Hydrogen Sulfide Is an Endogenous Potentiator of T Cell Activation*

Background: T cells encounter H2S in physiological and pathophysiological settings with unknown consequences. Results: Exogenous and endogenous H2S enhances T cell activation. Conclusion: T cell activation depends on H2S signaling. Significance: H2S can act as an autocrine or paracrine T cell activator and suggests a mechanistic link to inflammatory bowel disease progression. H2S is an endogenous signaling molecule that may act via protein sulfhydrylation to regulate various physiological functions. H2S is also a byproduct of dietary sulfate metabolism by gut bacteria. Inflammatory bowel diseases such as ulcerative colitis are associated with an increase in the colonization of the intestine by sulfate reducing bacteria along with an increase in H2S production. Consistent with its increased production, H2S is implicated as a mediator of ulcerative colitis both in its genesis or maintenance. As T cells are well established mediators of inflammatory bowel disease, we investigated the effect of H2S exposure on T cell activation. Using primary mouse T lymphocytes (CD3+), OT-II CD4+ T cells, and the human Jurkat T cell line, we show that physiological levels of H2S potentiate TCR-induced activation. Nanomolar levels of H2S (50–500 nm) enhance T cell activation assessed by CD69 expression, interleukin-2 expression, and CD25 levels. Exposure of T cells to H2S dose-dependently enhances TCR-stimulated proliferation with a maximum at 300 nm (30% increase, p < 0.01). Furthermore, activation increases the capacity of T cells to make H2S via increased expression of cystathionine γ-lyase and cystathionine β-synthase. Disrupting this response by silencing these H2S producing enzymes impairs T cell activation, and proliferation and can be rescued by the addition of 300 nm H2S. Thus, H2S represents a novel autocrine immunomodulatory molecule in T cells.


H 2 S is an endogenous signaling molecule that may act via protein sulfhydrylation to regulate various physiological functions. H 2 S is also a byproduct of dietary sulfate metabolism by gut bacteria. Inflammatory bowel diseases such as ulcerative colitis are associated with an increase in the colonization of the intestine by sulfate reducing bacteria along with an increase in H 2 S production. Consistent with its increased production, H 2 S is implicated as a mediator of ulcerative colitis both in its genesis or maintenance. As T cells are well established mediators of inflammatory bowel disease, we investigated the effect of H 2 S exposure on T cell activation. Using primary mouse T lymphocytes (CD3؉), OT-II CD4؉ T cells, and the human Jurkat T cell line, we show that physiological levels of H 2 S potentiate TCRinduced activation. Nanomolar levels of H 2 S (50 -500 nM) enhance T cell activation assessed by CD69 expression, interleukin-2 expression, and CD25 levels. Exposure of T cells to H 2 S
dose-dependently enhances TCR-stimulated proliferation with a maximum at 300 nM (30% increase, p < 0.01). Furthermore, activation increases the capacity of T cells to make H 2 S via increased expression of cystathionine ␥-lyase and cystathionine ␤-synthase. Disrupting this response by silencing these H 2 S producing enzymes impairs T cell activation, and proliferation and can be rescued by the addition of 300 nM H 2 S. Thus, H 2 S represents a novel autocrine immunomodulatory molecule in T cells.
H 2 S is emerging as an important member of the gasotransmitter family. At toxic environmental concentrations (Ͼ200 ppm), H 2 S inhibits mitochondrial cytochrome c oxidase (1). Lower nontoxic concentrations have potential physiological functions in neuromodulation (Ref. 2; for review, see Ref. 3), metabolic hibernation (4,5), protection from ischemia/reper-fusion injury (6 -10), oxygen sensing (11), vasodilatation (12,13), and promotion of angiogenesis (14). In common with nitric oxide, H 2 S is also implicated as both a pro- (15)(16)(17)(18) and antiinflammatory (18 -21) molecule in innate immune cells. Although the role of H 2 S signaling has been characterized in many other tissues and systems, it is unclear what role it plays in the regulation of the adaptive immune system. H 2 S is produced from dietary sulfate metabolism in the lumen of the large intestine by anaerobic sulfate-reducing bacteria (22). Clinical studies indicate that inflammatory bowel diseases such as ulcerative colitis are associated with an increase in the colonization of the intestine by sulfate-reducing bacteria along with an increase in the levels of H 2 S produced (23)(24)(25)(26)(27).
