A novel model including CXCR4 and CDK6 as potential predictive marker for docetaxel (doc) sensitivity in gastric cancer (GC)

Abstract

e15548 Background: A growing body of evidence has shed light on biomarkers for chemotherapeutic agent sensitivity in gastric cancer (GC) treatment. It has been demonstrated β-tubulin Ш (TUBB3) and microtubule associated protein tau (MAPT) might be considered as potential markers for taxanes treatment. However, an integrated predictive model for docetaxel (doc) sensitivity is still needed to customized chemotherapy in gastric cancer. Methods: Doc resistant gastric cancer cell line was established by exposure to gradually increasing doc concentration on an intermittent dosage schedule. Microarray was performed to explore different gene expression levels between parental cells and resistant cells. 11 gastrointestinal cancer cell lines were used to validate gene levels by real-time quantitative PCR as well as doc cytotoxicity. Histoculture drug response assay (HDRA) was adopted to examine doc sensitivity in 25 surgically dissected GC specimens. Discrimination analyses were used to determine which linear combinations of these genes best distinguished responder patients from nonresponder patients by adding one more genes each time. Results: After functional analysis of microarray, a panel of 11 genes were further validated, which resulting 4 genes, CXCR4 (P=0.047), CDK6 (P=0.014), USP15 (P<0.001) and CDH1 (P=0.001) that correlated to the doc sensitivity. The mRNA levels of CXCR4 and CDK6 were significantly higher in resistant GC specimens (P=0.0298, 0.022 respectively). A final response index including 5 genes (CXCR4, CDK6, USP15, TUBB3 and MAPT) showed potential predictive value for doc associated chemotherapy with an accuracy of 84% (P=0.001). Conclusions: A predictive model with five genes for doc therapy was generated and showed novel predictive role in doc sensitivity in GC. No significant financial relationships to disclose.