ORIGINAL ARTICLE Sphingosine-1-Phosphate Reduces CD4؉ T-Cell Activation in Type 1 Diabetes Through Regulation of Hypoxia-Inducible Factor Short Isoform I.1 and CD69 Suseela Srinivasan,1 David T. Bolick,1 Dmitriy Lukashev,2 Courtney Lappas,3 Michail Sitkovsky,2 Kevin R. Lynch,3 and Catherine C. Hedrick1,3

OBJECTIVES—Non-obese diabetic (NOD) mice develop spon- taneous type 1 diabetes. We have shown that sphingosine-1- phosphate (S1P) reduces activation of NOD diabetic endothelium phingosine-1-phosphate (S1P) is a bioactive lipid via the S1P1 receptor. In the current study, we tested the hypoth- that functions as an extracellular mediator and as esis that S1P could inhibit CD4ϩ T-cell activation, further reduc- an intracellular second messenger. S1P is synthe- ing inflammatory events associated with diabetes. Ssized by a wide variety of cell types, including ϩ lymphocytes, platelets, and macrophages in response to RESEARCH DESIGN AND METHODS—CD4 T-cells were isolated from diabetic and nondiabetic NOD mouse splenocytes growth factors and cytokines (1). S1P evokes diverse and treated in the absence or presence of S1P or the S1P1 cellular responses by binding to a group of five G-protein– receptor-specific agonist, SEW2871. Lymphocyte activation was coupled receptors of the endothelial differentiation gene examined using flow cytometry, cytokine bead assays, and a (Edg) family. S1P receptor expression varies among vas- lymphocyte:endothelial adhesion assay. cular cell types, with T-cells expressing only S1P1 and S1P4 (2). Recently, we reported an anti-inflammatory role RESULTS—Diabetic T-cells secreted twofold more ␥-interferon for S1P in aortic endothelial cells, most likely through (IFN-␥) and interleukin-17 than nondiabetic lymphocytes. Pre- S1P1 action (3). Studies have recently shown that S1P treatment with either S1P or SEW2871 significantly reduced receptor agonists, such as FTY720, reduce renal (4) and ϳ cytokine secretion by 50%. Flow cytometry analysis showed hepatic ischemic injury (5). FTY720 is currently being used increased expression of CD69, a marker of lymphocyte activa- in phase III clinical trials for multiple sclerosis (6) and has tion, on diabetic T-cells. Both S1P and SEW2871 prevented ϩ been shown to prevent onset of autoimmune diabetes in upregulation of CD69 on CD4 cells. Quantitative RT-PCR showed that lymphocytes from diabetic NOD mice had 2.5-fold mouse models in vivo (7). Nofer et al. (8) have recently lower hypoxia-inducible factor (HIF)-1␣ short isoform I.1 reported that FTY720 reduces atherosclerosis in LDL (HIF1␣I.1) mRNA levels than control. HIF1␣I.1 is a negative receptor–deficient mice. Other studies have shown the regulator of lymphocyte activation. S1P significantly increased importance of S1P and the S1P1 receptor in modulating HIF1␣ I.1 mRNA levels in both control and diabetic groups. IFN-␥ T-cell proliferation and homing to lymph nodes (9,10). production and surface CD69 expression was significantly in- Furthermore, S1P has recently been shown to regulate creased in lymphocytes of HIF1␣I.1-deficient mice. S1P did not lymphocyte egress from peripheral lymph nodes through reduce either CD69 or IFN-␥ expression in lymphocytes from interaction with the adaptor protein DOCK2 (11). FTY720 HIF1␣I.1-deficient mice. and SEW2871, two well-characterized S1P receptor ago- nists that act on the S1P1 receptor, cause dramatic lym- CONCLUSIONS—S1P acts through the S1P1 receptor and ␣ phopenia in mice when used in vivo. Goetzl and colleagues HIF1 I.1 to negatively regulate T-cell activation, providing a (12,13) have shown that S1P regulates CD4ϩ T-cell che- potential therapeutic target for prevention of diabetes and its vascular complications. Diabetes 57:484–493, 2008 motaxis and proliferation through action on S1P1. Thus, S1P analogs have the potential to be important drug therapies for inflammation and immune responses in a wide range of diseases. Hypoxia-inducible factor (HIF)-1␣ is a basic helix-loop- From the 1Robert M. Berne Cardiovascular Research Center, University of helix transcription factor that is induced under hypoxic Virginia, Charlottesville, Virginia; the 2New England Inflammation and Tissue conditions. HIF1␣ can also be induced during normoxic Protection Institute, Northeastern University, Boston, Massachusetts; and the 3Department of Pharmacology, University of Virginia, Charlottesville, Virginia. conditions by vascular endothelial growth factor and other Address correspondence and reprint requests to Catherine C. Hedrick, PhD, growth factors, various lipids, thrombin, and angiotensin II Cardiovascular Research Center, University of Virginia, P.O. Box 801394, 415 (14). The murine HIF1␣ gene contains two different first Lane Rd., MR5, Rm. G123, Charlottesville, VA 22908. E-mail: cch6n@virginia. ␣ edu. exons, termed I.1 and I.2 (15). Expression of HIF1 I.1 and Received for publication 23 June 2007 and accepted in revised form 31 I.2 is regulated via two distinct promoters, which give rise October 2007. to these two mRNA isoforms (15). HIF␣I.1 is inducible and Published ahead of print at http://diabetes.diabetesjournals.org on 14 No- vember 2007. DOI: 10.2337/db07-0855. differentially expressed, whereas the longer, classical iso- ELISA, enzyme-linked immunosorbent assay; FACS, fluorescence-activated form, HIF1␣I.2, is constitutively expressed. The isoform cell sorting; FITC, fluorescein isothiocyanate; HIF, hypoxia-inducible factor; HIF1␣ I.1 is an immediately early response gene in T-cells HIFBS, heat-inactivated fetal bovine serum; IFN-␥, ␥-interferon; IL, interleu- that is differentially upregulated upon T-cell receptor kin; mAb, monoclonal antibody; S1P, sphingosine-1-phosphate; TCR, T-cell receptor; TNF-␣, tumor necrosis factor-␣; UVA, University of Virginia. (TCR) activation of T-cells (16). Rapid accumulation of © 2008 by the American Diabetes Association. HIF1␣I.1 in activated T-cells has been observed in vivo The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance during cytokine-mediated inflammation (16). The first in- with 18 U.S.C. Section 1734 solely to indicate this fact. dications that HIF1␣ had both immunosuppressive and

