Prostaglandin D₂, its metabolite 15-d-PGJ₂, and peroxisome proliferator activated receptor-γ agonists induce apoptosis in transformed, but not normal, human T lineage cells

Abstract

Prostaglandin D₂ (PGD₂) is abundantly produced by mast cells, platelets, and alveolar macrophages and has been proposed as a key immunoregulatory lipid mediator. 15-Deoxy-Δ^12,14-PGJ₂ (15-d-PGJ₂), a key PGD₂ metabolite, is under intense study as a potential anti-inflammatory mediator. Little is known about PGD₂ or the role of 15-d-PGJ₂, if any, in regulating the activities of human T lineage cells. In this report we demonstrate that both PGD₂ and 15-d-PGJ₂ have potent antiproliferative effects, and in fact kill human T lymphocyte lines derived from malignant cells by an apoptotic mechanism. Interestingly, normal human T cells were not similarly affected. Although the T lymphocyte lines express mRNA for the PGD₂ receptor (DP-R), a potent DP receptor agonist, BW245C, did not inhibit the proliferation or viability of the cells, suggesting an alternative mechanism of action. PGD₂ and 15-d-PGJ₂ can bind to the peroxisome proliferator activated receptor-γ (PPAR-γ) which is implicated in lipid metabolism and apoptosis. Exposure to synthetic PPAR-γ ligands (e.g. ciglitazone, troglitazone) mimicked the inhibitory responses of PGD₂ and 15-d-PGJ₂, and induced apoptosis in the transformed T cells consistent with a PPAR-γ-dependent mechanism. These observations suggest that PPAR-γ ligands (which may include PGD₂) provide strong apoptotic signals to transformed, but not normal T lymphocytes. Thus, the efficacy of utilizing PPAR-γ and its ligands as therapeutics for human T cell cancers needs to be further evaluated.

Introduction

Prostaglandin D₂ (PGD₂) is produced by a variety of tissues including bone marrow, skin, brain, and spleen.¹ This lipid mediator is involved in key processes such as hormone release, sleep induction, and olfactory function in the central nervous system.¹ PGD₂ also regulates bronchoconstriction, platelet aggregation, and allergic reactions in the periphery.² Traditionally, PGD₂ has been thought to act on cells by binding the PGD₂ receptor (DP-R). The DP-R, like other prostaglandin receptors, is a G-protein coupled receptor. DP-R is coupled to a Gs protein, and therefore ligand binding induces an increase in cAMP intracellularly. Interestingly, Northern analysis and in situ studies have only localized the human DP-R to the retina and small intestine.³ This, and evidence showing that PGD₂ can bind to other prostaglandin receptors, including the FP receptor¹ suggests that some of the effects of PGD₂ do not necessarily occur through the DP-R.

In vivo, PGD₂ is dehydrated into the J-series of prostaglandins. Some of the most exciting recent work in the field of prostaglandin research has involved the study of 15-deoxy-Δ^12,14-PGJ₂ (herein referred to as 15-d-PGJ₂). 15-d-PGJ₂ is considered to be the key end product metabolite of PGD₂ and is involved in the inhibition of tumour growth, down-regulation of nitric oxide synthase expression in mouse macrophages, and attenuation of tumour necrosis factor-α (TNF-α) release from human monocytes.^4–8 One of its roles seems to be as a negative regulator of inflammation. 15-d-PGJ₂ is reported to exert its effects on cells by binding the peroxisome proliferator activated receptor-γ (PPAR-γ). PPARs are a family of ligand-activated nuclear transcription factors, and after ligand binding PPARs form a heterodimer with the retinoic X receptor, and the complex then binds to PPAR responsive elements (PPREs) in the promoter regions of target genes.^9–12 To date, PPAR-γ has been found primarily in adipose tissue, where it plays a key role in the regulation of adipogenesis. The receptor has also recently been found in epithelial and endothelial cells, macrophages, smooth muscle cells, and mouse T and B cells.^13–21

There has been recent controversy over the link between the action of 15-d-PGJ₂ and PPAR-γ binding. Initial reports focused on the ability of 15-d-PGJ₂ to bind and activate PPAR-γ.^11,22 Recently, it was reported that there may also be PPAR-γ-independent mechanisms of action by 15-d-PGJ₂ including, but not limited to, inhibition of IkB kinase (inhibitory protein that dissociates from NF-κB) and regulation of angiogenesis.^23,24 Although the exact mechanism of 15-d-PGJ₂ action in these systems is unknown, it has been postulated that 15-d-PGJ₂ binds and acts through the DP-R in some cases. In fact Wright²⁵ showed that 15-d-PGJ₂ could bind and activate through the DP-R almost as well as PGD₂ in HEK293 cells transfected with the rat DP-R. Thus the mechanisms of action of PGD₂ and/or 15-d-PGJ₂ are unclear at best, may vary system to system, and must be evaluated further.

