Thiazolidinedione Activation of Peroxisome Proliferator-activated Receptor γ Can Enhance Mitochondrial Potential and Promote Cell Survival*

Thiazolidinediones (TZDs) are widely used for treatment of type 2 diabetes mellitus. Peroxisome proliferator-activated receptor γ (PPARγ) is the molecular target of TZDs and is believed to mediate the apoptotic effects of this class of drugs in a variety of cell types, including B and T lymphocytes. The finding that TZDs induce lymphocyte death has raised concerns regarding whether TZDs might further impair immune functions in diabetics. To address this issue, we investigated the roles of PPARγ and TZDs in lymphocyte survival. PPARγ was up-regulated upon T cell activation. As previously reported, PPARγ agonists induced T cell death in a dose-dependent manner. However, the concentrations of TZD needed to cause T cell death were above those needed to induce PPARγ-dependent transcription. Surprisingly, at concentrations that induce optimal transcriptional activation, TZD activation of PPARγ protected cells from apoptosis following growth factor withdrawal. The survival-enhancing effects depended on both the presence and activation of PPARγ. Measurements of mitochondrial potential revealed that PPARγ activation enhanced the ability of cells to maintain their mitochondrial potential. These data indicate that activation of PPARγ with TZDs can promote cell survival and suggest that PPARγ activation may potentially augment the immune responses of diabetic patients.

Diabetes mellitus is one of the most common noncommunicable diseases. Approximately 100 million people worldwide, including 16 million people in the United States, suffer from diabetes. With about 54,000 deaths per year, it represents the seventh leading cause of death in the United States. The major causes of morbidity and mortality are results of long term complications of hyperglycemia, involving heart, kidneys, eyes, nerves, and the immune system. Among the diabetic population, 80 -90% are affected by type 2 diabetes, in which impaired tissue sensitivity to insulin is the primary metabolic defect. Thiazolidinediones (TZDs 1 ; "glitazones") are a new class of synthetic compounds that potentiate insulin action in the target tissues, alleviate hyperglycemia, and are efficacious in treating type 2 diabetes (see Refs. 1 and 2; reviewed in Ref. 3). At the molecular level, these compounds function as highly specific pharmacologic ligands of peroxisome proliferator-activated receptor ␥ (PPAR␥) (4).
PPARs are members of the nuclear hormone receptor superfamily. There are three isoforms of PPARs, ␣, ␦ (also known as ␤), and ␥. Among them, PPAR␥ is of particular interest, because, in addition to diabetes, it has been implicated in several other significant human pathological conditions, including atherosclerosis, cancer, and inflammation (5,6). Like other members of the nuclear receptor family, PPAR␥ serves as a transcription factor. Upon ligand binding, the receptor undergoes a conformational change. This disrupts a corepressor complex and leads to coactivator recruitment and transcriptional activation. PPAR␥ functions as a heterodimer with retinoid X receptor ␣, binding to a specific direct repeat DNA sequence and regulating many target genes involved in lipid metabolism.
PPAR␥ is most abundantly expressed in adipose tissue and colon epithelial cells. It is also broadly expressed in many other tissues including, but not limited to, bone marrow precursors, monocytes/macrophages, lymphocytes, pneumocytes, breast epithelial cells, hepatocytes, and myocytes (7). Studies of the cellular functions of PPAR␥ have thus far been focused mainly on adipose tissue and macrophages. In adipose tissue, PPAR␥ promotes adipogenesis by activating transcription of its target genes, such as adipocyte protein 2, lipoprotein lipase, and acyl-CoA synthetase (reviewed in Refs. 8 and 9). In foamy macrophages of the vasculature, PPAR␥ reduces lipid accumulation by affecting cholesterol flux in the cells through transcriptional regulation of scavenger receptors and reverse cholesterol transporters (10 -12).