484 DIABETES, VOL. 57, FEBRUARY 2008 S. SRINIVASAN AND ASSOCIATES

TABLE 1 Quantitative real-time PCR primers used in this study Name Sequence IFN-␥ Forward 5Ј-ACT CAA GTG GCA TAG ATG TGG AAG-3Ј Reverse 5Ј-GAC GCT TAT GTT GTT GCT GAT GG-3Ј IL-4 Forward 5Ј-ACA GGA GAA GGG ACG CCA TT-3Ј Reverse 5Ј-GAA GCC CTA CAG ACG AGC TCA-3Ј HIF1␣I.1 Forward 5Ј-CCC CGT CCA CCC ATT TCT-3Ј Reverse 5Ј-CTG GCT TCT TGC TTA CAG GAG AG-3Ј HIF1␣I.2 Forward 5Ј-CAC CGA TTC GCC ATG GA-3Ј Reverse 5Ј-TTT TCG ACG TTC AGA ACT CAT C-3Ј CD69 Forward 5Ј-TTC GAC GTT CAG AAC TCA TCT TTT-3Ј Reverse 5Ј-GCT GTT AAA TTC TTT GCC ATT TG-3Ј Cyclophilin Forward 5Ј-TGG AGA GCA CCA AGA CAG ACA-3Ј Reverse 5Ј-TGC CGG AGT CGA CAA TGA T-3Ј IL-17 Forward 5Ј-TCT CTG ATG CTG TTG CTG CT-3Ј Reverse 5Ј-AGG AAG TCC TTG GCC TCA GT-3Ј tissue-protective roles were provided by in vivo studies In vivo studies of SEW2871 action. Diabetic and nondiabetic NOD mice showing dramatically increased autoimmunity and tissue were bled retro-orbitally to obtain a prebleed sample. Mice were then injected damage in chimeric mice that possessed genetic deficiency intravenously with 2 mg/kg body wt SEW2871. After 4 h, blood was collected ␣ for measurement of plasma cytokine levels. Cytokines (IL-4, IL-10, and of HIF1 in their T- and B-cells (17). Direct evidence that ␥-interferon [IFN-␥]) were measured using ELISA. both isoforms of HIF1␣ are immunosuppressive in acti- CD4؉ T-cell isolation and activation. Spleens were harvested from female vated T-cells was provided by studying T-cells from mice diabetic and age-matched nondiabetic NOD mice, and the tissue was gently with complete genetic knockout of HIF1␣ isoforms (18). disrupted in PBS containing 5% heat-inactivated fetal bovine serum (HIFBS) Sitkovsky and colleagues (18) have shown that the on ice. Splenocytes were passed through a 40-␮m nylon cell strainer (BD ␣ Biosciences) and collected in PBS. Erythrocytes were removed using lysing HIF1 I.1 isoform is anti-inflammatory in T-cells, thereby ϩ preventing T-cell activation in vivo. Interactions between buffer (Sigma-Aldrich). CD4 T-cells were isolated with mouse CD4 subset column kit following the manufacturer’s protocol. The purity of CD4ϩ T-cells the immunosuppressive A2A adenosine receptor and Ͼ ␣ after isolation was 95% as measured by flow cytometry using anti-CD4 HIF1 in T-cells was first suggested by Sitkovsky and antibody. Cells were washed and resuspended in RPMI 1640 containing 10% colleagues (19,20). These studies led us to hypothesize HIFBS and 1% antibiotic-antimycotic solution. T-cells were activated in that the immunosuppressive effects of S1P could be re- 96-well plates coated with 2–10 ␮g/ml immobilized anti-CD3 mAb in the lated to HIF1␣ function in lymphocytes. presence and absence of 500 nmol/l S1P, 1 ␮mol/l S1P, or 1 ␮mol/l SEW2871 at 37°C in 5% CO for 16 h. In some studies, cells were incubated in the Despite several reports showing a protective role for 2 presence of 100 ng/ml pertussis toxin or 10 ␮mol/l VPC23019 for 16 h to inhibit S1P and S1P receptor agonists in inflammatory and im- S1P1 receptor signaling. mune diseases, the precise mechanisms for S1P action in Cytokine measurements. Splenic CD4ϩ T-cells were isolated and activated inhibiting inflammation and tissue damage are unclear. In as described above. IFN-␥, IL-4, IL-17, and IL-10 levels secreted into the media the current study, we report for the first time that S1P were measured by ELISA of cell supernatants after 16 h of treatment with 500 induces HIF1␣I.1 expression and activity to modulate nmol/l or 1 ␮mol/l S1P, 1 ␮mol/l SEW2871, or in some cases 100 ng/ml T-cell activation in NOD diabetic mice. We now show that pertussis toxin or 10 ␮mol/l VPC23019. ␣ T-cell:endothelial adhesion assay. Aortic endothelial cells were isolated regulation of HIF1 I.1 expression in lymphocytes is a and cultured as described previously (21) from nondiabetic NOD mice. primary mechanism by which the S1P-S1P1 axis regulates Activated CD4ϩ T-cells were labeled with Calcein-AM (Molecular Probes) lymphocyte activation. Thus, the action of pharmacologi- according to the manufacturer’s instructions. Fluorescently labeled CD4ϩ cal agonists of the S1P1 receptor should be beneficial to T-cells (50,000 per well) were added to a monolayer of endothelial cells in a reduce inflammatory events associated with several vas- 48-well tissue culture dish. After 45 min of incubation at 37°C, unbound T-cells cular diseases, including atherosclerosis and diabetes. were rinsed off, cells were fixed in 1% glutaraldehyde, and bound T-cells were counted within a 10 ϫ 10 grid using epifluorescence microscopy. Flow cytometry studies. T-cells (0.5 ϫ 106)in100␮l PBS with 1% HIFBS RESEARCH DESIGN AND METHODS (fluorescence-activated cell sorting [FACS] buffer) were labeled with 0.25 ␮g Female NOD/LtJ mice (stock no. 001976) were purchased from the Jackson each of FITC-labeled anti-mouse CD69 and PE-labeled anti-mouse CD4 Laboratories and were maintained in a pathogen-free barrier facility. All antibodies or FITC- and PE-conjugated isotype control antibodies at 4°C for 30 animal studies were performed following the approved guidelines of the min. Stained cells were washed with 10-fold excess volume of FACS buffer University of Virginia (UVA) Animal Care and Use Committee. Mice were and resuspended in PBS containing 1% paraformaldehyde. The fluorescence monitored for hyperglycemia by weekly measurement of blood glucose levels intensity was measured by collecting a minimum of 10,000 events using a BD using a One Touch Ultra glucometer (LifeScan). Mice were considered Biosciences FACS-Calibur dual laser benchtop flow cytometer using diabetic when two consecutive blood glucose levels were Ͼ13.8 mmol/l. Mice CellQuest software (Becton Dickinson) and analyzed using FlowJo software lacking HIF1␣I.1 were generated as described previously (18). (Treestar, San Carlos, CA). Reagents. S1P was purchased from Cayman Chemicals. Fluorescein isothio- RNA isolation and RT-PCR. CD4ϩ T-cells were harvested, and activated cyanate (FITC)-labeled anti-mouse CD69 (clone H1.2F3), phycoerythrin-la- using anti-CD3 mAb as described above. Total RNA was isolated using an beled anti-mouse CD4 (clone GK 1.5), and isotype control antibodies rat RNeasy minikit (Qiagen). Quantitative mRNA analysis was performed by IgG2b-PE and hamster IgG-FITC were all purchased from eBiosciences. real-time PCR using SYBR Green PCR master mix (Bio-Rad) in a Bio-Rad Enzyme-linked immunosorbent assay (ELISA) kits for murine interleukin MyIQ icycler. Primer sequences used are listed in Table 1. Data were analyzed (IL)-17, tumor necrosis factor-␣ (TNF-␣), and IL-10 were purchased from and presented as relative expression of mRNA of interest normalized to ⌬ eBiosciences. Mouse CD4 subset columns were obtained from R&D Systems. cyclophylin using the Ct method (22). Anti-CD3 monoclonal antibody (mAb)–coated plates and Th1/Th2 cytokine HIF1␣ siRNA transfection. CD4ϩ T-cells were isolated from nondiabetic Cytometric Bead Arrays were purchased from BD Biosciences. HIF1␣ smart mouse spleens as described above. Four million cells were transfected with pool siRNA was purchased from Dharmacon. Pertussis toxin was provided by 400 pmol HIF1␣ siRNA or scrambled siRNA by nucleofection using a mouse Dr. Erik Hewlett (UVA). T-cell nucleofector kit (Amaxa Biosystems). After 2 h, transfected cells were

