Role of Peroxisome Proliferator-Activated Receptor-γ and Its Coactivator DRIP205 in Cellular Responses to CDDO (RTA-401) in Acute Myelogenous Leukemia

Peroxisome proliferator-activated receptors (PPAR) are ligand-activated transcription factors that are members of the nuclear hormone receptor gene superfamily. To date, three PPAR subtypes have been identified, namely, PPARα, PPARβ/δ, and PPARγ (also termed as NR1C1, NR1C2, and NR1C3, respectively; ref. 1). PPARγ is an important regulator of lipid and glucose homeostasis, cellular differentiation, and inflammation. There are three PPARγ isoforms differing at their 5′ ends, each under the control of its own promoter (2). Most tissues express the PPARγ1 isoform, whereas the PPARγ2 isoform is specific to adipocytes. However, the receptor is expressed in many others tissues and cell types such as monocytes and macrophages, skeletal muscle, breast, prostate, colon, and type 2 alveolar pneumocytes.

PPARs form heterodimers with retinoid X receptors (RXR). Transcriptional regulation by nuclear receptors (NR), including PPARs and RXRs, involves the binding and recruitment of coactivators and/or mediators to target gene promoters. One component of the TRAP-Mediator complex, TRAP220 (the thyroid hormone receptor–associated protein, also known as ARC/DRIP205), is directly associated with the thyroid receptor, vitamin D receptor, and PPAR and facilitates NR-mediated transcription. Recent studies have further shown a functional role of TRAP220 in the optimal vitamin D receptor– and retinoic acid receptor–mediated myelomonocytic differentiation processes in hematopoietic cells (3). It was also found to act as a pivotal coactivator for GATA-1 in erythroid development (4) and has been shown to play a role in differentiation and proliferation of keratinocytes in response to vitamin D receptor stimulation (5, 6).

PPARγ ligands have been extensively investigated and are known to inhibit proliferation and induce differentiation or apoptosis in different cancer cell types, including hematologic malignancies (7). Our studies have shown that PPARγ ligands alone or in combination with RXR-specific activators can inhibit clonal proliferation and induce differentiation of HL-60, U937, and THP1 human myeloid leukemic cell lines (8). 2-Cyano-3,12-dioxooleana-1,9-dien-28-oic acid (CDDO), a synthetic triterpenoid, is a ligand for PPARγ (9) that induces growth arrest, apoptosis, and/or differentiation of a variety of tumor cell types. We have shown that CDDO induces Bcl-2 downregulation, mitochondrial depolarization, and caspase activation in myeloid leukemic cells (10). It also potently enhances apoptosis induced by tumor necrosis factor in myeloid leukemia cells (11) and refractory chronic lymphocytic leukemia B cells (12). In breast cancer, CDDO-induced growth inhibition correlates with PPARγ transactivation and is mediated, at least in part, by upregulation of P21 and downregulation of cyclin D1 (13, 14). Notably, CDDO and its derivatives also induce differentiation of leukemic cells (15–17), effects that in some cell types are synergistically enhanced by concomitant ligation of the RXR nuclear receptor. The multiple effects of CDDOs have been attributed to both PPARγ-dependent and PPARγ-independent mechanisms of action (18). One of the most powerful anti-inflammatory and anticarcinogenic activities induced by CDDO and related compounds is activation of the Nrf2/ARE signaling pathway, a major regulator of antioxidant cellular responses (for review, see ref. 19).

In this study, we investigated the role of PPARγ and its coactivator DRIP205 on the apoptosis- and differentiation-inducing properties of CDDO in leukemic cells. The results show that the PPARγ ligands CDDO and 15-deoxy-^12,14-15DPGJ₂ (15d15DPGJ₂) induce higher degrees of apoptosis in leukemic cells modified to overexpress PPARγ. Furthermore, we provide evidence that overexpression and recruitment of the coactivator DRIP205 contributes to the enhanced differentiation by this agent. Finally, we report induction of myeloid differentiation of PPARγ/DRIP205–coexpressing leukemic blasts from patients with acute myelogenous leukemia (AML) treated with CDDO (RTA-401) in a “first-in-man” clinical trial.

