Peroxisome Proliferator-Activated Receptor γ Overexpression Suppresses Growth and Induces Apoptosis in Human Multiple Myeloma Cells

Multiple myeloma is a neoplasm of differentiated plasma cells that is largely incurable, with a median survival duration of 3 to 5 years (1). It constitutes ∼10% of hematologic cancers and ranks as the second most frequent hematologic malignancy in the United States (2, 3). Current therapies for the disease include chemotherapy with or without stem cell transplantation, glucocorticosteroids, thalidomide, the proteosome inhibitor bortezomib, and combinations of these agents (1, 2). However, most of these treatments are still not curative; hence, newer treatments approaches are needed (3).

Peroxisome proliferator-activated receptors (PPAR) are ligand-activated transcription factors belonging to the nuclear hormone receptor family (4, 5). Three different isoforms have been identified: PPARα, PPARβ/δ, and PPARγ. On ligand binding, they form a heterodimer with 9-cis-retinoic acid retinoid X receptors. This complex binds to PPAR response elements (PPRE) located in the promoter regions of target genes (6, 7). A key role for PPARγ is in adipogenesis, where PPARγ activation leads to the generation of adipocyte-specific genes (8). PPARγ also participates in glucose homeostasis, cell cycle regulation, inflammation, atherosclerosis, apoptosis, and carcinogenesis (9). PPARγ can be activated by the naturally occurring prostaglandin D₂ metabolite 15-deoxy-Δ^12,14 prostaglandin J₂ (10, 11) as well as by lysophosphatidic acid (12) and nitrolinoleic acid (13). PPARγ is also activated by synthetic ligands including the thiazolidinedione class of clinically used anti-type II diabetic drugs (14).

PPARγ and PPARγ ligands have significant anti-inflammatory effects in immune cells (15). For example, macrophages treated with PPARγ ligands are inhibited in terms of activation and production of inflammatory cytokines (16–18). In T lymphocytes, activation of PPARγ inhibits proliferation and reduces their production of IFN-γ and tumor necrosis factor-α, as well as interleukin (IL)-2, by interfering with the transcription factor nuclear factor for activated T cells (19). In addition, our laboratory has shown that mouse and human T cells express PPARγ and treatment with PPARγ ligands induces T-cell apoptosis (20, 21). PPARγ also regulates B lymphocyte function. In PPARγ-haploinsufficient mice, B lymphocytes exhibit increased proliferation and survival and enhanced antigen-specific immune response and spontaneous nuclear factor-κB activation (15, 22). Furthermore, we and others have shown that PPARγ is expressed in human normal and malignant B cells and exposure to certain PPARγ ligands results in cell death (23–28).

The role of PPARγ in the onset and development of cancer has been the recent focus of much attention. Cancer cells, including hematologic cancers, display enhanced proliferation, in part, due to their failure to undergo terminal differentiation. The induction of differentiation and apoptosis by ligands of the nuclear hormone receptor family is a novel approach for cancer therapy; an example of this is the use of retinoic acid for the treatment of acute promyelocytic leukemia and other cancers (29–31). These studies have brought attention to the role of PPARγ in the development and progression of cancer. The efficacy of PPARγ ligands as potential anticancer drug therapies has been explored in several cancer models, including colon, breast, prostrate, and lung (32). Recently, we and others have shown that activation of PPARγ inhibits growth and induces apoptosis in hematologic cancers, including multiple myeloma (26, 33, 34).

The tumor microenvironment contributes significantly to multiple myeloma resistance to chemotherapy (35) via the release of cytokines, chemokines, and growth factors (e.g., IL-6) or by direct interaction between tumor cells and stromal cells (36, 37). Some of the effects of PPARγ ligands in the tumor microenvironment have been investigated in multiple myeloma (38, 39). In these studies, PPARγ ligands abrogated IL-6-dependent multiple myeloma cell growth by inhibiting the transactivation of IL-6/STAT3 (38). Moreover, PPARγ ligands inhibited adhesion of multiple myeloma cells to bone marrow stromal cells (BMSC) in a PPARγ-dependent fashion through inhibition of the transcription factors nuclear factor-κB and C/EBPβ (39).

A major limitation of these studies is that only PPARγ ligands were used to study the role of PPARγ in cancer cell biology. Because PPARγ ligands have been shown to have a variety of PPARγ-independent effects, the role of PPARγ remains unclear (40, 41). Indeed, research from our group and others exploring the viability of PPARγ as a therapeutic target in B-cell malignancies have used PPARγ ligands in their experiment design. To unequivocally determine the central role of PPARγ in the biology of multiple myeloma, we studied the functional significance of PPARγ overexpression in human multiple myeloma cells using a lentiviral PPARγ construct that we developed.