DIABETES, VOL. 57, FEBRUARY 2008 485 S1P AND HIF1␣ REGULATION IN LYMPHOCYTES

anti-CD3 antibody increased IFN-␥ release ϳ100-fold in both nondiabetic control and diabetic T-cells (Fig. 2, compare ST bars in B with A). After CD3 stimulation, there was a significant threefold increase in the amount of IFN-␥ produced by diabetic T-cells (see ST bar in Fig. 2B) compared with nondiabetic control lymphocytes. S1P at both 500 nmol/l and 1 ␮mol/l concentrations and the S1P1 receptor-specific agonist SEW2871 at 1 ␮mol/l had pro- found impacts in reducing IFN-␥ secretion by diabetic T-cells (Fig. 2B). S1P and SEW2871 also reduced IFN-␥ secretion by nondiabetic lymphocytes (Fig. 2B). We used S1P at 1 ␮mol/l for all remaining experiments for compar- ison with SEW2871 because we used SEW2871 at 1 ␮mol/l in vitro. The results using SEW2871 strongly suggest that the effects of S1P are through action on the S1P1 receptor. The S1P1 receptor couples solely through G␣i, so to confirm the role of S1P1, CD3 antibody-stimulated lym- FIG. 1. Plasma cytokine levels are elevated in diabetic NOD mice. Blood phocytes from nondiabetic and diabetic NOD mice were was collected from nondiabetic (CT) and diabetic (Diab) NOD mice. incubated with S1P or SEW2871 in the presence of either Plasma levels of IFN-␥, IL-4, TNF-␣, IL-5, and IL-2 were measured using ␣ a Cytometric Bead Array, and IL-10 was measured by ELISA. Diabetic pertussis toxin to inhibit G i coupling of S1P1, or with NOD mice had significantly higher levels of circulating IFN-␥ (*P < VPC23019, a specific antagonist of S1P1 (23). As shown in 0.005). IL-10 levels were significantly lower in diabetic NOD when Fig. 2B, both pertussis toxin and VPC23019 inhibit the ␥ compared with control mice (#P < 0.01). Data are shown as mean ؎ SE of six mice per group. effects of S1P on downregulation of IFN- secretion. Taken together, these data support the notion that S1P incubated with 1 ␮mol/l S1P, and the cells were activated using anti-CD3 mAb acts through the S1P1 receptor on T-cells to modulate as described above. After 24 h, medium was collected for IFN-␥ ELISA. Cells IFN-␥ secretion. We also examined IL-17 secretion, which were used for FACS analysis of CD69 and mRNA expression of IFN-␥, CD69, is indicative of a Th17 response, and IL-4 and IL-10 and HIF1␣. secretion, which are indicative of a Th2 response. As Lymphocyte studies using HIF1␣I.1-deficient mice. CD4ϩ T-cells were Ϫ Ϫ shown in Fig. 2C, IL-17 levels were increased approxi- isolated from the spleens of HIF1␣I.1 / and wild-type mice as described mately twofold in CD3 antibody-stimulated diabetic NOD above and were activated in the presence and absence of 1 ␮mol/l S1P for 16 h. Medium was collected for IFN-␥ measurement by ELISA. CD69 surface mouse T-cells compared with antibody-stimulated nondi- expression was analyzed by flow cytometry. abetic lymphocytes, and both S1P and SEW2871 signifi- Statistical analyses. Data were analyzed by ANOVA and Fisher’s protected cantly reduced IL-17 secretion. In contrast, we observed least significant difference test using StatView 6.0 software program. Results no changes in either IL-4 or IL-10 expression in the are expressed as the means Ϯ SE of eight mice per group unless otherwise diabetic NOD T-cells (Fig. 2C). S1P and SEW2871 had no noted in the figure legends. effect on secretion of IL-4 or IL-10. These data indicate that both Th17 and Th1 lymphocyte responses are activated in RESULTS type 1 diabetic mice, and that the S1P-S1P1 receptor axis Elevated levels of IFN-␥ in the plasma of diabetic impacts both types of lymphocyte responses. In addition NOD/LtJ mice. Recently, we reported that aortic endo- to inhibiting cytokine secretion, S1P and SEW2871 also thelial cells from diabetic NOD mice were highly activated decreased the expression of CD69, a marker of T-cell and that monocyte:endothelial interactions were dramati- activation. Surface CD69 expression was significantly in- cally increased in diabetic mice (3). In the current study, creased in diabetic lymphocytes compared with nondia- we explored T-cell activation in this diabetic NOD mouse betic control lymphocytes, averaging 72 Ϯ 2 and 61 Ϯ 2%, model. Plasma from hyperglycemic NOD mice and normo- respectively (Fig. 3). Both S1P and SEW2871 reduced glycemic littermate controls was analyzed for Th1/Th2 CD69 expression in diabetic lymphocytes to levels equal to cytokines (Fig. 1). Plasma IFN-␥ was approximately two- or below those of control, with SEW2871 having a dra- fold higher in the diabetic mice (15.95 Ϯ 0.19 pg/ml) matic impact on CD69 expression in both control and compared with their nondiabetic counterparts (8.54 Ϯ 0.19 diabetic lymphocytes (Fig. 3). Both pertussis toxin and the pg/ml). Plasma levels of TNF-␣, IL-2, IL-4, and IL-5 were S1P1 receptor antagonist VPC23019 prevented S1P action not significantly different between groups. However, dia- on CD69 expression, again implicating the S1P-S1P1 axis betic NOD mice showed a 50% reduction in plasma IL-10 in regulating lymphocyte activation. levels (P Ͻ 0.01). As another measure of T-cell activation, we measured CD4؉ T-cells from diabetic NOD mice are highly T-cell:endothelial adhesion. Our laboratory routinely uses activated, and S1P reduces T-cell activation. We a monocyte:endothelial adhesion assay that uses primary tested the hypothesis that S1P would reduce T-cell activa- mouse endothelial cells and monocytes (3,21,24). We tion and lymphocyte:endothelial interactions. CD4ϩ T- modified this method to study T-cell:endothelial adhesion. cells were isolated from spleens of control and diabetic Interestingly, we observed that diabetic T-cells adhered NOD mice using a mouse CD4 subset column, resulting in more readily to the endothelium than nondiabetic control a population that was about 95% pure CD4ϩ T-cells. The T-cells (Fig. 4). We observed a 50% increase in the number remaining 5% of cells were granulocytes, as evidenced by of adherent lymphocytes from diabetic mice compared 7/4ϩ expression (data not shown). Lymphocytes were with control mice. In some cases, CD4ϩ T-cells were studied either as naı¨ve cells or after stimulation with CD3 pretreated for 4 h with 1 ␮mol/l S1P or SEW2871 before antibody. Naı¨ve T-cells from diabetic mice showed a the lymphocytes were added to the endothelial cells for significant increase in IFN-␥ secretion compared with the adhesion assay. S1P and SEW2871 reduced lymphocy- nondiabetic littermates (Fig. 2A). Stimulation of cells with te:endothelial interactions by ϳ35% (Fig. 4). S1P also

