Abstract 342: Sub-group of HER2-negative breast cancer patients with hyperactive RAS network signaling identified: Dynamic pathway activity test identifies patients that may benefit from PI3K/mTOR or PI3K/mTOR/BCL inhibitors

Background: Dysregulated signaling through GPCRs is implicated in oncogenic RAS signaling in breast cancer (BC). The complex regulatory mechanisms that link RAS nodes can trigger an adaptive response (e.g. resistance) when a single RAS node is inhibited. Inhibition of multiple RAS nodes may thus be required to achieve durable antitumor responses. To identify patients with dysregulated RAS signaling tumors that may respond to RAS node inhibitors, an assay using an impedance biosensor was developed. The CELsignia RAS Activity Test measures GPCR-initiated signaling activity and PI3K, mTOR, and BCL's role in transducing this activity in live tumor cells. The current study set out to characterize the prevalence of dysregulated RAS signaling initiated through a GPCR in HER2-negative BC patients and the role played by PI3K, mTOR and BCL. Methods: Fresh tumor specimens from 69 HER2- BC patients were used to derive tumor cell samples. Dynamic live cell response to a GPCR agonist (LPA), a PI3K-α inhibitor (GDC-0077), a pan-PI3K inhibitor (copanlisib), a pan-PI3K/mTOR inhibitor (gedatolisib), and a BCL inhibitor (navitoclax) was measured using an xCELLigence impedance biosensor (Agilent Technologies). From these responses, the gross amount of GPCR-initiated signaling and corresponding participation of PI3K-α, all Class 1 PI3K-isoforms, mTORC1, and BCL was quantified and converted to a signaling score. Correlative analyses using FACS markers (RPS6, AKT, ERK, caspase 3) were performed. Results: Hyperactive RAS signaling (RASs+) initiated by GPCR pathways was found in 16 of 69 (23.1%; 95% CI=15%-34%) tumors. The signaling scores were bimodally distributed; for the 16 RASs+ and the 53 RASs- patient tumors, the mean scores were 486 (SD 248) and 91 (SD 73), respectively. For the 16 RASs+ tumors, inhibition of all Class 1 PI3K-isoforms attenuated, on average, 71% of the GPCR signaling, while inhibition of PI3K-α alone had only a nominal inhibitory effect. When mTOR inhibition was combined with pan-PI3K inhibition, GPCR signaling was further attenuated by 14%. The most complete attenuation of RAS-signaling occurred when PI3K, mTOR, and BCL were simultaneously inhibited. Early apoptosis markers (RPS6, AKT, ERK, caspase 3) were found most significantly in RASs+ tumors with pan-PI3K/mTOR inhibition and to a greater degree with pan-PI3K/mTOR/BCL inhibition. No apoptosis markers were found in RASs- tumors regardless of which RAS node was inhibited. Conclusions: These findings sug