Gene expression profiling of patient‐derived pancreatic cancer xenografts predicts sensitivity to the BET bromodomain inhibitor JQ 1: implications for individualized medicine efforts

• DOI: 10.15252/emmm.201606975
• Publication Date: 2017-03-09T01:10:46Z
• Authors: Benjamin Bian; Martin Bigonnet; Odile Gayet; Celine Loncle; Aurélie Maignan; Marine Gilabert; Vincent Moutardier; Stephane Garcia; Olivier Turrini; Jean‐Robert Delpero; Marc Giovannini; Philippe Grandval; Mohamed Gasmi; Mehdi Ouaissi; Veronique Secq; Flora Poizat; Rémy Nicolle; Yuna Blum; Laetitia Marisa; Marion Rubis; Jean‐Luc Raoul; James E Bradner; Jun Qi; Gwen Lomberk; Raul Urrutia; Andres Saul; Nelson Dusetti; Juan Iovanna
• Keywords: Adult; Male; 0301 basic medicine; Medicine (General); Cell Survival; [SDV]Life Sciences [q-bio]; JQ1; 610; [SDV.CAN]Life Sciences [q-bio]/Cancer; Antineoplastic Agents; QH426-470; Mice; 03 medical and health sciences; R5-920; [SDV.CAN] Life Sciences [q-bio]/Cancer; transcriptomic signature; [SDV.BBM] Life Sciences [q-bio]/Biochemistry, Molecular Biology; pancreatic adenocarcinoma; Genetics; bromodomains; Animals; Humans; [SDV.BBM]Life Sciences [q-bio]/Biochemistry, Molecular Biology; Precision Medicine; Research Articles; Cells, Cultured; Cancer; Aged; Aged, 80 and over; Gene Expression Profiling; Azepines; Middle Aged; Triazoles; Chromatin; 3. Good health; [SDV] Life Sciences [q-bio]; Pancreatic Neoplasms; Genomics & Functional Genomics; c-MYC; c‐MYC; Genomics and Functional Genomics; Heterografts; Epigenetics; Female

Research Article8 March 2017Open Access Source DataTransparent process Gene expression profiling of patient-derived pancreatic cancer xenografts predicts sensitivity to the BET bromodomain inhibitor JQ1: implications for individualized medicine efforts Benjamin Bian orcid.org/0000-0001-8691-5292 Centre de Recherche en Cancérologie Marseille (CRCM), INSERM U1068, CNRS UMR 7258, Parc Scientifique et Technologique Luminy, Aix-Marseille Université and Institut Paoli-Calmettes, Marseille, France Search more papers by this author Martin Bigonnet Odile Gayet Celine Loncle Aurélie Maignan Marine Gilabert Vincent Moutardier Hôpital Nord, CIC1409, AP-HM-Hôpital Université, Stephane Garcia Olivier Turrini Jean-Robert Delpero Marc Giovannini Philippe Grandval la Timone, Mohamed Gasmi Mehdi Ouaissi Veronique Secq Flora Poizat Rémy Nicolle Programme Cartes d'Identité des Tumeurs (CIT), Ligue Nationale Contre Le Cancer, Paris, Yuna Blum Laetitia Marisa Marion Rubis Jean-Luc Raoul James E Bradner Department Medical Oncology, Dana-Farber Cancer Institute, Harvard School, Boston, MA, USA Jun Qi Gwen Lomberk Laboratory Epigenetics Chromatin Dynamics, Departments Biochemistry Molecular Biology Medicine, Mayo Clinic, Rochester, MN, Raul Urrutia Andres Saul Interdisciplinaire Nanoscience Marseille-CNRS 7325, Nelson Dusetti Corresponding Author [email protected] orcid.org/0000-0002-6161-8483 Juan Iovanna orcid.org/0000-0003-1822-2237 Information Bian1, Bigonnet1, Gayet1, Loncle1, Maignan1, Gilabert1, Moutardier1,2,3, Garcia1,2, Turrini1,4, Delpero4, Giovannini4, Grandval5, Gasmi2,3, Ouaissi5, Secq2, Poizat4, Nicolle6, Blum6, Marisa6, Rubis1, Raoul4, Bradner7, Qi7, Lomberk8, Urrutia8, Saul9, *,1 1Centre 2Hôpital 3CIC1409, 4Institut 5Hôpital 6Programme 7Department 8Laboratory 9Centre *Corresponding author. Tel: [phone]; Fax: [phone]; E-mail: 828803; EMBO Mol Med (2017)9:482-497https://doi.org/10.15252/emmm.201606975 PDFDownload PDF article text main figures. Peer ReviewDownload a summary editorial decision including letters, reviewer comments responses feedback. ToolsAdd favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Abstract c-MYC controls than 15% genes responsible proliferation, differentiation, cellular metabolism in as well other cancers making transcription factor prime target