486 DIABETES, VOL. 57, FEBRUARY 2008 S. SRINIVASAN AND ASSOCIATES

FIG. 2. S1P and SEW2871 significantly reduce IFN-␥ and IL-17 secretion in diabetic T-cells. CD4؉ T-cells from nondiabetic control (CT) and diabetic (Diab) NOD mice were either studied freshly isolated (Naı¨ve) or after incubation on immobilized anti-CD3 mAb (ST) for 16 h in the absence or presence of 500 nmol/l or 1 ␮mol/l S1P (؉S1P) or 1 ␮mol/l SEW2871 (؉SEW2871). Some cells were also incubated in the presence of pertussis toxin (؉PTX) or with the S1P1 receptor antagonist VPC23019 (؉VPC23019) for 16 h. Medium was collected for the measurement of IFN-␥, IL-4, and IL-17. A: IFN-␥ production by naı¨ve lymphocytes. IFN-␥ production by freshly isolated naı¨ve lymphocytes from nondiabetic (CT) and diabetic (Diab) NOD mice was measured by ELISA. Diabetic T-cells secreted significantly higher levels of IFN-␥ compared with nondiabetic cells (*P < 0.001). B: Effects of S1P-S1P1 axis on IFN-␥ production in stimulated lymphocytes. Lymphocytes were isolated from nondiabetic (CT) and diabetic (Diab) NOD mice and stimulated via incubation on anti-CD3–coated plates in the presence or absence of S1P, SEW2871, VPC23019, or PTX. IFN-␥ was measured in medium by ELISA. S1P significantly reduced secretion of IFN-␥ in nondiabetic (#P < 0.01) and diabetic lymphocytes (*P < 0.0001) compared with CD3-incubated (ST) cells. SEW2871 also reduced IFN-␥ secretion by nondiabetic (#P < 0.01) and diabetic T-cells (*P < 0.0001). Incubation of cells with Pertussis toxin (PTX) or with VPC23019 blocked the action of S1P and SEW2871 on lymphocyte IFN-␥ secretion. C: Effects of S1P on lymphocyte Th17 and Th2 responses. Lymphocytes were isolated from nondiabetic (CT) and diabetic (Diab) NOD mice and stimulated via incubation on anti-CD3–coated plates in the presence or absence of 1 ␮mol/l S1P or SEW2871. IL-17, IL-10, and IL-4 were measured in cell supernatants by ELISA. *Significantly lower than CT, P < 0.05; **significantly lower than CT, P < 0.02; #significantly higher than CT, P < 0.001; ##significantly lower than Diab, P < 0.005. There were no significant changes in IL-4 or IL-10 production by the lymphocytes. slightly reduced nondiabetic mouse lymphocyte adhesion 0.01; Fig. 5). Taken together, these experiments in vitro to endothelium, although this did not reach statistical and in vivo using both S1P and SEW2871 strongly suggest significance. TNF-␣ stimulation of EC is shown as a that S1P reduces CD4ϩ T-cell activation through action on control for maximal adhesion by EC in our assay. the lymphocyte S1P1 receptor. SEW2871 reduces the circulating proinflammatory S1P inhibits T-cell activation by increasing cytokine IFN-␥ in vivo. The successful reduction of HIF1␣I.1␣ mRNA expression. HIF1␣ is a transcription IFN-␥ production by T-cells in vitro by the S1P1 receptor factor that is induced by both hypoxic and nonhypoxic agonist SEW2871 led us to test the effects of SEW2871 on pathways in T-cells. Two HIF1␣ mRNA isoforms are lymphocyte activation in vivo. SEW2871 was intravenously transcribed and translated, giving rise to HIF1␣ I.1 and I.2 injected into nondiabetic control and diabetic NOD mice. proteins. The short isoform HIFI.1 negatively regulates At 4 h after injection, plasma cytokine levels were mea- lymphocyte activation (18). Based on these findings, we sured by ELISA. Diabetic mice showed increased plasma asked whether S1P impacted HIF1␣ I.1 expression in IFN-␥ levels and reduced IL-10 levels (similar to those T-cells. To examine this, lymphocytes from control nondi- shown in Fig. 1). Plasma IL-4 levels did not change (data abetic littermate and diabetic NOD mice were activated in not shown). SEW2871 decreased plasma IFN-␥ levels in the absence or presence of either 1 ␮mol/l S1P or the diabetic mice by 50% to levels equivalent to those SEW2871. HIF1␣ I.1 mRNA expression was quantified by found in control mice, P Ͻ 0.005 (Fig. 5). Furthermore, quantitative real-time PCR. Interestingly, naı¨ve diabetic SEW2871 treatment in vivo significantly increased plasma NOD lymphocytes expressed 50% less HIF1␣ I.1 mRNA IL-10 levels in both nondiabetic and diabetic mice (P Ͻ than naı¨ve nondiabetic lymphocytes (Fig. 6A). Similarly,