Reagents and antibodies. Ciglitazone was purchased from Biomol. 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), DMSO, and anti-Flag M2 monoclonal antibody peroxidase conjugate were from Sigma. Antibodies against active caspase-3 and poly(ADP-ribose) polymerase (PARP) were purchased from BD Biosciences. A rabbit polyclonal anti-myeloid cell leukemia-1 (Mcl-1) antibody was purchased from Santa Cruz Biotechnology and the anti-X-linked inhibitor of apoptosis protein (XIAP) antibody was from R&D Systems. Total actin antibody was from Oncogene. The glyceraldehyde 3-phosphate dehydrogenase antibody and the caspase inhibitor VI (ZVAD-FMK) were purchased from Calbiochem.

Construction of lentiviral vectors. The pcDNA-Flag-PPARγ1-WT and the pcDNA-Flag-PPARγ1-L466A/E469A-DN human PPARγ1 plasmid was a kind gift from VKK Chatterjee (University of Cambridge). The cDNA encoding PPARγ1 protein tagged at its NH₂ terminus with a Flag epitope was amplified using Pfu DNA polymerase. The restriction enzymes NheI and NotI were introduced into the 5′ and 3′ PCR primers, respectively. The PCR product was then digested with NheI and NotI and subcloned into a NheI-NotI digested pCDH1-MCS1-EF1-copGFP vector (System Bioscience). These vectors were named LV-PPARγ and LV-PPARγ-DN and the vector without PPARγ cDNA was termed LV-empty. The insert sequences in all lentiviral vectors were verified by DNA sequencing.

Lentiviral vector production. Human embryonic kidney 293FT cells (Invitrogen) were grown to 50% to 70% confluency in DMEM (Life Technologies) supplemented with 10% fetal bovine serum in T-175 flasks. Subsequently, the VSVG pseudotyped HIV vector was generated by cotransfecting with 5 μg envelope vector (pCMV-VSVG), 14 μg transfer vector, and 14 μg packaging vector pCMV-Δ89.2 using LipofectAMINE LTX (Invitrogen). Cells were split into two T-175 flasks 6 h post-transfection. Supernatants were collected 48 and 72 h post-transfection. Virus was harvested by ultracentrifugation at 50,000 × g for 2 h at 4°C using a Beckman SW 28 rotor. The concentrated virus stocks were titered on RPMI 8226 cells based on green fluorescent protein (GFP) expression.

Cells and culture conditions. Human multiple myeloma RPMI 8226 cells were a kind gift from Dr. Robert G. Fenton (University of Maryland; ref. 26). The human multiple myeloma U266 cells and the BMSC HS-5 were purchased from the American Type Culture Collection. Both multiple myeloma cell lines are IL-6 independent. All cell lines were cultured in RPMI 1640 (Life Technologies) supplemented with 10% fetal bovine serum, 5 × 10⁻⁵ mol/L β-mercaptoethanol (Eastman Kodak), 10 mmol/L HEPES (US Biochemical), 2 mmol/L l-glutamine (Life Technologies), and 50 μg/mL gentamicin (Life Technologies). Human embryonic 293FT cells and human fetal lung fibroblasts (HFL-1) were purchased from the American Type Culture Collection (Manassas, VA) and grown in DMEM (Life Technologies) supplemented with 10% fetal bovine serum.

Lentiviral infections and microscopy. Human multiple myeloma cells were plated at a density of 1 × 10⁶ per well in a 12-well plate and infected with LV-empty or LV-PPARγ or LV-PPARγ-DN at multiplicities of infection (MOI) of 0.5 to 1. Twenty-four hours post-infection, the growth medium was replaced. For morphologic analysis, RPMI 8226 and U266 cells were examined by phase-contrast microscopy using an inverted microscope (Olympus IX81) and images were captured for subsequent analysis using ImagePro 4.5.1 (Media Cybernetics). Lentiviral transduced cells in cultures were identified based on the coexpressed GFP gene reporter.

PPARγ gene reporter analysis. Human embryonic kidney 293FT cells were plated at a density of 1.5 × 10⁵ per well in a 12-well plate. The following day, cells were cotransfected using LipofectAMINE LTX (Invitrogen). Briefly, 0.5 μg PPRE-luciferase reporter plasmid containing three copies of the ACO-PPRE (PPRE) from rat acyl-CoA oxidase (a gift from Dr. B. Seed, Massachusetts General Hospital) was cotransfected with 0.5 μg of either the LV-empty vector, the LV-PPARγ vector, or the LV-PPARγ-DN vector. Forty-eight hours after transfection, luciferase activity was assayed using the Promega Luciferase Assay System. Relative light units were determined with a Lumicount microplate luminometer (Packard Instrument).

BMSC and multiple myeloma coculture conditions. BMSC (HS-5) were plated in a 24-well plate at a density of 4 × 10⁴ per well and allowed to settle overnight. RPMI 8226 cells were infected with either LV-empty or LV-PPARγ (MOI = 1) for 16 h, after which viral supernatants were removed and new medium was added to the cells. Some of the LV-infected multiple myeloma cells were added to the BMSC and incubated at indicated time points. Seventy-two hours after incubation, cells were visualized microscopically and cell supernatants were collected to assess IL-6 secretion.