CDDO was manufactured under the RAID Program and kindly provided by Drs. E. Sausville (National Cancer Institute, Bethesda, MD) and M. Sporn (Dartmouth Medical College, Hanover, NH). 15d15DPGJ₂ was purchased from Cayman Chemical Company. N-(4′-Aminopyridyl)-2-chloro-5-nitrobenzamide (T007; ref. 20), a selective PPARγ antagonist, was synthesized as described previously (21). Annexin V-FITC was purchased from Roche Diagnostic Co., and CD14-FITC, CD11b-phycoerythrin (PE), and CD34-PE were from BD Biosciences. Rabbit polyclonal caspase-3 antibody (CPP32, 1:1,000) was from PharMingen; mouse monoclonal anti–caspase-8 (1:1,000) and rabbit polyclonal anti–caspase-9 (1:1,000) were purchased from Cell Signaling Technology; rabbit polyclonal anti-TRAP220, mouse monoclonal anti-PPARγ (1:500), mouse monoclonal poly(ADP-ribose) polymerase (anti-PARP) antibodies (1:200), and mouse monoclonal anti–proliferating cell nuclear antigen (anti-PCNA; 1:500) were from Santa Cruz Biotechnology; mouse monoclonal anti-p21 antibody (1:1,000) was from Calbiochem; mouse monoclonal anti-p27 (1:1,000), mouse monoclonal anti–HO-1 (1:1,000), and mouse monoclonal anti-Bip/GRP78 (1:1,000) antibodies were from BD Biosciences; mouse monoclonal anti-Flag and mouse monoclonal anti–β-actin (AC74) were from Sigma Chemical Co.; and goat anti-mouse and goat anti-rabbit horseradish peroxidase–conjugated secondary antibodies were from Bio-Rad.

HL-60, U937, MCF-7, and SW480 were purchased from The American Type Culture Collection. Bone marrow or peripheral blood samples from 260 AML or high-risk myelodysplastic syndrome (MDS) samples were assessed for PPARγ expression by reverse-phase protein array (Supplementary Methods; ref. 22). Samples were collected for the Leukemia Sample Bank at the University of Texas M.D. Anderson Cancer Center between January 15, 1998 and March 9, 2006, on Institutional Review Board (IRB)–approved protocol Lab01-473, and consent was obtained according to the Declaration of Helsinki. Samples were analyzed under an IRB-approved laboratory protocol (Lab05-0654).

Early apoptotic events were detected by the flow cytometric measurement of externalized phosphatidylserine with the Annexin-V-FLUOS Staining Kit from Roche Diagnostics. Cell cycle analysis was conducted using propidium iodide as described (23).

Cells were washed twice with PBS and resuspended in lysis buffer (10 mmol/L Tris-HCl, 1 mmol/L EDTA, 0.025% Triton X-100, pH 8.0) for 10 to 20 minutes on ice. Lysates were digested with RNase A (50 μg/mL) and then with proteinase K (2 mg/mL) for 1 hour each at 37°C. DNA samples were separated on 1.8% agarose gel containing ethidium bromide and visualized by UV light.

The differentiation of myeloid leukemic cells was determined by their ability to produce superoxide, as measured by the reduction of nitroblue tetrazolium (NBT) as described (24). The analysis of differentiation-specific cell surface antigens was measured by flow cytometry using the PE-conjugated anti-CD11b and FITC-conjugated anti-CD14 monoclonal antibodies.

SV40 β-galactosidase was obtained from Promega Corp.; pM vector was purchased from Clontech; and pMD, pMDm5, pMDm6, pMDm7, pMDm7Δ, and plasmids were generated by PCR (25). FLAG-tagged wild-type (wt) PPARγ construct was kindly provided by Dr. K. Chatterjee (Department of Medicine, University of Cambridge, Cambridge, United Kingdom). pcDNA3-DRIP205 expression plasmid was kindly provided by Dr. Leonard P. Freedman (Merck Research Laboratories, West Point, PA).

