Phospho-Proteomics Reveals Oncogenic Phospho-Tyrosine Signaling Networks in Cancers Lacking Mutated Or Amplified Tyrosine Kinases

Introduction: Many human tumors exhibit aberrant activation of tyrosine kinase signaling due to mutation or DNA amplification of tyrosine kinases (eg, EGFR). In contrast, prostate cancer and some subtypes of glioblastoma (GBM) lack mutated or amplified tyrosine kinases. To investigate the regulation of phospho-tyrosine signaling in the absence of tyrosine kinase mutations or amplifications, we applied quantitative, label-free mass spectrometry to a GBM cell line, a mouse model of prostate cancer and human metastatic castration resistant prostate cancer (CRPC) biopsies.

Materials and Methods: Human GBM cells (U87) were treated with glucose starvation, epidermal growth factor (EGF), hydrogen peroxide (H₂O₂) or the phosphatase inhibitor vanadate. Mouse models of prostate cancer were generated by expressing combinations of non-tyrosine kinase oncogenes commonly found in prostate cancer (constitutively active AKT, androgen receptor (AR), ERG and K-Ras G12V) in the prostate in vivo regeneration model system (Lawson et al, 2009). Human treatment naïve and metastatic castration resistant prostate cancer (CRPC) samples were obtained from the UCLA Translational Pathology Core Laboratory and the University of Michigan Rapid Autopsy Program. Quantitative, label-free mass spectrometry was performed on peptides immunoprecipitated using a pan-specific anti-phospho-tyrosine antibody (Drake et al, 2012; Graham et al, 2012).

Results and Discussion: U87 cells, which lack mutated or amplified tyrosine kinases, were treated with glucose starvation, EGF, H₂O₂ or vanadate. Each stimulus induced a unique signature of phospho-tyrosine signaling (Fig. 1a, n = 46 peptides). Unexpectedly, glucose starvation induced phosphorylation of proteins associated with focal adhesions, which suggested additional experiments that revealed a positive feedback loop between tyrosine kinase signaling and NADPH oxidase activity within focal adhesions (Graham et al, 2012). Next, analysis of mouse models of prostate cancer expressing non-tyrosine kinase oncogenes (eg, activated Akt and overexpressed ERG) (Drake et al, 2012) revealed oncogene-specific signatures of phospho-tyrosine signaling (Fig. 1b, n = 147 peptides), including an enrichment of EGFR substrates in Akt-ERG tumors. Finally, analysis of human prostate cancer patient samples demonstrated differential activation of tyrosine kinase signaling in treatment naïve and metastatic CRPC tissues (Fig. 1c, n = 126 peptides), including the presence of Src kinase substrates in metastatic tumors.

Figure 1. Hierarchical clustering of phospho-tyrosine peptides in a) U87 cells, b) mouse models of prostate cancer c) human CRPC patient (Pt) samples (LN, lymph node; Met, metastasis). Rows represent phospho-tyrosine peptides and columns represent samples run in duplicate. Red and green denote high and low phosphorylation, respectively.

Conclusions: We applied quantitative, label-free mass spectrometry to investigate regulation of phospho-tyrosine signaling in systems without mutated or amplified tyrosine kinases. This analysis revealed that a) metabolic perturbation (ie, glucose starvation) induces a positive feedback loop leading to focal adhesion-related tyrosine kinase signaling in U87 cells, b) the non-tyrosine kinase oncogenes Akt and ERG activate EGFR signaling in a mouse model of prostate cancer, and c) metastatic human CRPC biopsies exhibit elevated Src signaling compared to treatment naïve biopsies. Together, these data demonstrate that cell lines and tumors lacking tyrosine kinase mutations or DNA amplifications can exhibit elevated tyrosine kinase signaling, suggesting that tyrosine kinase inhibitors may be effective therapeutics even in these cancers.

Acknowledgements: N.A.G. is supported by the UCLA Scholars in Oncologic Molecular Imaging (SOMI) program, NIH grant R25T CA098010. J.M.D. is supported by the Department of Defense Prostate Cancer Research Program W81XWH-11-1-0504.

References: Drake, J.M., PNAS (2012), 109(5), 1643-1648; Graham, N.A., et al, Mol Syst Biol (2012), 8:589; Lawson, D.A., et al, PNAS (2009).