T cells also make an attractive target to study H 2 S signaling due to the lack of trans-sulfuration enzymes cystathionine ␥-lyase (CSE) 2 and cystathionine ␤-synthase (CBS) in naive cells (28,29) and the extensive literature characterizing the sensitivity of T cells to extracellular thiol redox status (30 -32). In contrast to naïve cells, activated T cells may be capable of making H 2 S (33), although the proliferation of a cytotoxic CD8 ϩ subset was inhibited by what we now consider supraphysiological doses of H 2 S (34). Therefore, we examine here whether T cell activation is affected by physiological concentrations of exogenous H 2 S and whether endogenous H 2 S production is required for T cell activation.

MATERIALS AND METHODS
Cells and Reagents-H 2 S refers to any of its various protonation states (H 2 S 3 HS Ϫ ϩ H ϩ 3 S 2Ϫ ϩ H ϩ ) with HS Ϫ being the predominant form at physiological pH (pK a ϭ 6.8). Na 2 S and NaHS are the corresponding sodium salts of these anionic forms of H 2 S and are thus considered H 2 S donors at physiological pH and are used as sources of H 2 S for this study. GYY4137 is also used as a slow-releasing H 2 S donor as characterized in Li * This work was supported, in whole or in part, by the National Institutes of et al. (35) and Lee et al. (36). C57Bl/6 and OT-II mice were anesthetized and sacrificed by cervical dislocation, and their spleens were harvested for T cell culture. OT-II mice and OVA  were kind gifts of the laboratory of Dr. Richard Morgan (NCI, NIH). The spleens were gently ruptured in a 40-m cell strainer (BD Biosciences) placed over a 50-ml Falcon tube using the back end of a 6-cc syringe plunger. The cells were rinsed-through with basal RPMI (0.1% BSA) and centrifuged at 200 ϫ g for 5 min. CD3ϩ T cells were purified by using a pan T cell isolation kit II and MACS MS columns (Miltenyi Biotec) without prior red-cell lysis according to the manufacturer's protocol. The cells were resuspended in 10 ml of RPMI containing 10% FBS, glutamine, and penicillin/streptomycin and plated in a flask for 30 min at 37°C and 5% CO 2 . Care and handling of animals was in accordance with the Animal Care and Use Committees of the National Cancer Institute (Protocol LP-012).
Unless specified otherwise, all chemicals were purchased from Sigma. GYY4137 was purchased from Cayman Chemicals (Ann Arbor, MI). For T cell activation, wells of either 6-or 96-well plates were coated overnight with a mixture of anti-CD3 and anti-CD28 antibodies (mouse cells, clones 17A2 and 37.51, respectively, BD Biosciences; Jurkat cells, OKT3 and CD28.2, functional grade, eBioscience) at 2 and 5 g/ml, respectively, in PBS without cations. The following day wells were washed twice with PBS to remove unbound antibody, and cells were added in growth medium for stimulation.
The concentrations of H 2 S used for this study are derived from the EC 50 concentrations of H 2 S (from NaHS or Na 2 S) used for stimulation of T cell proliferation (see Fig. 3). Also, the steady state concentration of H 2 S will depend on how much is added exogenously or made endogenously and the rate of its degradation. The degradation of H 2 S both enzymatically and nonenzymatically depends largely on O 2 concentration (38). Minimizing the O 2 levels maximizes available H 2 S. Due to this, we conducted the experiments at 1% O 2 and 94% N 2 using a hypoxic chamber (MIC-101, Billups-Rothenberg, Inc., Del Mar, CA) unless stated otherwise.
Flow Cytometry-T cells were washed with PBS supplemented with 0.1% BSA and 0.01% azide (FACS buffer) and stained using anti-CD69-FITC (BD Biosciences) or anti-CD25 (BD Biosciences) and analyzed with a FACS Caliber and Cell-Quest software (BD Biosciences).
Cell Adhesion Assay-To monitor cell adhesion we used an ACEA real time label-free cell monitoring system (Xcelligence, Roche Applied Science) (39). 50 l of basal media was allowed to incubate for 15 min in each well of an 8-well RT-CES plate at 37°C, 5% CO 2 . An additional 50 l containing 25,000 Jurkat cells and the specified treatments were added to the wells, and the cell index was measured at 30-s intervals. All data are nor-malized to 15 min after cell addition, when all cells have settled to the bottom of the well.