Although PPAR␥ is expressed in B and T cells, little is known about its primary functions in lymphocytes (13)(14)(15). Recent studies have suggested that PPAR␥ may play a role in modulating immune functions. It has been shown in human and murine T cells that treatment with PPAR␥ ligand leads to inhibition of T cell proliferation and a decrease in interleukin-2 (IL-2) production (14,16). In animal models, ligands of PPAR␥, including 15-deoxy-⌬ 12,14 -prostaglandin J 2 (15d-PGJ 2 ) and troglitazone, are also effective against inflammatory diseases, al-leviating rheumatoid arthritis and inflammatory bowel disease (17,18). It is, however, unclear whether the anti-inflammatory effects of TZD drugs are mediated through PPAR␥. Using PPAR␥-deficient embryonic stem cells, several groups have shown that in monocytes/macrophages, inhibition of cytokine production and proinflammatory gene expression by both 15d-PGJ 2 and TZDs is independent of PPAR␥ (10,11). In addition, non-TZD-derived PPAR␥ agonists do not inhibit the production of proinflammatory cytokines in activated macrophages (19).
In lymphocytes, some studies have reported that treatment of cells with certain PPAR␥ ligands induces apoptosis in both T and B cells (15,20), suggesting that PPAR␥ might be involved in the down-regulation of immune responses through the induction of lymphocyte death. These findings have significant implications, since millions of type 2 diabetic patients are currently taking TZDs to control their hyperglycemia and prevent complications. A serious concern is that TZDs might further impair the already compromised immune functions in these patients. We therefore set out to investigate the roles of PPAR␥ in lymphocyte survival.

EXPERIMENTAL PROCEDURES
Animals-C57BL/6 mice were purchased from the Jackson Laboratory (Bar Harbor, ME). 2C T cell receptor transgenic/RAG2Ϫ/Ϫ mice have been previously described (21). All mice were maintained in the University of Pennsylvania Animal Barrier Facilities (Philadelphia, PA). Animals received humane care in compliance with the "Guide for the Care and Use of Laboratory Animals" published by the National Institutes of Health (Bethesda, MD).
Chemicals-Rosiglitazone was obtained as a generous gift from GlaxoSmithKline, and ciglitazone was purchased from Biomol (Plymouth Meeting, PA). Sources of all other reagents are indicated below.
T Cell Purification and Culture-T cells were purified from human peripheral blood or mouse spleen by negative selection using Stem-Sep TM magnetic columns (StemCell Technologies, Vancouver, Canada) following the manufacturer's instructions. A purity of 95-99% was normally obtained with this method. Cells were then cultured in Dulbecco's modified Eagle's medium supplemented with glutamine, penicillin/streptomycin, HEPES buffer, minimal essential medium nonessential amino acids (Invitrogen), ␤-mercaptoethanol, and 10% fetal bovine serum (Mediatech, Inc., Herndon, VA).
T Cell Stimulation and Proliferation Assay-For T cell stimulation with anti-CD3 (145-2C11; BD PharMingen, San Diego, CA) alone or anti-CD3 plus anti-CD28 (37.51; BD PharMingen), the antibodies were covalently linked to tosyl-activated Dynabeads M-450 (Dynal, Great Neck, NY) according to the manufacturer's instructions. The antibodycoated beads were then added to purified T cells at a 3:1 bead/cell ratio for the time indicated for each experiment. For T cell stimulation with antigen-presenting cells, 2C T cell receptor transgenic T cells were cultured with irradiated C57BL/6 splenocytes plus 50 nM antigenic peptide SIYRYYGL (Multiple Peptide Systems, San Diego, CA) as described previously (21). For T cell proliferation assays, triplicate T cell cultures were plated at 2.5 ϫ 10 5 /ml in the presence of PPAR␥ agonists or Me 2 SO. Cells were stimulated with anti-CD3 and anti-CD28-coated beads for 3 days and pulsed with 5 nCi/l [ 3 H]thymidine for 6 -8 h at 37°C. Cells were harvested with a Tomtec harvester (Hamden, CT) on glass fiber filters and analyzed by a 1205 Betaplate liquid scintillation counter (Wallac, Turku, Finland).