DIABETES, VOL. 57, FEBRUARY 2008 487 S1P AND HIF1␣ REGULATION IN LYMPHOCYTES

(FIG. 3. Reduction in CD69 expression on murine CD4؉ T-cells by S1P and SEW2871. CD4؉ T-cells from control (CT) and diabetic NOD (NOD mice were activated by incubating on anti-CD3 mAb–coated plates in the absence or presence of 1 ␮mol/l S1P (؉S1P), 1 ␮mol/l SEW2871 ؉SEW2871), 100 ng/ml pertussis toxin (؉PTX), or 10 ␮mol/l VPC23019 (؉VPC23019). Cell surface expression of CD69 was assessed by flow) cytometry. Data are expressed as percentage of CD4؉ cells expressing CD69. CD69 was increased on diabetic NOD lymphocytes. Both S1P and SEW2871 reduced CD69 surface expression (*P < 0.05 vs. CT; #P < 0.01 vs. Diab; $P < 0.001 vs. Diab). PTX reversed the effect of S1P (**P < vs. CT ؉ S1P; ##P < 0.008 vs. Diab ؉ S1P). VPC23019 reversed the effect of S1P in Diab lymphocytes (##P < 0.008 vs. Diab ؉ S1P). Data 0.05 represent six mice per experimental condition. on TCR-stimulated activation, diabetic T-cells expressed HIF1␣ I.2 mRNA levels. TCR-stimulated control and dia- 50% less HIF1␣I.1 mRNA compared with nondiabetic betic lymphocytes had approximately a twofold increase T-cells (Fig. 6A). Both S1P and SEW2871 significantly in HIF1␣ I.2 mRNA levels over naı¨ve cells. However, no upregulated HIF1␣ I.1 mRNA in both control and diabetic significant differences were observed between control and lymphocytes, suggesting that S1P and S1P1 agonists could diabetic lymphocytes. Furthermore, neither S1P nor be beneficial for reducing T-cell activation through upregu- SEW2871 changed expression of HIF1␣ I.2 (Fig. 6B). lation of HIF1␣ I.1. Figure 6B shows the expression of These data suggest that action of S1P and SEW2871 through binding to the S1P1 receptor specifically regulates expression of anti-inflammatory HIF1␣I.1.

FIG. 4. Diabetic T-cells have increased adherence to endothelium. (CD4؉ T-cells were isolated from nondiabetic (CT) and diabetic (Diab NOD mice and activated using anti-CD3 mAb–coated plates in the absence or presence of 1 ␮mol/l S1P or 1 ␮mol/l SEW2871. Lympho- FIG. 5. S1P1 receptor agonism in vivo changes the plasma cytokine cytes were labeled with calcein-AM and were incubated with the profile. Nondiabetic (CT) and diabetic (Diab) mice were injected endothelial cell monolayer for an adhesion assay as described in intravenously with 2 mg/kg SEW2871 as described in RESEARCH DESIGN RESEARCH DESIGN AND METHODS. TNF-␣ (10 units/ml) was added to the AND METHODS. Plasma expression of cytokines was measured using endothelium as a positive control for maximal adhesion (#P < 0.0001 ELISA. Diabetic NOD mice showed increased IFN-␥ expression (**P < vs. CT). Diabetic NOD T-cells showed greater adhesion to EC than 0.0001 vs. CT); SEW2871 reduced IFN-␥ expression in NOD mice (*P < nondiabetic NOD T-cells (*P < 0.005 vs. CT). S1P and SEW2871 0.005). Diabetic NOD mice had reduced IL-10 secretion in plasma (#P < reduced adhesion (**P < 0.05 compared with Diab). Data represent the 0.009 vs. CT); SEW2871 increased IL-10 secretion into plasma of NOD .mice per group 6 ؍ mean ؎ SE of four experiments performed in duplicate. mice ($P < 0.01). Data represent n

488 DIABETES, VOL. 57, FEBRUARY 2008 S. SRINIVASAN AND ASSOCIATES

tative experiment, as well as a bar graph showing the means Ϯ SE of results using lymphocytes from six mice per condition. These data strongly suggest that HIF1␣ expression regulates CD69 activation in T-cells. Lymphocytes from HIF1␣I.1 knockout mice do not respond to S1P. Although we achieved compelling re- sults using siRNA, the siRNA was not specific for the HIF1␣I.1 isoform. To prove that HIF1␣I.1 is the signaling molecule for S1P in lymphocytes, we studied T-cells isolated from HIF1␣ I.1-deficient mice. HIF1␣I.1-deficient mice appear normal, and thymocytes and T-cells revealed no phenotypic differences compared with wild-type mice (18). CD4ϩ T-cells were isolated from littermate and HIF1␣I.1-deficient mice (18), and we found a significant pro-inflammatory induction of IFN-␥ secretion (Fig. 8A) and increased CD69 expression (Fig. 8B) in HIF1␣I.1 knockout mice (green line) compared with littermate mice (red line). Importantly, S1P was unable to reduce expres- sion of IFN-␥ and CD69 in the absence of HIF1␣I.1 (orange line), although S1P was effective in reducing IFN-␥ and CD69 in littermate control lymphocytes (blue line) (Fig. 8A and B). We also confirmed these data at the mRNA level, suggesting that HIF1␣I.1 is a primary transcription factor regulating expression of both IFN-␥ and CD69 in lympho- cytes (Fig. 8C). These data show a novel link between S1P and HIF1␣ I.1 signaling to regulate inflammation. Thus, S1P signaling through S1P1 in T-cells drives an anti- inflammatory phenotype via HIF1␣I.1 activation that may be an important therapeutic target for inflammation asso- ciated with diabetes.