For experiments where BMSC were infected with the lentiviral constructs, BMSC (HS-5) were plated in a 96-well plate at a density of 1.5 × 10⁴ per well and incubated overnight. The next day, HS-5 were infected with either LV-empty or LV-PPARγ (MOI = 1) with polybrene (6 μg/mL) for 24 h. Viral supernatants were then removed and the cells were incubated in the presence or absence of uninfected multiple myeloma RPMI 8226 cells (4 × 10⁵ per well). Some cells were treated with the PPARγ ligand ciglitazone (5 μmol/L). Forty-eight hours later, cell supernatants were collected to assess IL-6 secretion.

Annexin V and 7-amino-actinomycin D staining. Multiple myeloma cells infected with LV-empty and LV-PPARγ were collected 48 to 96 h after infection. Cells were washed twice in PBS and double-labeled with Annexin V-APC and 7-amino-actinomycin D (7-AAD) according to the manufacturer's protocol (BD PharMingen). Control populations consisted of unstained cells and cells stained with only Annexin V or 7-AAD. Cells were analyzed by flow cytometry using a Beckman Coulter flow cytometer. The fluorescence emission of 10,000 cells per treatment was collected in each channel.

Viability and proliferation assays. Cells were left untreated or were infected with LV-empty or LV-PPARγ for 72 h at a density of 4 × 10⁴ per well of a 96-well flat-bottomed microtiter plate. MTT assay was done to assess cell viability. Cell viability was also confirmed by trypan blue dye exclusion. MTT is taken up by cells via mitochondrial membrane potential and then reduced to a formazan residue by intracellular NAD(P)H-oxidoreductases in living cells with functional mitochondria (42). Ten microliters per well of a 5 mg/mL MTT (in 1× PBS) were added for the last 4 h of incubation. After incubation, the plate was centrifuged, the medium was removed, and DMSO was added to each well to dissolve the precipitate. The plate was read at 510 nm on a Benchmark microplate reader (Bio-Rad). For the proliferation assay, cells were cultured as described above and were infected for 48 h and 1 μCi/well [³H]thymidine was added for the last 18 h of culture. The cells were harvested onto a 96-well filter plate and the [³H]thymidine incorporation was detected as counts/min using a Topcount luminometer (Perkin-Elmer).

Propidium iodide analysis of DNA content. Multiple myeloma cells were infected with LV-empty and LV-PPARγ for 96 h, after which they were harvested, washed twice in PBS, and fixed with 2% paraformaldehyde at room temperature for 30 min. Cells were then washed twice with PBS and postfixed in ice-cold 70% ethanol for 10 to 14 h. After fixation, the cells were washed in PBS and resuspended in 0.1% Triton X-100, 0.2 mg/mL RNase A (Sigma-Aldrich), and 20 μg/mL propidium iodide (Sigma-Aldrich) in PBS. The cells were incubated for 30 min at room temperature and immediately analyzed on a BD Biosciences FACSCalibur flow cytometer. The percentage of cells with sub-G₀-G₁ DNA content was determined using FlowJo software (Tree Star).

Cell cycle analysis. Cell cycle analysis of human multiple myeloma cells was done using the APC bromodeoxyuridine (BrdUrd) Flow kit (BD PharMingen) according to the manufacturer's instructions. In the set of experiments using ciglitazone, cells were infected for 24 h and treated with ciglitazone (5 μmol/L) for an additional 48 h. Cells were then pulsed for 30 min at a final BrdUrd concentration of 10 μmol/L and stained the APC-BrdUrd flow kit.

IL-6 ELISA. IL-6 production was assessed in cell culture supernatants by an ELISA according to the manufacturer's instructions (BD PharMingen).

Western blots. One million human multiple myeloma cells were mock-infected or infected with LV-empty or LV-PPARγ. Seventy-two hours post-infection, whole-cell lysates were prepared using ELB buffer [50 mmol/L HEPES (pH 7), 250 mmol/L NaCl, 0.1% NP-40, 5 mmol/L EDTA, 10 mmol/L NaF, 0.1 mmol/L Na₃VO₄, 50 μmol/L ZnCl₂ supplemented with 0.1 mmol/L phenylmethylsulfonyl fluoride, 1 mmol/L DTT, and a mixture of protease and phosphatase inhibitors] and total protein was quantified using bicinchonic acid protein assay (BCA assay kit; Pierce). Protein (25 μg) was electrophoresed on an 8% to 16% Precise protein gel (Pierce) and transferred to a polyvinylidene difluoride membrane (Millipore). The membranes were analyzed for immunoreactivity with indicated primary antibody, washed, and then incubated with an appropriate horseradish peroxidase-conjugated secondary antibody. The membranes were then visualized by chemiluminescence using an ECL kit (Pierce).

Statistical analysis. Results are expressed as mean ± SD. Two-tailed Student's t test was done and P values < 0.05 were considered significant. All experiments were repeated at least three times.