As the data concerning the actions of PGD₂ and 15-d-PGJ₂ on human T lineage cells is sparse, we sought to determine the effects of these prostaglandins on cellular functions. We found that PGD₂ and 15-d-PGJ₂ inhibit the growth and viability of human T-cell lines by inducing apoptosis. The mechanism of action is consistent with PPAR-γ mediated effects. In addition, human T cells isolated from normal donors were not inhibited as were the established tumoral T-cell lines.

Materials and methods

Antibodies and reagents

Cells were cultured in RPMI-1640 tissue culture medium (Life Technologies, Grand Island, NY) supplemented with 10% fetal bovine serum (Life Technologies). 3-(4-Dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) and PGD₂ were purchased from Sigma Chemical Co. (St. Louis, MO). Troglitazone was kindly provided by Parke-Davis (Morris Plains, NJ). WY,14643, 15-d-PGJ₂, and ciglitazone were purchased from BioMol (Plymouth Meeting, PA). TriReagent was purchased from Molecular Research Center (Cincinnati, OH).

Human T-cell leukaemias, isolation of human T cells and culture conditions

Jurkat (human acute T-cell leukemia) and CCRF-CEM (human acute lymphoblastic T-cell leukemia) cells were purchased from the American Type Culture Collection (Rockville, MD). J-Jahn (human T-cell lymphoma) cells were kindly provided by Dr Stephen Dewhurst (University of Rochester, Rochester, NY). All three cell lines are long-term established tumoral cell lines. Human buffy coats were purchased from the American Red Cross. Human peripheral blood mononuclear cells (PBMC) were isolated using Ficoll according to the manufacturer's instructions. Human CD4 positive cells were positively selected with CD4 Dynabeads and Detachabeads (Dynal, Lake Success, NY) according to the manufacturer's instructions. Cell purity was assessed by flow cytometry with fluoroscein isothiocyanate (FITC)-labelled anti-human CD4 antibody (Pharmingen, San Diego, CA). Cells were greater than 98% CD4 positive. For each experiment, three different T-cell donors were used. T cells were plated into 96-well plates at a density of 5 × 10⁴ cells per well for 48 hr. At the initiation of culture, T cells were treated with various concentrations of prostaglandins in RPMI/0% serum. Normal human T cells were additionally incubated with either phorbol 12-myristate 13-acetate (PMA; 50 ng/ml) and ionomycin (500 ng/ml) or α-CD3 (1 µg/ml) and α-CD28 (250 ng/ml). Serum was added back to the cultures after 8 hr to a concentration of 10%. Viability and proliferation were then measured after 48 hr. Cell viability was measured by MTT (Sigma). MTT was added for the final 4 hr of culture. After incubation, dimethyl sulphoxide (DMSO; Sigma) was added to each well to dissolve the formazan product. Absorbance at 510 nm was recorded using a Titretek Multiskan enzyme-linked immunosorbent assay (ELISA) plate reader (Flow Laboratories, McLean, VA). Proliferation was measured by ³H-thymidine incorporation (Amersham, Arlington Heights, IL). ³H-thymidine was added at a final concentration of 1 µCi/well for the final 24 hr of culture. Plates were then harvested with a Micromate 196 cell harvester (Packard Co., Meriden, CT) and incorporation of ³H-thymidine determined with a Matrix 96 direct beta counter (Packard).