FIG. 6. S1P and SEW2871 increase HIF1␣ I.1 mRNA expression in diabetic T-cells. CD4؉ T-cells from nondiabetic (CT) and diabetic DISCUSSION (Diab) NOD mice were either studied freshly isolated (Naı¨ve) or CD4ϩ T-cells have been shown to contribute to the ath- incubated on immobilized anti-CD3 mAb (ST) in the presence of 1 ␮mol/l S1P (؉S1P) or 1 ␮mol/l SEW2871 (؉SEW2871). A: HIF1␣I.1 erosclerotic burden in mice (25). S1P has been shown to mRNA expression. HIF1␣ I.1 mRNA expression was measured by regulate lymphocyte homing to lymph nodes from circu- quantitative real-time PCR. Diabetic T-cells expressed 50% less lation through action on the lymphocyte S1P1 receptor HIF1␣I.1 mRNA compared with nondiabetic T-cells (*P < 0.005). Both ␥ S1P and SEW2871 significantly upregulated HIF I.1 mRNA expression (12). Studies have suggested that S1P also reduces IFN- in control and diabetic lymphocytes, **P < 0.009. B: HIF1␣I.2 mRNA production by lymphocyte cell lines (26), but studies to expression. HIF1␣I.2 mRNA expression was measured by quantitative clearly define the role of S1P in regulating lymphocyte real-time PCR. Expression was increased on lymphocyte stimulation activation have not been performed to date. In the current (*P < 0.009 vs. naı¨ve); however, there was no effect of either S1P or SEW2871 on HIF1␣I.2 mRNA expression. study, we examined the role of S1P in reducing T-cell activation in vivo. We show for the first time that S1P HIF1␣ negatively regulates the activation of T-cells. reduces T-cell activation in type 1 diabetic mice by reduc- To directly test the contribution of HIF1␣ I.1 on T-cell ing IFN-␥ and CD69 expression in primary mouse CD4ϩ activation, we used siRNA to HIF1␣. The siRNA was not T-cells. Furthermore, we identified that S1P acts via the specific for the HIF1␣I.1 isoform in that knockdown of lymphocyte S1P1 receptor to upregulate HIF1␣ isoform HIF1␣ I.1 mRNA was 53% and knockdown of HIF1␣I.2 was I.1, an anti-inflammatory transcription factor that regulates 43% in primary CD4ϩ T-cells from control nondiabetic IFN-␥ and CD69 expression in primary T-cells. Thus, we NOD mice (Fig. 7A). The knockdown of HIF1␣ corre- have identified a novel link between the transcription sponded to a significant 2.7-fold increase in IFN-␥ produc- factor HIF1␣ I.1 and the anti-inflammatory lipid S1P in tion by lymphocytes (Fig. 7B). S1P reduced IFN-␥ lymphocytes. production in lymphocytes transfected with scrambled We found that T-cells isolated from diabetic NOD mouse siRNA, but not in lymphocytes transfected with HIF1␣ spleens were highly activated and secreted approximately siRNA (Fig. 7B), suggesting that HIF1␣ is important for twofold more IFN-␥ and IL-17 than lymphocytes isolated S1P action in lymphocytes. Moreover, knockdown of from control, nondiabetic littermate mice. CD69, a marker HIFI␣ caused a dramatic induction of surface CD69 ex- of early lymphocyte activation, was highly upregulated in pression on lymphocytes, illustrating activation of T-cells diabetic NOD lymphocytes. Taken together, these data when HIF1␣ levels are low (Fig. 7C). As shown in the illustrate that CD4ϩ T-cells are “preactivated ” in diabetic magenta and light blue lines in Fig. 7C, siRNA to HIF1␣ NOD mice and are capable of significantly contributing to increased CD69 expression compared with either scram- the inflammatory state in vivo. bled siRNA (dark blue and dark red lines) or untreated We observed significant reductions in plasma IL-10 nondiabetic lymphocytes (light green and orange lines). levels in type 1 diabetic mice (Figs. 1 and 5). In vivo An isotype control antibody is used in each case to administration of SEW2871, the S1P1 receptor-specific illustrate specificity of the fluorescent signal. Also shown agonist, increased expression of IL-10 (Fig. 5). IL-10 is in Fig. 7C are corresponding dot plots from this represen- secreted by CD4ϩCD25ϩ T regulatory cells and by macro-

DIABETES, VOL. 57, FEBRUARY 2008 489 S1P AND HIF1␣ REGULATION IN LYMPHOCYTES

FIG. 7. S1P acts through HIF1␣ to regulate CD4؉ T-cell activation. CD4؉ T-cells from nondiabetic NOD mice were transfected with either scrambled, control siRNA or HIF1␣ siRNA. Untransfected CD4؉ T-cells from nondiabetic NOD mice were used as controls for comparison to rule out any nonspecific siRNA effects. A: HIF1␣I.1 and I.2 mRNA. Expression of HIF1␣I.1 and I.2 mRNAs were measured by quantitative RT-PCR. Significantly lower than scrambled siRNA and untransfected cells, P < 0.001. B: IFN-␥ secretion. CD4؉ T-cells from nondiabetic NOD mice were* transfected with either scrambled, control siRNA or HIF1␣ siRNA and treated in the absence or presence of S1P (؉S1P). Untransfected CD4؉ T-cells from nondiabetic NOD mice were used as controls. IFN-␥ secretion was measured by ELISA. #Significantly lower than scrambled and untransfected, P < 0.05; *Significantly higher than scrambled and untransfected, P < 0.0001. S1P had no effect in reducing IFN-␥ secretion in the absence of HIF1␣. C: CD69 surface expression. CD69 surface expression on CD4؉ T-cells of nondiabetic NOD mice was quantified by flow cytometry. HIF1␣ siRNA significantly increased the percentage of surface CD69 expression (magenta lines), compared with scrambled siRNA (SCsiRNA) control (dark blue lines). Control, untransfected nondiabetic lymphocytes (CT) are shown in green. Isotype controls were also run with each sample: CT ؉ isotype is shown in orange; the scrambled siRNA ؉ isotype is shown in dark red; and the HIF1␣ siRNA ؉ isotype is shown in light blue. The right panel shows representative dot plots for the corresponding histogram. The bar graph shows the means ؎ SE of the .percentage of CD4؉CD69؉ T-cells from six mice per group. *Significantly higher than CT or SCsiRNA, P < 0.008

490 DIABETES, VOL. 57, FEBRUARY 2008 S. SRINIVASAN AND ASSOCIATES

FIG. 8. HIF1␣ I.1 is required for the anti-inflammatory action of S1P in T-cells. CD4؉ T-cells were isolated from littermate controls (WT) and ␥-HIF1␣ I.1-deficient (HIFI.1KO) mice and incubated with S1P (؉S1P) as described in RESEARCH DESIGN AND METHODS. A: IFN-␥ secretion. IFN secretion was measured by ELISA. Significantly higher than littermate controls, *P < 0.01. Significantly lower than littermate controls, #P < 0.05. .S1P had no effect in the HIF1␣I.1 KO lymphocytes. B: CD69 expression. CD69 surface expression on CD4؉ T-cells was analyzed by flow cytometry HIF1␣I.1-deficient mice had a significant increase in CD69 surface expression (green line) compared with littermate controls (red line). S1P significantly reduced CD69 expression on littermate lymphocytes (blue line) but did not reduce CD69 expression on HIF1␣I.1 lymphocytes ,yellow line). The bar graph represents the means ؎ SE of CD69 expression in eight mice per group. *Significantly lower than littermate controls) P < 0.05; **significantly higher than littermate controls, P < 0.001. C: IFN-␥ and CD69 mRNA expression. IFN-␥ and CD69 mRNA expression in littermate controls and HIF1␣I.1-deficient lymphocytes was measured by quantitative real-time PCR. S1P did not reduce expression of either IFN-␥ or CD69 mRNA in HIF1␣I.1-deficient lymphocytes. *Significantly lower than littermate controls, P < 0.001; **significantly higher than littermate controls, P < 0.0001 by ANOVA. phages. As shown in Fig. 2C, we observed no changes in overexpressed murine S1P4 in a D10G4.1 T-cell line that IL-10 secretion in diabetic T-cells, and S1P or SEW2871 did lacked endogenous S1P receptors. They found that addi- not change secretion of IL-10 by these lymphocytes. We tion of 1 ␮mol/l S1P to the S1P4-D10G4.1 cells caused a have not yet examined IL-10 secretion in response to S1P significant reduction in IFN-␥ secretion (13). We used a or SEW2871 by T-cell subsets, including T regulatory cells. S1P4/S1P1 dual agonist, VPC23153 (27), and found similar In preliminary studies, we found that both S1P and reductions in lymphocyte activation, especially at doses SEW2871 stimulate macrophage secretion of IL-10 through Ͼ1 ␮mol/l (data not shown). We attributed these findings agonism of the S1P1 receptor on macrophages (data not to agonism at S1P1 rather than S1P4 because of the dose shown). Increased IL-10 levels are indicative of an anti- required for maximal effect on lymphocyte function (data inflammatory response, again supporting the notion that not shown). SEW2871 is specific for S1P1 and has no the S1P-S1P1 axis promotes an anti-inflammatory pheno- action on S1P4, and we found that SEW2871 was quite type in vivo. Experiments to characterize the role of S1P in effective in reducing lymphocyte activation, both in vitro T-cell subset activation will be important follow-up studies and in vivo. Thus, our data suggest that S1P1 is the primary to this current work. receptor through which S1P regulates lymphocyte activa- Lymphocytes express the S1P receptors S1P1 and S1P4. tion, although we cannot completely rule out some contri- We found that the activation state of CD4ϩ T-cells was bution of S1P4. reduced by S1P through action on the S1P1 receptor using The murine HIF1␣ gene contains two different first the S1P1 receptor-specific agonist SEW2871 both in vivo exons, termed I.1 and I.2 (15). Expression of HIF1␣I.1 and and in vitro. This is somewhat in contrast to a recent study -I.2 is regulated via two distinct promoters, which give rise by Goetzl and colleagues (13) that indicated that S1P4 to these two mRNA isoforms (15). Sitkovsky and col- regulated IFN-␥ secretion by lymphocytes. This group leagues (18) have recently shown that the HIF1␣I.1 iso-