Fluorescence Microscopy-Antigen-presenting cells often vary in the distribution of major histocompatibility complexes, making the use of an antibody-coated surface advantageous in these experiments as it provides a simpler and more homogenous stimulus necessary for quantifying cytoskeletal changes (37). The glass surface also allows for a single orientation of activated cells and a high degree of cell polarization. The microtubule organizing complex (MTOC) of Jurkat cells that are maximally activated by the antibody-coated surface should be at the center of the cell contact point (x,y plane) and directly above the contact surface (x,z plane) (40). Likewise, previous studies have indicated that TCR-specific antibody-coated microspheres can serve as artificial antigen-presenting cells by providing a focal stimulus capable of inducing MTOC reorientation to the immunological synapse (41). Both are used to assess cytoskeletal changes in this study.
8-Chambered Lab-tek coverslips (Nunc) were coated with either 0.01% w/v poly-L-lysine solution (Sigma) or anti-CD3/ anti-CD28 antibodies (1 and 5 g/ml) in PBS overnight at 4°C and then washed with PBS. EYFP-tubulin-or enhanced GFPactin-expressing Jurkat cells (50,000 cells) in basal RPMI medium were added to the wells and allowed to incubate at 37°C, 5% CO 2 with specified treatments before imaging on the confocal microscope. Confocal images were acquired on a Zeiss LSM 510 NLO confocal system (Carl Zeiss, Thornwood, NJ) with a 63ϫ Plan-Apochromat 1.4 NA oil immersion objective. Z-stacked images were acquired at 512 ϫ 512 pixels per frame using an 8-bit pixel depth for each channel and line averaging set to 4 collected sequentially in a multi-track, 3 channel mode.
Gene Expression Studies-In 6-well plates, 1-3 ϫ 10 6 cells for each condition were maintained in 1 or 20% O 2 for the specified amount of time. The cells were harvested, and mRNA was extracted using TRIzol (Invitrogen) according to the manufacturer's protocol. cDNA was synthesized from 1-5 g of mRNA using superscript first-strand RT-PCR reagents (Invitrogen) according to the manufacturer's protocol. Quantitative realtime-PCR was then performed using the SYBR Green kit (Thermo) on the following gene/primer sets: hypoxanthine phosphoribosyltransferase, CD69, IL-2, CSE, CBS, SQR (sequences Table 1). Hypoxanthine phosphoribosyltransferase was used as the internal control for expression based on previous reports of its superior stability over other commonly used control genes (42). Results were calculated based on the ⌬Ct method and normalized to hypoxanthine phosphoribosyltransferase. T Cell Proliferation-Cells were seeded at 100,000 cells per well in 96-well plates uncoated or precoated with anti-CD3/ CD28 as specified above and treated as indicated. Proliferation was assessed using cell titer-96 MTS reagent (Promega) after 72 h of growth according to the manufacturer's protocol. The MTS signal for the cells on day 0 was subtracted from the signal after 72 h to quantify net proliferation. Proliferation was also assessed by CFSE dilution (Invitrogen) in murine CD3ϩ cells using flow cytometry. CFSE-based proliferation was analyzed using Flow Jo software. Cells were maintained in 1 or 20% O 2 as indicated.
IL-2 Transcription Factor Luciferase Assays-The reporter plasmid for the IL-2 promoter elements CD28RE-Luc, NFAT-Luc, STAT5-Luc, CRE-Luc, NFB-Luc, and AP1-Luc were kind gifts of Dr. Kevin Gardner (43) constructed in PGL3-basic luciferase plasmids (Promega). Jurkat T cells (2 ϫ 10 6 ) were cotransfected with 5 g of each reporter plasmid and 1 g of the Renilla luciferase construct using an Amaxa nucleofector system (program X-001). After transfection, the cells were cultured in RPMI containing 10% FBS, glutamine, and penicillin/ streptomycin overnight followed by stimulation in antibodycoated 96-well plates for the indicated amount of time (determined by time-course for maximum signal). Cells were lysed by adding 20 l of passive lysis buffer to 100 l of cells for 5 min after triturating. The dual luciferase reporter assay system (Promega) was used for the quantification of reporter gene expression. Luciferase activities were determined using a Turner Biosystems 20/20N luminometer. Transcription factordriven firefly luciferase activity was normalized by dividing firefly luciferase units by the units corresponding to Renilla luciferase activity.