RNA Preparation and RT-PCR-Total RNA was isolated from peripheral T cells 24 h poststimulation using TRIzol reagent (Invitrogen) following the manufacturer's instructions. Reverse transcription was carried out using oligo(dT) primer and Superscript TM II (Invitrogen) according to the manufacturer's instructions. PCR was performed using AmpliTaq DNA polymerase and buffers from Applied Biosystems (Branchburg, NJ). The reactions were conducted in a 9600 GeneAmp PCR system (PerkinElmer Life Sciences) using the following conditions: denaturation at 94°C for 3 min, followed by 25 cycles of 94°C for 30 s, 52°C for 30 s, and 72°C for 1 min, followed by extension at 72°C for 5 min. The forward and reverse primers for the amplification of the human PPAR␥ cDNA are as follows: 5Ј-TAT CAA GCC CTT CAC TAC TG-3Ј and 5Ј-CTG ATG GCA TTA TGA GAC AT-3Ј, which generate a 443-bp PCR product.
Cell Lines and Culture-FL5.12 cells, an IL-3-dependent hematopoietic cell line, were used for transfection experiments. Cells were transfected with pcDNA3.1-hPPAR␥1 with electroporation as described previously (22). G418-resistant clones transfected with PPAR␥ were identified by Western blot as described above. Each of the cell survival experiments was performed with at least two PPAR␥ lines and two empty vector control lines. All cell lines were cultured in complete RPMI supplemented with 10% fetal bovine serum (Mediatech), 0.3 pg/ml recombinant IL-3 (BD PharMingen), and 1 mg/ml Geneticin (Invitrogen).
Luciferase Assay-A reporter construct, acyl-CoAx3-TK-LUC (23), and a ␤-galactosidase expression plasmid were co-transfected by electroporation into a FL5.12-based PPAR␥ cell line and an empty vector control line. Different doses of rosiglitazone or Me 2 SO, the drug vehicle, were added 16 h after transfection. Cells were then collected 24 h after the drug addition. Luciferase (Promega, Madison, WI) and ␤-galactosidase assays (CLONTECH Laboratories, Palo Alto, CA) were performed according to the manufacturer's instructions. Results of PPAR␥ transcriptional activity were normalized to ␤-galactosidase expression.
IL-3 Withdrawal, Glucose Withdrawal, and Cell Viability Determination-To withdraw IL-3, FL5.12 cells were washed three times in RPMI and resuspended in complete RPMI lacking recombinant IL-3. To limit glucose, cells were washed in glucose-free RPMI three times and resuspended in RPMI supplemented with 10% dialyzed fetal bovine serum (Mediatech) and 50 M of glucose, which is one-two hundredth of the standard glucose used in regular RPMI media. Cell viability was determined by cellular exclusion of 2 g/ml propidium iodide followed by flow cytometric analysis of 1 ϫ 10 5 events.

PPAR␥ Is Expressed in Peripheral T Lymphocytes and Is
Up-regulated upon T Cell Activation-It has been reported that PPAR␥ is expressed in murine T cell clones and human peripheral T lymphocytes (14,16). To confirm these findings, PPAR␥ mRNA levels in purified human peripheral T lymphocytes were assayed under three different activation conditions using RT-PCR. T cells were either resting, treated with anti-CD3 to stimulate the T cell receptor pathway, or treated with anti-CD3 plus anti-CD28 to activate the co-stimulation pathway as well. As shown in Fig. 1A, a trace amount of PPAR␥ mRNA was detected in resting T lymphocytes. The expression was upregulated in T cells stimulated with anti-CD3 antibody and even more so in T cells optimally stimulated with both anti-CD3 and anti-CD28 antibodies. As a loading control, ␤-actin expression levels were similar in all of these specimens.
To further verify that PPAR␥ is up-regulated in activated T lymphocytes with a different system, Western blot analysis was performed on murine primary T cells. T cells were either unstimulated or stimulated in vitro with antigen-presenting cells for 1 day or 2 days. Consistent with the RT-PCR results on human cells, the Western blot analysis showed that PPAR␥ protein expression was increased in activated T cells after 2 days of stimulation (Fig. 1B). Taken together, these data demonstrate that PPAR␥ is up-regulated following T cell activation.