Construction of a lentiviral-based vector for PPARγ1 overexpression in multiple myeloma cell lines. We have shown previously that both normal and malignant human B cells, including multiple myeloma cells, modestly express PPARγ and that exposure to certain PPARγ agonists induces cell death (25–28), suggesting that PPARγ maybe a therapeutic target for multiple myeloma. To begin to determine both the physiologic role of PPARγ in multiple myeloma and the functional consequences of its overexpression, we constructed a lentivirus for PPARγ gene delivery. Although human B lineage cells are generally difficult to transduce, our pilot experiments revealed that a lentivirus would provide an efficacious gene delivery system. We constructed a lentiviral vector that contains a Flag-tagged PPARγ1 cDNA under the control of the CMV promoter. The vector also expresses copGFP under the control of the elongation factor-1α promoter. As additional controls, the backbone lentiviral vector, which expresses copGFP, but not the PPARγ1 gene (LV-empty), and the lentiviral vector expressing a dominant-negative form of PPARγ1 (LV-PPARγ-DN) were used (Fig. 1A). This dominant-negative contains two amino acid substitutions that inhibit the release of corepressors and reduce coactivator recruitment to the PPAR promoter (43).

To evaluate the function of PPARγ in human multiple myeloma cells, we used the well-characterized multiple myeloma cell RPMI 8226 (44). First, we transduced 1 × 10⁶ RPMI 8226 cells with LV-empty or LV-PPARγ or LV-PPARγ-DN at a MOI of 1. Forty-eight hours post-infection, GFP expression was assessed by flow cytometry. More than 80% of the cells were transduced (Fig. 1B). We next measured the presence of Flag-tagged PPARγ protein using an anti-Flag antibody and found that infection with the LV-PPARγ and LV-PPARγ-DN, but not the LV-empty, resulted in Flag expression, indicating that transduction with LV-PPARγ and LV-PPARγ-DN resulted in increased levels of exogenous PPARγ-WT or PPARγ-DN protein, respectively (Fig. 1C). To test whether this overexpressed PPARγ protein was functional and could then bind to its promoter sequence, we used a luciferase reporter assay generated by cotransfecting human 293FT cells with a luciferase reporter construct containing three PPRE elements (PPRE-LUC) and either LV-empty, LV-PPARγ, or LV-PPARγ-DN. Forty-eight hours post-transfection, cell lysates were collected. Cells transfected with LV-PPARγ vector exhibited a dramatic increase in promoter activity, as assessed by an increase in luciferase, compared with that of LV-empty and LV-PPARγ-DN (Fig. 1D), indicating that the PPARγ that is overexpressed is indeed functional.

Lentiviral PPARγ overexpression reduces cell proliferation and induces cell cycle arrest. We next determined the effect of PPARγ overexpression on cell proliferation by assessing cell number and [³H]thymidine incorporation. In Fig. 2A the first column indicates the cell number before infection (0 h). For multiple myeloma cells that were uninfected or infected with LV-empty, the cell number almost doubled by 48 h post-infection. In contrast, the cells infected with LV-PPARγ showed a substantial reduction in cell number compared with that of LV-empty-infected cells. To assess whether this was due, at least in part, to a decrease in proliferation, [³H]thymidine incorporation was measured. Figure 2B shows that proliferation was not significantly influenced by LV-empty transduction (P = 0.06). However, RPMI 8226 proliferation was reduced by ∼50% in LV-PPARγ-expressing cells compared with that of RPMI 8226 transduced with LV-empty (P = 0.02). Given that PPARγ overexpression results in a decrease in proliferation, as assessed by [³H]thymidine incorporation, we next analyzed the effects of PPARγ overexpression on cell cycle progression using BrdUrd and 7-AAD staining (Fig. 2C). Cells positive for BrdUrd represent cells that have entered the S phase, whereas 7-AAD reveals cell subsets in the G₁ and G₂-M phases. Multiple myeloma cells infected with LV-PPARγ displayed a 4-fold decrease of cells in S phase compared with that of LV-empty-infected cells (7% in LV-PPARγ-infected cells versus 28% in LV-empty-infected cells) and a 2-fold increase in cells in G₂-M phase (60% in LV-PPARγ-infected cells versus 30% in LV-empty-infected cells) by 48 h. Taken together, PPARγ overexpression results in a decrease in RPMI 8226 cell proliferation as a consequence of a G₂-M cell cycle arrest.

PPARγ overexpression reduces multiple myeloma cell viability by eliciting apoptosis. We next determined whether the decrease in proliferation of RPMI 8226, as a result of PPARγ overexpression, was due to the induction of apoptosis. Cells undergoing apoptosis show characteristic morphologic features such as shrinkage and membrane blebbing (45). To determine whether overexpression of PPARγ induced apoptosis, RPMI 8226 cells were infected with LV-PPARγ, LV-empty, or LV-PPARγ-DN. Cellular morphology was then assessed by light and fluorescence microscopy. Seventy-two hours after infection, cells infected with either LV-empty or LV-PPARγ-DN showed their characteristic round morphology, whereas PPARγ-overexpressing cells (infected with LV-PPARγ) showed morphologic changes characteristic of apoptosis, such as plasma membrane blebbing (Fig. 3A, arrow) and cytoplasmic convolution (Fig. 3A, arrowhead). To test whether PPARγ overexpression has the same effect in another multiple myeloma cell line, we infected U266 multiple myeloma cells at a MOI of 1. These cells showed similar changes in cellular morphology as RPMI 8226 multiple myeloma cells (Fig. 3B). To further determine whether the increased expression of PPARγ influenced multiple myeloma cell viability, RPMI 8226 cells were infected with LV-empty or LV-PPARγ and cell viability was assessed by MTT assay. Seventy-two hours post-infection, viability was reduced by 50% in cells overexpressing PPARγ compared with that of uninfected and LV-empty control infected cells (Fig. 3C). Taken together, these results indicate that PPARγ overexpression has proapoptotic effects resulting in a deleterious effect on multiple myeloma cell survival. Interestingly, PPARγ overexpression did not induce cell death in human fibroblasts nor in embryonic kidney cells (data not shown), suggesting that this effect is cell type specific.

