100 CHAPTER 4 The 22 phosphosites upregulated in resistant patients, 4 of which were uniquely identified in this group, were linked to various immune processes by gene ontology analysis, such as response to interleukin-18, immune response and immune effector process. The 56 phosphosites upregulated in sensitive patients (of which 35 uniquely identified) were linked to various cellular regulatory and signaling processes, such as enzyme linked receptor protein- and transmembrane receptor protein tyrosine kinase signaling pathways, peptidyl-tyrosine autophosphorylation, positive regulation of cell motility and VEGFR and Epidermal Growth Factor Receptor (EGFR) signaling pathways (Supplementary Figure 3). Supplementary Table 4 lists the role of proteins corresponding to the candidate phosphosite signature according to available literature. Since tyrosine kinase inhibitors such as sunitinib specifically target aberrant kinase signaling, a functional analysis of activated kinases is essential for a good understanding of sensitivity to sunitinib treatment. To this end, we performed INKA24,46-48 analysis to further explore the differences in tumor biology between individual sensitive and resistant patients. Overall, 51 unique tyrosine kinases were identified in 23 patients. For each patient, the top-20 most activated kinases were ranked (Supplementary Figure 4). Mitogen-activated Protein Kinase (MAPK3) (p = 0.028) and EGFR (p = 0.045) showed significantly higher activity in sensitive patients compared to resistant patients. INSR/IGF1R was exclusively activated in a substantial number of sensitive patients (Figure 1d). To gain further insight in the biological differences between the groups, a post-translational modifications (PTM) signature enrichment analysis (SEA) was performed. As opposed to gene set enrichment analysis (GSEA), PTM-SEA takes into account the specific combinations of sites of phosphorylation, making it more suitable for analyzing phosphoproteomics data. PTM-SEA showed that three phosphosite-centric signatures were significantly enriched (p < 0.05) in resistant patients: “FGF1 and prolactine pathways” and “EPHA substrates”. Fifteen signatures were enriched in sensitive patients, among which “insulin, VEGF and FGF2 treatment” and “KIT receptor pathway” (Figure 1e). PROTEOME ANALYSIS Expression proteomics was successfully performed on lysate of 25 (17 sensitive and eight resistant) out of 26 samples. In total, 6097 unique proteins were identified (Supplementary Table 3), of which 173 were differentially expressed (p < 0.05 & FC > 2 & ≥ 50% data presence in group with highest abundance) (Figure 2); 109 were upregulated in sensitive and 64 in resistant patients. Of these, FOSL2 was uniquely found in resistant tumors and seven proteins were unique in sensitive tumors (AGMAT, DMGDH, BHMT2, ABCC1, UGT2A3, MEM263 and RBP5). These 173 robust differential proteins are visualized in Figure 2a, split by group. Gene ontology mining revealed that highly abundant proteins in resistant tumors were associated with vesicle mediated transport and excretion from cell processes, while in sensitive tumors, proteins with highest abundance were associated with multiple metabolic processes, such as small molecule -, carboxylic acid -, oxoacid - and glucoronate metabolic processes (Figure 2c).
RkJQdWJsaXNoZXIy MTk4NDMw