For transient transfection assays, SW480 cells were cotransfected with DNA using FuGENE. In coactivation experiments, cells were cotransfected with 250 ng of SV40 β-galactosidase (used as an internal control) and 1 μg of the reporter construct containing three copies of the acyl-CoA oxidase PPARγ response element (PPRE) cloned upstream of the TK-LUC reporter, PPREx3-LUC reporter (kindly provided by Dr. Ronald M. Evans, The Salk Institute, La Jolla, CA), and various amounts of DRIP205 or deletion mutant constructs. After 24 hours, transfected cells were treated with DMSO or 0.75 μmol/L CDDO for another 48 to 72 hours. Relative luciferase activity was calculated by dividing luciferase activity by β-galactosidase activity for each well.

Cells were lysed at a density of 1 × 10⁶/50 μL in protein lysis buffer (0.25 mol/L Tris-HCl, 2% sodium dodecylsulfate, 4% β-mercaptoethanol, 10% glycerol, 0.02% bromophenol blue) supplemented with a protease inhibitor cocktail (Roche Diagnostic). For preparation of nuclear lysates, cells were washed in cold PBS and digitonin-permeabilized for 5 minutes on ice at a density of 20 million cells/mL in extraction buffer (250 mmol/L sucrose, 70 mmol/L KCl; 137 mmol/L NaCl; 4.3 mmol/L Na₂HPO4; 1.4 mmol/L KH₂PO4 (pH 7.2), 200 μg/mL digitonin, and protease inhibitors). Cells were centrifuged at 1,000 × g for 5 minutes at 4°C, and pellet containing nuclear fraction was lysed as above. Cell lysates were then loaded onto a 10% to 12% SDS-PAGE gel (Bio-Rad). After electrophoresis, proteins were transferred onto Hybond-P membranes (Amersham Pharmacia Biotech), followed by immunoblotting. Signals were detected using a PhosphorImager (Storm 860, version 4.0, Molecular Dynamics).

Patients with refractory/relapsed AML were treated with CDDO (RTA-401, Reata Pharmaceuticals) from 0.6 to 75 mg/m²/h × 5 days, in a phase 1 clinical trial, following informed consent according to the University of Texas M.D. Anderson Cancer Center and Princess Margaret Hospital-Ontario Cancer Institute guidelines (Supplementary Table S1). Bone marrow or peripheral blood sample mononuclear cells were separated by Ficoll-Hypaque (Sigma Chemical) density-gradient centrifugation. Apoptosis was examined by simultaneous staining with CD34 allophycocyanin (APC; BD) and tetramethylrhodamine methyl ester (TMRM; Invitrogen) to measure mitochondrial inner transmembrane potential (Δψ_m). Cell differentiation was examined by simultaneous staining with CD34-APC (BD), CD33-PE-Cy7 (BD), CD14-FITC (BD), and CD11b-PE (BD) at the indicated time points.

Bone marrow and peripheral blood samples obtained from AML patients were lysed with RNA Stat 60 (Tel-Test). The RNA was subjected to purification in an RNeasy ion-exchange column (Qiagen) with on-column DNase treatment. cDNA was prepared from 1.0 μg of total RNA per 20 μL of mix containing 0.07 μg/μL random-sequence hexamer primers, 0.5 mmol/L deoxynucleotide triphosphates, 5 mmol/L DTT, 0.2 units/μL SuperAsin RNase inhibitor (Ambion), and 10 units/μL SuperScript III (Invitrogen). Real-time PCR was carried out using an ABI Prism 7700 instrument as described (26). For primer and probe sets to detect PPARγ, p21, NQO1, and housekeeping gene ABL1, we used TaqMan Gene Expression Assays Hs00234592_m1, Hs00355782_m1, Hs00168547_m1, and Hs00245445_m1, respectively. The abundance of each transcript of interest relative to that of ABL1 was calculated as follows: relative expression = 100 × 2 exp [−ΔCt], where ΔCt is the mean Ct of the transcript of interest less the mean Ct of the transcript for ABL.