Lentiviral overexpression of PPARγ induces apoptosis and activation of caspases in human multiple myeloma cells. A characteristic of the early stages of apoptosis is the exposure of phosphatidylserine from the inner leaflet of the membrane to the external surface (45). To further characterize the cytopathic effects caused by PPARγ overexpression on apoptotic cell death, double staining was done on LV-empty and LV-PPARγ-infected cells using anti-Annexin V-APC monoclonal antibody and the DNA stain 7-AAD followed by analysis using flow cytometry (Fig. 4A). Cells infected with LV-PPARγ exhibited a dramatic increase in Annexin V/7-AAD double-positive population compared with that of LV-empty control (Fig. 4A, top right quadrants). These results are even more evident at the 96 h time point, where the cells infected with LV-PPARγ showed a further increase in Annexin V/7AAD double-positive population, supporting the fact that PPARγ overexpression induces apoptosis. We also observed an increase on 7-AAD single-positive cells at 96 h time point, which may suggest secondary necrosis. Rates of apoptosis after LV-PPARγ infection were also determined by DNA staining with propidium iodide and flow cytometric analysis. Apoptotic cells exhibit reduced DNA content that is detected by the appearance of a sub-G₀-G₁ cell population by flow cytometry. As shown in Fig. 4B, LV-empty control cells display a background level between 7% and 12%. In contrast, the number of sub-G₀-G₁ events in RPMI 8226 cells infected with LV-PPARγ increased over time from 23% at 72 h to 40% at 96 h post-infection. These results further confirm the induction of apoptosis by PPARγ overexpression.

Cells undergoing apoptosis also display an increase in nuclear chromatin condensation; these compact nuclei become brightly fluorescent when labeled with nuclear stains, such as 4′,6-diamidino-2-phenylindole (46). To further support PPARγ-induced cell death, multiple myeloma cells left uninfected or infected with either LV-empty or LV-PPARγ were stained with 4′,6-diamidino-2-phenylindole. Fluorescence microscopy revealed nuclear condensation and fragmentation 96 h after LV-PPARγ infection (Fig. 4C, right, arrows). In contrast, nuclei from noninfected and LV-empty-infected cells predominantly appeared uniformly stained without condensation (Fig. 4C, left and middle). Quantification of apoptotic nuclei in cultures exposed to LV-PPARγ exhibited 37% of cells with apoptotic nuclei, whereas only 4% of cells with apoptotic nuclei were detected in cells infected with LV-empty (Fig. 4C). These results further support the findings that apoptotic changes of cell and nuclear morphology are induced by PPARγ overexpression. Together, these changes reveal an induction of a nonproliferating, apoptotic phenotype in cells overexpressing PPARγ.

Caspase-3 has a key role in apoptosis, being responsible for the proteolytic cleavage of many key proteins, including the nuclear enzyme PARP (47). To examine the molecular basis for the PPARγ-dependent induction of apoptosis, we measured caspase activity after LV-PPARγ transduction. Human multiple myeloma RPMI 8226 cells were infected with either LV-empty or LV-PPARγ, and 72 h later, cell lysates were collected. The active form of caspase-3, as well as the cleavage of PARP, was determined by Western blot analysis. Cells that were infected with LV-PPARγ showed an increased level of active caspase-3 compared with that of uninfected and LV-empty-infected cells (Fig. 4D). PARP cleavage was determined by the appearance of an 85-kDa fragment by Western blot analysis. Cells that were left uninfected or that were infected with LV-empty control virus contain predominantly the full-length (116 kDa) PARP, whereas the 85 kDa cleaved PARP appeared in the cells that were infected with LV-PPARγ (Fig. 4D). These results suggest that PPARγ overexpression induces caspase activation.

