Abstract P3-10-18: Elucidating the role of functional signal transduction pathway activity in sensitivity and response of triple negative breast cancer to PI3K inhibition

Abstract

Abstract Introduction The PI3K signaling pathway is frequently active in triple negative breast cancer (TNBC), but response to PI3K inhibition is still poorly understood. To gain insights in therapy response and resistance, it is important not only to assess genetic alterations of these tumors, but also their phenotypic characteristics. Furthermore, not only is the functional status of the PI3K pathway important to know, but also other signal transduction pathways that may drive tumor growth, as they may explain primary or acquired therapy resistance. Material & Method We analyzed 17 patient derived xenograft (PDX) models of TNBC with varying response to buparlisib treatment. Long-term response (˜30 days) was assessed by comparing growth rates between buparlisib and vehicle treated PDXs. For characterization, mRNA expression analysis was performed on tumor material post 3 days of vehicle or buparlisib treatment of each PDX model. Functional pathway activity was determined using computational Bayesian networks that look at mRNA levels of pathway target genes resulting from activation (Verhaegh et al, Cancer Res 2014), amongst others for the FOXO-PI3K (van Ooijen et al, Am J Pathol, in press), ER, AR, Hedgehog, TGFbeta and Wnt pathways. These computational networks were calibrated on samples with known pathway activity, and biologically validated on various healthy and diseased cell and tissue types. Results On an initial set of 6 PDX models, pathway activity clearly varied between different TNBC PDXs, and between vehicle and buparlisib treatment. ER pathway activity was low in all samples, as expected in TNBC. Two PDX models with the most growth reduction by buparlisib showed high PI3K activity, of which one based on low FOXO activity and the other on oxidative stress. The former, best responding PDX showed a clear reduction in PI3K activity (restoring FOXO activity), when comparing 3 days buparlisib to vehicle treatment. Of the two PDX models with least growth reduction, one had low PI3K activity, while the other one, carrying a PIK3CA mutation, did show high PI3K activity (low FOXO), but this remained after treatment. Other differences in pathway activity that were found, included slightly elevated AR and Wnt activity in one PDX with good response, and somewhat higher TGFbeta activity in four PDXs (good, medium and poor response). Analysis of the remaining 11 PDXs is ongoing, including other signal transduction pathways, to investigate the variation in pathway activity across the entire panel and to shed more light on the differences in tumor biology between the PDX models. Conclusion Our computational Bayesian networks measured differences in functional activity of signal transduction pathways across a collection of TNBC PDX models, which may be expected due to the large variation between TNBC tumors. Functional PI3K activity was related to growth inhibition by buparlisib treatment, and reduction of PI3K pathway activity was observed upon treatment in responding PDX models. Other pathways also showed variation across PDXs. Further clinical evaluation of our signal transduction pathway activity measurement is ongoing. RT-qPCR based analysis is available, optimized for FFPE tissue and small samples. Citation Format: Guo Z, Verhaegh W, Hoog J, Rajput S, van Strijp D, van de Stolpe A, Li S, Ma C. Elucidating the role of functional signal transduction pathway activity in sensitivity and response of triple negative breast cancer to PI3K inhibition [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P3-10-18.