RNA isolation and reverse transcription–polymerase chain reaction (RT-PCR)

RNA was isolated by TriReagent according to the manufacturer's instructions. Two µg of RNA were added to 1 µg of oligo d(T) (Pharmacia, Piscataway, NJ) in diethyl pyrocarbonate (DEPC)-treated water and incubated at 60° for 5 min and 4° for 3 min One mm each of the four deoxynucleic acids (dNTPs) (Pharmacia), 8 µl 5× cDNA synthesis buffer (Gibco BRL, Grand Island, NY) and 400 units of Maloney murine leukaemia virus (MMLV) reverse transcriptase (Gibco BRL) were added and the reactions incubated at 37° for 1 hr, 95° for 5 min, and stored at 4°. For each cDNA synthesis reaction, a reaction was performed without RT and used as a negative control in the PCR. An aliquot of cDNA synthesis reaction was added to 5 µl 10× PCR buffer (Boehringer Mannheim Biochemica, Indianopolis, IN), 1 mm dinucleotide triphosphates (dNTPs), oligonucleotide primers specific for PPAR-γ, DP-R, or the housekeeping gene β-actin at a concentration of 1 µm, 2·5 units of Taq DNA polymerase (Boehringer), and water to a final volume of 50 µl. Primer sequences were as follows, β-actin: 5′-GTGGGGCGCCCCAGGCACCA and 3′-CTCCTTAATGTCACGCACGCACGATTTC; DP-R: 5′-TCTCAAAGAGGGGTGTGACC and 3′-AGACTCCGGTTCTGAGCGTA; PPAR-γ: 5′-TCTGGCCCACCAACTTTGGG and 3′-CTTCACAAGCATGAACTCCA. PCR samples were run for 35 cycles (94° for 50 s, 60° for 50 s, 72° for 90 s) with a final extension at 72° for 7 min in a DNA thermal cycler. Samples were analysed by gel electrophoresis on 1% agarose gels and stained with ethidium bromide. The identity of the RT–PCR products was confirmed by DNA restriction enzyme analysis. A normal strain of human lung fibroblast (L828) was used as a positive control for PPAR-γ. HMC-1, a human mast cell line, was used as a positive control for the DP-R.

Immunocytochemistry

CCRF-CEM, J-Jahn, Jurkat, and normal human peripheral blood T cells were stained with a monoclonal antihuman PPAR-γ antibody (Santa Cruz Biotechnology, Santa Cruz, CA). For normal human T-cell stimulations, cells were treated with α-CD3 antibody at 1 µg/ml and α-CD28 antibody at 250 ng/ml for 24 and 48 hr prior to staining. Briefly, cells were cytospun, treated with 3% hydrogen peroxide to quench intrinsic peroxidase activity, and blocked with normal horse serum (5%). Slides were then treated with monoclonal mouse anti-human PPAR-γ antibody at 10 µg/ml (Santa Cruz Biotechnology) or monoclonal immunoglobulin G1 (mIgG1) isotype control at 10 µg/ml (Caltag, Burlingame, CA). After an overnight incubation, biotin-labelled horse anti-mouse IgG (Vector Laboratories, Burlingame, CA) secondary antibody was added at a 1 : 200 dilution, followed by streptavidin–horseradish peroxidase (HRP; Jackson Laboratories, West Grove, PA) at 1 : 1000. Slides were developed with AEC Reagent (Zymed Laboratories, San Francisco, CA) and counterstained with haematoxylin (Sigma). Cells were viewed and photographed under bright field.

Detection of apoptosis

For apoptosis detection using annexin-V–FITC (Pharmingen), T cells were treated with DMSO (diluent), 15-d-PGJ₂, troglitazone, ciglitazone, or WY,14643. After incubation, cells were washed two times in cold PBS and stained with annexin-V–FITC according to the manufacturer's instructions and as described in Harris et al.¹⁹ Cells were analysed on a Becton-Dickinson FACSscan, gating on the live cell population. For the detection of apoptosis using the TdT-mediated biotin–dUTP nick-end labelling (TUNEL) method, cells were treated as above, but spread onto slides and stained with TdT (Gibco-BRL) and Biotin-16-dUTP (Boehringer-Mannheim) as previously described.¹⁹ Cells were viewed and photographed under bright field.

