Peroxisome Proliferator-Activated Receptor ; and Retinoid X Receptor Ligands Are Potent Inducers of Differentiation and Apoptosis in Leukemias

Molecular Cancer Therapeutics 1249

Peroxisome proliferator-activated receptor ; and retinoid X receptor ligands are potent inducers of differentiation and apoptosis in leukemias

Marina Konopleva,1 Elena Elstner,4 (i.e., all trans-retinoic acid) enhanced differentiating and Teresa J. McQueen,1 Twee Tsao,1 growth-inhibitory effects. 2-Cyano-3,12-dioxooleana-1,9- Andrey Sudarikov,1 Wei Hu,1 Wendy D. Schober,1 dien-28-oic acid induced differentiation and apoptosis Rui-Yu Wang,1 David Chism,1 Steven M. Kornblau,1 with much greater potency than the other PPAR; ligands Anas Younes,2 Steven J. Collins,5 H. Phillip Koeffler,6 in established cell lines and primary chronic lymphocytic and Michael Andreeff1,3 leukemia samples. Exposure to 2-cyano-3,12-dioxooleana-1,9-dien-28-oic acid induced mitochondrial depo- Section of Molecular Hematology and Therapy and Departments larization and caspase activation, which was associated of 1Blood and Marrow Transplantation, 2Lymphoma, and 3Leukemia, University of Texas M.D. Anderson Cancer Center, with apoptosis induction. In Bcl-2-overexpressing chronic Houston, Texas; 4Department of Medicine (Charite´), Division lymphocytic leukemia cells, the small-molecule Bcl-2 of Hematology/Oncology, Humboldt University, Berlin, Germany; inhibitor HA14-1 sensitized these cells to 2-cyano-3,12- 5Fred Hutchinson Cancer Research Center, Seattle, Washington; 6 dioxooleana-1,9-dien-28-oic acid–induced apoptosis. and Division of Hematology/Oncology, Cedars-Sinai Medical ; Center/University of California at Los Angeles School of Medicine, These results suggest that PPAR ligation alone and in Los Angeles, California combination with retinoids holds promise as novel therapy for leukemias by activating the transcriptional activity of target genes that control apoptosis and differentiation in Abstract leukemias. [Mol Cancer Ther 2004;3(10):1249–62] The peroxisome proliferator-activated receptor ; (PPAR;) is a member of the nuclear receptor family that forms heterodimers with retinoid X receptor. These heterodimers Introduction bind to DNA and activate the transcription of target genes. Recent progress in the molecular and cellular biology of the Here, we report that the PPAR; receptor protein is nuclear steroid receptor superfamily has identified certain expressed in primary myeloid and lymphoid leukemias members as molecular targets for cancer therapy (1). They and in lymphoma and myeloma cell lines. In this study, we include estrogen receptors, retinoic acid receptors, retinoid compared the activity of several PPAR; ligands including X receptors (RXR; the RXR-specific ligands are termed BRL49653 (rosiglitazone), 15-deoxy-#12,14-prostaglan- ‘‘rexinoids’’), and the vitamin D receptor (the vitamin D– specific ligands are termed ‘‘deltanoids’’). These nuclear din J2, and the novel triterpenoid 2-cyano-3,12-dioxooleana-1,9-dien-28-oic acid on leukemia cells. Exposure to receptors are putative cancer therapy targets because they these PPAR; ligands induced apoptosis in myeloid (U937 function as transcription factors that control the expression and HL-60) and lymphoid (Su-DHL, Sup-M2, Ramos, Raji, of many genes related to cell differentiation (1, 2). The Hodgkin’s cell lines, and primary chronic lymphocytic strongest evidence for the therapeutic potential of this leukemia) cells. A similar exposure to these PPAR; ligands approach comes from the efficacy of retinoic acid receptor a induced the differentiation of myeloid leukemic cells. A activation in the treatment of acute promyelocytic leukemia combination of PPAR; ligands with a retinoid X receptor (3). Over 90% of patients with acute promyelocytic agonist (i.e., LG100268) or a retinoic acid receptor agonist leukemia achieve complete remission following treatment with the naturally occurring retinoid all trans-retinoic acid (ATRA; ref. 4). The peroxisome proliferator-activated receptor g (PPARg) is a recent addition to the nuclear steroid receptor Received 7/14/03; revised 7/2/04; accepted 7/22/04. superfamily that is believed to be involved in the regulation Grant support: Leukemia and Lymphoma Society (M. Konopleva), of lipid metabolism. Like other nuclear steroid receptors, National Cancer Institute grants RO1 CA89346 and 1 P50 CA100632-01, (M. Andreeff) Deutsche Forschungsgemeinschaft grant and J-Carreras the dysregulation of PPARg has been implicated in (E. Elstner). aberrant differentiation. PPARg is an important mediator The costs of publication of this article were defrayed in part by the of the differentiation of preadipocytes to adipocytes and is payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to a target in the treatment of diabetes, where receptor indicate this fact. activation increases sensitivity to endogenous insulin. To Requests for reprints: Michael Andreeff, Department of Blood and Marrow activate transcription, PPARg must form a heterodimer Transplantation, University of Texas M.D. Anderson Cancer Center, with RXRa (5). The PPARg/RXR heterodimers can then be 1515 Holcombe Boulevard, Unit 448, Houston, TX 77030. Phone: 713-792-7260; Fax: 713-794-4747. activated by the ligation of PPARg or RXR ligands (6), but E-mail: [email protected] the simultaneous ligation of both PPARg and RXR max- Copyright C 2004 American Association for Cancer Research. imizes their activation. PPARg can be activated by

Mol Cancer Ther 2004;3(10). October 2004

naturally occurring ligands such as the long-chain fatty (27), a selective PPARg antagonist, was synthesized and acids 15-deoxy-D12,14-prostaglandin J2 (15-d-PGJ2) and 9- kindly provided by Dr. Stephen H. Safe (Texas A&M and 13-cis-hydroxyoctadecadienoic acid (7–9). They are University, College Station, TX). HA14-1 (ethyl 2-amino-6- also activated by synthetic ligands including the antidia- bromo-4-[1-cyano-2-ethoxy-2-oxoethyl]-4H-chromene-3- betic agents troglitazone, BRL49653 (rosiglitazone), and carboxylate), a novel nonpeptidic ligand of the Bcl-2 pioglitazone; L-Tyr-based PPARg ligands (10); Fmoc-L-Leu BH3-binding pocket (28, 29) that inhibits Bcl-2 function, was amino acids (11); and certain triterpenoids (12). kindly provided by Dr. Ziwei Huang (University of Illinois Although PPARg ligands have mainly been used for the at Urbana-Champaign, Urbana, IL). The Fas signaling anti- treatment of diabetes, they have also been shown to induce body CH11 was purchased from Immunotech (Miami, FL) lipogenic differentiation of malignant cells in vivo in and tumor necrosis factor–related apoptosis-inducing li- patients with liposarcomas (13). In hematopoiesis, PPARg gand (TRAIL) was purchased from Alexis (San Diego, CA). expression is increased during differentiation of monocytes Cell Lines to macrophages and the activation of PPARg/RXR-regu- HL-60, U937, Jurkat, Daudi, Raji, Ramos, 8226, and IM-9 lated genes induces macrophage differentiation (9). PPARg cell lines were obtained from the American Type Culture expression has been detected in human breast, lung, Collection (Rockville, MD). HL-60 is a myeloid leukemia bladder, colon, prostate, and pancreatic cancer cells and cell line; U937 is a monocytic leukemia cell line; Daudi, the ligand-induced activation of PPARg/RXR induced Ramos, and Raji are human Burkitt lymphoma cell lines; g apoptosis in these cell types (14–21). PPAR /RXR signal- Jurkat is a T-lymphoblast cell line; 8226 and IM-9 are ing is involved in the induction of apoptosis in T myeloma cell lines. The Sup-M2 and Su-DHL-1 cell lines lymphocytes (22) and myeloid leukemias (23). Interesting- were derived from CD30+ anaplastic large-cell lymphomas ly, a recent report showed that synthetic PPARg agonists and carry the t(2;5) chromosomal translocation (30). The used at high concentrations inhibited the translation human Hodgkin and Reed-Sternberg–derived cell lines initiation of eukaryotic initiation factor-2 by phosphoryla- HD-MYZ, HD-LM-2, L-428, and KM-H2 were provided by tion in PPARg-deficient embryonic stem cells, suggesting the German Collection of Microorganisms and Cell the existence of a receptor-independent antiproliferative Cultures (Braunschweig, Germany). The phenotypes and mechanism (24). genotypes of these cell lines have been published previ- In this study, we investigated the effects of structurally ously (31). The HL-60-derived subline CDM-1, which different PPARg ligands on a variety of myeloid and expresses low levels of PPARg, was kindly provided by lymphoid leukemic cell lines and primary leukemia Dr. Peter Davies (University of Texas, Houston, TX; ref. 23). g samples. The PPAR ligands decreased viability and Subjects induced apoptosis, with the novel triterpenoid 2-cyano- Bone marrow or peripheral blood samples were obtained 3,12-dioxooleana-1,9-dien-28-oic acid (CDDO) being the for in vitro studies from patients with chronic lymphocytic most potent cytotoxic agent. Concomitant RXRa ligation leukemia (CLL), acute lymphoblastic leukemia, and acute enhanced the effects of PPARg ligands and the combination myelogenous leukemia (AML). Samples were obtained of a PPARg ligand with ATRA induced maturation of during routine diagnostic procedures after informed myelomonocytic cells. Furthermore, the proapoptotic effects of CDDO could be further enhanced by inhibition consent was obtained in accordance with regulations and of Bcl-2. Hence, the apoptotic and differentiation-inducing protocols approved by the Human Subjects Committee of properties of PPARg ligands and their low toxicity make the University of Texas M.D. Anderson Cancer Center them promising candidates for antileukemia therapy. (Houston, TX). Mononuclear cells were separated by Ficoll- Hypaque (Sigma Chemical) density gradient centrifugation. Normal peripheral blood, bone marrow, or mobilized Materials and Methods peripheral blood stem cells were obtained from healthy Reagents donors or from cryopreserved specimens following in- CDDO was synthesized by Dr. Takashi Honda and formed consent. Low-density cells were separated by colleagues (Dartmouth Medical College, Hanover, NH; ref. Ficoll-Hypaque density gradient centrifugation and + 25) and kindly provided through the Rapid Access to Inter- enriched by magnetic activated cell sorting for CD34 cells vention Development mechanism by Dr. Edward Sausville as described previously (32). (National Cancer Institute/Cancer Therapy Evaluation Inhibition of DNA Synthesis and Cell Viability 3 5 Program, Bethesda, MD). ATRA was purchased from [ H]Thymidine Uptake. Triplicate samples of 1 Â 10 cells Sigma Chemical Co. (St. Louis, MO). RXR-specific ligand suspended in 200 AL RPMI were cultured in the presence or LG100268 (26) was kindly provided by Dr. Reid Bissonnette absence of PPARg ligands in 96-well flat-bottomed micro- (Ligand Pharmaceuticals, San Diego, CA). The agonists for titer plates (Costar, Cambridge, MA) in a humidified j PPARg included 15-d-PGJ2 (Cayman Chemical Co., Ann atmosphere of 5% CO2 in air at 37 C. After 2 days, 0.1 ACi Arbor, MI; refs. 7, 8), BRL49653 (provided by SmithKline [3H]thymidine (specific activity, 6.7 Ci/mmol, DuPont, Beecham Pharmaceuticals, Harlow, United Kingdom), and Wilmington, DE) was added to each well and the cultures troglitazone (provided by Drs. W. Johnson and A. Salteil, were incubated overnight. The cells were then deposited Parke-Davis/Warner-Lambert, Ann Arbor, MI). T0070907 onto harvester filters using a semiautomatic cell harvester.

