PD42-07: Investigating the synthetic lethality of Enhancer of zeste homolog 2 (EZH2) inhibition in AT-rich interactive domain-containing protein 1A (ARID1A) mutant bladder cancer

James Ferguson, Hasib Rehman, Darshan Chandrashekar, George Netto, Soory Varambally

Introduction: Next-gen sequencing of bladder cancer (BCa) promises to move the field towards better risk stratification, therapeutic target identification, and more personalized therapies for patients. Specifically, genes involved in epigenetic modifications such as ARID1A are frequently mutated in BCa.  Previously, we have shown that EZH2 is over-expressed and required for cancer growth. EZH2 functions as the catalytic subunit of the polycomb repressive complex 2 (PRC2) which methylates lysine 27 on histone 3 (H3K27me), resulting in transcriptional silencing. Herein, we show that ARID1A mutations sensitize BCa cells in vitro and in vivo to EZH2 inhibition with the small molecule GSK-126.

Methods: In silico analysis (TCGA) compared disease-free survival (DFS) between ARID1A mutant (ARID1Amut) and wildtype (wt) tumors. Western blot compared EZH2, ARID1A, and H3K27me3 protein levels between matched pairs of BCa and normal urothelium from cystectomy specimens. Cell proliferation assays were performed with/without EZH2 inhibitor GSK-126. Stable ARID1A knockdown cell lines were also tested. Murine xenograft experiments were performed using ARID1Amut and wt BCa cell lines to compare tumor growth inhibition by GSK-126. RNA microarray evaluated differentially expressed genes in ARID1Amut and wt BCa cell lines with/without GSK-126. Western blot and chromatin-immunopreciptation (ChIP) analysis confirmed the microarray findings.

Results: In silico analysis revealed that ARID1Amut BCa has worse DFS compared to ARID1Awt BCa. EZH2 and resultant H3K27me3 protein levels are dramatically increased in BCa compared with normal urothelium, with a decrease in ARID1A protein levels. In vitro studies showed that only ARID1Amut and not wt cell lines are inhibited by EZH2 inhibitor GSK-126. ARID1A knockdown in 5637 cells resulted in de novo GSK-126 sensitivity. In vivo murine xenograft models confirmed this effect. Transcriptomic analysis comparing ARID1Awt and mut cell lines with/without GSK-126 revealed differential expression of several genes that could explain the mechanism GSK-126 sensitivity. Western blot and ChIP confirmed that these genes are EZH2 transcriptional targets and are strong candidates to explain the molecular mechanisms underlying EZH2 and ARID1A crosstalk.

Conclusions: ARID1A mutation is a biomarker for EZH2 inhibitor sensitivity in BCa cells and may represent a new epigenetic therapeutic target for patients. Source of

Funding: None