Intracellular cGMP Assay-Jurkat cells were assayed at 5 ϫ 10 5 cells per condition in 0.5 ml of basal media (RPMI, glutamine, penicillin/streptomycin, 0.01% BSA). The cells were incubated for 4 h with the indicated concentrations of NaHS and diethylenetriamine/NONOate at 37°C. The cells were then lysed, and total intracellular cGMP levels were measured via an immunoassay using a cGMP kit (GE/Amersham Biosciences) according to the manufacturer's instructions. cGMP levels were normalized to control non-diethylenetriamine/NONOate-treated samples.
siRNA Knockdown Experiments-siRNAs were synthesized using the Silencer siRNA Construction kit (AP Biosystems). Three different siRNAs were made to target each gene (CBS and CSE) based on primers selected from the Ambion siRNA target finder web application. Each siRNA transfection represented in the text is an equal mixture of each of the following three individual siRNAs targeting unique exons for maximum potency (sequences in Table 2).
Mouse T cells were harvested as above, and CD3 ϩ T cells were purified by using a pan T cell isolation kit II and MACS MS columns (Miltenyi Biotec) without prior red-cell lysis according to the manufacturer's protocol. Purified T cells were then transfected with 30 pmol of siRNA cocktails using a mouse T cell nucleofection kit and an Amaxa nucleofector (program X-01) according to the manufacturer's protocol. Jurkat cells were maintained in antibiotic-free RPMI growth medium and transfected using nucleofection kit V and program X-01 according to the manufacturer's protocol. After transfection, the cells were cultured in RPMI containing 10% FBS, glutamine, and penicillin/streptomycin 16 h followed by stimulation in antibody-coated 96-well plates for 72 h with specified analyses.

RESULTS
H 2 S Enhances TCR-stimulated T Cell Activation-We first assessed the effect of physiological (nM) H 2 S concentrations on T cell activation. CD69 expression is one of the earliest events following TCR-mediated T cell activation (44). Using quantitative real-time-PCR to assess CD69 mRNA expression, anti-CD3/CD28 stimulation of freshly isolated mouse T cells led to a time-dependent increase in CD69 expression (Fig. 1A). CD69 expression was significantly (p Ͻ 0.05) increased at 4 and 10 h post-stimulation in the presence of 300 nM H 2 S. Mouse T cell CD69 expression was also dose-dependently increased at 6 h post-stimulation in the presence of H 2 S (50, 500 nM) by 30 and 100%, respectively (Fig. 1B). Other markers of T cell activation were also enhanced in the presence of 300 nM H 2 S such as IL-2 and CD25 mRNA expression (Fig. 1, C and D). Human Jurkat T-lymphoma cells serve as a continuous cell line that is widely used to study T cell activation. CD69 and IL-2 induction were similarly enhanced in Jurkat cells activated by anti-CD3 and CD28 antibodies (Fig. 1, E and F). Consistent with increases in CD69 mRNA, CD69 protein expression was enhanced by polyclonal activation (Fig. 1G) and was dose-dependently enhanced by H 2 S in activated Jurkat cells (Fig. 1H). CD69 expression was enhanced by 27% in the presence of 100 nM H 2 S and 33% in the presence of 50 M of the H 2 S donor molecule GYY4137 corresponding to a steady state H 2 S concentration of 200 nM (Fig. 1I). In addition to polyclonal stimulation of T cell activation, we tested the ability of H 2 S to enhance antigen-specific T cell activation. CD4ϩ T cells freshly isolated from OT-II mice were co-stimulated with 1 g/ml OVA-2 peptide and 5 g/ml anti-CD28 antibody with and without H 2 S treatment (300 nM). H 2 S significantly enhanced OVA-2 peptide stimulation of IL-2 expression at 6 and 24 h (Fig. 2A). In addition, 100 nM H 2 S enhanced OVA-2 stimulation of CD25 protein levels by 30% (Fig. 2B).