At High Concentrations, 15d-PGJ 2 and Ciglitazone Inhibit T Cell Proliferation and Induce Cell Death-Previously, it was reported that activation of PPAR␥ with certain ligands inhibits T cell proliferative responses (14,16) and induces apoptosis in B and T lymphocytes (15,20). It was suggested that decreased lymphocyte proliferation is a consequence of the cell death (15,20). To confirm these findings, we treated activated T cells with PPAR␥ ligands, examined T cell proliferation by [ 3 H]thymidine incorporation, and determined cell viability by propidium iodide staining. As shown in Fig. 2A, both 15d-PGJ 2 and ciglitazone inhibited T cell proliferation in a dose-dependent fashion. The effective concentrations for the inhibition were Ն5 M for 15d-PGJ 2 and Ն80 M for ciglitazone, respectively. In contrast, Wy 14643, an agonist of PPAR␣, had no significant effects on cell survival at concentrations up to 160 M. Treatment of these cells with 15d-PGJ 2 and ciglitazone also increased cell death (Fig. 2B, left panel), which correlated with the inhibition on T cell proliferation. These findings are similar to those previously reported by others in the other systems (14,16). As a control, we also examined the effects of PPAR␥ on the survival of FL5.12 cells, an IL-3-dependent lymphocyte line that expresses little or no PPAR␥. The FL5.12 cell line has been previously studied extensively for multiple cell death and survival pathways (24 -27). Compared with the primary T cells, higher doses of 15d-PGJ 2 and ciglitazone were required to cause significant death of FL5.12 cells, indicating that this cell line was less sensitive to the apoptotic effects of these drugs (Fig. 2B).
At Low Concentrations, PPAR␥ Ligands Lead to Increased Cell Survival Rather than Death in a PPAR␥-dependent Fashion-If PPAR␥ is responsible for the cell death-inducing effects of 15d-PGJ 2 and ciglitazone, we reasoned that introduction of PPAR␥ would restore the sensitivity of FL5.12 cells to death induced by these drugs. To test this possibility, FL5.12 cell lines that express human PPAR␥1 were established. Western blot analysis confirmed that PPAR␥ protein was overexpressed in the stably transfected cell lines (Fig. 3A). We compared cell survival of the vector control and PPAR␥ lines treated with or without 15d-PGJ 2 and ciglitazone. Unexpectedly, PPAR␥ lines died at the same rate and over the same dose range as vector control lines when treated with the drugs (data not shown), indicating that overexpression of PPAR␥ did not further sensitize cells to the death-inducing effects of 15d-PGJ 2 and ciglitazone.
To test whether TZD-activation of PPAR␥ might alter the sensitivity of cells to a different death stimulus, we subjected the cells to IL-3 withdrawal. FL5.12 cells undergo rapid apoptotic death in 18 -48 h upon IL-3 withdrawal (22). Under this condition, we compared survival of a PPAR␥ and a vector control cell line treated with or without PPAR␥ ligands. Surprisingly, the PPAR␥ line, treated with low concentrations of ciglitazone (10 M) or rosiglitazone (0.5 M), survived ϳ2 or 3 times better than Me 2 SO-treated cells, respectively, at 24 h after IL-3 withdrawal. The survival benefits depended both on the presence of PPAR␥ (Fig. 3B, compare Me 2 SO-treated PPAR␥ cells with Me 2 SO-treated control cells) and the activation of PPAR␥ with agonists (compare ciglitazone-or rosiglitazone-treated PPAR␥ cells with Me 2 SO-treated cells). The presence of unidentified physiological activators of PPAR␥ in the serum or cells may account for the better survival of the PPAR␥ line with Me 2 SO treatment than the control cell line with the same treatment. With PPAR␥ lines of different expression levels (Fig. 3A, g2 versus g14), there was a close correlation between the dosage of PPAR␥ and degree of cell survival-enhancing effects; e.g. the g14 line (Fig. 3A) exhibited reproducible protection but less protection than the g2 line (data not shown). The dosage effects were also observed in subsequent mitochondrial membrane potential assays and glucose limitation assays (see below).
Using a luciferase reporter construct carrying PPAR response elements, we confirmed that the PPAR␥ introduced into the cells was functionally active (Fig. 3C). It was activated to a near maximal level by 0.5 M rosiglitazone, the same concentration at which we observed a survival phenotype. There was a good quantitative correlation between the -fold level of transcriptional activation (ϳ3-fold) and the degree of the cell survival promoted by 0.5 M rosiglitazone (ϳ3-fold; Fig. 3, compare B with C), suggesting that enhanced viability by rosiglitazone is mediated through the transcriptional activity of PPAR␥.
