Abstract A28: Evaluation of CDK12 protein expression as a novel biomarker for PARP1/2 inhibitor sensitivity in breast cancer

Abstract

Abstract The CycK/CDK12 complex is known to protect normal cells from genomic instability by regulating the DNA damage response (DDR) and has been postulated as a tumor suppressor gene in high-grade serous ovarian cancer. We have previously shown that in breast cancer, CDK12 is recurrently targeted by both DNA rearrangements (13% of HER2-amplified breast cancers) and recurrent point mutations (2.6% of unselected breast cancers). Furthermore, we have shown that loss of CDK12 confers sensitivity to PARP1/2 inhibitors, and therefore may constitute a biomarker of response. The aims of this study were i) to determine the frequency of CDK12 protein loss in unselected breast cancers, and any potential correlations with prognosis and/or molecular subtype; ii) assess the frequency of CDK12 protein loss in HER2-positive breast cancer, given the high frequency of CDK12 breakpoints in HER2-amplified tumors and correlate this with response to anti-HER2 therapy and iii) to identify additional genomic mechanisms of CDK12 disruption. We stained two tissue microarrays (TMA) with a validated antibody against CDK12 (Clone 57311, Abcam). The first cohort comprised 794 unselected primary breast cancers and the second comprised 121 HER2-positive breast cancers that had been treated with Herceptin. CDK12 expression was quantified using a modified Allred score, with a score of 0 being considered negative and a score of 7 or 8, positive. Correlation statistics were calculated using SPSS with a Pearson's Chi Square or Fisher's Exact Test p-value of <0.05 indicating significance. In order to corroborate our findings, we re-analysed publicly available data from The Cancer Genome Atlas (TCGA) and METABRIC datasets, to ascertain the frequency of CDK12 copy number breakpoints, somatic mutations and methylation and correlate these with RNA expression levels. In the TMA tumor cohort, 73 tumors were CDK12 negative (9%), and 146 were CDK12 positive (18%). In both these groups, a significant correlation with HER2 expression was seen, with 96% of CDK12 negative tumors being HER2 negative (p = 0.006) and only 39% of CDK12 positive tumors being HER2 positive (p < 0.001). CDK12 positive tumors showed a significantly poorer breast cancer specific (p = 0.023) and disease specific (p = 0.046) survival as compared with tumors with low or absent CDK12. In addition, CDK12 loss significantly correlated with both PARP1 cleaved (p = 0.002) and PARP1 non-cleaved (p = 0.007) expression. Interestingly, 60% of CDK12 positive tumors expressed both estrogen (ER) and progesterone receptors (PR; p = 0.039), whilst loss of CDK12 significantly correlated with a triple negative phenotype (p = 0.009). In the HER2-positive cohort, four tumors were CDK12 negative (2%), and 71 tumors were CDK12 positive (59%). Loss of CDK12 expression did not improve survival in this cohort following Herceptin treatment. In the METABRIC cohort HER2-amplified tumors that had a breakpoint within CDK12 (n = 26) had significantly lower levels of mRNA expression than CDK12-amplified tumors (n = 175; p<0.001). Mining of TCGA data identified CDK12 mutations seen in 1.5% of unselected breast cancers; 1.4% of ER positive tumors; 2% of triple negative cancers; and 1.2% of HER2-positive tumors (which also expressed ER). Of these, 68% were predicted to disrupt protein function, which correlated with a significant decrease in CDK12 mRNA expression (p = 0.046). Furthermore, mutations in CDK12 showed a degree of mutual exclusivity with mutations in other DNA repair genes including BRCA1, ATM, ATR and PALB2, indicating some degree of epistasis. CDK12 promoter methylation was a rare event seen in 1/641 (1.6%) unselected breast cancers. In conclusion, our results highlight that CDK12 is disrupted via multiple genomic mechanisms and is a potential tumor suppressor gene in breast cancer. Loss of CDK12 protein expression is seen in 9% of unselected breast cancers and 18% of triple negative breast cancers, and may be a worthwhile treatment option for triple negative patients. Routine diagnostic quantification of CDK12 protein in tumors using immunohistochemistry is feasible, making it a viable biomarker for response to PARP1/2 inhibitor therapy. Citation Format: Kalnisha Naidoo, Patty Wai, Sarah Maguire, Frances Daley, Yinyin Yuan, Emad A. Rakha, Ian O. Ellis, Christopher J. Lord, Andrew R. Green, Rachael C. Natrajan. Evaluation of CDK12 protein expression as a novel biomarker for PARP1/2 inhibitor sensitivity in breast cancer. [abstract]. In: Proceedings of the AACR Special Conference on Advances in Breast Cancer Research; Oct 17-20, 2015; Bellevue, WA. Philadelphia (PA): AACR; Mol Cancer Res 2016;14(2_Suppl):Abstract nr A28.