Cells were transfected by the Amaxa electroporator Nucleofector I from Amaxa Biosystems, using the Nucleofector Kit C (program W-001). Small interfering RNA (siRNA) PPARγ transfection was performed using validated Stealth RNAi VHS 40941 for duplex 1 (siRNA1) and VHS 40944 for duplex 2 (siRNA2; Invitrogen). Cells (10⁶) were resuspended in 100 μL of Nucleofector Solution C and transfected by electroporation with scramble LO GC Duplex Stealth RNAi Negative Control (12 935-200; Invitrogen) or with PPARγ siRNA. Forty-eight hours after transfection, DMSO or CDDO was added to the cells for 24 hours, and protein expression was monitored by immunoblotting.

We have previously shown that PPARγ protein is expressed in both myeloid and lymphoid leukemic cell lines (8). Expression of PPARγ protein was analyzed in 260 newly diagnosed leukemia-enriched AML/MDS samples. PPARγ protein was variably expressed in different AML subtypes (Supplementary Fig. S1). There was no difference in overall survival between patients expressing high or low levels PPARγ (not shown).

We next characterized the relationship between PPARγ expression and the ability of pharmacologic PPARγ ligands to affect the growth and survival of leukemic cells. To this end, we established stable transfectants of U937 cells expressing empty vector (pcDNA3) or Flag-tagged wt-PPARγ (Fig. 1A–C). Overexpression of PPARγ facilitated growth of cells but the differences did not reach statistical significance (P = 0.1; Supplementary Fig. S2A). Exposure of cells to the PPARγ ligand CDDO further enhanced PPARγ protein expression in both vector control and transfected cells (Fig. 1C). As shown in Fig. 1D, CDDO induced PPARγ mRNA expression wt-PPARγ–overexpressing cells or in the same cells transfected with scrambled siRNA (2.3- and 3-fold, respectively), and this induction was largely abrogated in PPARγ siRNA–transfected cells. These data suggest that PPARγ ligation with CDDO induces expression of its cognate receptor in a PPARγ-dependent fashion.

Next, we examined the functional consequences of PPARγ induction and overexpression in response to the receptor activation by CDDO or by the structurally different PPARγ agonist 15dPGJ₂. Vector control (pcDNA3) or wt-PPARγ–transfected U937 cells were treated with 1 μmol/L CDDO for 24 hours. Overexpression of PPARγ significantly enhanced the sensitivity of leukemic cells to apoptosis induced by CDDO or 15dPGJ₂ (Fig. 2A and Supplementary Fig. S2B). We next analyzed modulation of apoptosis and cell cycle–regulating proteins in PPARγ-transfected clones. Following treatment with CDDO for 24 hours, cells overexpressing wt-PPARγ exhibited significantly decreased procaspase-3 and increased cleavage of caspase-3 and its substrate PARP (Fig. 2B). In accordance with these results, CDDO induced a higher degree of endonucleolytic DNA cleavage in PPARγ-overexpressing cells. Analysis of the upstream (initiator) caspases showed the appearance of the active (p18) fragment of caspase-8 in these cells, whereas caspase-9 was similarly cleaved in both vector control and PPARγ-overexpressing cells (Fig. 2B). In a time-course experiment, no caspase-8 cleavage and DNA fragmentation were observed in vector-treated cells, whereas both ligands induced caspase-8 cleavage and DNA fragmentation at 5 and 24 hours in PPARγ-overexpressing cells (Fig. 2C). In contrast, caspase-9 cleavage did not substantially differ between control and PPARγ-transfected cells. Activation of caspase-8 may proceed through CHOP-dependent transcriptional upregulation of death receptor 5 (DR5)/tumor necrosis factor-related apoptosis-inducing ligand receptor 2 (TRAIL-R2) triggered by endoplasmic reticulum stress and unfolded protein response (27), and CDDO slightly increased the level of the endoplasmic reticulum stress marker protein Bip/GRP78 in PPARγ-overexpressing cells (Fig. 2B). Knocking down exogenously transfected PPARγ by two different siRNA constructs abolished apoptosis induced by lower CDDO concentration (0.5 μmol/L) but did not significantly affect it at higher CDDO concentration (1 μmol/L; Fig. 2D), indicating the contribution of PPARγ-dependent and PPARγ-independent mechanisms of cell death. Likewise, the pharmacologic PPARγ inhibitor T007 partially protected PPARγ-overexpressing cells, but not control cells, from CDDO-induced apoptosis (Supplementary Fig. S2C). CDDO potently induced expression of the stress-responsive inducible enzyme hemeoxygenase-1 (HO-1; Fig. 2B) and of one of the critical Nrf2 target genes, NADPH quinine oxidoreductase-1 (NQO1; Supplementary Fig. S3A), in both control and PPARγ-overexpressing cells.