We next determined the time course of the survival-promoting effect of PPAR␥ (Fig. 3D). Although activation of PPAR␥ protected cells from death up to 24 h after IL-3 withdrawal, almost all cells died by 48 h after IL-3 removal irrespective of the presence or activation status of PPAR␥. This cannot be attributed to loss of drug activity, since the addition of fresh PPAR␥ agonists at 24 h did not prolong survival beyond 48 h (data not shown). Thus, the cell survival-promoting effect of PPAR␥ is short-lived compared with that of other antiapoptotic genes, such as bcl-x L , suggesting that PPAR␥ promotes cell survival by a different mechanism.
The cell survival-promoting effects depended on the dosage of rosiglitazone (Fig. 3E). Importantly, at a concentration of 50 nM, which is comparable with its dissociation constant (K D ) for PPAR␥ (48 -100 nM), rosiglitazone enhanced cell survival to a near maximal level following IL-3 deprivation, providing another piece of evidence that the effects of rosiglitazone are mediated through binding to PPAR␥.
PPAR␥ Maintains Mitochondrial Potential during Cell Death-Growth factor withdrawal results in a rapid decline in glycolysis and mitochondrial potential (reviewed in Ref. 28). Therefore, we investigated the effects of PPAR␥ activation on the mitochondrial membrane potential. The membrane potential was measured by mitochondrial incorporation of tetramethylrhodamine ethyl ester, a potentiometric dye, and analyzed by flow cytometry. In a vector control cell line, mitochondrial membrane potential was significantly reduced, even in the presence of rosiglitazone, at 12 h after IL-3 withdrawal. In contrast, the reduction was largely prevented by the activation of PPAR␥ in a PPAR␥-transfected cell line (Fig. 4A). Maintenance of mitochondrial membrane potential depended on the activation of PPAR␥, since the maintenance was observed with rosiglitazone treatment but not with Me 2 SO treatment in the PPAR␥ line (Fig. 4B). From these experiments, we conclude that PPAR␥ activation results in maintenance of the mitochondrial membrane potential following IL-3 deprivation.

PPAR␥ Promotes Cell Survival and Maintains Mitochondrial Potential When Glucose Is Limited-Previous studies have
shown that a primary role of IL-3 in cell survival is to permit cells to take up and utilize glucose, and glucose limitation has adverse effects on cell survival similar to IL-3 withdrawal (27). Based on the role of PPAR␥ to activate genes involved in lipid metabolism, we hypothesized that PPAR␥ might promote the maintenance of mitochondrial potential by making alternative substrates available to mitochondria. To investigate whether PPAR␥ could protect cell death caused by a decline in glycolysis, we examined cell survival under the condition of glucose restriction. As shown in Fig. 5A, when cells were cultured in media containing 50 M glucose, which is one-two hundredth of the concentration in standard RPMI, the PPAR␥ transfected cell line demonstrated better survival in both a PPAR␥-dependent (Fig. 5A, compare Me 2 SO-treated PPAR␥ cell line with Me 2 SO-treated vector control line) and activation-dependent manner (Fig. 5A, compare the rosiglitazone-treated with the Me 2 SO-treated PPAR␥ cell line). Furthermore, activation of PPAR␥ prevented the decline of mitochondrial potential in cells cultured in low glucose (Fig. 5B). Taken together, these data suggest that PPAR␥ promotes cell survival by allowing cells to maintain a mitochondrial potential when substrates are limited by either IL-3 removal or glucose restriction.
Increased Cell Survival Is a Specific Response to PPAR␥ Activation-To test whether the effects on cell survival were specific for PPAR␥, we generated a FL5.12 cell line stably transfected with PPAR␣. Recently, it was shown that the ␣ isoform is present in resting murine lymphocytes (29). In contrast to the PPAR␥-transfected lines, the PPAR␣-transfected cells, when activated by an ␣-ligand, Wy 14643, did not show enhanced cell survival (Fig. 6).