Results

PGD₂ and its metabolite 15-d-PGJ₂ inhibit the proliferation of transformed, but not normal, human T cells

To determine if PGD₂ or its metabolite 15-d-PGJ₂ affected the proliferation of human T cells, the transformed T-cell lines CCRF-CEM, J-Jahn, and Jurkat were treated with the prostaglandins for 48 hr at the indicated concentrations. As a control, cells were also treated with vehicle (DMSO). Figure 1 clearly shows that PGD₂ and 15-d-PGJ₂ inhibited the proliferative response of CCRF-CEM, J-Jahn, and Jurkat T cells in a dose-dependent manner. In all cases, there was complete inhibition of proliferation by a concentration of 10 µm. Interestingly, 15-d-PGJ₂ had a more potent antiproliferative effect than PGD₂ on the T-cell lines. The EC₅₀ for 15-d-PGJ₂ is < 1 µm, while the EC₅₀ for PGD₂ is closer to 5 µm. We next evaluated the effects of these prostaglandins on normal T cells isolated from healthy donors. T cells isolated from PBMC were cultured with PGD₂ or 15-d-PGJ₂ in the absence or presence of stimulation. Unstimulated human T cells do not proliferate in vitro, so no conclusions can be reached with regard to these responses. T cells stimulated with α-CD3 and α-CD28 antibodies are shown in Fig. 1(d). The T cells isolated from normal subjects did not show any inhibition of proliferation in response to PGD₂ or 15-d-PGJ₂. Similar results were found when cells were stimulated with PMA and ionomycin (data not shown).

Figure 1.

PGD₂ and its metabolite 15-d-PGJ₂ inhibit the proliferation of transformed but not normal human T cells. CCRF-CEM (a), J-Jahn (b), Jurkat (c), or normal human T cells (d) were incubated for 48 hr in the presence of either PGD₂ or 15-d-PGJ₂. Non-transformed cells were also treated with anti-CD3 and anti-CD28 as described in Materials and Methods. Cells were pulsed with 1 µCi/µl of ³H-thymidine for the final 8 hr of culture. Data is graphed as percent DMSO control response. Each individual experiment was performed in triplicate, and repeated three times with similar results each time. Asterisks indicate P < 0·05 based on a two-tailed Student's t-test.

PGD₂ and 15-d-PGJ₂ inhibit the viability of T lymphocyte lines

There are many possible causes of an inhibition of proliferation in T lymphocytes, including arrest at a certain stage of the cell cycle, anergy, or cell death. To determine if the antiproliferative activities of the prostaglandins were due to induction of cell death, MTT assays were performed as described in the Materials and Methods. Figure 2 clearly shows that both PGD₂ and 15-d-PGJ₂ are inhibiting the viability of CCRF-CEM, J-Jahn, and Jurkat T cells, explaining their lack of proliferative potential. Again, 15-d-PGJ₂ was much more effective at inducing cell death than PGD₂, with an approximate log difference in the EC₅₀ values. Also, it was determined if 15-d-PGJ₂ or PGD₂ had any effect on the viability of T lymphocytes isolated from normal donors. Concurrent with the findings in Fig. 1(d), the prostaglandins did not affect the viability of stimulated normal T cells (Fig. 2d). No inhibition of viability was noted when unstimulated normal T lymphocytes were treated with PGD₂ or 15-d-PGJ₂ (data not shown). These data demonstrate that PGD₂ and its metabolite 15-d-PGJ₂ inhibit the viability of human T-cell lines, but not normal human T cells.

Figure 2.

PGD₂ and its metabolite 15-d-PGJ₂ Inhibit the viability of transformed but not normal human T cells. CCRF-CEM (a), J-Jahn (b), Jurkat (c), or normal human T cells (d) were incubated for 48 hr in the presence of either PGD₂ or 15-d-PGJ₂. Non-transformed T cells were additionally treated with anti-CD3 and anti-CD28 as described in Materials and Methods. Cells were treated with MTT for the final 8 hr of culture. Data is graphed as percentage DMSO control response. Each individual experiment was performed in triplicate, and repeated three times with similar results each time. Asterisks indicate P < 0·05 in a two-tailed Student's t-test.