treating patients. The transcriptome 55 show that 30% them share an exacerbated profile MYC transcriptional targets (MYC-high). This cohort is characterized high level Ki67 staining, lower differentiation state, shorter survival time compared MYC-low subgroup. To define classifier signature, we selected group 10 transcripts which increased MYC-high six group. We validated ability these markers panel identify from both: discovery validation cohorts primary cell cultures same then showed cells patients are sensitive JQ1 treatment cells, monolayer, 3D cultured spheroids vivo xenografted tumors, due cycle arrest followed apoptosis. Therefore, results provide new potentially novel therapeutic modalities distinct subgroups tumors may find application future management within setting clinics. Synopsis large characterize tumor phenotypes. Sustained growth aggressivity subgroup ductal adenocarcinoma (PDAC) strongly depend on program c-MYC. A 16-transcript signature discriminates MYC-dependent (MYC-high) PDAC learning xenografts. highly proliferative undifferentiated tumors. treatment, well-described inhibitor. Introduction Pancreatic one most lethal major public health issue since there approximately 230,000 cases per year worldwide with number deaths (Jemal al, 2005). Like others malignant diseases, complex combination genetic, epigenetic, environmental factors gives rise particularly heterogeneous disease, having different set symptoms, predisposition early metastasis, (Yachida Iacobuzio-Donahue, 2013; Dunne Hezel, 2015; Waddell 2015). heterogeneity highlights necessity stratify goal predicting better therapies (Heller Koay 2016; Noll 2016). One strategy discover potential patient stratification focus identifying pathways deregulated when absolutely keeping alterations (e.g., oncogene "dependence" survive grow; Mancias Kimmelman, 2011; Cohen Consequently, it logical assume blockage specific inhibitors should lead arrest, death, regression. Using rational, would be possible select, means few markers, particular "addicted" pathways, modern medicine. frequently deregulated, although insufficiently therapeutically exploited pathway involves (Mertz 2011). influences significant involved growth, apoptosis (Dang, 1999, 2012; Prendergast, 1999; Schmidt, 1999). In fact, has been implicated pathogenesis one-third all human malignancies. As relates cancer, disease current study, was found originally amplified (Schleger 2002) using interphase fluorescence situ hybridization, overexpressed 40% 2002). However, recently, whole-exome sequencing microdissected revealed percentage gene 12% (Witkiewicz Early studies confirmed oncogenic role genetically engineered mouse models, upon overexpression display tumorigenesis (Morton Sansom, 2013). addition, variety experimental later shown upregulation sufficient induce formation without additional genetic manipulation any (Lin 2013), deletion allele decelerates development (Walz 2014), targeted RNAi approach blocks (Saborowski subsequent increase PGC-1α key determinant OXPHOS dependency stem (Sancho Interestingly, recent work, Wirth Schneider propose use marker (Wirth Schneider, All features indicate behaves driver PDAC. many have dedicated potent options (Soucek 2008; Annibali 2014; McKeown Bradner, Fletcher Prochownik, Key extraterminal family proteins (BET), efficiently inhibited compound, necessary activity (Nesbit Delmore Kandela Notably, suppresses mice inhibiting both inflammatory signals (Mazur Conversely, inhibition thought also essential mechanism suppress progression hematological malignancies (Knoechel Roderick Trabucco Thus, based their status testing response timely paramount medical importance. Several focused predictive inhibitors. Puissant al (2013) reported amplification MYCN medulloblastoma robust those JQ1. Moreover, certain rare called NUT