PPARγ agonists are known to induce expression of the cell cycle inhibitory protein p21^Waf1/CIP1, and we have previously shown that CDDO induced p21 mRNA and protein in breast cancer cells (14). Consistent with these findings, CDDO induced p21 mRNA in leukemic cells (Fig. 3A). This increase in P21 transcription was evident both in control cells and in cells overexpressing PPARγ (Fig. 3A and B). The expression of p27^KIP1 was unchanged, and no significant differences in cell cycle distribution were noted between PPARγ-overexpressing and control cells (Supplementary Fig. S2D). Although P21 promoter contains a potential conserved consensus PPRE, CDDO may also increase transcription of P21 indirectly through increased binding of Sp1, Sp3, and Sp4 transcription factors to the GC-rich regions of the P21 promoter. To determine the mechanism of transcriptional activation of P21 by CDDO, we conducted promoter assays by transient transfection of SW480 cells expressing endogenous PPARγ (Fig. 3C, inset) with full-length P21 luciferase promoter plasmid and plasmids containing point mutations in GC-rich elements 1 to 6 (Mut1–Mut6) of the proximal P21 promoter (28). CDDO induced an ∼6-fold increase in P21-wt promoter activity (Fig. 3C, p21-Luc wt). Interestingly, CDDO was capable of transactivating all constructs containing point mutations. To determine if CDDO induces p21 expression via PPARγ, we measured p21 levels following depletion of PPARγ by siRNA. As shown in Fig. 3D, inhibition of PPARγ using siRNA failed to block CDDO-induced levels of p21 mRNA or protein expression. These data indicate that CDDO causes induction of p21 levels in a PPARγ-independent fashion. Consistent with documented properties of triterpenoids and other electrophilic compounds to activate the Keap1/Nrf2 antioxidant pathway, CDDO robustly induced HO-1 (Fig. 3D) and NQO1 (Supplementary Fig. S3B) in cells with functional or silenced PPARγ, indicating that these responses are likewise PPARγ independent. No appreciable change in Bip/GRP78 expression was noted in control or transfected cells (Fig. 3D), indicating no significant contribution of endoplasmic reticulum stress to the proapoptotic effects of CDDO in this cell system.

PPARγ ligands recruit the coactivator DRIP205 to PPARγ. Furthermore, nuclear factor coactivators are known to mediate tissue-specific effects (29). To ascertain whether DRIP205 overexpression will affect the transcriptional activity of PPARγ, the ability of CDDO and 15dPGJ2 to induce (PPRE)3-tk-luc reporter was examined in MCF-7 cells transfected with full-length DRIP205 plasmid. Transient cotransfection with (PPRE)3-tk-luc reporter and full-length pcDNA3-DRIP205 plasmid resulted in higher levels of PPARγ transactivation induced by 15dPGJ₂ and CDDO (1.9- and 2.7-fold, respectively; Fig. 4A).