PGD₂ and 15-d-PGJ₂ induce apoptosis in human T-cell lines

To determine if the observed mechanism of cell death is apoptotic or necrotic in nature, two different assays were employed. During the early phases of apoptosis, phosphatidylserine is translocated to the outer leaflet of the plasma membrane. The protein annexin-V binds phosphatidylserine with high affinity and a fluorescent version has been widely used to detect phosphatidylserine on the surface of apoptotic cells.¹⁹ Figure 3(a, b, c) shows that treatment of Jurkat T cells with PGD₂ or 15-d-PGJ₂ for 18 hr results in an increase in annexin-V binding in comparison to control treated cells supporting an apoptotic mechanism of cell death. To confirm that PGD₂ and 15-d-PGJ₂ were inducing apoptosis, TUNEL staining was also performed. Figure 3(d, e, f) demonstrates that 15-d-PGJ₂ and PGD₂ treatment of Jurkat cells results in positive TUNEL staining after 24 hr, indicative of apoptosis. Therefore, the inhibition of proliferative response (Fig. 1) and viability (Fig. 2) by PGD₂ and one of its metabolites 15-d-PGJ₂ in Jurkat T cells is a result of an induction of apoptosis. The general caspase inhibitors Z-VAD-fmk and BOD-D-fmk did not inhibit the prostaglandin-mediated cell death suggesting a caspase-independent mechanism of apoptosis (data not shown).

Figure 3.

PGD₂ and 15-d-PGJ₂ induce apoptosis in Jurkat cells. DMSO treated (a, d), 15-d-PGJ₂ treated (b, e) at 10 µm, or PGD₂ treated (c, f) at 25 µm were stained for annexin–FITC (a–c) or TUNEL (d–e). For annexin, cells were incubated for 18 hr with the indicated concentration, then stained according to manufacturer's instructions, and analyzed on a BD FACSscan. TUNEL staining was performed according to manufacturers directions after 24-h incubation with prostaglandins. Cells were photographed under bright field.

Transformed but not normal human T lymphocytes express DP-R mRNA

There are several possible mechanisms whereby both PGD₂ and 15-d-PGJ₂ could be acting on transformed, but not normal T lymphocytes to induce apoptosis. 15-d-PGJ₂ can bind and activate via the PGD₂ receptor (DP-R). Thus PGD₂ and 15-d-PGJ₂ both could bind DP-R, and induce apoptosis of the transformed T cells. Another possibility is that PGD₂ is being broken down into 15-d-PGJ₂ in culture, and binding to PPAR-γ (the nuclear hormone receptor that 15-d-PGJ₂ exerts most of its effects on cells through), and this is inducing T-cell apoptosis. Additionally, because PGD₂, like 15-d-PGJ₂, does bind PPAR-γ, it could be directly binding PPAR-γ and initiating apoptosis. The induction of apoptosis could also involve a combination of these effects. Therefore, the general mechanism most likely involves either a DP-R- or PPAR-γ-mediated pathway, or a combination of both. To start to tease apart the actions of these prostaglandins, we first tested the DP-R hypothesis. As it has never been shown whether or not malignant or normal human T cells express the DP-R, we first looked by RT-PCR for mRNA expression. Figure 4 clearly shows that transformed human T cells do express DP-R mRNA. Therefore it is possible that PGD₂ and 15-d-PGJ₂ are acting through the DP-R. Interestingly, peripheral blood T cells isolated from normal donors did not express mRNA for the DP-R, possibly explaining the lack of effect of the prostaglandins on these cells.

Figure 4.

Malignant human T cell lines, but not normal T cells, express mRNA for the DP-R. Total RNA was isolated from cells and reverse transcribed into cDNA. The cDNA was amplified with primers specific for β-actin (as a control) or DP-R. HMC-1 (human mast cell line) was used as a positive control for DP-R. Reverse transcriptase (−) samples were negative in all cases.

DP-R activation is not the mechanism for PGD₂ inhibition of proliferation and viability

To directly determine whether or not the DP-R is involved in the inhibition of proliferation and viability in T cells, we employed a potent DP-R agonist, BW245C.^26,27 CCRF-CEM, J-Jahn, and Jurkat T cells were treated with either PGD₂ or BW245C for 48 hr, and proliferation and MTT assays performed as described in Materials and Methods. As Fig. 5 shows, the DP-R agonist does not inhibit the proliferation or viability of any of the three transformed T-cell lines tested. In contrast, PGD₂ inhibits the proliferation and viability of all the T-cell lines. To further rule out a role for DP-R in affecting the T lineage cells, a DP-R antagonist, AH6809, was used in conjunction with PGD₂ to determine if the antagonist could block the effects of PGD₂. AH6809 had no effect on the inhibition of proliferation and viability induced by PGD₂ (data not shown). These studies suggest that the mechanism by which 15-d-PGJ₂ and PGD₂ are killing T lymphocyte lines is not via the DP-R.