H 2 S Enhances Activation-dependent Cytoskeletal Dynamics in T Cells-T cell activation induces cytoskeletal rearrangements necessary for full activation and polarization toward an antigen-presenting cell (45,46). Activation-dependent spreading and attachment of T cells depends on cytoskeletal dynamics in the MTOC and actin cytoskeleton. The key cytoskeletal elements actin and ␤-tubulin were identified as direct sulfhydrylation targets of endogenous H 2 S produced in mouse liver (47). H 2 S sulfhydrylation of actin in vitro resulted in enhanced polymerization. However, this was only observed at supraphysiological (100 -200 M) levels of H 2 S. We addressed the function of these targets at physiological H 2 S concentrations. Jurkat T cells were plated on a surface coated with anti-CD3/CD28, and adhesion was quantified by measuring impedance at this interface expressed as a normalized cell index. As expected, cells attached more firmly on anti-CD3/CD28-coated surfaces as compared with poly-L-lysine-coated surfaces (Fig. 3A). This signal was dose-dependently enhanced in the presence of 10 nM to 1 M Na 2 S (Fig. 3B). Thus, physiological doses of H 2 S enhance T cell interaction with a TCR activating surface.
MTOC translocation was assessed in Jurkat cells stably transfected with EYFP-␣-tubulin and stimulated using anti-CD3/CD28-coated microspheres to induce an immunological synapse. Previous studies have indicated that TCR-specific antibody-coated microspheres can serve as artificial antigen-presenting cells by providing a focal stimulus capable of inducing MTOC reorientation to the immunological synapse (41). Microsphere activation alone resulted in MTOC translocation directly adjacent to the microsphere in 34 Ϯ 2% of the cells (Fig. 4A). Co-treatment with 300 nM Na 2 S significantly enhanced this translocation (53 Ϯ 5 versus 34 Ϯ 2, p Ͻ 0.05, Fig. 4, A and B).
The enhancement of MTOC translocation by H 2 S was further analyzed by imaging the interaction of EYFP-␣-tubulinexpressing Jurkat cells with control or anti-CD3/CD28 coated glass surfaces. Antigen-presenting cells often vary in the distribution of major histocompatibility complexes, making use of an antibody-coated surface advantageous in these experiments, as it provides a simpler and more homogenous stimulus necessary for quantifying cytoskeletal changes (37). The glass surface also allows for a single orientation of activated cells and a high degree of cell polarization. The MTOC of Jurkat cells that are maximally activated by the antibody-coated surface should be at the center of the cell contact point (x,y plane) and directly above the contact surface (x,z plane) (40). Although MTOCs had mostly random orientation on the control surfaces, greater than 60% were centrally oriented in the x,y plane on the activating surface (Fig. 4C). Na 2 S treatment significantly increased this positioning to 90% (Fig. 4D). Interestingly, in the three-dimensional rendering of the x,z plane, the MTOC was properly located at the activating surface for both Na 2 S, treated and untreated, but the tubulin cytoskeleton of the Na 2 S-   FEBRUARY 3, 2012 • VOLUME 287 • NUMBER 6 JOURNAL OF BIOLOGICAL CHEMISTRY 4215 treated cells was much more uniformly elongated than untreated cells (Fig. 4E). Thus, exogenous H 2 S enhances MTOC dynamics in activated T cells.

H 2 S Potentiates T Cell Activation
Likewise, the response of the actin cytoskeleton to anti-CD3/ CD28 activation was notably altered by treatment with 300 nM Na 2 S (Fig. 4F). Control and Na 2 S-treated cells form a distinct actin collar within the first 5 min after plating on anti-CD3/ CD28 surfaces. This actin collar significantly increased in intensity until 15 min (Fig. 4G). Subsequently, breakdown in the actin collar occurs in the Na 2 S-treated cells after 15 min, whereas the control cells maintain this collar. These data are consistent with the Jurkat cell adhesion data in Fig. 3B showing greater dynamics in the presence of H 2 S. Dynamic regulation of the actin cytoskeleton at the immunological synapse enhances downstream signaling from the TCR (48), which can be disrupted by actin polymerization targeting drugs such as cytochalasin D (49,50).
We also investigated several previously defined targets of H 2 S in other systems, including K ϩ ATP channels, cGMP-dependent phosphodiesterases (51), and Akt (52) and found that these are not responsible for the effects of H 2 S on T cell activation (supplemental Fig. 1).
