Ribociclib&Belinostat In Patients w Metastatic Triple Neg Breast Cancer & Recurrent Ovarian Cancer w Response Prediction By Genomics

We hypothesize that ribociclib plus belinostat will be a well-tolerated and demonstrate activity in women with metastatic triple-negative breast or recurrent ovarian cancers. This trial combines two targeted agents, a CDK inhibitor and a lysine deacetylase (DAC) inhibitor for use as the first targeted treatment for metastatic triple-negative breast cancer. This in itself would be a leap forward for women who currently endure sequential treatment with cytotoxic chemotherapies. However, we will also be using samples from this trial in NIH-funded research on clonal evolution and geographic heterogeneity in breast cancer. Blockade of CDK4/6 and of lysine deacetylases (DAC) are both promising treatments for breast cancer. DAC inhibitors have been shown to have single-agent activity against triple-negative breast cancer in vivo, possibly through decreasing EMT, decreasing migration, and decreasing metastasis \[6, 7\] CDK4/6 inhibitor monotherapy causes cell cycle arrest and tumor regression in triple-negative breast cancer models. \[8, 9\] However, neither approach has been successfully translated into clinical use as a monotherapy. There are several lines of evidence demonstrating that inhibiting DAC increases the sensitivity of breast cancer to CDK4/6 inhibition. Recently, we showed that inhibitors of cell cycle progression, including CDKN1A, CDKN1C, CDKN2B, and CDKN2D, are increased by DAC inhibition in breast cancer cell lines, including triple-negative cell lines. \[10\] Others have shown cell cycle arrest by DAC inhibitors due to increases in p21 or decreases in CDK2 and cyclin A. \[11-13\] We also see an increase in RB1 expression in breast cancer cell lines. In the TCGA breast cancer analysis, RB1 was frequently down-regulated but not mutated or lost in triple-negative breast cancer. \[14\] Therefore, we hypothesized that DAC inhibition would increase the sensitivity of breast cancer to CDK4/6 inhibition. We first tested this hypothesis in a small panel of cell lines. We showed and published that all showed synergy between a CDK4/6 inhibitor (PD-0322991) and DAC inhibitors. \[10\] Since then, we have tested the combination of the DAC inhibitor vorinostat and the CDK4/6 inhibitor PD-322991 in a larger panel of cell lines and have shown at least additivity in most triple-negative cell lines. Indeed, it seems that those cancers least sensitive to CDK4/6 inhibition may benefit the most from co-treatment with a DAC inhibitor. Recently, we investigated ribociclib and belinostat in six triple-negative breast cancer cell lines. All 6 cell lines showed growth inhibition with belinostat and panobinostat. Synergy was seen in four of the six cell lines at multiple doses. We have also recently completed an analysis on a window of opportunity trial of the DAC inhibitor high-dose valproic acid, in breast cancer.\[15\] In that study, women were treated with valproic acid for one week before definitive therapy began. Fifty percent of women with triple-negative breast cancers had a significant decrease in Ki-67 with one week of valproic acid, providing proof of principle biologic activity of DAC inhibition in triple-negative breast cancer. (By contrast, no HER2 positive tumor had a decrease in Ki-67 with valproic acid treatment.) Given the genomic similarities between serous ovarian cancer and triple-negative breast cancer seen in the TCGA and pan-cancer analyses \[16\], we have chosen to include ovarian cancers in the dose-escalation component to increase accrual rate but not in the dose-expansion. Neither CDK4/6 inhibition nor DAC inhibition have a well-defined predictive biomarker to determine which cancers will respond and which will not. Because of the complex genomic effects of inhibition of the 11 DACs and the multiple targets of DACs, including histones, p53, etc, it is unlikely that a single gene or protein alteration will predict sensitivity. We have published a method for predicting sensitivity to targeted drugs using genome-wide gene expression analysis. \[17\] After demonstrating proof of principle by accurately predicting sensitivity to vemurafenib, we developed a genomic biomarker for the DAC inhibitor valproic acid. We then validated that this biomarker predicted sensitivity in vitro and in vivo. In the window of opportunity discussed above, our genomic signature had an AUC of 0.66 for predicting decreased Ki-67 with one week of treatment with valproic acid. We will develop a similar multi-gene gene expression based biomarker for the combination of ribociclib and belinostat.