Figure 5.

PGD₂ does not inhibit proliferation and viability of T cells through the DP-R. CCRF-CEM (a), J-Jahn (b), or Jurkat (c) cells were cultured for 48 hr with either PGD₂ or BW245C. For proliferation assays (a(i)–c(i)) cells were pulsed with 1 µCi/µl of ³H-thymidine for the final 8 h of incubation. For MTT (b(i)–c(i)) cells were treated with MTT for the final 8 hr of culture. Data is graphed as percentage DMSO control response. Each individual experiment was performed in triplicate, and repeated three times with similar results each time. Asterisks indicate P < 0·05 based on a two-tailed Student's t-test.

Malignant human T-cell lines express mRNA and protein for PPAR-γ

To test the hypothesis that PGD₂ and 15-d-PGJ₂ could be acting through PPAR-γ to exert their effects, we first determined if normal or malignant human T cells express PPAR-γ. Total RNA was isolated from normal or malignant human T cells, and reverse transcribed as described in Materials and Methods. cDNA was then subject to PCR with control primers or primers specific for human PPAR-γ. Figure 6 shows that CCRF-CEM, J-Jahn, and Jurkat cells all contain mRNA for PPAR-γ. However, no mRNA for PPAR-γ was present in T cells obtained from healthy donors. The presence of PPAR-γ in human T-cell lines was further examined by immunocytochemistry. Figure 7 shows that the transformed T-cell lines J-Jahn and Jurkat contain PPAR-γ protein, confirming the mRNA data. The staining pattern of the cells is both cytoplasmic and nuclear. Normal peripheral blood T cells, however, do not express detectable protein for PPAR-γ. When normal T lymphocytes were stimulated with α-CD3 and α-CD28 however, there was a slight induction of PPAR-γ protein expression after 24 and 48 hr of stimulation (Fig. 8), although less than that found in transformed T-cell lines. Therefore, as the transformed human T-cell lines contain both mRNA and protein for PPAR-γ, the possibility exists that the inhibitory effects of these prostaglandins on T cells is mediated through this receptor.

Figure 6.

Malignant human T-cell lines, but not normal T cells, express PPAR-γ mRNA. Total RNA was isolated and reverse transcribed into cDNA as discussed in Materials and Methods. The cDNA was amplified with control primers (β-actin) or primers specific for PPAR-γ. L828 human lung fibroblasts were used as a positive control for PPAR-γ. Reverse transcriptase (−) samples were negative in all cases.

Figure 7.

Malignant human T-cell lines express PPAR-γ protein. Malignant human T cells were spread onto slides and stained for PPAR-γ as described in Materials and Methods. Cells were then viewed and photographed under bright field. mIgG1 was used as an isotype control.

Figure 8.

Stimulation of peripheral human T cells from normal donors modestly induces PPAR-γ expression. CD4 positive T lymphocytes were isolated from buffy coats of normal donors as described in Materials and Methods. Cells were stimulated with or without anti-CD3 and anti-CD28 for 24 and 48 hr. Cells were cytospun, stained for PPAR-γ expression, then photographed under bright field. Unstimulated cells did not express PPAR-γ at any time point analysed.

PPAR-γ agonists mimic the PGD₂ and 15-d-PGJ₂ inhibition of proliferation and viability

15-d-PGJ₂ is known to bind and activate PPAR-γ. There is also evidence that 15-d-PGJ₂ may work through a PPAR-γ-independent mechanism. Therefore, we employed PPAR-γ agonists to determine if the inhibition of proliferation and viability in transformed T cells are potentially mediated through PPAR-γ. The PPAR-γ agonists ciglitazone and troglitazone specifically bind and activate PPAR-γ. We also included the negative control compound WY,14643 that specifically binds the PPAR-α receptor to further show specificity of response. Proliferation and MTT viability assays were performed on all three cell lines as previously described. The EC₅₀ values presented in Table 1 show that the PPAR-γ agonists ciglitazone and troglitazone exert similar effects on transformed T cells as 15-d-PGJ₂. The PPAR-α agonist, WY,14643, does not inhibit these T cells, supporting the concept that PGD₂, 15-d-PGJ₂, ciglitazone, and troglitazone act via PPAR-γ. In addition, to verify that freshly isolated human T cells were not affected by PPAR-γ agonists, proliferation and viability assays were performed as above. Neither the PPAR-γ agonists ciglitazone or troglitazone inhibited the proliferation or viability of primary human T lymphocytes (data not shown).