H 2 S Biosynthetic Capacity Is Increased with T Cell Activation-As a diffusible gasotransmitter, we considered whether H 2 S could act as an autocrine and paracrine mediator of TCRstimulated proliferation in a manner similar to IL-2. Expression of the two major H 2 S biosynthetic genes, CBS and CSE was notably increased over that in resting T cells after activation, 150 and 250%, respectively (Fig. 5, A and B). Steady state levels of H 2 S also depend on its rate of degradation via SQR, an O 2 -dependent mitochondrial enzyme (53,54). SQR was slightly decreased in activated versus resting T cells (Fig. 5C), implying that higher levels of H 2 S are present in activated T cells. Consistent with this concept, uptake of the CBS and CSE substrate cysteine is dramatically increased upon T cell activation (55). Thus, the capacity of activated T cells to make H 2 S is enhanced, implying that H 2 S acts as an autocrine regulator of T cell activation.
Endogenous H 2 S Regulates T Cell Activation-Unlike naïve T cells, Jurkat cells express CBS (28,29). To address the role of endogenous H 2 S produced by CBS on cell adhesion, Jurkat cells were transfected with CBS siRNA. The decreased expression of CBS mRNA (Fig. 6A) resulted in decreased adhesion, which could be partially rescued by the addition of exogenous H 2 S (300 nM Na 2 S, Fig. 6B). Gene expression levels were reduced by greater than 50% compared with scrambled control RNA in Jurkat cells treated with CBS-specific siRNA. siRNA knockdown of endogenous H 2 S-producing capacity via CBS significantly impaired both CD69 and IL-2 expression in TCR-activated Jurkat cells (Fig. 6, C and D). Because CD69 and IL-2 protein levels correlate well with mRNA levels (56), our data indicate that physiological levels of exogenous and endogenous H 2 S control critical aspects of T cell activation and signaling.
To further address the role of endogenously produced H 2 S in activated T cells, we assessed the activation-dependent proliferation of purified mouse CD3ϩ T cells in the presence of H 2 S. TCR stimulation of T cells induces proliferation and supports their in vitro polyclonal expansion in part via autocrine and paracrine IL-2 signaling (57,58). Proliferation of TCR-stimulated mouse spleen-derived T cells was enhanced in a biphasic manner by 30 -300 nM NaHS (Fig. 6E) as assessed using an MTS proliferation assay. These results are supported by a CFSE staining experiment showing a similar increase in proliferation index of CD8 cells treated with 300 nM H 2 S (supplemental Fig. 2). In a 20% O 2 atmosphere, maximal enhancement of proliferation was 30% at 300 nM NaHS. We compared the effects of H 2 S on proliferation at 1% O 2 because enzymatic and nonenzymatic degradation of H 2 S is O 2 -dependent (38). We then assessed proliferation due to endogenous H 2 S via siRNA knockdown of CBS and CSE. Targeting either CBS or CSE with siRNAs decreased the proliferation of TCR-activated T cells by about 20% (Fig. 6F). However, targeting both enzymes inhibited TCR-driven proliferation as compared with scrambled siRNA control by nearly 40%. In all cases proliferation was rescued by adding back 300 nM NaHS.

DISCUSSION
These data demonstrate physiological functions of endogenous and exogenous H 2 S in T cell activation. TCR-dependent T cell activation and IL-2 expression are enhanced by H 2 S, and activation in turn increases expression of H 2 S biosynthetic enzymes without altering expression of the major H 2 S degradative enzyme SQR (Fig. 7). This function of H 2 S was validated for general and antigen-specific T cell activation of murine T cells and for the human Jurkat T cell line. The effects of H 2 S are mediated at least in part by enhanced microtubule and actin cytoskeletal dynamics that, along with work by the Snyder laboratory, suggest that the cytoskeleton is a direct and sensitive physiological target of H 2 S (47).
