CtBP-a targetable dependency for tumor-initiating cell activity and metastasis in pancreatic adenocarcinoma

Abstract

Ctbp2 is a uniquely targetable oncogenic transcriptional coregulator, exhibiting overexpression in most common solid tumors, and critical to the tumor-initiating cell (TIC) transcriptional program. In the "CKP" mouse pancreatic ductal adenocarcinoma (PDAC) model driven by mutant K-Ras, Ctbp2 haploinsufficiency prolonged survival, abrogated peritoneal metastasis, and caused dramatic downregulation of c-Myc, a known critical dependency for TIC activity and tumor progression in PDAC. A small-molecule inhibitor of CtBP2, 4-chloro-hydroxyimino phenylpyruvate (4-Cl-HIPP) phenocopied Ctbp2 deletion, decreasing tumor burden similarly to gemcitabine, and the combination of 4-Cl-HIPP and gemcitabine further synergistically suppressed tumor growth. Pharmacodynamic monitoring revealed that the 4-Cl-HIPP/gemcitabine combination induced robust and synergistic tumor apoptosis and marked downregulation of the TIC marker CD133 in CKP PDAC tumors. Collectively, our data demonstrate that targeting CtBP represents a fruitful avenue for development of highly active agents in PDAC that cooperate with standard therapy to limit both primary and metastatic tumor burden.

Fig. 1. Expression and role of CtBP in human and murine PDAC tumor progression.

a, b Allred score distribution for CtBP1 (a) and CtBP2 (b) in PDAC tumor specimens obtained from PDAC patients across TNM stages I–IV and stained for CtBP1 and 2 by IHC. c Kaplan–Meier survival curves for CKP and CKP2 mice, n = 15/group; median survivals were 8.1 and 9.5 weeks, respectively; p < 0.001. d Box plot of tumor weights obtained from CKP and CKP2 mice, n = 10/group; *p < 0.001. (inset) Representative images of CKP vs. CKP2 tumors. e Quantitation of the number of malignant peritoneal implants in CKP and CKP2 mice at 6–8 weeks of age. f Mesentery of representative CKP and CKP2 mice. Arrows indicate peritoneal implants in CKP mesentery (left panel), while no implants are observed in CKP2 mesentery (right panel). g H&E section of metastatic lesion obtained from a CKP mesenteric implant; arrows indicate pleomorphic nuclei (×400). h Quantitation of ascitic fluid volumes in each genotype; *p < 0.02. Methods: An ATA27 automated tissue microarrayer was used to construct tissue microarrays (TMAs) that were prepared from tumor-containing portions of pancreatic resection samples of 64 patients identified in the VCU Health Pathology archives with a diagnosis of PDAC, and the TMAs were analyzed for CtBP1 and CtBP2 protein abundance by using IHC. All retrospective human tumor material and clinical information was obtained under a VCU Institutional Review Board approved protocol. For IHC, tumors derived from pancreatic cancer patients were formalin fixed and paraffin embedded on slides, and were subjected to deparaffinization and hydration steps followed by quenching and peroxidase reaction steps. Antigen retrieval was performed at pH 6 by using a pressure cooker, followed by blocking for an hour with 5% goat serum and incubation overnight with primary antibodies at a dilution of 1:100 in blocking buffer. Antibodies—CtBP1 (Cat no. 612042, BD Biosciences); CtBP2 (Cat no. 612044, BD Biosciences, Cambridge, MA). After 3 washes in PBS, secondary antibodies (Dako Envision Systems HRP-conjugated anti-mouse or anti-rabbit IgG, Cat no. K4000, Carpinteria, CA, USA) were incubated for 1 h followed by 3 washes in PBS and slide development by using DAB chromogen substrate from Dako Envision Systems (Cat no. K3467). Nuclear staining was performed by incubation in hematoxylin (Electron Microscopy Sciences, Hatfield, PA, USA) for 2 min and was followed by dehydration with gradient alcohols and coverslipping steps. For the TMA-staining analysis, the Allred scoring system was used with protein expression rated on a scale from 0 to 8, with 8 being the highest level of expression. Scores for each stain were the average of values obtained from two independent pathologists. All animal studies performed for this paper were approved by VCU Institutional Animal Care and Use Committee. C57BL/6J male mice heterozygous for the Ctbp2 allele (Ctbp2^+/−) were purchased from Jackson Laboratory (Bar Harbor, ME) and mated with wild-type C57BL/6J female mice with further backcross of the mutant genotypes to C57BL/6J at least 6 generations before initiating experiments. Allele-specific PCR assays were used to genotype for Ctbp2 deletion (end-point PCR: Ctbp2^+/− Internal Control, F- 5′-CAAATGTTGCTTGTCTGGTG-3′ and R-5′-GTCAGTCGAGTGCACAGTTT-3′, Mutant F-5′-CCAGTGGGGATCGACGGTATC-3′, and Mutant R-5′-CACTCCAACGCAGCACCATC-3′). For pancreatic tumor mouse models, CKP2 mice were generated by breeding Ctbp2 ^+/− mice (as above) with C57BL/6J Pdx-cre; LSL-K-ras; p53 ^fl/fl mice (CKP). CKP and CKP2 mice were killed at indicated ages, and tumor weights, peritoneal implants, and ascites were measured. Allele-specific PCR assays were performed for Ctbp2 haploinsufficiency as mentioned above

Fig. 2. Expression of pancreatic TIC markers requires physiologic levels of Ctbp2.

a Representative IHC analysis of c-Myc expression in CKP and CKP2 mouse pancreatic tumors; the dashed line divides the area of normal acini (N) from tumor (T). b Representative IHC analysis of cancer stem cell markers CD44 and CD133 in CKP and CKP2 mouse pancreatic tumors. Methods: For IHC of mouse pancreatic tumors from CKP and CKP2 mice refer to Fig. 1 IHC methods. The primary antibodies were used at a dilution of 1:100 in blocking buffer and incubated overnight. Antibodies—c-Myc (N-262, Cat no. sc-764, Santa Cruz Biotechnology), CD44 (Cat no. ab157107, Abcam, Cambridge, MA), and CD133 (Cat no. 18470-1-AP, Proteintech, Rosemont, IL)

Fig. 3. Combination therapy with CtBP2 inhibitor induces apoptosis and significantly reduces pancreatic tumor burden.

a Schema for treatment of CKP mice with vehicle, gemcitabine, 4-Cl-HIPP, or the combination, 3× per week for 3 weeks or until humane endpoint. b Representative images of tumors obtained from CKP mice treated with vehicle, gemcitabine, 4-Cl-HIPP, or gemcitabine and 4-Cl-HIPP. c Quantification of tumor weights obtained from mice treated with above-listed drugs, n = 5 mice/treatment group. d Quantification of ascites aspirated from CKP mice treated with above-mentioned drugs, n = 5 mice/group. e, f Representative IHC staining of pancreatic tumors for TIC marker CD133 (e) and apoptosis indicator cleaved caspase 3 (f) obtained from CKP mice treated with above-mentioned drugs. Methods: CKP mice were injected with vehicle, gemcitabine (10 mg/kg, IP), 4-Cl-HIPP (100 mg/kg, IP) or a combination of both, intraperitoneally three times a week for 3 weeks, or until humane endpoint. At killing, tumor weights and ascites were measured. Refer to Fig. 1 Methods for IHC details; CD133 staining is detailed in Fig. 2; cleaved caspase 3 antibody (dilution 1:200; Cat no. 9661 Cell Signaling, Danvers, MA)