A recent study of the mutation landscape of >3000 cancers across 12 major cancer types from the Cancer Genome Atlas (TCGA) program revealed PIK3CA as the second most commonly mutated gene, occurring at >10% frequency across most cancer types, with mutations occurring mostly in the PIK3CA kinase (39% overall, 52% for breast cancer) and helical (45% overall, 39% for breast cancer) domains and their most frequent hotspot mutations being H1047R and E545K, respectively. Limited preclinical evidence has suggested that mutations affecting the kinase and helical domains can produce different phenotypic consequences; and our recent analysis of the TCGA data indicates that helical domain mutations, relative to kinase domain mutations, are preferentially associated with pathway/PARADIGM activity features enriched for cell motility mechanisms (e.g. RHO GTPases) and loss of gap junctions, suggesting that helical domain mutations might serve to enhance cell invasiveness and metastatic motility. To validate the functional differences predicted by these two classes of PIK3CA mutations, we turned to the non-malignant breast epithelial cell line MCF10A bearing targeted pathogenic knock-ins of H1047R or E545K at the PIK3CA locus; and we compared RNA-seq expression profiles from the wildtype (wt) and knockin isogenic clones. RSEM was used to estimate gene and isoform abundance; and EBseq was used to perform differential expression analysis. In a three-group comparison, over 2500 differentially expressed genes were identified as differentiating wt cells and the PIK3CA mutated MCF10A clones, with the greatest transcriptional changes (2043 genes) observed between either kinase or helical domain mutations and wt cells. Functional enrichment analysis (DAVID) revealed enrichment of 17 pathway categories among these 2043 genes, representing 3 major functional groups, the most prominent being protein translation. Only 409 genes were differentially expressed across all three groups, suggesting the transcriptome effect differentiating E545K from H2047R bearing MCF10A clones was less pronounced. In a direct comparison of the kinase and helical domain clones, >900 differentially expressed genes were identified that showed enrichment in 64 pathways representing altered cell morphology, motility, apoptosis and phosphorylation, but without enrichment in protein translation function. In particular, MCF10A E545K clones showed significant loss of expression in 19 tight junction-associated genes linked to oncogenic EMT (e.g. OCLN, CLDN-7, TJP1, MYL9). Thus, consistent with our pathway conclusions drawn from the TCGA clinical dataset, helical domain mutations appear to confer an EMT-like expression phenotype beyond the effects of both helical and kinases domain mutations on protein translation.

Citation Format: Christina Yau, Joe Garay, Stephen Benz, Joe Gray, Ben Park, Christopher C. Benz. An EMT-like gene expression profile differentiates helical from kinase domain PIK3CA mutations. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 1084. doi:10.1158/1538-7445.AM2015-1084