Growth inhibition and induction of apoptosis by ciglitazone treatment. Thiazolidinediones, a class of antidiabetic drugs, are high-affinity synthetic ligands of PPARγ (14). To assess whether a model PPARγ ligand further enhanced human multiple myeloma cell death in cells that overexpress PPARγ, multiple myeloma cells were infected at a lower MOI of 0.5 with either LV-empty or LV-PPARγ. At this MOI, ∼50% of the cells were transduced as assessed by analysis of GFP by flow cytometry (data not shown). We hypothesized that RPMI 8226 cells would proliferate less with modest PPARγ overexpression. We further speculate that treatment with ciglitazone would augment apoptosis in cells expressing higher PPARγ protein levels. Twenty-four hours after transduction, cells were either untreated or treated with 5 μmol/L ciglitazone for an additional 48 h. We showed previously that this concentration of ciglitazone does not induce apoptosis in multiple myeloma cells (26). Cells were then pulsed with BrdUrd for 30 min, fixed, permeabilized, and stained intracellularly with anti-BrdUrd-APC antibody. Cells were then analyzed for both GFP expression and BrdUrd incorporation by flow cytometry. As Fig. 5A shows, LV-PPARγ-infected cells showed somewhat reduced incorporation of BrdUrd, especially the higher expressers of PPARγ (GFP-high, top right quadrant). When cells were treated with 5 μmol/L ciglitazone, there was an enhanced inhibition in DNA synthesis in cells that overexpressed PPARγ but not in the cells infected with LV-empty or uninfected cells. At the same time, the percentage of apoptotic cells (7-AAD positive) increased from 16% in LV-empty-infected cells treated with ciglitazone to 27% in cells infected with LV-PPARγ and treated with ciglitazone, whereas it remained almost the same in LV-empty-infected cells (Fig. 5B). These data support the hypothesis that increased levels of PPARγ further sensitize the cells to PPARγ ligand-induced growth arrest and cell death.

Mechanism whereby PPARγ overexpression inhibits multiple myeloma cell proliferation and induces cell death. In the next set of experiments, we determined whether PPARγ overexpression-induced caspase activation led to multiple myeloma cell apoptosis by using a broad-spectrum caspase inhibitor, ZVAD-FMK. Briefly, multiple myeloma RPMI 8226 cells were left uninfected or were infected with LV-empty or LV-PPARγ. Six hours after infection, cells were treated every day with either DMSO or with 50 μmol/L ZVAD-FMK. ZVAD-FMK treatment alone results in no toxicity to the cells (data not shown). Cells were collected at different time points and apoptosis was analyzed by the DNA stain 7-AAD. This 7-AAD solution incorporates into cells that have their plasma membrane integrity disrupted. Multiple myeloma cells that were infected with LV-PPARγ exhibited a time-dependent increase in the percentage of 7AAD-positive cells in comparison with that of uninfected and LV-empty-infected cells (Fig. 6A). The percentage of 7-AAD-positive cells in ZVAD-FMK-treated cells was 4-fold lower than that of the LV-PPARγ-infected cells without ZVAD-FMK treatment. ZVAD-FMK treatment, however, did not completely inhibit the induction of apoptosis by PPARγ overexpression (Fig. 6A). We then evaluated whether treatment of PPARγ-overexpressing cells with ZVAD-FMK inhibited caspase activation. After 5 days of infection, cells were lysed and probed for the active form of caspase-3 and PARP cleavage. PARP cleavage was observed in the PPARγ-overexpressing cells and treatment with ZVAD-FMK slightly reduced PARP cleavage. Moreover, cells that were transduced with LV-PPARγ exhibited a dramatic increase in active caspase-3. However, when these cells were treated with ZVAD-FMK, this activation was completely abolished (Fig. 6B). We next evaluated the levels of XIAP, which regulates cell death by inhibiting caspases (48). We found that XIAP levels were down-regulated in the cells overexpressing PPARγ, and treatment with the caspase inhibitor restored XIAP levels back to normal (Fig. 6B). We also examined the levels of the antiapoptotic protein Mcl-1 in PPARγ-overexpressing cells. Multiple myeloma cells infected with LV-PPARγ down-regulated Mcl-1 levels; however, when cells were treated with the caspase inhibitor, the levels of Mcl-1 were restored (Fig. 6B). Taken together, these results indicate that PPARγ overexpression induces cell death by activating caspases and by down-regulating antiapoptotic proteins.

PPARγ overexpression-induced multiple myeloma cell death can not be overcome by BMSC. Tumor cell proliferation and survival are regulated by the local microenvironment. Stromal cells are a critical component of this tumor microenvironment. Some characteristics include multiple myeloma interaction with stromal cells as well as growth factors secreted into the local environment. Importantly, adhesion of multiple myeloma cells to stromal cells confers drug resistance (37). Therefore, we tested whether coincubation with BMSC would rescue multiple myeloma cells from undergoing apoptosis mediated by PPARγ overexpression. Multiple myeloma RPMI 8226 cells were infected with either LV-empty or LV-PPARγ and cultured them with BMSC. Similar morphologic changes were observed as described earlier, where RPMI 8226 cells transduced with LV-PPARγ showed plasma membrane blebbing (Fig. 7A, middle) and cytoplasmic convolution (Fig. 7A, bottom), whereas LV-empty-transduced cells kept their characteristic round morphology (Fig. 7A, top). Multiple myeloma cells were further analyzed for cell death by using 7-AAD staining. To distinguish between multiple myeloma cells and BMSC, an anti-CD38 antibody was used, which is expressed by multiple myeloma cells but not by BMSC. Again, multiple myeloma cells infected with LV-PPARγ had a time-dependent increased in 7-AAD-positive cells in comparison with LV-empty or uninfected cells (Fig. 7B). When cells were cocultured with BMSC, there was a modest baseline increase in 7-AAD-positive cells in uninfected cells and cells infected with LV-empty. However, when PPARγ-overexpressing multiple myeloma cells were coincubated with BMSC, there was a time-dependent increase in 7-AAD-positive cells; this 7-AAD-positive population was higher than the 7-AAD-positive population observed in the absence of BMSC (Fig. 7B). These results indicate that coincubation of multiple myeloma cells with BMSC does not rescue multiple myeloma cells from undergoing apoptosis induced by PPARγ overexpression.