Table 1.

PPAR-γ agonists inhibit the proliferation and viability of malignant human T-cell lines

	EC50 (μm)
Treatment	CCRF-CEM	J-Jahn	Jurkat
(a) Proliferation
15-d-PGJ2	1	1	2
Ciglitazone	4	5	3
Troglitazone	6	6	4
WY, 14643	6	6	4
(b) Viability
15-d-PGJ2	1	1	2
Ciglitazone	4	5	3
Troglitazone	6	6	4
WY, 14643	6	6	4

CCRF, J-Jahn, or Jurkat cells were incubated for 48 h in the presence of PPAR-γ agonists or a PPAR-α agonist as a control. EC₅₀ values are shown for T cell proliferation (a) and viability (b) assays. In the proliferation assays, cells were pulsed with 1 µCi/µl ³H-thymidine for the final 8 h of culture. For MTT assays, cells were cultured with MTT for the final 8 h of incubation. Each experiment was performed in triplicate and repeated three times with similar results each time.

PPAR-γ agonists induce apoptosis in transformed human T lymphocyte lines

To verify that the induction of apoptosis is consistent with a PPAR-γ dependant mechanism, annexin V staining (Fig. 9a) and TUNEL assays (Fig. 9b) were performed as described in Materials and Methods. Jurkat cells were in this case treated with 15-d-PGJ₂, the PPAR-γ agonists ciglitazone and troglitazone, or the PPAR-α agonist WY,14643. Figure 9 shows that the PPAR-γ, but not PPAR-α, agonists induce apoptosis in Jurkat human T lymphocytes in a manner similar to that of PGD₂ and 15-d-PGJ₂. Similar results were found with CCRF-CEM cells (data not shown). Therefore, PGD₂ and 15-d-PGJ₂ kill malignant T cells via inducing apoptosis, most likely in a PPAR-γ-dependent manner.

Figure 9.

PPAR-γ agonists induce apoptosis in Jurkat cells. Jurkat cells were treated as described in Materials and Methods and stained for annexin–FITC (a) or TUNEL (b). For annexin staining, cells were incubated for 18 h with the treatments shown above, stained according to the manufacturer's instructions, and analysed by flow cytometry. For TUNEL staining, cells were treated for 24 h prior to staining, then viewed and photographed under bright field.

Discussion

This is the first report demonstrating that PGD₂ and 15-d-PGJ₂ inhibit the proliferation and viability of malignant human T-cell lines. Although it was possible that the effects of these prostaglandins were mediated through the DP-R, as both PGD₂ and 15-d-PGJ₂ can bind this receptor²⁵ we show that the more likely mechanism is via PPAR-γ (Table 1 and Fig. 9). There are several possible ways to explain the action of PGD₂. First, it could be an indirect activation. PGD₂ could be broken down in vitro to 15-d-PGJ₂, and the metabolite could be acting on PPAR-γ. It is possible that 15-d-PGJ₂ could enter the cell via a carrier transporter protein similar to ones described by Narumiya's group^28,29 for other cyclopentenone prostaglandins (e.g. Δ-12 PGJ₂). Once inside the cell, an additional mechanism allows transport of the cyclopentenone prostaglandins into the nucleus where they could act directly on nuclear PPAR-γ. Although it has never been directly shown that 15-d-PGJ₂ can use these transporters, its similar structure to the prostaglandins evaluated make this mechanism possible. The need for PGD₂ to be metabolized into 15-d-PGJ₂ prior to uptake may explain why higher concentrations of PGD₂ are needed to induce the same response as 15-d-PGJ₂. Another possibility is that there could be direct binding of PGD₂ to PPAR-γ followed by activation of the transcription factor. Several studies have shown that PGD₂ can directly bind PPAR-γ.^11,12,22 It is possible that PGD₂ could enter the cell via a prostaglandin transporter molecule. Although the presence of such a molecule has never been documented in human T cells, it has in other cell types such as neuronal cells.³⁰ Further study to identify the exact mechanism of PGD₂ activity is needed to clarify these possibilities.