Mol Cancer Ther 2004;3(10). October 2004

Radioactivity was measured in a scintillation counter and were washed in PBS and resuspended in 100 ALbinding the results are expressed as the counts per minute for the buffer containing Annexin V (Roche Diagnostic Corp., respective samples. The inhibition of DNA synthesis was Indianapolis, IN). Cells were analyzed by flow cytometry expressed as the percentage of the counts per minute after the addition of propidium iodide (38). Annexin V binds measured in treated sample divided by the counts per to those cells that express phosphatidylserine on the outer minute in the control (DMSO-treated) sample. layer of the cell membrane, and propidium iodide stains Cell viability was determined by counting triplicate the cellular DNA of cells with a compromised cell samples for trypan blue dye–excluding cells or by the membrane. In CLL samples, Annexin V positivity was ViaLight proliferation/cytotoxicity kit (LumiTech Ltd., determined on CD19+ B cells after appropriate gating. Nottingham, United Kingdom). In this assay, the biolumi- For analyzing time-dependent induction of apoptosis, a nescence of ATP is a measure of viability (33). Briefly, ATP three-color apoptosis assay combining measurement of was first extracted from cells using nucleotide releasing phosphatidylserine/Annexin V, analysis of the mitochon- reagent, after which ATP monitoring reagent was added drial membrane potential (MMP, Dcm) using chloromethyl and light emission was measured using a luminometer. X-rosamine (39), and detection of active caspases by the Colony Formation in Soft Agar CaspaTag assay was developed and validated (40). Briefly, Cells were seeded in a two-layer soft agar system as cells were washed in PBS and resuspended in 200 AL described previously (34). Briefly, the bottom layer con- chloromethyl X-rosamine (Molecular Probes, Eugene, OR) sisted of 0.5% agar in which the test substances were working solution (300 nmol/L in RPMI) followed by the mixed; the top layer was 0.3% agar (Difco Laboratories, addition of 5 AL CaspaTag (FAM-DEVD-FMK, Intergen, Detroit, MI) in which 1,000 to 5,000 cells were mixed per Inc., Purchase, NY) working solution (150Â in PBS). plate. All experiments were done in triplicate. After 10 days After a 1-hour incubation at 37jC, cells were washed in PBS of incubation at 37jC in a humidified atmosphere contain- and the pellet was resuspended in 100 AL binding buffer ing 5% CO2 in air, colonies (>40 cells) were counted using containing Annexin V-biotin (100Â, Roche Diagnostic an inverted microscope. Corp.). Cells were incubated in the dark at room tempera- Studies of Induction of Differentiation ture for 15 minutes, washed once with Annexin binding The differentiation of myeloid leukemic cells was buffer, and stained with 10 AL avidin-phycoerythrin on ice determined by their ability to produce superoxide, as for 30 minutes. After washing with the Annexin binding measured by the reduction of nitroblue tetrazolium (NBT), buffer, the cell pellet was resuspended in binding buffer and which occurs when cells undergo either monocyte or analyzed by a FACSCalibur flow cytometer (BD Biosciences, granulocyte differentiation (35). Staining the cells for a- San Jose, CA) using appropriate compensation. Untreated naphthyl acetate esterase (NSE, Sigma Chemical) was also cells were stained using the same protocol and served as used as a cytosolic indicator of differentiation. Differenti- controls. ation-specific cell morphology was assessed on cytospin Western Blot Analysis preparations stained with the Diff-Quick Stain Set (Baxter Cell lysates derived from f2 Â 105 cells were subjected to Healthcare Corp, Miami, FL). The analysis of surface SDS-PAGE in 12% polyacrylamide gels followed by protein differentiation–specific cell surface antigens was measured transfer to a Hybond-P membrane (Amersham Pharmacia by flow cytometry. The phycoerythrin-conjugated anti- Biotech, Little Chalfont, United Kingdom) and immunoblot- CD11b and FITC-conjugated anti-CD14 monoclonal anti- ting. Antibody against PPARg was obtained from Santa Cruz bodies (Becton Dickinson, San Jose, CA) were used at a 1:10 Biotechnology (Santa Cruz, CA), anti-Bcl-2 from DAKO Corp. dilution. The percentage of positive cells was calculated by (Carpinteria, CA), and caspase-8 and -9 and a specific subtracting the percentage of cells with a fluorescence antibody recognizing only the p20-processed caspase-3 band intensity greater than the set marker using the isotype were from Cell Signaling Technology Inc. (Beverly, MA). control (background) from the percentage of cells with a Anti-h-actin blots were run in parallel as loading controls. fluorescence intensity greater than the same marker using Signals were detected by a PhosphorImager (Storm 860, the specific antibody (36). version 4.0, Molecular Dynamics, Sunnyvale, CA). Phagocytosis Transfection Assays After exposure to PPARg ligands and/or ATRA, HL-60-CDM-1 cells were transfected with empty expres- leukemic cells were washed twice in PBS and cultured in sion vector (pcDNA3), FLAG-tagged wild-type PPARg,or regular medium for 1 day and their ability to phagocytose FLAG-tagged L466A/E469A dominant-negative (DN) yeast was tested as described elsewhere (37). Briefly, PPARg mutant together with a selectable marker (neo). Candida albicans were opsonized in 10% human AB serum. Wild-type human PPARg and the DN-PPARg mutant A 5:1 ratio of C. albicans to leukemic cells was incubated at were kindly provided by Dr. Krishna K. Chatterjee 37jC for 30 minutes. The cells were stained with a Diff- (University of Cambridge, Cambridge, United Kingdom; Quick Stain Set, and the percentage of cells containing three ref. 41). Stable transfectants of the leukemic cells were or more yeast particles was determined microscopically. obtained using the calcium phosphate method (42) and FlowCytometric Analysis of Apoptosis neomycin-resistant clones were selected by treatment with In primary AML samples, apoptosis was determined by G418. Lysates from the selected clones were evaluated phosphatidylserine/Annexin V flow cytometry. Briefly, cells for transgene expression by immunoblot analyses using