Despite having several clearly defined signaling roles, the levels of endogenously produced H 2 S that mediate these functions are in dispute, and the most recent measurement in tissue homogenates is 15 nM (59). This lower estimate calls into question several proposed physiological targets of H 2 S that require micromolar H 2 S levels to modulate their activities (60 -62). Mustafa et al. (47) identified actin and ␤-tubulin, key elements of the cytoskeleton, as direct targets of endogenous H 2 S produced in murine liver. H 2 S sulfhydrylation of actin resulted in enhanced polymerization and a visible change in the actin cytoskeleton; however, this was only observed in vitro at supraphysiological levels of exogenously added H 2 S (100 -200 M).  We sought to examine the effect of H 2 S on the cytoskeleton at physiological levels in T cells. During antigen-dependent T cell activation, cytoskeletal rearrangement is essential for full activation and polarization of the T cell toward the antigenpresenting target cell (45,46). The actin cytoskeleton polymerizes to provide a physical scaffold for the assembly of T cell receptor proteins as well as for the associated supramolecular activation complex. The tubulin cytoskeleton reorganizes to radiate vectorially outward from the center of the supramolecular activation complex to its MTOC. Compounds that disrupt microtubule or actin polymerization, such as taxol, colchicine, or cytochalasin D, inhibit antigen-dependent and IL-2-driven T cell proliferation (49,50). T cells also make an attractive target due to the lack of transsulfuration enzymes (CSE and CBS) in naive cells (28,29) and the extensive literature characterizing the sensitivity of T cells to extracellular thiol redox status (30 -32).
Although our studies present the cytoskeleton as one possible target of H 2 S-dependent T cell activation, the proliferation data are biphasic, suggesting that additional concentration-dependent H 2 S targets exist. Also, the maximal effect was greater at low O 2 , consistent with an O 2 -mediated degradative mechanism. This confirms numerous other reports of a complex relationship between O 2 and H 2 S in the cell. Because we did not observe a simple leftward shift in H 2 S effect with a decrease in O 2 , we suggest that there may not be a simple single ligand receptor relationship for H 2 S signaling in these cells.
In contrast to naïve cells, activated T cells are capable of making H 2 S. Peripheral blood lymphocytes were previously reported to produce H 2 S, although this biosynthetic activity was not specifically attributable to T cells (33). Furthermore, H 2 S inhibited proliferation of cytotoxic CD8ϩ T cells; however, this occurred at what we now consider to be a supraphysiological concentration of H 2 S (34). In contrast, our data establish that proliferating T cells depend on their endogenous capacity to generate H 2 S, and this capacity is enhanced by T cell activation. In addition to this autocrine role, H 2 S may have a paracrine function in T cell activation. This idea is supported by the report of increased H 2 S output from activated antigen presenting cells (63). Peripheral T cell activation and proliferation must be tightly controlled to prevent autoimmune destruction of cells and tissue. To achieve this, resting T cells are limited in their ability to take up basic amino acids such as cysteine (or cystine). T cells lack efficient transport of cystine and rely on a steady production by neighboring antigen-presenting cells to support their proliferation. Although not completely understood, the cysteine requirement of activated T cells is reportedly for making GSH and also participates in changing the extracellular thiol redox status to a more reducing environment, thus altering the activity of exofacial membrane-bound signaling proteins. We postulate that the additional cysteine demand could also satisfy the up-regulated H 2 S synthesis necessary for full T cell proliferation. More recently, Garg et al. (64) reported that CBS protein levels are increased in T cells from BALBc mice activated by dendritic cells and an anti-CD3 antibody. Their results also indicate that the transsulfuration pathway is intact in naïve T cells as a source of cysteine to make glutathione, as cysteine and cystine uptake is limited. Con-sumption of cysteine for biosynthesis from the transsulfuration pathway would preclude the generation of H 2 S from cysteine. We hypothesize that during T cell activation, when cysteine is no longer limited, that the transsulfuration enzymes CBS and CSE supply H 2 S to support T cell activation and function while consuming cysteine.
This work could have important implications to the pathogenesis of inflammatory bowel disease. Inflammatory bowel disease is a chronic inflammatory disease caused by the generation and persistence of colitogenic CD4ϩ effector and memory T cells that react to antigens of commensal bacteria. H 2 S is continually produced by luminal sulfate-reducing commensal bacteria in the colon and is normally detoxified by rhodanese in the surrounding mucosal cells to thiosulfate (65,66). An increase in the steady state H 2 S levels either from overproduction or reduced consumption is thought to play a role in the etiology of inflammatory bowel disease and related cancers (67). Our data provide a logical link between these reports in that the excess H 2 S could contribute to unwanted T cell activation toward commensal H 2 S-producing bacteria. Combined with the novel physiological signaling function for H 2 S as a co-stimulator of T cell activation and proliferation, these findings establish H 2 S as an endogenous and exogenous co-regulatory signal for T cells.