Abstract 4409: Prospective identification of RB pathway alterations predict response to SY-1365, a selective CDK7 inhibitor, in a panel of high-grade serous ovarian cancer (HGSOC) patient derived xenograft (PDX) models

Abstract

Introduction: CDK7, a key regulator of transcription and cell cycle progression, has been implicated in the pathogenesis of ovarian cancer (OC). SY-1365, a potent and selective inhibitor of CDK7, has been shown to induce tumor growth inhibition (TGI), including complete regressions, in OC PDX models, with an association observed between responses and oncogenic alterations in the RB pathway. Here we report on the predictive value of RB pathway alterations in an independent set of OC PDX models. Methods: Tumor growth inhibition (TGI) was evaluated in 21 independent HGSOC PDX models by comparing growth in SY-1365-treated mice (30-40 mg/kg, twice weekly, 4 weeks) versus growth in vehicle treated mice (n=3 per group). DNA and RNA were extracted from FFPE tumor tissue collected from untreated mice for all 21 PDXs. Gene mutations and copy number (CN) were evaluated by DNA sequencing (OncoCODE410, WuXi). Gene expression was assessed using the Nanostring Pan Cancer Pathway panel. RB pathway changes were defined per The Cancer Genome Atlas analysis of HGSOC - RB1 deletion or mutation, CDKN2A downregulation (DR) or deletion, CCNE1 amplification (Amp), CCND1 Amp, CCND2 upregulation (UR) - which identified these changes in 67% of HGSOC patients. Genes with homozygous deletions (HD), high-level Amp (CN≥6), or deleterious mutations were considered altered; genes with normalized expression values in the upper and lower ranges (Z score±1.5) across all models were considered altered. Models with 1 or more alterations were classified as RB-incompetent and were predicted to respond to SY-1365. Results: Assessments of SY-1365-induced-TGI and RB pathway changes were conducted independently in a blinded manner. Of the 21 models tested, 12 (57%) showed responses (greater than 50% TGI, range 51-87%). RB pathway changes were observed in 7 (33%) of the PDX models: CCNE1 Amp (n=2); CCND2 UR (n=2); CCND2 UR and CCND1 Amp (n=1); CCND2 UR and CDKN2A DR (n=1); CDKN2A HD/DR and RB1 mutation (L335*) resulting in a premature stop codon (n=1). All RB-incompetent models (predicted responders; n=7) showed anti-tumor responses to SY-1365 (100% positive predictive value). Of the 14 predicted non-responder PDX models, 5 showed responses suggesting the presence of additional undetected RB pathway changes or alternative mechanisms conferring SY-1365 sensitivity. Overall, a balanced accuracy of 79.2% (sensitivity=58.3%, specificity=100%) was observed (p Conclusions: RB pathway alterations are predictive of response to SY-1365 in HGSOC PDX models. The results support exploration of RB pathway changes as predictive biomarkers of SY-1365 clinical activity in patients with ovarian cancer. SY-1365 is currently being assessed in a phase 1 trial in adult patients with ovarian and breast cancers (NCT03134638). Citation Format: Nan Ke, Liv Johannessen, Nisha Rajagopal, David Orlando, John Carulli, Graeme Hodgson. Prospective identification of RB pathway alterations predict response to SY-1365, a selective CDK7 inhibitor, in a panel of high-grade serous ovarian cancer (HGSOC) patient derived xenograft (PDX) models [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 4409.