midline carcinomas carrying tandem fusion BRD4 (nuclear protein testis) important (Stathis outside relatively examples very difficult predict efficient genomic approaches. overcome issue, tumoral can effective way develop signatures notably terms chemosensitivity. transcriptomic classifies PDAC, whose appears result prospective 16 independent determined third xenograft bear thus were likely respond treatment. Indeed, matching monolayer confirm prediction. conclude tools determine clinical interest select suggesting similar useful aimed at stratifying treatments. Results Selection or order patients, 30 obtained surgery 25 biopsy samples taken EUS-FNA implanted subcutaneously into preserved (PDX). histopathologic characteristics displayed Table 1. anatomopathological nuclear shape staining intensity, nucleocytoplasmic ratio, eosinophilia, mucins production, degree) after least successive passages (Duconseil Growth rates reach volume 1 cm3 ranged 2 6 months PDX. Total RNA PDX, performed Affymetrix platform. Subsequently, 239 RNAs regulated accordance v1 v2 list Signatures Database (MSigDB). Figure 1A represents hierarchical clustering heatmap top significantly high-expressed dendrogram showing distance between indicates presence two red blue colors, respectively. observe 17/55 (30.9%) 134/239 (P-value = 0.01996 q-value (FDR) 0.044). rank-listed available Appendix S1. gain insight biological processes enriched vs. subgroups, Set Enrichment Analysis (GSEA). Fig 1B, low differentiated phenotype associated (Normalized Score 3.60 FDR 0.00) DNA replication genome maintenance 3.24 0.00). contrast, reflect state such digestion −2.23 glycoprotein −1.76 complete statistically Biological Process, Curated Geneset Enriched, Hallmarks Enriched presented Datasets EV1, EV2, EV3, EV4, EV5, EV6. give PDX index, IHC-based scoring epithelial compartment 1C (left part), semi-quantitative reveals proliferate (Ki67 mean score 2.88 ± 0.25 [n 17] 2.06 0.12 38], P 0.0018). degree H&E paraffin-embedded tissues sections. (right shows (differentiation 0.77 0.2 1.82 0.08 0.0002). scores provided S1A B, analyzed outcome Kaplan–Meier analysis considering overall relapsing free cohort. 1D, median 9.2 18.8 (HR 2.43 [1.1–5.1]). relapse-free 5.6 11.5 subgroup, respectively 2.7 [1.3–5.8]). Altogether, observations activity. they poor expectancy. Combined, constitute solid characterization molecular, biological, xenografts, build trajectory toward Clinicopathological parameters Patient distribution (learning cohort) (%) Resectable Unresectable n Sex Male 34 (62) 18 Female 21 (38) 12 9 Age Mean 64 66 61 Min–Max 41–86 45–86 41–83 Other No 44 (80) 20 24 Yes 11 (20) Tumor location Head 32 (58) 19 13 Undefined (11) 0 Body 3 (5.5) Tail 14 (25.5) 5 Specimen type Primary 46 (84) Hepatic metastasis (9) Carcinomatosis 4 (7) diagnosis Localized 29 (53) 28 Locally advanced 8 (14.5) Metastasis (23.5) Identification cancer-derived Hierarchical non-supervised method. present patterns probe sets corresponding (Hugene 2.0 ST Array, Genechips). (n 17 patients) 38 patients). RMA normalized represented color relative (high red, blue). GSEA expression. Top groups (MYC-high MYC-low) represented; 825 MSigDB collections used. NES enrichment score, corresponds false rate. degree: Samples IHC scored (0 negative maximal staining) lowest differentiation). Pictures representing Appendix. **P 0.0018; ***P 0.0002 (mean SEM, 38, unpaired t-test two-tailed). curves (upper graph) (lower subgroups. P-values calculated log-rank test. data online figure. Data [emmm201606975-sup-0009-SDataFig1.pdf] Download figure PowerPoint used classifying subtypes classify subtypes, total genes. first (Figs 2C) identified upregulated obtain downregulated top-score whole profiles (Fig 2A B). 2. For each patient, 60 ratios computed centered normalization (see method) revealing p