NGS and medically-driven integrative bioinformatics applications in gastrointestinal stromal tumors to overcome the TKI resistance

The complex molecular biology involved in gastrointestinal stromal tumors (GIST) progressive disease led to a growing urgency and interest in developing new strategies to overcome TKI resistance. In this study, we have used Next generation sequencing data and several bioinformatics approaches to investigate the GITS infiltrating immune-cell subpopulations and to perform an integrated molecular characterization of D842V mutant GIST, with the aim evaluating the potential of an anti-PD-L1 treatment in combination with tyrosine kinase inhibitors (TKI) and to identify new possible target. Gene expression profile and immunohistochemistry (IHC) of a cohort of 31 GIST KIT and PDGFRA mutated supported the presence of immune infiltrate in GIST, with dominance of CD4+ and CD8+ T cells and M2 macrophages showing a remarkable similarity with melanoma microenvironment. The expanded IFN-γ-induced immune signature (EIIS) genes were expressed in most of GIST samples and positively correlated with PD-L1 abundance (p < .0001). Moreover, the median TIS score for GIST was between the 65th and 70th percentile of the Cancer Genome Atlas dataset, in the same range of tumors responding to anti-PD-1/PD-L1. Analysis of the Gene Expression Omnibus database GIST samples pre- and post-treatment confirmed that imatinib downregulates PD-L1 and IRF1 expression through the inhibition of KIT and PDGFRA, thus contributing to counteract the suppressed adaptive immune response against GIST. The presence of a rich immune infiltrate in GIST along with the presence of TIS and EIIS suggests that GIST may benefit from immunotherapy along with tyrosine kinase inhibitors. Analysis of whole exome sequencing data of 19 D842V mutant GIST samples did not show any actionable recurrent molecular events, beyond D842V, of therapeutic significance. Molecular modeling however, suggests that the D842V mutant protein binds imatinib with lower affinity with respect to wild-type structure, showing higher stability during the interaction with other type I TKIs (like crenolanib).