PPARγ overexpression in multiple myeloma or BMSC inhibits basal and multiple myeloma-induced IL-6 production by BMSC and the inhibitory effect is enhanced by ciglitazone. We next analyzed whether PPARγ overexpression had any effect on multiple myeloma-induced IL-6 production by BMSC. Multiple myeloma cell adhesion to BMSC is known to trigger IL-6 production from BMSC (49). IL-6 is a cytokine that regulates growth and survival of multiple myeloma (49). We infected multiple myeloma cells with either LV-empty or LV-PPARγ and then added them to BMSC-coated plates. Culture supernatants were collected 72 h after coincubation. We observed an increase in IL-6 production when BMSC were coincubated with either uninfected or LV-empty-infected multiple myeloma cells. However, when BMSC were coincubated with LV-PPARγ-overexpressing multiple myeloma cells, there was a significant decrease in IL-6 production by BMSC. Interestingly, the levels of IL-6 were 2-fold lower than the levels produced in the absence of multiple myeloma cells (Fig. 8A). These results suggest that PPARγ overexpression in multiple myeloma cells inhibits IL-6 production by BMSC.

We next investigated the effects of PPARγ overexpression in BMSC. We infected BMSC HS-5 cells with LV-empty or LV-PPARγ. BMSC were cultured alone or were cultured with multiple myeloma cells in the presence or absence of ciglitazone (5 μmol/L). There was a slight decrease in the basal production of IL-6 from BMSC that were infected with LV-PPARγ (Fig. 8B). In the presence of ciglitazone, basal IL-6 production was further reduced from PPARγ-overexpressing BMSC. When these BMSC were coincubated with multiple myeloma, LV-empty-infected BMSC up-regulated IL-6 production. However, when PPARγ-overexpressing BMSC were coincubated with multiple myeloma cells, BMSC failed to produce IL-6 (Fig. 8B). Taken together, these results suggest that increasing the levels of PPARγ in either multiple myeloma or BMSC decreases both basal and multiple myeloma-induced IL-6 production.

Multiple myeloma is a prevalent hematologic cancer with poor prognosis, with an average survival of only 3 to 5 years. PPARγ and PPARγ ligands represent new targets in the treatment of B-cell malignancies. In addition to the well-known adipogenic role of PPARγ, we propose that PPARγ is a tumor suppressor. There is emerging evidence for a direct role of PPARγ functional mutations in the initiation of several common human cancers, including lung, breast, prostate, pancreas, and ovary (50, 51). Moreover, the expression levels and/or transactivation of PPARγ are impaired in certain cancers. For example, in human lung cancer, decreased expression of PPARγ correlates with poor prognosis (52). In addition, well-differentiated adenocarcinomas have higher PPARγ expression compared with poorly differentiated samples, implying that PPARγ regulates tumor progression (53). The PPARγ gene is found in the human chromosome 3, band 3p25 (54). Chromosomal abnormalities, such as 3p deletions, have been identified in several hematologic cancers (55). Here, we report that PPARγ overexpression in multiple myeloma cells reduces proliferation and induces apoptosis. Importantly, treatment with ciglitazone, a selective PPARγ agonist, augments these effects. These observations suggest that loss of PPARγ expression and/or transactivation potential via post-translational modifications or inactivating mutations may augment the risk of developing multiple myeloma. Therefore, determining whether the expression levels and/or activation status of PPARγ changes during progression from normal plasma cells to monoclonal gammopathy of undetermined significance and finally to multiple myeloma may be an important prognostic indicator.

The role, however, of PPARγ in the development of hematologic malignancies has not been extensively studied. We have shown previously that both normal and malignant human B cells, including multiple myeloma cells, express moderate levels of PPARγ and that exposure to certain PPARγ agonists induces apoptosis (25–28). However, the biological relevance of PPARγ in multiple myeloma has, to date, not been explored.