Mol Cancer Ther 2004;3(10). October 2004

anti-FLAG antibodies. The clones expressing high levels of the specific proteins were further subcloned by limiting dilution. Representative subclones of each of the HL-60- CDM-1 transfectants were passaged twice per week and used for the studies. Data are representative of those derived from at least two independent clonal transfectants of CDM-1/PPARg cells. Forluciferasereporterassays,1AgTK-PPRE-Luc plasmid (a gift from Ron M. Evans, Salk Institute, La Jolla, CA) was transfected into MCF-7 cells using Fugene 6 (Boehringer-Mannheim, Mannheim, Germany). Cells were then cultured for 24 hours in complete medium and treated with CDDO or 15-d-PGJ2 alone or in combination with T007. Luciferase activity was assayed with luciferase assay system (Promega, Madison, WI). Statistics Results are expressed as the means F SD of triplicate samples. Statistical significance was determined by two- tailed, paired, Student’s t test with a P < 0.05 confidence Figure 1. PPARg protein expression in leukemic cells. Expression of interval. The combination index (CI) for experimental PPARg in leukemic cell lines, eight primary myelodysplastic syndrome/ treatment combinations was calculated to determine the AML and seven primary CLL samples, was studied by Western blot analysis using a monoclonal antibody to PPARg. The monoclonal antibody synergistic, additive, or antagonistic effects of the combi- reacts with both PPARg1 and PPARg2 isoforms. To confirm the specificity nations using the Chou-Talalay method (43) and Calcusyn of the reaction, we used the monoclonal antibody to PPARg or antibody software (Biosoft, Ferguson, MO). When CI is 1, the preadsorbed with blocking peptide for competition studies. Disappearance of the specific band after preadsorption with the blocking peptide (top left, equation represents the conservation isobologram and for Daudi, Sup-M2, and U937 cells) allowed us to confirm the correct indicates additive effects. CI values of <1.0 indicate a more position of the PPARg band. than expected additive effect (i.e., synergism).

PPAR; Ligands Decrease Viability of Myeloid and Results Lymphoid Leukemic Cells PPAR; Is Expressed in Lymphoid and Myeloid Leuke- The effects of several PPARg ligands on the viability of mic Cells leukemic cells are shown in Fig. 2. In U937 cells with The expression of PPARg was examined by Western blot increased expression of PPARg, BRL49653 decreased analysis in myeloid and lymphoid leukemic cell lines and viability in a dose-dependent fashion (Fig. 2A). However, in primary samples from leukemia patients. To confirm the a concentration of 50 Amol/L BRL49653 was required for a specificity of the reaction, a monoclonal antibody to PPARg 50% inhibition of growth in these cells. BRL49653 at 25 or antibody preadsorbed with blocking peptide was used Amol/L decreased viability by f50% in Raji, Su-DHL, Sup- for competition studies. PPARg was found to be expressed M2 cells and Hodgkin’s cells (Fig. 2B and C), but no effect in B and T lymphoid leukemias, lymphomas, Hodgkin’s was seen in HL-60 and Ramos cells (Fig. 2A and B). 15-d- disease, myeloma cell lines, and primary AML and CLL PGJ2 at 5 Amol/L consistently killed both lymphoid and cells (Fig. 1). PPARg was expressed at higher levels in myeloid cell lines. CDDO was the most potent agent in this myelomonocytic U937 than in myeloid HL-60 cells, group and decreased viability markedly at 1 Amol/L and confirming data published previously (44). PPARg protein almost completely eliminated the viable cell population at was also expressed in 9 of 11 primary AML samples with 2 Amol/L. high blast count (>50%; see examples in Fig. 1). Low To determine if PPARg ligands decrease viability by expression was noted in two of four samples from patients inducing apoptosis, we conducted a time course experi- with advanced myelodysplastic syndrome (refractory ment in U937 cells with CDDO (2 Amol/L). Discrete events anemia with excess blasts). We then compared PPARg of apoptosis were determined by multiparametric flow mRNA expression in normal cells and in primary AML cytometry. After 4-hour exposure to CDDO, 21% of cells samples. In normal peripheral blood (n = 3) and in CD34+ had lost their MMP (green cell population) and a small cells separated by apheresis (n = 2), PPARg was expressed proportion (f12%) was Annexin V positive. There was no at low levels (peripheral blood, mean 5.2 F 3.3 compared evidence of caspase activation at this time. After 12 hours, with a normal calibrator sample; CD34+ cells, 1.7 F 1.2). In 68% cells were MMP-low and f14% of these cells were 15 of 20 primary AML samples, PPARg expression was CaspaTag positive (blue cell population) and 12% were increased (>2-fold) compared with normal CD34+ cells. In Annexin V positive, respectively. After 24 hours, half of the particular, PPARg was highly expressed in six of seven MMP-low cells became CaspaTag positive and Annexin V AML-M4 specimens, with 10,000-fold increased expression positive. The other half of this population lost MMP but in two cases (Table 1). showed no caspase activation and did not express

Mol Cancer Ther 2004;3(10). October 2004

Table 1. Relative PPARg mRNA expression levels in primary AML independent mechanism(s). This effect was more pro- samples nounced at higher concentrations where CDDO (z1 Amol/L) killed 84% of vector controls, 95% and 96% of Source % Blasts FAB PPARg Expression PPARg-overexpressing cells, and 74% of DN transfectants.

PB 49 M1 0.58 BM 80 M1 0.87 BM 87 M1 1.2 BM 53 M1 319.6 PB 86 M1 189.9 BM 96 M1 6.59 BM 54 M2 0.4 PB 36 M2 5.06 Ph 98 M1 8.78 Ph 97 M3 91.4 BM 89 M4 4.18 BM 92 M4 7.11 BM 58 M4 2.23 BM 73 M4 115.4 Ph 92 M4 36.3 BM 91 M4 10,405.0 BM 44 M4 10,016.0 PB 38 M6 17.44 PB 89 AML (FAB ND) 252.34 BM 48 AML (FAB ND) 13.7

NOTE: PPARg expression was analyzed by quantitative Taqman reverse transcription-PCR. A comparative C T method was used to calculate the relative amount of PPARg in the sample (X) normalized to endogenous h reference ( 2-microglobulin) and relative to the amount of PPARg in ÀDDCT calibrator sample (Y): 2 , where C T is the difference in threshold cycles h for PPARg and 2-microglobulin and DDC T for sample X = DC T ,X À DC T ,Y . One of the normal peripheral blood samples was designated as calibrator or the 1Â sample. FAB, French-American-British classification; PB, peripheral blood; BM, bone marrow; Ph, pheresis; ND, not determined. phosphatidylserine/Annexin V (Fig. 3A). These data suggested that CDDO induces mitochondrial depolarization followed by caspase activation in leukemic cells. Western blot analysis revealed that CDDO induced the cleavage of caspase-8, -9, and -3 after 24-hour exposure (Fig. 3B). Similar data were obtained in lymphoid Ramos cells (data not shown). Examining the Relationship between PPAR; Expres- sion and Apoptosis Induction by PPAR; Ligands To characterize the correlation between PPARg expression and antileukemic activity of certain PPARg ligands, we tested the sensitivity of PPARg-overexpressing leukemic cells to CDDO-induced killing. We employed HL-60- CDM-1 cells for these studies because of their low endogenous levels of PPARg protein (23). Vector control (pcDNA3) or wild-type PPARg-transfected HL-60-CDM-1 cells (two different clones) were treated with 0.3 and 0.5 Amol/L CDDO for 48 hours. Apoptosis was determined by Annexin V staining using flow cytometry. Whereas CDDO induced apoptosis in the vector control cells, forced Figure 2. PPARg ligands decrease viability of myeloid and lymphoid g leukemic cells. U937, HL-60 (A), non-Hodgkin’s (B), or Hodgkin’s (C) overexpression of PPAR doubled the sensitivity of HL- lymphoma cell lines were treated with BRL49653 (BRL;25or50Amol/L), 60-CDM-1 cells to CDDO-induced killing (Fig. 4A). HL-60- 15-d-PGJ2 (PGJ2;5Amol/L), or CDDO (1 or 2 Amol/L) for 5 days. Viability CDM-1 cells transfected with a DN-PPARg mutant that was determined by the ViaLight assay as described in Materials and functions as an inhibitor (41) were much less sensitive than Methods. Columns, mean percentage compared with control (untreated) cultures, with each experimental point involving triplicate wells; bars, SD. the wild-type transfected cells (Fig. 4A). However, low DMSO was used as solvent control. In all four non-Hodgkin’s lymphoma levels of apoptosis were still observed suggesting PPARg- cell lines, <1% of cells were viable after incubation with 15-d-PGJ2.