Previous studies have shown that the NR boxes in DRIP205/TRAP220 contribute to the physical and functional interactions of these coactivators with NRs (30, 31). Their role in coactivation of PPARγ was further investigated in SW480 cotransfected with (PPRE)3-tk-luc reporter and NH₂- or COOH-terminal GAL4-DRIP205 deletion constructs (Fig. 4B). The full-length DRIP205 expression plasmid encodes for 1,566 amino acids, which are identical to amino acids 16–1,581 of the TRAP220 coding sequence. CDDO significantly induced activity in cells transfected with (PPRE)3-tk-luc and GAL4-DRIP205 (wild-type). In addition, significant coactivation, albeit to a lesser degree, was observed for several GAL4-DRIP205 chimeras (pMDm5, pMDm7, pMDm7Δ, and pMDm6), two of which express NH₂-terminal (pMDm5) or COOH-terminal (pMDm6) regions of DRIP205 but do not contain the central NR box sequences (Fig. 4C). These results confirm that the NR boxes of DRIP205 may contribute to, but are not required for, coactivation and indicate that multiple domains of DRIP205 are involved in interactions with PPARγ.

To determine the functional role of DRIP205 in the context of PPARγ ligation in leukemic cells, we investigated the effects of CDDO in HL-60 cells stably transfected with DRIP205 plasmid (three separate clones; Fig. 5A and Supplementary Fig. S4A). No difference in cell growth (Supplementary Fig. S4B), cell cycle progression (Supplementary Fig. S4C), or apoptosis (Supplementary Fig. S4D) was found in cells overexpressing DRIP205, cultured alone or exposed to CDDO. In contrast, CDDO induced a higher degree of myelomonocytic differentiation in DRIP205-overexpressing cells as shown by increased expression of CD11b (P < 0.01; Fig. 5B) and by NBT assay (36.7 ± 6.1% versus 79.7 ± 4.7%, P < 0.001; Fig. 5C).

To determine if the enhanced differentiation was mediated through PPARγ, we assessed CD11b expression in cells pretreated with the pharmacologic PPARγ antagonist T007. As shown in Fig. 5D, blocking PPARγ transactivation significantly diminished CDDO-induced myelomonocytic differentiation in DRIP205-overexpressing cells but not in vector-transduced control cells.

In a first-in-man clinical phase I trial of increasing doses of CDDO (RTA-401; escalated from 0.6 to 75 mg/m²/h) in patients with relapsed/refractory AML, we investigated the in vivo differentiating and proapoptotic effects of CDDO and correlated these changes with PPARγ and DRIP205 expression in cells from nine patients by quantitative TaqMan PCR (PPARγ) and immunoblotting (PPARγ and DRIP205). Clinical characteristics of the patients are summarized in Supplementary Table S1. PPARγ mRNA was expressed in all samples at baseline albeit at different levels. PPARγ and DRIP205 proteins were expressed in samples from seven of nine patients studied; no expression of either protein was detected in samples from patients #307 and #309 (Fig. 6A). After 6 days of continuous CDDO administration, PPARγ mRNA was induced >2-fold in four patient samples (Supplementary Table S2). In four of the nine patients, an increase in CD11b⁺ and CD14⁺ cells and a concomitant reduction of immature cells expressing CD34 or CD33 were observed (#301, #304, #305, and #306, Fig. 6B). Examples of flow cytometric profiles are shown in Supplementary Fig. S5. Baseline expression of PPARγ protein was highest in samples from patients #301 and #304 (Fig. 6A), and in all four patients increase in PPARγ mRNA was shown (1.5-, 2.4-, 1.8-, and 2.2-fold, respectively; Supplementary Table S2). In these, p21 mRNA was induced >2-fold in samples #304, #305, and #306. No change in differentiation markers was observed in patients #307 and #309 with no detectable baseline PPARγ or DRIP205 proteins. Moderate induction of apoptosis, documented as loss of mitochondrial membrane potential in circulating CD34⁺ cells, was observed in samples from three patients (#301, #303, and #305); in sample #303, corresponding apoptosis induction was seen in day 6 bone marrow CD34⁺ cells (Fig. 6C). Examples of flow cytometric profiles are shown in Supplementary Fig. S6. Clinically, patients did not fulfill protocol response criteria: differential counts did not change significantly and maximum tolerated dose was not reached at the low dose levels in this phase I study.