Although no-one knows what the in vivo levels of 15-d-PGJ₂ are, PGD₂ has been found in the micromolar range in several systems.¹ Because cells in close proximity to T cells like mast cells and macrophages produce PGD₂, and thus 15-d-PGJ₂, it is likely that high enough concentrations will be available to affect T-cell function. In addition, Narumiya's group has shown that when cyclopentenone prostaglandins enter cells, they are concentrated approximately 20-fold. For example, when 10 µm of a J-series prostaglandin was added to murine leukaemia cells, 200 µm was measured inside the cell.^28,29 Therefore, the concentrations of 15-d-PGJ₂ used in these studies could be found in, and thus regulate T-cell function, in vivo.

The expression of PPAR-γ mRNA and protein in transformed human T-cell lines is a novel finding. Interestingly, immunocytochemistry (Fig. 7) shows that PPAR-γ is expressed in the cytoplasm as well as in the nucleus. This unusual localization pattern of PPAR-γ expression has been noted in other cell types such as mouse T cells and synoviocytes.^19,31 It is possible that PPAR-γ is transported into the nucleus during various stages of the cell cycle, or after certain stimuli. For example, treatment with PPAR-γ ligands such as 15-d-PGJ₂ or synthetic agonists may induce translocation of PPAR-γ to the nucleus. Once in the nucleus, the receptor could bind to DNA and activate transcription of target genes.³² This compartmentalization of PPAR-γ may give an added regulatory mechanism governing its responsiveness. Certainly other transcription factors, such as nuclear factor κB (NFκB) are also regulated in a similar fashion, lending credence to this as a possible pathway of PPAR-γ activation in T cells.

T lymphocytes from normal human donors were not inhibited in their proliferation and viability by PPAR-γ agonists, and did not express PPAR-γ mRNA nor protein. When the T cells were stimulated with α-CD3 and α-CD28, however, a modest induction of PPAR-γ protein expression was noted by immunocytochemistry after 24 and 48 hr of stimulation. Even though the stimulation up-regulated PPAR-γ expression, there was no accompanying inhibition of growth of the cells. It is entirely possible that by the time PPAR-γ was present in the cells, the prostaglandins (PGD₂ and 15-d-PGJ₂) had already metabolized into inactive metabolites, as they both have a relatively short half-life in vitro. Therefore, they were not able to act on the newly produced PPAR-γ to exert inhibitory effects. Alternatively, after PPAR-γ was produced, it could have been subsequently inactivated because of phosphorylation by map kinases. Map kinase-mediated phosphorylation of PPAR-γ leads to inactivity of the transcription factor.³³ Because stimulation in T lymphocytes through the T-cell receptor (mimicked by α-CD3 and α-CD28 binding) results in a map kinase cascade, it is likely that any PPAR-γ present or made during the stimulation would be inactivated. Additionally, the subcellular staining pattern of PPAR-γ may be involved in its activity or lack thereof. In activated human T cells, PPAR-γ stains mostly in the cytoplasm, while there is both nuclear and cytoplasmic staining in transformed T cells. This difference in expression patterns may play a role in the functionality of PGD₂ and 15-d-PGJ₂ in the cells. These explanations are, of course, not mutually exclusive.

The data showing that PGD₂ and 15-d-PGJ₂ inhibit the proliferation and viability of transformed human T cells but not normal human T cells are intriguing (Figs 1 and 2). The lack of PPAR-γ mRNA and protein in unstimulated normal human T lymphocytes suggests a reason for the differences in activity of the compounds (Figs 6 and 7). Perhaps during tumorigenesis of T cells, PPAR-γ is dysregulated, resulting in high expression of this transcription factor. This hypothesis is supported by the fact that PPAR-γ is highly expressed in other malignancies. The reason for increased PPAR-γ expression may be random, or may induce unknown survival signals for the tumour. Another scenario is that populations of T cells are heterogeneous in terms of PPAR-γ expression. Individuals with T-cell cancers may have expressed PPAR-γ in at least some of their normal T cells, and by an unidentified mechanism, some of these PPAR-γ expressing normal T cells become transformed. Certainly more extensive analysis of a larger population is needed for the generalization of these results, but the implications for possible therapies for T-cell cancers are evident. The ability to selectively induce the apoptosis of T-cell cancers, and not healthy T cells would be extremely beneficial as an aid to current therapies as T-cell malignancies are notoriously refractory to treatment. Therefore, targeting this receptor may prove useful for drug therapies in the future.