumin (Ova) peptide (0.5 μg/μl) or vehicle (Al(OH)₃, 2.5 μg/μl) on two occasions, and CD62L T cells were isolated from the spleen 1 week after the final injection. Pterins were measured using HPLC as described under “Experimental Procedures.” The p values represent results of unpaired two-sided t tests following Bonferroni correction for three comparisons (n = 4–5).

DISCUSSION

In the current study, we show that T cell receptor ligation causes a coordinated increase in both NOS2 and NOS3 while also markedly stimulating expression of the rate-limiting enzyme for BH₄ production, GTPCH-1. This concomitant stimulation of these three enzymes allows T cell production of NO and plays an important role in modulating T cell function. The increase in BH₄ occurs in both CD4 and CD8 cells and in the former, regulates cytokine production, in large part by preventing O₂^˙̄ production from the NOSs. We further show that T cell activation in vivo stimulates T cell BH₄ levels. These studies have identified a previously unknown role of BH₄ in modulating T cell function upon activation.

We found that the newly expressed GTPCH-1 is phosphorylated on serine 72. In prior studies, we found that the mechanical force of shear stress stimulates GTPCH-1 phosphorylation by casein kinase IIα′. As in the case with shear in the endothelium, the casein kinase II inhibitor TBB reduces phosphorylation of the T cell GTPCH-1, indicating a role of casein kinase II in T cells. Thus, the ultimate effects of T cell receptor ligation and laminar shear in the endothelium on GTPCH-1 phosphorylation and activity are quite similar. In both cell types, resting BH₄ and GTPCH-1 activity are very low, but are stimulated by more than 20-fold by the respective stimuli of T cell receptor ligation and endothelial laminar shear stress. It is of interest that both stimuli affect the actin cytoskeleton in the respective cell types and share similar signaling mechanisms, including activation of tyrosine kinases, G proteins, and MAP kinases (35, 36). In endothelial cells, proline-rich tyrosine kinase 2 (Pyk2) is induced rapidly upon exposure to shear stress in a redox-sensitive fashion (37). Likewise, Pyk2 is induced by T cell receptor ligation and promotes production of TH₂ cytokines in vivo (38, 39). The precise roles of these signaling pathways leading to GTPCH-1 activation remain undefined. Of interest, the tyrosine kinase Lck is upstream of Pyk2 activation in T cells (35). In the current studies, we found that the multikinase inhibitor dasatinib, which potently inhibits Lck (25), completely prevented the increase in GTPCH-1 in response to anti-CD3 exposure.

In the present study, we found that prevention of BH₄ synthesis in stimulated cells further increased T cell O₂^˙̄ production and prevented an increase in NO production (Fig. 4), compatible with uncoupling of NOS within these cells. In memory CD4⁺ cells, this led to an increase in the TH₂ cytokine IL-4 and a decrease in the TH₁ cytokine IFN-γ. These perturbations of cytokine production were prevented by co-treatment of cells with PEG-SOD, suggesting that O₂^˙̄ or a product of O₂^˙̄ such as the peroxynitrite anion contribute to TH₂ polarization of T cells. These findings are in accord with a prior study from Jackson et al., who showed that stimulation of T cell blasts with anti-CD3 caused production of reactive oxygen species as detected by the fluorescent dye dichlorofluorescein diacetate (30). In cells from mice lacking the NADPH oxidase subunit gp91^phox, these investigators showed that T cell receptor ligation augmented IFN-γ and diminished production of IL-4 and IL-5, suggesting that an absence of reactive oxygen species from this source promoted TH₁ polarization. More recently, King et al. showed that induction of intracellular O₂^˙̄ by a redox cycling agent promoted TH₂ polarization of human T cells (31). These data are in accord with our present study supporting a role of O₂^˙̄ in suppressing IFN-γ and stimulating IL-4 production.

It is conceivable that uncoupled NOS could have affected cytokine production by also reducing T cell levels of NO. Ibiza et al. have shown that upon T cell receptor ligation with an antigen-presenting cell, NOS3 moves with the Golgi apparatus to the region of the immunological synapse (3). These investigators further showed that NOS3 overexpression enhances phosphorylation of the CD3z chain and ZAP70 and increases IFN-γ production. We found that supplementation with an NO donor had no effect on cytokine production after DAHP treatment, suggesting that the predominant effect of NOS uncoupling was mediated by increased O₂^˙̄ rather than reduced NO production. Of note, DAHP did not increase IL-5, but PEG-SOD lowered its levels below those observed at base line. This suggests that uncoupled NOS does not affect IL-5 production, but that O₂^˙̄ from other sources likely modulates expression of this cytokine. These findings are in keeping with the notion that various sources of cellular reactive oxygen species can be highly localized and can differentially affect responses to extracellular stimuli.

It is interesting to speculate how T cell levels of BH₄ might be regulated in the setting of an inflammatory stimulus in vivo. In keeping with our cell culture studies, we found that ovalbumin peptide immunization of OT-II mice increased T cell levels of BH₄. For the short term, this should promote NO production and TH₁ polarization. In regions of inflammation, the production of reactive oxygen species is commonly increased. Others and we have shown that BH₄ is susceptible to degradation by oxidants such as peroxynitrite, the carbonate radical and hydroxyl (16, 40). In preliminary studies, we also found that incubation of T cells with the peroxynitrite-generating compound SIN-1 caused oxidation of intracellular BH₄ (data not shown). This would likely uncouple T cell NOS enzymes, increase NOS-derived O₂^˙̄, and promote TH₂ cytokine production, leading to resolution of the inflammatory state.

In summary, our current findings have two important implications. First, they illustrate that GTPCH-1 expression and activity are markedly enhanced by T cell receptor ligation and that this markedly increases T cell BH₄ levels and production of NO. Second, in conditions where BH₄ levels are reduced, the NOS enzymes can become an important source of O₂^˙̄ production, which in turn can alter T cell cytokine production.