Mol Cancer Ther 2004;3(10). October 2004

Figure 3. CDDO induces caspase activation and mitochondrial depolarization in U937 cells. Apoptosis was determined in U937 cells after 4, 12, and 24 hours of treatment with 2 Amol/L CDDO. A, in a multicolor flow cytometry assay, chloromethyl X-rosamine (CMXRos) measures changes in MMP; Annexin V staining determines changes in phosphatidylserine expression on the plasma membrane; and CaspaTag examines caspase activation. Y axis, chloromethyl X-rosamine; X axis, CaspaTag (top) and Annexin V (bottom). Red, live cells preserving their MMP; they are negative for CaspaTag and Annexin V. Green, cells that lose their MMP; at 4 hours, they are CaspaTag negative, but a proportion are Annexin V positive (in vehicle-containing cultures, 5.5% cells were Annexin V positive and 4.1% CaspaTag positive). At 12 hours, the activation of caspases is detected by the CaspaTag assay (blue, 14%). These cells have low MMP and are Annexin V positive. At 24 hours, massive caspase activation and Annexin positivity is noted, but a proportion of chloromethyl X-rosamine-low CaspaTag/Annexin-negative cells exist. B, in the same experiment, the cleavage of caspase-8, -9, and -3 was studied by Western blot analysis at 24 hours. Actin was used as loading control.

In an alternative approach, we used the novel selective PPAR; and RXR Ligand Combinations Are More Po- PPARg antagonist T007 (27). T007 completely abrogated tent than PPAR; Ligands Alone in Inhibiting Prolifera- the CDDO- and 15-d-PGJ2-induced transactivation of tion and Inducing Differentiation and Apoptosis in PPRE-Luc in MCF-7 cells transfected with PPRE-Luc Leukemic Cells expression plasmid (Fig. 4B). Pretreatment of U937 cells PPARg forms heterodimers with RXR and the ligation with 2 Amol/L T007 partially inhibited myeloid differen- of both receptors enhances their antidiabetic effect in vivo tiation induced by CDDO or 15-d-PGJ2. CD11b was (45). RXR is expressed at ubiquitously high levels in all expressed on >90% of control cells. However, CDDO and hematopoietic cells examined in this study (data not 15-d-PGJ2 increased the mean fluorescence intensity of shown). We therefore evaluated the effects of combined CD11b staining (mean fluorescence intensity, CDDO 73 PPARg and RXR ligation. First, we examined the ability of versus DMSO 23, mean fluorescence intensity = 80 for 15-d- synthetic PPARg ligands alone to induce differentiation in PGJ2) and T007 diminished it (mean fluorescence intensity HL-60 cells. CDDO, 15-d-PGJ2, and BRL49653 induced = 53). 15-d-PGJ2, but not CDDO, increased the percentage myelomonocytic differentiation in HL-60 cells as shown by of cells expressing CD14 differentiation marker (55% morphology, NBT reduction assay, and induction of the compared with 0.1% in DMSO) and this effect was cell surface CD11b and CD14 differentiation markers. In abrogated by T007 (5% CD14+ cells). this respect, CDDO was much more potent on a molar basis Although T007 significantly diminished apoptosis in- than 15-d-PGJ2 and BRL49653 (Fig. 5). duced by low concentrations of CDDO and 15-d-PGJ2 We then combined 15-d-PGJ2 or CDDO with the RXR- (Fig. 4C), it only marginally decreased cell death when specific ligand LG100268 (100 nmol/L). In HL-60 cells, PPAR ligands were used at higher concentrations, suggest- the combination of LG100268 with 15-d-PGJ2 or CDDO ing again the existence of a potential PPAR-independent induced pronounced myelomonocytic differentiation mechanism of apoptosis. (Table 2), with CDDO being much more potent than

Mol Cancer Ther 2004;3(10). October 2004

cell growth minimally (13-20%), whereas in combination with LG100268 (100 nmol/L) inhibition increased to 50%. The induction of cell differentiation also resulted in decreased viability. The trypan blue exclusion assay showed no decrease in cell viability when BRL49653 or LG100268 were used alone, whereas a decrease of >50% was noted when both ligands were combined. Similarly, RXR ligation potentiated the growth-inhibitory effect of 15- d-PGJ2 and CDDO. The RXR-specific ligand LG100268 also enhanced CDDO-induced apoptosis in five primary AML samples (DMSO, 24.7 F 6.3%; Annexin V positivity: LG100268, 32.4 F 6.7%; 2 Amol/L CDDO, 57.4 F 6.1%; CDDO plus LG100268, 73 F 9%; P < 0.02 compared with CDDO alone). Combined PPAR; andRetinoicAcidReceptorA Liga- tion (Troglitazone and ATRA) Induces Differentiation and Inhibits Clonogenic Growth of Myeloid Leukemic Cells Because ATRA has been shown to enhance the growth- inhibitory effects of PPARg ligands (44, 46, 47), we examined the combined effects of ATRA and the PPARg ligand troglitazone in myeloid leukemias. A 3-day exposure of U937 cells to either ATRA (1 Amol/L) or troglitazone (10 Amol/L) resulted in 28% and <5% NBT- positive cells, respectively. In contrast, the combination of troglitazone and ATRA induced 78% NBT-positive cells. Whereas CD14 expression did not change, CD11b was induced by ATRA (57%) and more so by ATRA plus troglitazone (95%; data not shown). These phenotypical changes were associated with morphologic evidence of differentiation (Fig. 7A). Controls and troglitazone-exposed U937 cells displayed the characteristic morphology of myelomonocytic leukemic cells, with round nuclei containing condensed chromatin surrounded by a rim of baso- philic cytoplasm. Cells cultured with ATRA alone exhibited morphologic signs of differentiation in that these cells had irregularly shaped nuclei, limited chromatin condensation, and an apparent decrease in the nuclear to cytoplasmic ratio. In contrast, cells cultured with both ATRA and Figure 4. Apoptosis induced by CDDO and 15-d-PGJ2 is regulated by troglitazone developed condensed, lobulated nuclei char- PPARg expression. A, HL-60-CDM-1 cells were stably transfected with acteristic of cells undergoing granulocyte-like differentia- wild-type (wt) or DN-PPARg (DN) and treated with 0.3 (white bars) and 0.5 (black bars) Amol/L CDDO, and apoptosis was determined by Annexin tion. In addition, the nuclei were surrounded by the V flow cytometry. Columns, net apoptosis induction after subtracting abundant foamy cytoplasm frequently observed in macro- background apoptosis in the presence of vehicle (0.1% DMSO). pcDNA3, phage-differentiated monocytes. Similar changes were vector control; two different clones of each wild-type and DN trans- observed in THP-1 leukemic cells (data not shown). fectants were tested. Inset, expression of FLAG epitope detected by f immunoblotting. B, MCF-7 cells were transiently transfected with 1 Agof NSE is a monocyte-specific enzyme expressed in 3% TK-PPRE-Luc construct. After 24 hours, transfected cells were treated of U937 and 13% of THP-1 cells. No significant change in with 2 Amol/L CDDO or 5 Amol/L 15-d-PGJ2 alone or in combination with enzyme activity was observed after incubation of cells 20 Amol/L T007. Columns, mean luciferase activity (n = 3); bars, SD. C, U937 cells were treated with 2 Amol/L PPARg antagonist T007 followed with either ATRA or troglitazone. However, the combina- by exposure to 1 Amol/L CDDO or 3 Amol/L 15-d-PGJ2. Apoptosis was tion of both compounds greatly increased the percentage determined by Annexin V/propidium iodide flow cytometry after 48 hours of NSE-positive cells (Table 3). Likewise, the ability U937 of treatment. and THP-1 cells to phagocytose C. albicans particles increased significantly after a combined treatment of ATRA 15-d-PGJ2, at much lower concentrations. Accordingly, and troglitazone (Fig. 7B; Table 3) as compared with the when PPARg ligands were combined with LG100268, the effects of either compound alone. The two-layer soft agar proliferation of HL-60 cells was more inhibited than by any system was used to study the effects of troglitazone and of the ligands alone (Fig. 6). As an example, BRL49653 ATRA on clonogenic growth of leukemic cells. Whereas (rosiglitazone) alone, used at 10 and 25 Amol/L, decreased troglitazone alone used at 10À8 to 10À5 mol/L did not

Mol Cancer Ther 2004;3(10). October 2004

Figure 5. PPARg ligands induce maturation of HL-60 cells. Left, morphology of HL-60 cells in control cultures and after incubation with indicated concentrations of PPARg ligands (7 days, Â600 mag- nification). Right, NBT staining with percentage of NBT-positive cells counted manually. Percentage of cells expressing CD11b and CD14 differentiation markers was analyzed by flow cytometry.

significantly affect clonal proliferation of U937 and THP-1 We then analyzed the interactions between ATRA and cells, colony formation was significantly diminished by CDDO in U937 cells using isobologram Calcusyn software. the combination of troglitazone with ATRA (Table 3; CDDO and ATRA were used at a fixed 1:1 ratio at 0.5, 0.75, troglitazone plus ATRA versus troglitazone or ATRA, and 1 Amol/L concentrations for 48 hours. The effects on P < 0.02). cell proliferation and apoptosis were assessed by cell

Table 2. Induction of differentiation in HL-60 cells by PPAR; and RXR ligation

+ CD11b/CD14 Vehicle 15-d-PGJ2 (1 Amol/L) 15-d-PGJ2 (3 Amol/L) 15-d-PGJ2 (5 Amol/L) CDDO (0.3 Amol/L) CDDO (0.6 Amol/L)

Alone 2.7 F 0.1 3.6 F 0.4 15.3 F 0.1 21.5 F 5.1 15.4 F 2.6 25.3 F 2.6 + LG 100268 4.9 F 1.3 10.3 F 4.7 19.3 F 2.9 24.8 F 5.9 55.1 F 3.4 96.2 F 1.6

NOTE: Induction of differentiation was analyzed using the phycoerythrin-conjugated anti-CD11b and FITC-conjugated anti-CD14 monoclonal antibodies. Data represent means F SD of the percentage of double-positive cells compared with isotype control antibodies from three independent experiments.

