Abstract PI3Kα is highly mutated in cancer resulting in hyperactivation of lipid kinase activity and downstream AKT signaling. H1047 is the most common site of oncogenic mutation and occurs in ~14% of all breast cancers. Initial therapeutic benefit of targeting PI3Kα was established with alpelisib, an alpha-selective PI3K inhibitor that is equipotent against wild-type and mutant forms. However, wild-type PI3Kα inhibition results in frequent dose-limiting toxicities including hyperglycemia, restricting the full potential of this drug. Selective targeting of H1047X-mutant PI3Kα is expected to both improve anti-tumor activity and reduce toxicity. STX-478 is an allosteric, CNS-penetrant, selective PI3Kα H1047X inhibitor, having excellent drug-like properties and exceptional kinome selectivity. STX-478 demonstrated minimal inhibition of CYP enzymes in vitro, supporting the potential for combinations with a wide range of therapeutics in breast cancer and a variety of other tumor types. STX-478 selectivity extended to the inhibition of other activating kinase domain mutations in biochemical assays. In a diverse panel of PI3Kα H1047X mutant cell lines, STX-478 selectively reduced the cellular levels of pAKT (S473) with a strong correlation between pAKT inhibition and cell viability (R = 0.8). In a high-throughput viability screen of 467 cancer cell lines, the presence of PIK3CA H1047X and other kinase domain mutations were the single strongest predictor of STX-478 sensitivity with potency superior to alpelisib. STX-478 also selectively inhibited the proliferation of cell lines with PI3Kα helical domain mutations, potentially due to the selective dependency of these cells on mutant PI3Kα. When combined with fulvestrant, lapatinib, or abemaciclib, STX-478 demonstrated synergistic anti-proliferative activity in cell lines with relevant ER/HER2 status. Unlike alpelisib, STX-478 did not impair glucose metabolism or cause insulin resistance at efficacious doses. In the T47D (PI3Kα H1047R) breast cancer model, STX-478 (100 mg/kg) monotherapy caused tumor regression whereas alpelisib caused only stasis. STX-478 combination with fulvestrant was well-tolerated, with more consistent and deeper tumor regression. Similar results were observed in a PI3Kα H1047R mutant ER+/HER2- PDX model, where fulvestrant monotherapy showed minimal activity, while combination with STX-478 yielded tumor regressions. In an ER+/HER2+ PDX model (PI3Kα H1047R/R108H), palbociclib and STX-478 (100 mg/kg) monotherapy resulted in similar efficacy while the combination was well tolerated and yielded tumor regression. Together these data indicate robust STX-478 monotherapy activity that was well tolerated and improved when dosed in combination with fulvestrant or CDK4/6 inhibitors. Finally, we investigated the effect of STX-478 treatment in an ER+ PDX model carrying a helical domain mutation. STX-478 treatment resulted in tumor growth inhibition at doses that did not result in metabolic dysfunction, suggesting that STX-478 may also be efficacious in treating PIK3CA mutant tumors with helical domain mutations. In summary, STX-478 efficacy was superior to alpelisib at a dose level that exceeds the clinically relevant exposure in mice without causing metabolic dysfunction. STX- 478 has a predicted low human dose, CNS exposure, low risk of DDI, and a predicted long half-life with minimal variation in peak-to-trough plasma concentrations which further supports a favorable therapeutic index. STX-478 has the potential to provide a best-in-class profile to improve outcomes in patients harboring tumors with prevalent PI3Kα H1047X mutations as well as other kinase and helical domain mutant tumors. The significant CNS exposure of STX-478 is expected to enable this treatment for patients with brain tumors and brain metastases not afforded by existing options. STX-478 is currently in IND enabling studies and is expected to enter human clinical trials in 2023. Citation Format: Leonard Buckbinder, David J. St. Jean, Brendon Ladd, Trang Tieu, Philip Jonsson, Jacob Alltucker, Samantha Manimala, Weixue Wang, Angel Guzman-Perez, Darrin D. Stuart, Gregory Dowdell. STX-478, a mutant-selective PI3Kα H1047X inhibitor clinical candidate with a best-in-class profile: Pharmacology and therapeutic activity as monotherapy and in combination in breast cancer xenograft models [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P4-07-04.