PPARγ ligands inhibit cancer cell proliferation and induce apoptosis and/or differentiation in multiple tumor types, and these effects have been attributed to both PPARγ-dependent and PPARγ-independent mechanisms. In this study, we evaluated the role of PPARγ and one of its cellular coactivators, DRIP205, in the proapoptotic and differentiating properties of PPARγ agonists CDDO and 15dPGJ₂. A high-throughput reverse-phase protein array technique showed high levels of PPARγ expression in 260 primary AML samples. To functionally characterize the relationship between baseline PPARγ levels and cellular effects of PPARγ agonists in leukemic cells, we generated stably transfected myeloid leukemic cells overexpressing the receptor. U937 cells induced to overexpress wt-PPARγ were more sensitive to the proapoptotic effects of PPARγ ligands CDDO and 15dPGJ₂ compared with vector-transduced cells. These proapoptotic effects were significantly inhibited by silencing PPARγ with siRNA or by blocking PPARγ activation with the pharmacologic antagonist T007, consistent with previously published findings of PPARγ-dependent and PPARγ-independent mechanisms of action of this class of agents. Time-course analysis showed that high PPARγ levels facilitated cleavage of caspase-8 and caspase-3 (but not of caspase-9), resulting in accelerated PARP cleavage, DNA fragmentation, and apoptosis. Of note, several reports indicated the ability of CDDOs to activate the extrinsic apoptotic pathway and sensitize to TRAIL via diverse molecular mechanisms including FLIP downregulation (32), c-jun NH₂-terminal kinase–mediated induction of TRAIL receptor expression (33), and inhibition of NF-κB–dependent antiapoptotic proteins (11). Conversely, data reported by us and others show that CDDO and its more potent derivative CDDO-Me promoted the release of cytochrome c from isolated mitochondria, suggesting that CDDOs directly target the mitochondria to trigger the intrinsic pathway of cell death (34, 35). Data presented here suggest a proximal role for caspase-8 downstream of ligand-activated PPARγ, whereas direct mitochondrial effects of CDDO observed at higher concentrations are likely PPARγ independent, possibly by modifying the mitochondrial proteins through nucleophilic attack and Michaels addition (36). Unlike in non–small lung cancer cells (27). CDDO did not induce significant endoplasmic reticulum stress response, hence making upregulation of DR5 an unlikely mechanism of caspase-8 activation. The exact mechanistic link between PPARγ transactivation and activation of the extrinsic apoptotic pathway in AML remains to be determined.

It has recently been shown that synthetic triterpenoids are potent activators of Nrf2/ARE signaling in a variety of cell types, resulting in marked induction of a variety of antioxidative genes and detoxifying enzymes (37–39). In our studies, CDDO promptly upregulated expression of HO-1 and NQO1 in leukemia cells engineered to overexpress or silence PPARγ. These observations are consistent with the notion that these responses are likely mediated by the chemical structure of CDDO and other electrophilic compounds capable of modifying cysteine residues on KEAP1 protein (37) and represent important PPARγ-independent activities of this class of compounds.

PPARγ agonists including CDDO modulate cell cycle progression in multiple tumor types (13, 14). Our present data show that CDDO induced expression of p21waf1/CIP protein in leukemic cells. This induction was observed in parental cells, in cells overexpressing PPARγ, or in cells transfected with PPARγ siRNA. These findings indicate that the ability of CDDO to activate P21 promoter is likely mediated via PPARγ-independent mechanisms. Because induction of p21 expression is frequently mediated via increased binding of Sp1, Sp3, and Sp4 transcription factors to the GC-rich regions of the P21 promoter, we used the constructs containing point mutations in the GC elements. In contrast to our previous study in pancreatic cells (40), we were unable to identify the specific site required for PPARγ-dependent activation of P21. Of note, p21 is regulated by many different pathways and transcription factors, and CDDO could conceivably mediate its effects on p21 expression through an alternate pathway. Surprisingly, induction of p21 protein expression did not translate into a discernible cell cycle arrest in leukemic cells. The observation that CDDO preferentially induces apoptosis rather than cell cycle arrest in AML cells attests to the cell type–dependent properties of these agents, likely related to the distinct mitochondrial architecture of leukemic cells compared with solid tumor cells. Whether functional consequences of p21 overexpression other than control of cell cycle, such as regulation of apoptosis, differentiation, or transcriptional activation (41), are operational in leukemic cells remains to be investigated. Notably, p21 mRNA induction was observed in samples from three of the nine patients treated with very low doses of CDDO (RTA-401) in a phase I clinical trial.