Mol Cancer Ther 2004;3(10). October 2004

additional CLL samples with the combination of CDDO and a Fas agonistic antibody (CH11) or TRAIL, which induce apoptosis through the extrinsic pathway (49, 50), and the small-molecule Bcl-2 inhibitor HA14-1 that is believed to regulate the intrinsic pathway (51–53). Six of the primary CLL samples were from patients that relapsed after treatment with fludarabine combined with other agents, and sample 8 was derived from a patient with refractory T-cell CLL. CDDO alone at 1 Amol/L induced apoptosis in five of nine samples tested, including samples from two fludarabine-resistant patients and one patient with refractory T-cell CLL (Table 4). The RXR-selective ligand LG100268, being nontoxic alone, enhanced the induction of apoptosis by CDDO in five of nine CLL samples, including two samples in which CDDO alone did not induce cell death. Activation of the extrinsic apoptotic pathway by CH11 or TRAIL induced moderate cell death in only two of nine CLL samples. When combined with CDDO, CH11 Figure 6. Effect of PPARg ligands either alone or in combination with did not affect cytotoxicity, but TRAIL enhanced CDDO- LG100268 on the viability of HL-60 cells. HL-60 cells were incubated with induced killing in two of nine samples. In contrast, HA14-1 the indicated concentrations of 15-d-PGJ2, BRL49653, or DMSO (control) for 6 days, either alone (5) or combined with 100 nmol/L LG100268 (n). was cytotoxic in four of nine samples and potentiated the The effect on proliferation was examined using the [3H]thymidine CDDO-induced apoptosis in all samples tested (Fig. 9A). incorporation assay. Furthermore, the combined treatment induced pronounced cell death in the four samples in which CDDO alone counts and Annexin V flow cytometry, respectively. was ineffective. In all CLL samples, we found expression Isobologram analysis (43) showed that the proapoptotic of Bcl-2 protein, as shown in Fig. 9B, and there was no interaction between CDDO and ATRA was synergistic change in Bcl-2 expression with any of the treatments (ED50, CI = 0.861; ED75, CI = 0.588; ED90, CI = 0.683). (Fig. 9B). To further define the mechanisms by which cells PPAR; Ligands Induce Apoptosis in Primary Leukemic are sensitized to the proapoptotic effects of CDDO, we but not in Normal Cells compared the cleavage of caspase-8 and -9 in four CLL We used samples from CLL patients to examine the samples treated with the CDDO/HA14-1 combination. In sensitivity of primary leukemic cells to the growth- all four samples, CDDO efficiently cleaved caspase-8. inhibitory and apoptogenic effects of PPARg ligands. The Nevertheless, CDDO alone did not induce apoptosis in PPARg ligands 15-d-PGJ2 (5 Amol/L) and BRL49653 (25 two of these samples, suggesting that the additional Amol/L) significantly decreased the viability of CLL cell cleavage of caspase-9, facilitated by functional Bcl-2 samples by inducing apoptosis (Fig. 8A and B) as shown by blockade, was required for full proapoptotic effect of the Annexin V flow cytometry. drug. Examples of CDDO-sensitive (CLL1) and CDDO- We then tested the effects of CDDO on primary CLL. CDDO resistant (CLL2) samples are shown in Fig. 9B. This reportedly induced apoptosis in CLL cells, including samples observation in CLL suggests that CDDO induces apoptosis from patients with fludarabine-resistant CLL, by activating the via the activation of components of both extrinsic and caspase-8 and -3 (48). In our study, a similar response was intrinsic apoptotic pathways. observed in a total of seven primary CLL samples. CDDO induced dose-dependent inhibition of proliferation in CLL cells (data not shown), which was accompanied by the concomitant induction of apoptosis (Fig. 8C). Discussion In contrast, viability of normal CD34+ cells in the This study was conducted to characterize the activity of presence of 15-d-PGJ2 (5 Amol/L) was only moderately PPARg nuclear receptor signaling pathway in leukemia decreased, whereas BRL49653 at pharmacologic concentra- cells. Leukemic cell lines showed high expression of PPARg tion (25 Amol/L) was ineffective (data not shown). These protein in both myeloid and lymphoid cells including data were confirmed by Annexin V flow cytometry Hodgkin’s disease and multiple myeloma cells. Although (Fig. 8D). CDDO induced significantly less apoptosis in the presence of PPARg has been reported in leukemic cell CD34+ cells compared with CLL, suggesting differential lines (44, 54, 55), we report here the frequent expression of killing of leukemic versus normal progenitor cells (Fig. 8D). PPARg in primary leukemia samples including AML and CDDO-Induced Apoptosis in CLL Samples Is Enhanced CLL. PPARg mRNA expression in primary AML was by Simultaneous RXRLigation or Bcl-2 Blockade highest in myelomonocytic AML (M4) subtypes. These data To further delineate the mechanism(s) by which cells are suggest that CDDO and other PPARg ligands might be sensitized to CDDO-induced apoptosis, we treated nine particularly useful in the treatment of myelomonocytic

Mol Cancer Ther 2004;3(10). October 2004

Figure 7. PPARg ligand troglitazone (TGZ) combined with ATRA induce differentiation and phagocytic ability of leukemic cells. A, morphology of U937 cells in control cultures and after incubation with ATRA (1 Amol/L), troglitazone (10 Amol/L), or their combination for 6 days. B, ability of THP-1 and U937 cells to phagocytose C. albicans.

subtypes of AML/myelodysplastic syndrome. This was myelomonocytic differentiation of HL-60 cells, with CDDO also noted in a recent clinical trial in which PPARg ligands being the most potent. CDDO and 15-d-PGJ2 also induced exerted antileukemia activity in patients with chronic apoptosis in leukemic cells, whereas rosiglitazone only myelomonocytic leukemia (56). seemed to inhibit proliferation. It has been reported that PPARg ligands can induce The expression of PPARg in lymphoid malignancies leukemic cells to differentiate toward macrophages (44, 57). provides a potential target for novel therapeutic strategies In our study, 15-d-PGJ2, BRL49653, and CDDO induced for these malignancies as well. Although the proapoptotic

Mol Cancer Ther 2004;3(10). October 2004

Table 3. Combination of troglitazone and ATRA induces differ- which is dysregulated in leukemias, provides an attractive entiation and inhibits clonogenic growth of myeloid leukemic target of CDDO-mediated cytotoxicity, the simultaneous cells inhibitions of Bcl-2 expression or function resulted in

