Abstract P3-08-04: The genomic and transcriptomic landscape of PIK3R1 mutated breast cancers

Abstract

Abstract Background: Human breast cancers may harbor pathogenic somatic alterations in PIK3R1, which encodes the regulatory subunit—p85a—of the PI3K signaling complex. Prior studies developed an isogenic cellular system lacking p85a and investigated therapeutic approaches for breast cancers that lack functional p85a. For instance, somatic loss of PIK3R1 may sensitize breast cancer cells to MEK inhibition; work in patient-derived xenograft models confirmed this observation. Here, we wanted to investigate the significance of PIK3R1 mutations (PIK3R1MUT) in human breast cancer by using real-world data to characterize the genomic landscape of breast cancer patients with PIK3R1MUT. Additionally, our methodology allows us to assess the effect of PIK3R1MUT on corresponding mRNA expression-levels, tumor mutational burden (TMB), and microsatellite instability (MSI) to better understand the molecular-level effects of this important gene. Methods: We used the Tempus LENS platform to retrospectively analyze next-generation sequencing (NGS) data from 3400 HER2-negative (HER2-) advanced or metastatic breast cancer patients with confirmed hormone receptor status (HR+/-). Our cohort consisted of molecularly profiled, deidentified breast cancer cases using the Tempus xT solid tumor assay (DNA-seq of 595-648 genes at 500x coverage, and full transcriptome RNA-seq). This assay assesses mutations in both germline and somatic tissue, and characterizes nucleotide variants, insertions/deletions, and copy number variations. Results: The prevalence of PIK3R1MUT in HER2- breast cancer was 2.3% (77/3400). PIK3R1MUT occurred more frequently than expected in HR- breast cancer relative to the proportion we observed for wild-type PIK3R1 (Chi-squared test, p<0.001). In the HR+ subtype, high TMB (defined as 10 mutations/MB) occurred more frequently in PIK3R1MUT tumors than in samples with wild-type PIK3R1 (PIK3R1WT) status (Fisher’s exact test, p=0.039). Further, MSI-high status was absent in PIK3R1MUT tumors but was observed in PIK3R1WT samples. We assessed co-mutational patterns and found that pathogenic or likely pathogenic mutations in PTEN, TP53. and NF1 were more frequent in PIK3R1MUT samples, whereas mutations in PIK3CA were more frequent in the PIK3R1WT cohort (Table 1). At the transcriptional-level, PIK3R1 expression was similar in HR+ samples regardless of PIK3R1 mutational status but PIK3R1 transcript expression was significantly higher among PIK3R1MUT samples (Wilcoxon rank-sum test, p<0.001). Conclusions: Our study used real-world evidence to build on previous pre-clinical studies, and illustrates the importance of PIK3R1MUT in breast cancer. We found that certain mutations associated with poor outcomes and endocrine therapy resistance (e.g., PTEN, and NF1) were more frequent in PIK3R1MUT tumor samples. Interestingly, we did not find evidence that PIK3R1MUT results in decreased PIK3R1 mRNA expression but instead observed that in the HR- subtype PIK3R1MUT gene expression was significantly higher. Overall, this study shows that PIK3R1 may be an important therapeutic target in breast cancer. Table 1.- Co-mutations and PIK3R1 transcript abundance in PIK3R1MUT vs PIK3R1WT samples.Correlation of Gene Co-mutations with PIK3R1MUTGenesPIK3R1MUT (n=77)n (%)PIK3R1WT (n=3323)n (%)p-valuePositively correlatedPTEN21 (27%)176 (5.3%)<0.001NF113 (17%)134 (4.0%)<0.001TP5344 (57%)1289 (39)0.001Negatively CorrelatedPIK3CA9 (12%)861 (26%)0.005PIK3R1 gene expressionSub-typeNormalized expressionNormalized expressionp-valueHR+3.34 (n=30)3.28 (n=1720)0.2HR-3.45 (n=33)3.23 (n=678)<0.001 Citation Format: Melody Cobleigh, Emmanuel Okeke, Brett Mahon, Elizabeth Mauer, Alex Barrett, Abde Abukhdeir. The genomic and transcriptomic landscape of PIK3R1 mutated breast cancers [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P3-08-04.