Abstract 1213: Comparing protein pathway activation mapping portraits between gliobastoma patient-matched primary tumor, xenografts and neurospheres: implications for precision medicine

Abstract

Abstract Background: Most current targeted therapeutics are directed against the activated protein kinase-driven signaling pathway architecture. While guiding therapy selection based on protein phosphorylation is biochemically rational given these proteins represent the direct drug targets it is not certain which protein networks/signaling protein targets are the causal driver of any patient's tumor. Thus, the availability of a system that recapitulates the human in vivo setting and allows for credentialing and identification of which activated signaling networks are causally necessary for tumor viability would be extremely important. Methods: To test two such systems and determine how they resemble the cell signaling architecture of the human primary tumor we compared eight snap-frozen laser capture microdissected (LCM) primary glioblastoma multiforme (GBM) samples with tumor-matching LCM mouse xenografts as well as matching neurospheres (NP). Reverse phase protein microarrays were utilized wherein the total level or phosphorylated level of 70 key signal transduction proteins known to be the direct targets for a number of important targeted therapies and implicated in GBM tumor biology were measured. Results: Overall, the signaling architecture of the xenografts more closely resembled the matched primary LCM tumors in this study set. However, unsupervised hierarchical clustering of the data revealed instances where pathway information such as total c-MET, phospho RAF (S259), phospho GSK3αβ (S21/9) and total EGFR was retained across matched LCM primary tumor, xenografts and NP. In addition there were a number of instances where the signaling architecture of the primary human tumor did not resemble the matched NP and xenograft signature. Conclusion: These pilot data indicate that 3D cell culture conditions may effectively recapitulate in vivo signaling network activation in certain instances, although these preliminary results also suggest that xenografts more closely resemble the cell architecture of matching primary GBM tumors. These data highlight the importance and impact of the microenvironment and culture conditions on tumor cell signal transduction network activation. However, two caveats to this approach remain: (a) drug studies in GBM xenografts and NP are limited to those targeting signaling pathways that are faithfully reproduced from the human in vivo setting and (b) the time required to develop a patient matched xenograft model for individualized drug testing may be too long for a fast-progressing disease such as GBM. Citation Format: Claudius Mueller, Ana C. deCarvalho, Tom Mikkelsen, Laila Poisson, Valerie Calvert, Andrew Borgman, David M. Cherba, Mary E. Winn, Emanuel F. Petricoin. Comparing protein pathway activation mapping portraits between gliobastoma patient-matched primary tumor, xenografts and neurospheres: implications for precision medicine. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 1213. doi:10.1158/1538-7445.AM2014-1213