c highly increased induction of programmed cell death in NSE* Phagocytosis Colony lymphoid cells with high Bcl-2 levels. Formationb Because CDDO reportedly binds to and transactivates the nuclear receptor PPARg (12), we explored the correla- U937 Control 3 11 100 tion between PPARg expression levels and sensitivity to Troglitazone 10À5 mol/L 5 15 100 F 19 ATRA 10À6 mol/L 6 10 95 F 21 CDDO. Our experiments using genetically modified HL-60- Troglitazone + ATRA 12 28 60 F 17 CDM-1 cells showed sensitization to the cytotoxic effects of THP-1 Control 13 8 100 CDDO in cells transfected with wild-type PPARg. This Troglitazone 10À5 mol/L 14 16 90 F 19 effect was significantly diminished in DN-PPARg trans- ATRA 10À6 mol/L 5 38 50 F 16 fectants. The DN-PPARg receptor reportedly retains ligand Troglitazone + ATRA 40 80 22 F 9 and DNA binding but exhibits reduced transactivation and activity due to impaired coactivator recruitment (41). NOTE: Results represent the mean of three independent experiments. Alternatively, pretreatment of U937 cells with the specific *Percentage of NSE-positive cells. c Percentage of cells phagocytosing C. albicans. PPARg antagonist, T007, diminished CDDO- and 15-d- b Percentage of leukemic colonies compared with vector control (100%). PGJ2-induced apoptosis and differentiation, suggesting that these effects are PPARg dependent. However, at higher properties of 15-d-PGJ2 have been described previously (54, concentrations, CDDO was capable of inducing apoptosis 55, 58, 59), here we report that CDDO is more potent than in cells carrying the DN mutant, and blockade of PPARg 15-d-PGJ2 in triggering apoptosis in lymphoid cells did not prevent cytotoxic effects of the compound, including Hodgkin’s and myeloma cells. Moreover, CDDO suggesting existence of other, yet unidentified, cellular was effective in primary refractory T-cell CLL, confirming targets. The induction of apoptosis by PPARg agonists may recently published data showing that PPARg agonists therefore be independent of PPARg modulation as was exhibit selectivity in apoptosis induction in transformed, reported for some other biological effects of PPARg ligands but not normal, T-lineage cells (59). Apoptosis seems the (60, 61). For example, a recent study showed that, at primary mechanism of CDDO. Whereas activation of concentrations that induce optimal transcriptional activa- certain components of the extrinsic pathway (i.e., caspase-8), tion of PPARg, thiazolidinedione protected T cells from

Figure 8. PPARg ligands decrease viability of primary CLL cells. A, cells from six primary CLL samples were cultured in the presence of BRL49653 (25 Amol/L) or 15-d-PGJ2 (5 Amol/L) for 3 days. Viability was determined by ViaLight assay as described in Materials and Methods. Col- umns, mean percentage compared with DMSO-treated control cultures; bars, SD. B, induction of apoptosis was determined by Annexin V flow cytometry as described in Materials and Methods. C, cells from seven primary CLL samples were cultured in the presence of indicated concentrations of CDDO. The induction of apoptosis was determined by Annexin V flow cytometry. Columns, mean of cell numbers or the percentage of apoptotic cells in seven different samples compared with untreated control cultures; bars, SD. D, effects of PPARg ligands on normal cells. For BRL49653 and 15-d-PGJ2 experiments, purified CD34+ cells (n = 4) were used. For CDDO experiments, purified CD34+ cells from six apheresis samples were examined. Induction of apoptosis was determined by Annexin V flow cytometry.

Mol Cancer Ther 2004;3(10). October 2004

Table 4. Induction of apoptosis by CDDO is enhanced by Bcl-2 inhibitor HA14-1 in primary CLL samples

Patient Vehicle CDDO LG 100268 CH11 TRAIL HA14-1 CDDO+ CDDO+ CDDO+ CDDO+ (DMSO) (1 Amol/L) (1 Amol/L) (50 ng/mL) (50 ng/mL) (15 Amol/L) LG1002 68 CH11 HA14-1 TRAIL

1 14.8 20.1 14.9 8.8 13 26.7 21.6 12.9 88.1 43 2 3.6 6.1 4.9 2.6 2.8 7.7 15.1 5.4 35 54.6 3 7.2 5.1 8.6 ND 20 32.8 10.5 14.3 42.2 13.8 4 5.4 8.0 5.5 4.5 5 6.7 12.8 9.9 53.5 12.4 54.542.6 5.2 5.5 5.9 75.7 48.3 39.8 93.6 42.3 63.174.1 6.4 2.5 0.8 12.7 21.8 4.8 99.3 6.5 7 5.6 10.5 10.3 11.7 6.5 58.4 22.2 15.1 83.5 12.9 8 18.2 53.5 28.2 28.3 30.9 19.6 63.2 59.2 59.9 59.1 92.369 4.4 2.8 2 6.9 80.9 64.6 95.7 59.4 Mean F SEM 7.2 F 1.8 32.1 F 9.3 9.8 F 2.6 8.3 F 3.1 9.7 F 3.3 27.5 F 8.2 32.9 F 8.4 25.1 F 7.8 72.3 F 8.3 33.8 F 7.4

NOTE: Induction of apoptosis in CLL B-cells was determined by double staining of CD19 surface antigen and phosphatidylserine/Annexin V. Samples 1, 3, and 4 were obtained from previously untreated CLL patients and samples 2, 5, 6, and 9 from fludarabine-resistant CLL patients. Sample 8 was from T-cell CLL patient. In samples 1, 2, and 6, spontaneous apoptosis in medium only was similar to DMSO-containing medium (19.1%, 3%, and 2.5%, respectively). ND, not done (increased induction of apoptosis is denoted in bold).

growth factor deprivation, whereas it induced apoptosis when used at higher concentrations (62). Similarly, 15-d- PGJ2 or thiazolidinedione were capable of decreasing the release of inflammatory cytokines (60) or causing G1 cell cycle arrest (24) in PPARg-null embryonic stem cells. Conclusive proof of PPARg dependence will come from testing PPARg ligands in vivo in a PPAR-deficient background or in vitro by silencing expression of PPARg using small interfering RNA. The combination of the PPARg ligand troglitazone and certain retinoids reportedly suppressed colony formation of leukemic cells in an additive fashion (44). Because PPARg and RXR are known to function as heterodimers, we hypothesized that combined treatment with PPARg and RXR-specific ligands would significantly potentiate the antileukemic effect of PPARg ligands alone. Indeed, the combination of PPARg and RXR activation (using the specific ligand LG100268) profoundly induced differentiation in HL-60 cells. Furthermore, LG100268 enhanced CDDO- induced apoptosis in an additive fashion in primary AML and CLL samples. Of note, the RXR agonist was used at only 100 nmol/L, a concentration at which the compound alone was only minimally active in all cell systems tested. The combined use of ligands for PPARg and RXR has been shown to induce the synergistic transactivation of re- porter genes and enhance adipocyte differentiation and differentiation of liposarcoma in vivo (63, 64). The anti- tumor activity of LGD1069, which is an analogue of LG100268, has been reported in clinical trials (65–67). Further studies are needed to determine if the cooperative Figure 9. Blockade of Bcl-2 with HA14-1 synergistically enhances CDDO-induced apoptosis in primary CLL cells via caspase activation. A, effects of CDDO and rexinoids result from the recruitment cells from nine primary CLL samples were cultured with 50 ng/mL CH11, 50 of selective coactivators [i.e., p160 by RXR and DRIP205 ng/mL TRAIL, 15 Amol/L HA14-1 (HA), 1 Amol/L CDDO, either alone or in by PPARg (68)]. combination, for 72 hours. Apoptosis was determined by Annexin V flow cytometry after gating on CD19+ CLL cells. B, Western blot analysis of The combination of ATRA and troglitazone was required caspase activation in two CLL samples treated with CDDO, HA14-1, or their to induce terminal differentiation in myeloid leukemic cell combination. Cells from patient 1 were sensitive, whereas cells from patient lines as shown by a variety of functional assays (NBT, NSE, 2 were resistant to CDDO-induced apoptosis. Bottom, percentage of CD19/ Annexin V-positive cells. The cleavage of caspase-8, -9, and -3 was CD11b, and the ability to phagocytose C. albicans). Trogli- determined as described in Materials and Methods. Bcl-2 protein levels tazone and ATRA combined synergistically decreased the showed no change with either treatment. Actin was used as loading control. clonogenic capacity of myelomonocytic leukemia lines.