DISCUSSION
Although it was discovered several years ago that PPAR␥ is expressed in hematopoietic cells (13), the role PPAR␥ plays in these cells remains largely unknown. Recent studies have shown that PPAR␥ is expressed in lymphocytes and that acti-vation of PPAR␥ by its ligands inhibits T cell proliferation and IL-2 production (14,16). It has also been reported that PPAR␥ induces apoptosis in both B and T lymphocytes (15,20). In the present study, we confirmed that PPAR␥ is expressed in both human and murine peripheral T cells. Moreover, we have dem- onstrated for the first time that PPAR␥ is up-regulated at both the messenger RNA and protein levels following T cell activation. It is possible that this up-regulation reflects PPAR␥ involvement in allowing the cell to meet the increased metabolic demand and/or synthetic events that occur as a result of T cell activation. An effective immune response requires both the proliferative expansion of reactive T cells and the stimulation of the cells' ability to perform specific immune functions, such as cytokine production (30). Both proliferation and cytokine production place significant bioenergetic demands on the cell and are likely to affect glucose utilization, mitochondrial function, and lipid biosynthesis.
Using high concentrations of 15d-PGJ 2 and ciglitazone, we reproduced the cell death-inducing effects observed by others. Surprisingly, under conditions of growth factor withdrawal or glucose limitation, PPAR␥-overexpressing cells survive better when treated with either ciglitazone or rosiglitazone at concentrations that are comparable with their K D values for PPAR␥. The survival benefit depends on both the presence and the activation of PPAR␥. The survival-promoting effects are specific for PPAR␥, since PPAR␣ did not affect survival of FL5.12 cells. The ability of PPAR␥ to maintain mitochondrial membrane potential suggests that it sustains survival by providing mitochondria with alternative substrates for coupled electron transport and oxidative phosphorylation. In support of this notion, PPAR␥ enhances cell viability when glucose is limited. It is also possible that PPAR␥ promotes cell survival by reducing the fatty acid-dependent uncoupling of mitochondrial potential through the uncoupling proteins. PPAR␥ promotes cell survival by a different mechanism from antiapoptotic genes in the Bcl-2 family. Although PPAR␥ delays cell death, cells die eventually by 48 h after IL-3 withdrawal. In contrast, FL5.12 cells transfected with Bcl-x L live up to several days under the same conditions (24). Moreover, under IL-3 withdrawal condition, the mitochondrial potential in Bcl-x L -overexpressing cells declines. Bcl-x L prolongs cell life by keeping the voltage-dependent anion channel on the outer membrane of mitochondrial open, sustaining ADP/ATP exchange and so maintaining mitochondrial homeostasis (31). In contrast, activated PPAR␥ maintains the mitochondrial membrane potential.
Opposite to the cell survival effects observed in this study, it has been reported that activation of PPAR␥ with its ligands induces apoptosis in several types of normal or tumor tissues and cell lines. These include endothelial cells (32), vascular smooth muscle cells (33), B lymphocytes and B lymphoma cell lines (15), T lymphocytes (20), breast carcinoma (34), lung carcinoma (35), gastric carcinoma (36), pancreatic carcinoma (37), choriocarcinoma (38), and hepatoma cell lines (39). Based on these studies, it has been suggested that TZDs may represent a new class of drugs for the treatment of lymphomas and carcinomas.
Several factors may account for the discrepant observations between the present study and previous reports. In this study, whereas TZDs do induce a dose-dependent induction of apoptosis, these effects occur at much higher concentrations than those required for activation of transcriptional activity of PPAR␥. Furthermore, the levels of PPAR␥ neither correlate with the ability of TZDs to induce death nor affect the dose range in which cell death is observed. PPAR␥ agonists, such as 15d-PGJ 2 , ciglitazone, and troglitazone, have been shown to possess effects independent of activation of PPAR␥ (10, 11, 40 -43). In keeping with this possibility, many of the previous studies suggesting that TZDs induce cell death utilize concentrations of TZDs several orders of magnitude higher than their K D for PPAR␥.
We have found that TZDs can promote cell survival at doses that induce optimal PPAR␥ transcriptional activity. Taken together, our results suggest that type 2 diabetic patients taking these drugs may not be at risk for further impairment of their immune function. The ability of PPAR␥ to promote cell survival under conditions of growth factor withdrawal might even improve immune cell functions at avascular or necrotic sites such as diabetic ulcers.