PIK3CA and ARID1A co-altered tumors are sensitive to AKT inhibitor, capivasertib

Background: PIK3CA (PK) and ARID1A (AD) co-alterations occur in 1% of cancers. These cancers had enhanced sensitivity to PI3K inhibition with copanlisib (cop) in the MATCH trial arm Z1F. Given the recent FDA approval of the AKT inhibitor, capivasertib (capi), here we aim to evaluate the potential benefit of capi treatment in the setting of PK and AD alterations and the mechanism by which enhanced sensitivity might be occurring. Methods: CRISPR/Cas9 was used to knockout AD from SW48PIK3CA-H1047R (SW48PK), and PIK3CAH1047R mutant locally advanced rectal cancer (LARC) patient-derived cancer organoids (PDCOs). PK or PKAD Colorectal (CRC), ovarian (OC), and renal (RC)PDCOs were obtained from the NCI repository and were treated with 2 µM capi for 48 hrs. Percent relative change in the longest diameter (PRC) was measured prior to and after treatment in(capi or control) and effect size was measured using Glass's delta (GD) . Western blot was used to determine levels of PI3K proteins (RPS6, AKT), apoptotic proteins (MCL-1, BCL-xL, Bax, Bims), and mTORC1 protein, PRAS40. Athymic nude mice were flank injected with SW48PK and SW48PKAD and treated with 100mg/kg capi (10% DMSO in 90% of 20% SBE-beta-CD in saline) or vehicle b.i.d for 28 days or until moribund. Ki-67 and cleaved caspase-3 in these tumors were assessed by immunohistochemistry (IHC) and quantified as positive cells/high powered field (HPF). Results: Capi treatment responses were observed in PKAD co-altered CRC PK (PRC 4.8%, GD 2.0), LARC-PK (0.5%, 0.9), OC-PKAD(10%, 1.03), RC-PKAD (1.52%, 1.52), and LARC-PKAD (2.91%, 1.78) PDCOs, indicating sensitivity across cancer types and enhanced sensitivity in co-altered lines relative to PK alone for LARC. Western blot showed greater reduction in p-RPS6 in LARC-PKAD by 6h compared to LARC-PK, but no significant differences were seen in 2D lines. p-AKT increased upon capi treatment across all 2D and 3D lines. Among apoptotic proteins, BCL-xL steadily increased over time in co-mutant cultures. Bims (cytotoxic isoform of Bim) had a stronger induction in 6h in LARC-PKAD compared to LARC-PK. Levels of p-PRAS40, MCL-1, Bax, and Bak remained largely comparable across lines. In vivo, PKAD tumors showed a significant tumor growth delay (p = 0.01) with capi treatment, while PK tumors did not (p = 0.06). Conclusion: PIK3CA and ARID1A co-altered tumors have enhanced sensitivity to capi relative to PIK3CA mutant tumors, potentially through induction of Bims. Capi should be investigated further clinically in PIK3CA and ARID1A co-altered cancers. Additionally, further studies identifying novel drug combinations with capi are warranted to further enhance patient response