Mol Cancer Ther 2004;3(10). October 2004

ATRA binds to all the three retinoic acid receptors and 3,12-dioxooleana-1,9-dien-28-oic acid (CDDO), is a ligand for the peroxisome proliferator-activated receptor g. Mol Endocrinol 2000;14:1550 – 6. directly activates them (69); ATRA does not bind to RXRs 13. Demetri GD, Fletcher CD, Mueller E, et al. Induction of solid tumor but shows RXR-stimulating activity in transactivation differentiation by the peroxisome proliferator-activated receptor-g ligand assays perhaps due to its isomerization to 9-cis-retinoic troglitazone in patients with liposarcoma. Proc Natl Acad Sci U S A acid (70, 71). This may explain the observed synergism 1999;96:3951 – 6. 14. Mueller E, Sarraf P, Tontonoz P, et al. Terminal differentiation of between troglitazone and ATRA. Alternatively, ‘‘activated’’ human breast cancer through PPAR g. Mol Cell 1998;1:465 – 70. g PPAR /RXR heterodimers are able to cross-bind and 15. Chang T-H, Szabo E. Induction of differentiation and apoptosis by activate a retinoic acid response element h (72, 73). Whereas ligands of peroxisome proliferator-activated receptor g in non-small cell ATRA alone did not show activity in clinical trials in lung cancer. Cancer Res 2000;60:1129 – 38. patients with AML (other than acute promyelocytic 16. Tsubouchi Y, Sano H, Kawahito Y, et al. Inhibition of human lung cancer cell growth by the peroxisome proliferator-activated receptor-g leukemia; refs. 74, 75), the combination of retinoids with agonists through induction of apoptosis. Biochem Biophys Res Commun PPARg ligands may be beneficial. In summary, our results 2000;270:400 – 5. suggest that PPARg ligands combined with clinically 17. Guan YF, Zhang YH, Breyer RM, Davis L, Breyer MD. Expression of available RXR ligands have potential utility for the peroxisome proliferator-activated receptor g (PPARg) in human transitional bladder cancer and its role in inducing cell death. Neoplasia 1999;1:330 – 9. treatment of a broad spectrum of hematologic malignan- 18. Padilla J, Kaur K, Harris SG, Phipps RP. PPAR-g-mediated regulation of cies. Of importance, significantly less toxicity against normal and malignant B lineage cells. Ann N Y Acad Sci 2000;905:97 –109. normal progenitor cells compared with leukemias was 19. Hisatake J-I, Ikezoe T, Carey M, Holden S, Tomoyasu S, Koeffler HP. observed, suggesting differential killing of leukemic cells. Down-regulation of prostate-specific antigen expression by ligands for peroxisome proliferator-activated receptor g in human prostate cancer. A clinical phase I trial combining BRL49653 (rosiglitazone) Cancer Res 2000;60:5494 – 8. with the RXR-ligand bexarotene (Targretin) was recently 20. Motomura W, Okumura T, Takahashi N, Obara T, Kohgo Y. Activation initiated at University of Texas M.D. Anderson Cancer of peroxisome proliferator-activated receptor g by troglitazone inhibits cell Center and will test the safety and biological and clinical growth through the increase of p27Kip1 in human pancreatic carcinoma cells. Cancer Res 2000;60:5558 – 64. effects of this concept in hematopoietic malignancies. 21. Kitamura S, Miyazaki Y, Shinomura Y, Kondo S, Kanayama S, Matsuzawa Y. Peroxisome proliferator-activated receptor g induces growth arrest and differentiation markers of human colon cancer cells. Jpn J Cancer Res 1999;90:75 – 80. Acknowledgments 22. Szondy Z, Reichert U, Bernardon JM, et al. Induction of apoptosis by We thank Dr. Numsen Hail for the critical review of the article and Rose retinoids and retinoic acid receptor g-selective compounds in mouse Lauzon for the excellent administrative support. thymocytes through a novel apoptosis pathway. Mol Pharmacol 1997;51: 972 – 82. 23. Nagy L, Thomazy VA, Shipley GL, et al. Activation of retinoid X receptors References induces apoptosis in HL-60 cell lines. Mol Cell Biol 1995;15:3540– 51. 1. Sporn MB, Suh N, Mangelsdorf DJ. Prospects for prevention and 24. Palakurthi SS, Aktas H, Grubissich LM, Mortensen RM, Halperin JA. treatment of cancer with selective PPARg modulators (SPARMs). Trends Anticancer effects of thiazolidinediones are independent of peroxisome Mol Med 2001;7:395 – 400. proliferator-activated receptor g and mediated by inhibition of translation initiation. Cancer Res 2001;61:6213 – 8. 2. Mangelsdorf DJ, Thummel C, Beato M, et al. The nuclear receptor superfamily: the second decade. Cell 1995;83:835 – 9. 25. Honda T, Rounds BV, Gribble GW, Suh N, Wang Y, Sporn MB. Design and synthesis of 2-cyano-3,12-dioxoolean-1,9-dien-28-oic acid, a novel Castaigne S, Chomienne C, Daniel MT, et al. All- retinoic acid as a 3. trans and highly active inhibitor of nitric oxide production in mouse macro- differentiation therapy for acute promyelocytic leukemia. I. Clinical results. phages. Bioorg Med Chem Lett 1998;8:2711 – 4. Blood 1990;76:1704 – 9. 26. Boehm M, Zhang L, Zhi L, et al. Design and synthesis of potent 4. Warrell RP Jr, Frankel SR, Miller WH Jr, et al. Differentiation therapy of retinoid X receptor selective ligands that induce apoptosis in leukemia acute promyelocytic leukemia with tretinoin (all-trans-retinoic acid). N cells. J Med Chem 1995;38:3146 – 55. Engl J Med 1991;324:1385 – 93. Lee G, Elwood F, McNally J, et al. T0070907, a selective ligand for 5. Nolte RT, Wisely GB, Westin S, et al. Ligand binding and co-activator 27. peroxisome proliferator-activated receptor g, functions as an antagonist of assembly of the peroxisome proliferator-activated receptor-g. Nature 1998;395:137 – 43. biochemical and cellular activities. J Biol Chem 2002;277:19649 – 57. 6. Forman BM, Chen J, Evans RM. Hypolipidemic drugs, polyunsaturated 28. Andreeff M, Leysath C, Konopleva M, Huang Z. Induction of fatty acids, and eicosanoids are ligands for peroxisome proliferator- apoptosis in AML by HA14-1, A cell-permeable organic compound activated receptors a and y. Proc Natl Acad Sci U S A 1997;94:4312 – 7. identified by protein structure-based computer screening that binds the BH1 to BH3 surface pocket of Bcl-2. Blood 2000;96:502a. 7. Forman BM, Tontonoz P, Chen J, Brun RP, Spiegelman BM, Evans RM. 15-Deoxy-12,14 prostaglandin J2 is a ligand for the adipocyte determi- 29. Wang JL, Liu D, Zhang ZJ, et al. Structure-based discovery of an nation factor PPAR. Cell 1995;83:803 – 12. organic compound that binds Bcl-2 protein and induces apoptosis of tumor cells. Proc Natl Acad Sci U S A 2000;97:7124 – 9. 8. Kliewer SA, Lenhard JM, Willson TM, Patel I, Morris DC, Lehman JM. A prostaglandin J2 metabolite binds peroxisome proliferator-activated 30. Morris SW, Kirstein MN, Valentine MB, et al. Fusion of a kinase gene, receptor and promotes adipocyte differentiation. Cell 1995;83:813 – 9. Alk, to a nucleolar protein gene, Npm, in non-Hodgkin’s-lymphoma. Science 1994;263:1281 – 4. 9. Nagy L, Tontonoz P, Alvarez JG, Chen H, Evans RM. Oxidized LDL regulates macrophage gene expression through ligand activation of PPAR. 31. Drexler HG. Recent results on the biology of Hodgkin and Reed- Cell 1998;93:229 – 40. Sternberg cells. II. Continuous cell lines. Leuk Lymphoma 1993;9:1 – 25. 10. Henke BR, Blanchard SG, Brackeen MF, et al. N-(2-benzoylphenyl)-L- 32. Miltenyi S, Muller W, Weichel W, Radbruch A. High gradient magnetic tyrosine PPARg agonists. 1. Discovery of a novel series of potent cell separation with MACS. Cytometry 1990;11:231 – 8. antihyperglycemic and antihyperlipidemic agents. J Med Chem 1998; 33. Slater K. Cytotoxicity tests for high-throughput drug discovery. Curr 41:5020 – 36. Opin Biotechnol 2001;12:70 – 4. 11. Rocchi S, Picard F, Vamecq J, et al. A unique PPARg ligand with potent 34. Munker R, Norman A, Koeffler HP. Vitamin-D compounds—effect on insulin-sensitizing yet weak adipogenic activity. Mol Cell 2001;8:737 – 47. clonal proliferation and differentiation of human myeloid cells. J Clin Invest 12. Wang Y, Porter WW, Suh N, et al. A synthetic triterpenoid, 2-cyano- 1986;78:424 – 30.

