Gene expression analyses determine two different subpopulations in KIT-negative GIST-like (KNGL) patients

Oncotarget. 2018 Apr 3;9(25):17576-17588. doi: 10.18632/oncotarget.24799.

Abstract

Introduction: There are limited findings available on KIT-negative GIST-like (KNGL) population. Also, KIT expression may be post-transcriptionally regulated by miRNA221 and miRNA222. Hence, the aim of this study is to characterize KNGL population, by differential gene expression, and to analyze miRNA221/222 expression and their prognostic value in KNGL patients.

Methods: KIT, PDGFRA, DOG1, IGF1R, MIR221 and MIR222 expression levels were determined by qRT-PCR. We also analyzed KIT and PDGFRA mutations, DOG1 expression, by immunohistochemistry, along with clinical and pathological data. Disease-free survival (DFS) and overall survival (OS) differences were calculated using Log-rank test.

Results: Hierarchical cluster analyses from gene expression data identified two groups: group I had KIT, DOG1 and PDGFRA overexpression and IGF1R underexpression and group II had overexpression of IGF1R and low expression of KIT, DOG1 and PDGFRA. Group II had a significant worse OS (p = 0.013) in all the series, and showed a tendency for worse OS (p = 0.11), when analyzed only the localized cases. MiRNA222 expression was significantly lower in a control subset of KIT-positive GIST (p < 0.001). OS was significantly worse in KNGL cases with higher expression of MIR221 (p = 0.028) or MIR222 (p = 0.014).

Conclusions: We identified two distinct KNGL subsets, with a different prognostic value. Increased levels of miRNA221/222, which are associated with worse OS, could explain the absence of KIT protein expression of most KNGL tumors.

Keywords: DOG1; IGF1R; KIT; KNGL; miRNA221/222 cluster.