PPARγ overexpression has been used by other groups to study its effects on inflammatory diseases such as colitis, where PPARγ levels were reduced as a consequence of inflammation (56). Moreover, when PPARγ was overexpressed in thyroid carcinoma cells, it induced cell cycle arrest and cell death (57). In concordance with these studies, we showed that increased PPARγ expression leads to inhibition of proliferation and induction of apoptosis in multiple myeloma cells even in the absence of an exogenous PPARγ ligand. Because transduction with LV-PPARγ alone achieved therapeutic effects, we suggest that increasing the levels of PPARγ in these cells may allow for interactions with endogenous PPARγ ligands that activate the antiproliferative actions associated with PPARγ. Moreover, in the presence of the PPARγ ligand ciglitazone, cells overexpressing PPARγ had an even greater reduction in proliferation, further supporting the concept that PPARγ plays an important role in regulating multiple myeloma cell growth, as also reported in other cell types (32, 58–60).

A reduction in [³H]thymidine incorporation and BrdUrd incorporation (Fig. 2) suggests that PPAR-overexpressing cells do not progress through S phase. Moreover, this reduction in S-phase transit occurs before the cells display apoptotic changes, suggesting the inhibition of cell proliferation precedes apoptosis. Rapid cell shrinkage and increased granularity are the most evident morphologic changes associated with apoptosis, which also accompany nuclear condensation and membrane blebbing (45). In addition, cells at early stages of apoptosis expose phosphatidylserine from the inner side of the membrane to the membrane surface (45). Herein, we observed dramatic changes in multiple myeloma cellular morphology as well as an increased number of Annexin V/7-AAD double-positive subset after infection with LV-PPARγ. Coincident with these events, we observed an increase in cells displaying sub-G₀-G₁ DNA content, as well as an induction of nuclear condensation in the cells that overexpressed PPARγ, but not in the LV-empty control infected cells. These results are indicative that apoptotic induction is PPARγ dependent, as infection with LV-empty control virus or the dominant-negative form of PPARγ did not cause any changes in cellular or nuclear morphology.

Cells undergoing apoptosis experience mitochondrial damage that leads to activation of caspases (47). Caspases are cysteine proteases responsible for cleaving cellular substrates during apoptosis, including DNA cleavage, which is characteristic of apoptosis (47). The PPARγ-dependent morphologic changes correlated with a loss of mitochondrial membrane integrity and cleavage of pro-caspase-3 and PARP, confirming activation of the intrinsic mitochondrial cell death pathway. Furthermore, coincubation with a general caspase inhibitor inhibited PPARγ overexpression-induced cell death and inhibited caspase-3 activation as well as reduced PARP cleavage.

The IAPs are a family of caspase inhibitors that specifically inhibit caspase-3, caspase-7, and caspase-9 to prevent apoptosis (48). One of the members of this family is XIAP. Increased expression of the IAP family proteins, including XIAP, correlate with poor prognosis and chemotherapy-induced drug resistance in multiple myeloma (61). Mcl-1 is an antiapoptotic protein belonging to the Bcl-2 family (62). Mcl-1 protects multiple myeloma cells against spontaneous and chemotherapy-induced apoptosis (62). Moreover, down-regulation of Mcl-1 induces multiple myeloma cell death (62). Herein, the levels of both antiapoptotic protein XIAP and Mcl-1 were down-regulated in the cells overexpressing PPARγ but not in the cells treated with caspase inhibitor or the LV-empty-infected cells. Although the exact mechanism of PPARγ overexpression-induced cell death remains uncertain, these results indicate that PPARγ overexpression appears to target several pathways in multiple myeloma, including the activation of caspase-dependent and caspase-independent pathways and the down-regulation of inhibitors of apoptosis.

Bone marrow microenvironment is a very important factor for multiple myeloma drug resistance and chemoresistance (37). BMSC produce IL-6, a cytokine that protects multiple myeloma cells from drug-mediated apoptosis (36, 63). Inhibition of IL-6 signaling sensitizes multiple myeloma cells to bortezomib-induced cell death (64). The involvement of PPARγ ligands in influencing IL-6 production has been shown by Farrar et al., where PPARγ ligands inhibited both multiple myeloma cell adhesion to BMSC and multiple myeloma-induced IL-6 production by BMSC (38, 39). Here, we show that PPARγ overexpression reduces IL-6 production by BMSC even in the absence of exogenous ligand (Fig. 8). Thus, PPARγ not only regulates multiple myeloma cells but also can affect the ability of BMSC to produce factors such as IL-6 that promotes tumor progression. Moreover, PPARγ overexpression-induced multiple myeloma cell death was not able to be rescued by BMSC, indicating that PPARγ overexpression might be able to overcome bone marrow microenvironment-mediated resistance.

Collectively, these results indicate that therapeutic attempts to induce PPARγ expression in multiple myeloma cells may show synergistic activity with the use of PPARγ ligands for therapy. Although appropriate targeting and delivery mechanisms would be required before PPARγ gene therapy could become a reality, the observation made here does suggest that pharmacologic means of up-regulating PPARγ in cancer cells and multiple myeloma cells, in particular, may be a potential new therapeutic target.

No potential conflicts of interest were disclosed.

We thank Dr. Stephen Dewhurst for kindly providing the pCMV-VSVG and pCMV- Δ89.2 vectors for lentivirus production, Dr. Robert G. Fenton for providing the RPMI 8226 multiple myeloma cell line, and Dr. Carolyn J. Baglole for comments and review of this article.