Mol Cancer Ther 2004;3(10). October 2004

35. Collins SJ, Ruscetti FW, Gallagher RE, Gallo RC. Normal functional lymphomas express PPAR-g and are killed by PPAR-g agonists. Clin characteristics of cultured human promyelocytic leukemia cells (HL-60) Immunol 2002;103:22 – 33. after induction of differentiation by dimethylsulfoxide. J Exp Med 1979; 56. Denzlinger C, Dittman K, Mohle A, et al. Peroxisome proliferator 149:969 – 74. activated receptor g (PPARg) agonists as antileukemic agents in CMML. 36. Drach J, McQueen T, Engel H, Andreeff M, Maise R, Braylan R. Blood 2001;98:625a. Retinoic acid-induced expression of CD38 antigen in myeloid cells is 57. Tontonoz P, Nagy L, Alvarez JGA, Thomazy VA, Evans RM. PPAR mediated through retinoic acid receptor-a. Cancer Res 1994;54:1746 – 52. promotes monocyte/macrophage differentiation and uptake of oxidized 37. Territo MC. Monocyte function in man. J Immunol 1997;118:187– 92. LDL. Cell 1998;93:241 – 52. 38. Clodi K, Kliche K-O, Zhao S, et al. Cell-surface exposure of 58. Schlezinger JJ, Jensen BA, Mann KK, Ryu HY, Sherr DH. Peroxisome phosphatidylserine correlates with the stage of fludarabine-induced proliferator-activated receptor g-mediated NF-nB activation and apoptosis apoptosis in chronic lymphocytic leukemia (CLL) and expression of in pre-B cells. J Immunol 2002;169:6831 – 41. apoptosis-regulating genes. Cytometry 2000;40:19 – 25. 59. Harris SG, Phipps RP. Prostaglandin D(2), its metabolite 15-d-PGJ(2), 39. Poot M, Pierce RH. Detection of changes in mitochondrial function and peroxisome proliferator activated receptor-g agonists induce apoptosis during apoptosis by simultaneous staining with multiple fluorescent dyes in transformed, but not normal, human T lineage cells. Immunology and correlated multiparameter flow cytometry. Cytometry 1999;35:311– 7. 2002;105:23 – 34. 40. Schober-Ditmore W, Leysath CE, Konopleva M, Andreeff M. 60. Chawla A, Barak Y, Nagy L, Liao D, Tontonoz P, Evans RM. PPAR-g Simultaneous analysis of mitochondrial membrane potential, caspase dependent and independent effects on macrophage-gene expression in cleavage, phosphatidylserine (PS) and cell surface antigen expression of lipid metabolism and inflammation. Nat Med 2001;7:48 – 52. cells undergoing apoptosis. Cytometry 2002;77. 61. Moore KJ, Rosen ED, Fitzgerald ML, et al. The role of PPAR-g in 41. Gurnell M, Wentworth JM, Agostini M, et al. A dominant-negative macrophage differentiation and cholesterol uptake. Nat Med 2001;7:41– 7. peroxisome proliferator-activated receptor g (PPARg) mutant is a consti- 62. Wang YL, Frauwirth KA, Rangwala SM, Lazar MA, Thompson CB. tutive repressor and inhibits PPARg-mediated adipogenesis. J Biol Chem Thiazolidinedione activation of peroxisome proliferator-activated receptor 2000; 275:5754 – 9. g can enhance mitochondrial potential and promote cell survival. J Biol 42. Ausubel FM, Brent R, Kingston RE, et al. Current protocols in Chem 2002;277:31781 – 8. molecular biology. New York: Wiley Interscience; 2003. 63. Tontonoz P, Singer S, Forman BM, et al. Terminal differentiation of 43. Chou TC, Talalay P. Quantitative analysis of dose-effect relationships: human liposarcoma cells induced by ligands for peroxisome proliferator- the combined effects of multiple drugs or enzyme inhibitors. Adv Enzyme activated receptor and the retinoid X receptor. Proc Natl Acad Sci U S A Regul 1984;22:27 – 55. 1997;94:237 – 41. 44. Asou H, Verbeek W, Williamson E, et al. Growth inhibition of myeloid 64. Schulman IG, Shao G, Heyman RA. Transactivation by retinoid X leukemia cells by troglitazone, a ligand for peroxisome proliferator activated receptor-peroxisome proliferators-activated receptor (PPAR) heterodimers: receptor, and retinoids. Intl J Oncol 1999;15:1027 – 93. intermolecular synergy requires only the PPAR hormone-dependent 45. Mukherjee R, Davies PJ, Crombie DL, et al. Sensitization of diabetic activation function. Mol Cell Biol 1998;18:3483 – 94. and obese mice to insulin by retinoid X receptor agonists. Nature 1997; 65. Rizvi NA, Marshall JL, Dahut W, et al. A Phase I study of LGD1069 in 386:407 – 10. adults with advanced cancer. Clin Cancer Res 1999;5:1658 – 64. 46. Zang C, Wachter M, Liu H, et al. Ligands for PPARg and RAR cause 66. Khuri FR, Rigas JR, Figlin RA, et al. Multi-institutional phase I/II trial of induction of growth inhibition and apoptosis in human glioblastomas. oral bexarotene in combination with cisplatin and vinorelbine in previously J Neurooncol 2003;65:107 – 18. untreated patients with advanced non-small-cell lung cancer. J Clin Oncol 47. Elstner E, Muller C, Koshizuka K, et al. Ligands for peroxisome 2001;19:2626 – 37. proliferator-activated receptor g and retinoic acid receptor inhibit growth 67. Duvic M. Treatment of cutaneous T-cell lymphoma from a dermatol- and induce apoptosis of human breast cancer cells in vitro and in BNX ogist’s perspective. Clin Lymphoma 2000;1:S51 – 5. mice. Proc Natl Acad Sci U S A 1998;95:8806 – 11. 68. Yang W, Rachez C, Freedman LP. Discrete roles for peroxisome 48. Pedersen IM, Kitada S, Schimmer A, et al. The triterpenoid CDDO proliferator-activated receptor g and retinoid X receptor in recruiting nuclear induces apoptosis in refractory CLL B cells. Blood 2002;100:2965 – 72. receptor coactivators. Mol Cell Biol 2000;20:8008 – 17. 49. Yonehara S, Ishii A, Yonehara M. A cell-killing monoclonal antibody 69. Giguere V, Ong ES, Segui P, Evans RM. Identification of a receptor for (anti-Fas) to a cell surface antigen co-downregulated with the receptor of the morphogen retinoic acid. Nature 1987;330:624 – 9. tumor necrosis factor. J Exp Med 1989;169:1747 – 56. 70. Mangelsdorf DJ, Ong ES, Dyck JA, Evans RM. Nuclear receptor that 50. Suliman A, Lam A, Datta R, Srivastava RK. Intracellular mechanisms identifies a novel retinoic acid response pathway. Nature 1990;345: of TRAIL: apoptosis through mitochondrial-dependent and -independent 224 – 9. pathways. Oncogene 2001;20:2122 – 33. 71. Heyman RA, Mangelsdorf DJ, Dyck JA, et al. 9-cis retinoic acid is a 51. Hao JH, Yu M, Liu FT, Newland AC, Jia L. Bcl-2 inhibitors sensitize high affinity ligand for the retinoid X receptor. Cell 1992;68:397 – 406. tumor necrosis factor-related apoptosis-inducing ligand-induced apoptosis 72. Kliewer SA, Umesono K, Noonan DJ, Heyman RA, Evans RM. by uncoupling of mitochondrial respiration in human leukemic CEM cells. Convergence of 9-cis retinoic acid and peroxisome proliferator signaling Cancer Res 2004;64:3607 – 16. pathways through heterodimer formation of their receptors. Nature 52. An J, Chen Y, Huang Z. Critical upstream signals of cytochrome c 1992;358:771 – 4. release induced by a novel Bcl-2 inhibitor. J Biol Chem 2004;279: 73. Han S, Wada RK, Sidell N. Differentiation of human neuroblastoma by 19133 – 40. phenylacetate is mediated by peroxisome proliferator-activated receptor g. 53. Chen J, Freeman A, Liu J, Dai Q, Lee R-M. The apoptotic effect of Cancer Res 2001;61:3998 – 4002. HA14-1, a Bcl-2-interacting small molecular compound, requires Bax 74. Bassan R, Chiodini B, Lerede T, et al. Prolonged administration of all- translocation and is enhanced by PK11195. Mol Cancer Ther 2002;1: trans retinoic acid in combination with intensive chemotherapy and G-CSF 981 – 7. for adult acute myelogenous leukemia: single-center pilot study in different 54. Padilla J, Kaur K, Cao HJ, Smith TJ, Phipps RP. Peroxisome proliferator risk groups. Hematol J 2002;3:193 – 200. activator receptor-g agonists and 15-deoxy-D(12,14)-PGJ(2) induce 75. Estey EH, Thall PF, Pierce S, et al. Randomized phase II study of apoptosis in normal and malignant B-lineage cells. J Immunol 2000;165: fludarabine F cytosine arabinoside F idarubicin F all-trans retinoic acid F 6941 – 8. granulocyte-colony stimulating factor in poor prognosis newly-diagnosed 55. Padilla J, Leung E, Phipps RP. Human B lymphocytes and B non-APL AML and MDS. Blood 1999;93:2478 – 84.

Mol Cancer Ther 2004;3(10). October 2004

Downloaded from mct.aacrjournals.org on September 24, 2021. © 2004 American Association for Cancer Research. Peroxisome proliferator-activated receptor γ and retinoid X receptor ligands are potent inducers of differentiation and apoptosis in leukemias

Marina Konopleva, Elena Elstner, Teresa J. McQueen, et al.

Mol Cancer Ther 2004;3:1249-1262.

Updated version Access the most recent version of this article at: http://mct.aacrjournals.org/content/3/10/1249

Cited articles This article cites 71 articles, 31 of which you can access for free at: http://mct.aacrjournals.org/content/3/10/1249.full#ref-list-1

Citing articles This article has been cited by 18 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/3/10/1249.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/3/10/1249. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.