DIABETES, VOL. 57, FEBRUARY 2008 491 S1P AND HIF1␣ REGULATION IN LYMPHOCYTES form negatively regulates T-cell activation. In Fig. 7, we 2. Hla T, Maciag T: An abundant transcript induced in differentiating human used siRNA to HIF1␣ and found approximately a ϳ50% endothelial cells encodes a polypeptide with structural similarities to reduction in HIF1␣I.1 mRNA and HIF1␣I.2 mRNA expres- G-protein-coupled receptors. J Biol Chem 265:9308–9313, 1990 ␣ 3. Whetzel AM, Bolick DT, Srinivasan S, Macdonald TL, Morris MA, Ley K, sion in T-cells. Interestingly, the HIF1 siRNA completely Hedrick CC: Sphingosine-1 phosphate prevents monocyte/endothelial in- abolished the action of S1P on IFN-␥ production (Fig. 7B). teractions in type 1 diabetic NOD mice through activation of the S1P1 This was somewhat surprising in that we observed a total receptor. Circ Res 99:731–739, 2006 loss of the S1P effect with only a 50% loss of HIF1␣I.1. 4. Awad AS, Ye H, Huang L, Li L, Foss FW Jr, Macdonald TL, Lynch KR, There are two possible explanations for this outcome. Okusa MD: Selective sphingosine 1-phosphate 1 receptor activation re- First, HIF1␣I.2 plays a minor role in regulating T-cell duces ischemia-reperfusion injury in mouse kidney. Am J Physiol Renal activation as we have shown previously (18). Thus, reduc- Physiol 290:F1516–F1524, 2006 ␣ 5. Man K, Ng KT, Lee TK, Lo CM, Sun CK, Li XL, Zhao Y, Ho JW, Fan ST: tions in expression of both isoforms of HIF1 may con- FTY720 attenuates hepatic ischemia-reperfusion injury in normal and tribute to the loss of the S1P effect. Second, antibodies to cirrhotic livers. Am J Transplant 5:40–49, 2005 accurately quantify protein levels of the two isoforms are 6. Baumruker T, Billich A, Brinkmann V: FTY720, an immunomodulatory not available, so we are unsure as to the level of expres- sphingolipid mimetic: translation of a novel mechanism into clinical sion of HIF1␣I.1 protein in the siRNA studies. It could be benefit in multiple sclerosis. Expert Opin Investig Drugs 16:283–289, 2007 that there is very little HIF1␣I.1 protein remaining in the 7. Yang Z, Chen M, Fialkow LB, Ellett JD, Wu R, Brinkmann V, Nadler JL, cell after HIF1␣ siRNA treatment. However, we anticipate Lynch KR: The immune modulator FYT720 prevents autoimmune diabetes that S1P primarily acts through the HIF1␣I.1 isoform to in nonobese diabetic mice small star, filled. Clin Immunol 107:30–35, 2003 8. Nofer JR, Bot M, Brodde M, Taylor PJ, Salm P, Brinkmann V, van Berkel T, regulate T-cell activation, in that mice deficient solely in Assmann G, Biessen EA: FTY720, a synthetic sphingosine 1 phosphate the HIF1␣I.1 isoform showed increased T-cell activation, analogue, inhibits development of atherosclerosis in low-density lipopro- and S1P had no ability to reduce this cell activation (Fig. tein receptor-deficient mice. Circulation 115:501–508, 2007 8). Levels of HIF1␣I.2 are normal in these knockout mice, 9. Graler MH, Huang MC, Watson S, Goetzl EJ: Immunological effects of allowing us to specifically study the I.1 isoform in vivo. transgenic constitutive expression of the type 1 sphingosine 1-phosphate Thus, our HIF1␣I.1 knockout mouse studies indicate that a receptor by mouse lymphocytes. J Immunol 174:1997–2003, 2005 primary pathway of regulation of T-cell activation by S1P 10. Sawicka E, Dubois G, Jarai G, Edwards M, Thomas M, Nicholls A, Albert R, Newson C, Brinkmann V, Walker C: The sphingosine 1-phosphate is through HIF1␣I.1. ϩ ␣ receptor agonist FTY720 differentially affects the sequestration of CD4 / HIF1 is known for its regulation by hypoxia but can CD25ϩ T-regulatory cells and enhances their functional activity. J Immu- also be induced by nonhypoxic stimuli (28). Karliner et al. nol 175:7973–7980, 2005 (29) observed that 10 ␮mol/l S1P protected rat neonatal 11. Nombela-Arrieta C, Lacalle RA, Montoya MC, Kunisaki Y, Megias D, cardiomyocytes from hypoxic cell death in vitro, although Marques M, Carrera AC, Manes S, Fukui Y, Martinez A, Stein JV: Differen- the mechanisms for this protection were not identified. tial requirements for DOCK2 and phosphoinositide-3-kinase gamma during Further evidence for a link between S1P and hypoxia T and B lymphocyte homing. Immunity 21:429–441, 2004 comes from a recent study by Pyne and colleagues (30) 12. Dorsam G, Graeler MH, Seroogy C, Kong Y, Voice JK, Goetzl EJ: Trans- duction of multiple effects of sphingosine 1-phosphate (S1P) on T cell that illustrates that hypoxia increased sphingosine kinase functions by the S1P1 G protein-coupled receptor. J Immunol 171:3500– 1 expression to produce more S1P in pulmonary smooth 3507, 2003 muscle cells. Thus, S1P synthesis could be a protective 13. Wang W, Graeler MH, Goetzl EJ: Type 4 sphingosine 1-phosphate G mechanism induced by the SMC when exposed to chronic protein-coupled receptor (S1P4) transduces S1P effects on T cell prolifer- hypoxic conditions. Taken together, our results on lym- ation and cytokine secretion without signaling migration. FASEB J 19: phocytes and the hypoxia studies of others show a novel 1731–1733, 2005 link between HIF1␣ and the protective action of S1P. 14. Richard DE, Berra E, Pouyssegur J: Nonhypoxic pathway mediates the ϩ induction of hypoxia-inducible factor 1alpha in vascular smooth muscle In summary, we have found that CD4 T-cells from type cells. J Biol Chem 275:26765–26771, 2000 1 diabetic mouse models are “preactivated ” in vivo to 15. Wenger RH, Rolfs A, Spielmann P, Zimmermann DR, Gassmann M: Mouse release pro-inflammatory cytokines. The anti-inflamma- hypoxia-inducible factor-1alpha is encoded by two different mRNA iso- tory sphingolipid S1P reduces T-cell activation in diabetic forms: expression from a tissue-specific and a housekeeping-type pro- mice in vivo. Action of S1P on lymphocyte activation is moter. Blood 91:3471–3480, 1998 16. Lukashev D, Caldwell C, Ohta A, Chen P, Sitkovsky M: Differential primarily through action on the lymphocyte S1P1 receptor regulation of two alternatively spliced isoforms of hypoxia-inducible through downstream target activation of the anti-inflam- factor-1 alpha in activated T lymphocytes. J Biol Chem 276:48754–48763, matory transcription factor HIF1␣ I.1. Thus, specific tar- 2001 geting of the S1P1-HIF1␣I.1 axis in lymphocytes may be an 17. Kojima H, Gu H, Nomura S, Caldwell CC, Kobata T, Carmeliet P, Semenza important regulator of immune responses in chronic dis- GL, Sitkovsky MV: Abnormal B lymphocyte development and autoimmu- eases, such as diabetes and atherosclerosis. nity in hypoxia-inducible factor 1alpha -deficient chimeric mice. Proc Natl Acad SciUSA99:2170–2174, 2002 18. Lukashev D, Klebanov B, Kojima H, Grinberg A, Ohta A, Berenfeld L, ACKNOWLEDGMENTS Wenger RH, Ohta A, Sitkovsky M: Cutting edge: hypoxia-inducible factor 1alpha and its activation-inducible short isoform I. 1 negatively regulate C.C.H. has received research grants from the Juvenile functions of CD4ϩ and CD8ϩ T lymphocytes. J Immunol 177:4962–4965, Diabetes Research Foundation and the National Institutes 2006 of Health (HL079621). 19. Sitkovsky MV, Lukashev D, Apasov S, Kojima H, Koshiba M, Caldwell C, We thank Joanne Lannigan and Mike Solga in the UVA Ohta A, Thiel M: Physiological control of immune response and inflamma- Flow Cytometry Core Facility for their technical assis- tory tissue damage by hypoxia-inducible factors and adenosine A2A receptors. Annu Rev Immunol 22:657–682, 2004 tance, Dr. Joel Linden (UVA) for helpful discussions, and 20. Sitkovsky M, Lukashev D: Regulation of immune cells by local-tissue Dr. Margaret A. Morris (UVA) for providing some of the oxygen tension: HIF1 alpha and adenosine receptors. Nat Rev Immunol NOD mice. 5:712–721, 2005 21. Hatley ME, Srinivasan S, Reilly KB, Bolick DT, Hedrick CC: Increased production of 12/15 lipoxygenase eicosanoids accelerates monocyte/endo- REFERENCES thelial interactions in diabetic db/db mice. J Biol Chem 278:25369–25375, 1. Rosen H, Goetzl EJ: Sphingosine 1-phosphate and its receptors: an 2003 autocrine and paracrine network. Nat Rev Immunol 5:560–570, 2005 22. Livak KJ, Schmittgen TD: Analysis of relative gene expression data using

