BET inhibition decreases HMGCS2 and sensitizes resistant pancreatic tumors to gemcitabine.

Efforts to develop targetable molecular bases for drug resistance for pancreatic ductal adenocarcinoma (PDAC) have been equivocally successful. Using RNA-seq and ingenuity pathway analysis we identified that the superpathway of cholesterol biosynthesis is upregulated in gemcitabine resistant (gemR) tumors using a unique PDAC PDX model with resistance to gemcitabine acquired in vivo. Analysis of additional in vitro and in vivo gemR PDAC models showed that HMG-CoA synthase 2 (HMGCS2), an enzyme involved in cholesterol biosynthesis and rate limiting in ketogenesis, is overexpressed in these models. Mechanistic data demonstrate the novel findings that HMGCS2 contributes to gemR and confers metastatic properties in PDAC models, and that HMGCS2 is BRD4 dependent. Further, BET inhibitor JQ1 decreases levels of HMGCS2, sensitizes PDAC cells to gemcitabine, and a combination of gemcitabine and JQ1 induced regressions of gemR tumors in vivo. Our data suggest that decreasing HMGCS2 may reverse gemR, and that HMGCS2 represents a useful therapeutic target for treating gemcitabine resistant PDAC.

The BET Inhibitor JQ1 Augments the Antitumor Efficacy of Gemcitabine in Preclinical Models of Pancreatic Cancer.

Gemcitabine is used to treat pancreatic cancer (PC), but is not curative. We sought to determine whether gemcitabine + a BET bromodomain inhibitor was superior to gemcitabine, and identify proteins that may contribute to the efficacy of this combination. This study was based on observations that cell cycle dysregulation and DNA damage augment the efficacy of gemcitabine. BET inhibitors arrest cells in G1 and allow increases in DNA damage, likely due to inhibition of expression of DNA repair proteins Ku80 and RAD51. BET inhibitors (JQ1 or I-BET762) + gemcitabine were synergistic in vitro, in Panc1, MiaPaCa2 and Su86 PC cell lines. JQ1 + gemcitabine was more effective in vivo than either drug alone in patient-derived xenograft models (P < 0.01). Increases in the apoptosis marker cleaved caspase 3 and DNA damage marker γH2AX paralleled antitumor efficacy. Notably, RNA-seq data showed that JQ1 + gemcitabine selectively inhibited HMGCS2 and APOC1 ~6-fold, compared to controls. These proteins contribute to cholesterol biosynthesis and lipid metabolism, and their overexpression supports tumor cell proliferation. IPA data indicated that JQ1 + gemcitabine selectively inhibited the LXR/RXR activation pathway, suggesting the hypothesis that this inhibition may contribute to the observed in vivo efficacy of JQ1 + gemcitabine.

Development of gemcitabine-resistant patient-derived xenograft models of pancreatic ductal adenocarcinoma.

AIM: Gemcitabine is a frontline agent for locally-advanced and metastatic pancreatic ductal adenocarcinoma (PDAC), but neither gemcitabine alone nor in combination produces durable remissions of this tumor type. We developed three PDAC patient-derived xenograft (PDX) models with gemcitabine resistance (gemR) acquired in vivo, with which to identify mechanisms of resistance relevant to drug exposure in vivo and to evaluate novel therapies. METHODS: Mice bearing independently-derived PDXs received 100 mg/kg gemcitabine once or twice weekly. Tumors initially responded, but regrew on treatment and were designated gemR. We used immunohistochemistry to compare expression of proteins previously associated with gemcitabine resistance [ribonucleotide reductase subunit M1 (RRM1), RRM2, human concentrative nucleoside transporter 1 (hCNT1), human equilibrative nucleoside transporter 1 (hENT1), cytidine deaminase (CDA), and deoxycytidine kinase (dCK)] in gemR and respective gemcitabine-naive parental tumors. RESULTS: Parental and gemR tumors did not differ in tumor cell morphology, amount of tumor-associated stroma, or expression of stem cell markers. No consistent pattern of expression of the six gemR marker proteins was observed among the models. Increases in RRM1 and CDA were consistent with in vitro-derived gemR models. However, rather than the expected decreases of hCNT1, hENT1, and dCK, gemR tumors expressed no change in or higher levels of these gemR marker proteins than parental tumors. CONCLUSION: These models are the first PDAC PDX models with gemcitabine resistance acquired in vivo. The data indicate that mechanisms identified in models with resistance acquired in vitro are unlikely to be the predominant mechanisms when resistance is acquired in vivo. Ongoing work focuses on characterizing unidentified mechanisms of gemR and on identifying agents with anti-tumor efficacy in these gemR models.

The combination of BET and PARP inhibitors is synergistic in models of cholangiocarcinoma.

Our previous finding that the BET inhibitor (BETi) JQ1 increases levels of the DNA damage marker γH2AX suggested that JQ1 might enhance the sensitivity of tumor cells to PARP inhibitors (PARPi), which are selectively toxic to cells that harbor relatively high levels of DNA damage. To address this hypothesis, we evaluated the effect of a BETi (JQ1 or I-BET762) combined with a PARPi (olaparib or veliparib) in KKU-055 and KKU-100 cholangiocarcinoma (CCA) cell lines and of JQ1 with olaparib in a xenograft model of CCA. Each combination was more effective than any of the four drugs as single agents. Combination indices ranged from 0.1 to 0.8 at the ED50 for all combinations, indicating synergy and demonstrating that synergy was not limited to a specific combination. Mechanistically, downregulation of BETi molecular targets BRD2 or BRD4 by shRNA sensitized CCA cells to BETi as single agents as well as to the combination of a BETi + a PARPi. Our data indicate that combinations of a BETi with a PARPi merit further evaluation as a promising strategy for CCA.

FTY720 enhances the anti-tumor activity of carboplatin and tamoxifen in a patient-derived xenograft model of ovarian cancer.

Ovarian cancer is the fifth leading cause of cancer-related deaths among women in the United States. Although most patients respond to frontline therapy, virtually all patients relapse with chemoresistant disease. This study addresses the hypothesis that carboplatin or tamoxifen + FTY720, a sphingosine analogue, will minimize or circumvent drug-resistance in ovarian cancer cells and tumor models. In vitro data demonstrate that FTY720 sensitized two drug-resistant (A2780. cp20, HeyA8. MDR) and two high-grade serous ovarian cancer cell lines (COV362, CAOV3) to carboplatin, a standard of care for patients with ovarian cancer, and to the selective estrogen receptor modulator tamoxifen. FTY720 + tamoxifen was synergistic in vitro, and combinations of FTY720 + carboplatin or + tamoxifen were more effective than each single agent in a patient-derived xenograft model of ovarian carcinoma. FTY720 + tamoxifen arrested tumor growth. FTY720 + carboplatin induced tumor regressions, with tumor volumes reduced by ∼86% compared to initial tumor volumes. Anti-tumor efficacy was concomitant with increases in intracellular proapoptotic lipid ceramide. The data suggest that FTY720 + tamoxifen or carboplatin may be effective in treating ovarian tumors.