Emerging evidence suggests the critical importance of the cellular context, in particular the composition of tissue-specific coactivators and corepressors, in the biological responses to NR agonists. PPARγ is known to interact with both the p160/SRC-1 family of coactivators and the multisubunit DRIP/Mediator coactivator complex. Our results show that CDDO induced significant activity in cells transfected with (PPRE)3-tk-luc and full-length DRIP205. In addition, significant coactivation was observed using several NH₂- and COOH-terminal domain mutants of DRIP205, and this coactivation did not require the NR boxes (Fig. 4B). These results suggest that multiple domains of DRIP205 are involved in interactions with PPARγ, similar to findings reported for estrogen receptor-α coactivation by DRIP205 (25). Interestingly, recent structural and functional analyses indicate that a direct interaction of PPARγ with DRIP/Mediator complex through the NR motifs of DRIP205 is not required for PPARγ-stimulated adipogenesis (42).

DRIP205 is involved in the vitamin D–triggered regulation of gene transcription during keratinocyte differentiation (6), and overexpression of DRIP205 was observed in some cancer cell lines (43). CDDOs have been shown to induce differentiation in myeloid leukemia cells (16, 17, 44), and in this study, CDDO induced a higher degree of myelomonocytic differentiation in DRIP205-overexpressing HL-60 cells, a process mediated through PPARγ. Whereas we recently reported that one of the mechanisms of differentiation induction by CDDO involves modulation of CEBPα expression and function (45), our data shown here provide first evidence that high cellular levels of the coactivator DRIP205 can enhance the differentiation induced by PPARγ ligation and is therefore an important determinant of tissue-specific effects of PPARγ agonists. We here report that leukemic blasts from patients treated in a phase I clinical trial of CDDO (RTA-401) express DRIP205 in seven of nine samples, all of which expressed PPARγ mRNA and protein. Further, sequential studies showed increased expression of the differentiation markers CD11b and/or CD14 in four patients. In these patients, CDDO induced PPARγ transcription. Although the numbers are too small to draw definitive conclusions, the data suggest that CDDO activates PPARγ in a subset of patients with AML in vivo, whose cells express DRIP205. We did not observe correlation between PPARγ levels and apoptosis induction, possibly due to very low levels of CDDO in this phase I study. Alternatively, this finding may indicate that PPARγ-dependent functions of CDDO may manifest primarily through differentiation induction rather than apoptotic responses in primary AML cells. Recently, the RXR agonist bexarotene was shown to induce differentiation in non–acute promyelocytic leukemia patients with AML who were treated with this agent in a phase I trial (46). Taking into consideration multiple studies showing that addition of RXR ligands synergistically enhances the differentiating and growth-suppressive effects of PPARγ ligands (8, 47), the combined use of these agents seems to be worth testing in the therapy of AML. The ongoing efforts by the Nuclear Receptor Signaling Atlas consortium to profile coactivators/corepressors in primary AML may assist in identifying patients who are likely to benefit from PPARγ/RXR ligation strategies.

M. Andreeff and M. Konopleva: ownership interest and consultants, Reata Disc. The other authors disclosed no potential conflicts of interest.

Grant Support: Leukemia and Lymphoma Society grant R64149-07 (M. Konopleva); Leukemia Specialized Program of Research Excellence, National Cancer Institute grant 1 P50 CA100632-01 (M. Andreeff); Leukemia and Lymphoma Society grant 6089 and NIH PO1 grant CA-55164 (S.M. Kornblau); and Cancer Center Support Grant P30 CA016672 34 (K.R. Coombes).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.