492 DIABETES, VOL. 57, FEBRUARY 2008 S. SRINIVASAN AND ASSOCIATES

real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods 27. Clemens JJ, Davis MD, Lynch KR, Macdonald TL: Synthesis of benzimid- 25:402–408, 2001 azole based analogues of sphingosine-1-phosphate: discovery of potent, 23. Davis MD, Clemens JJ, Macdonald TL, Lynch KR: Sphingosine 1-phosphate subtype-selective S1P(4) receptor agonists. Bioorg Med Chem Lett 14: analogs as receptor antagonists. J Biol Chem 280:9833–9841, 2005 4903–4906, 2004 24. Bolick DT, Srinivasan S, Kim KW, Hatley ME, Clemens JJ, Whetzel A, 28. Zhou J, Brune B: Cytokines and hormones in the regulation of hypoxia Ferger N, Macdonald TL, Davis MD, Tsao PS, Lynch KR, Hedrick CC: inducible factor-1alpha (HIF-1alpha). Cardiovasc Hematol Agents Med Sphingosine-1-phosphate prevents tumor necrosis factor-{alpha}-mediated Chem 4:189–197, 2006 monocyte adhesion to aortic endothelium in mice. Arterioscler Thromb 29. Karliner JS, Honbo N, Summers K, Gray MO, Goetzl EJ: The lysophospho- Vasc Biol 25:976–981, 2005 lipids sphingosine-1-phosphate and lysophosphatidic acid enhance sur- 25. Robertson AK, Hansson GK: T cells in atherogenesis: for better or for vival during hypoxia in neonatal rat cardiac myocytes. J Mol Cell Cardiol worse? Arterioscler Thromb Vasc Biol 26:2421–2432, 2006 33:1713–1717, 2001 26. Fieger CB, Huang MC, Van Brocklyn JR, Goetzl EJ: Type 1 sphingosine 30. Ahmad M, Long JS, Pyne NJ, Pyne S: The effect of hypoxia on lipid 1-phosphate G protein-coupled receptor signaling of lymphocyte functions phosphate receptor and sphingosine kinase expression and mitogen- requires sulfation of its extracellular amino-terminal tyrosines. FASEB J activated protein kinase signaling in human pulmonary smooth muscle 19:1926–1928, 2005 cells. Prostaglandins Other Lipid Mediat 79:278–286, 2006

DIABETES, VOL. 57, FEBRUARY 2008 493