Co-overexpression of AXL and c-ABL predicts a poor prognosis in esophageal adenocarcinoma and promotes cancer cell survival

Background: Esophageal adenocarcinoma (EAC) is highly aggressive and characterized by poor prognosis. AXL expression has been linked to Barrett's tumorigenesis and resistance to chemotherapy, which is associated with c-ABL intracellular localization. However, the molecular and functional relationship between AXL and c-ABL and the clinical significance of the co-expression of these proteins in EAC remain unclear. Methods: We used immunohistochemical analysis (IHC) on tissue microarrays containing human EAC samples (n=53) and normal esophageal tissues (n=11) in combination with corresponding deidentified clinicopathological information to evaluate the expression and the prognostic significance of AXL and c-ABL in EAC. The data were statistically analyzed using Kruskal-Wallis, the chi-square, the Fisher's exact, and Pearson tests. The Kaplan-Meier method and Cox proportional hazards regression model were used to evaluate cancer patient survival. We used a serum deprivation EAC cell model to investigate the pro-survival function of AXL and c-ABL using cell viability, apoptosis, and lactate dehydrogenase activity assays. We performed in vitro assays, including Western blotting, quantitative real-time PCR, and translational chromatin immunoprecipitation (TrIP-Chip) to study the molecular relationship between AXL and c-ABL in EAC cells. Results: IHC analysis revealed that AXL and c-ABL were overexpressed in 55% and 66% of EAC samples, respectively, as compared to normal tissues. Co-overexpression of the two proteins was observed in 49% of EAC samples. The chi-square test indicated a significant association between AXL and c-ABL expression in the EAC samples (χ(2) = 6.873, p = 0.032), and the expression of these proteins was significantly associated with EAC patient age (p < 0.001), tumor stage (p < 0.01), and lymph node status (p < 0.001). AXL and c-ABL protein expression data analysis exhibited an identical clinicopathological association profile. Additionally, we found a significant association between expression of AXL (χ(2) = 16.7, p = 0.002) or c-ABL (χ(2) = 13.4, p = 0.001) and survival of EAC patients. The Cox proportional hazards model and log rank test predicted a significant increase in mortality of patients with high expression of AXL [hazard ratio (HR): 2.86, 95% confidence interval (CI): 1.53 - 5.34, p = 0.003] or c-ABL [HR: 3.29, 95% CI: 1.35 - 8.03, p = 0.001] as compared to those patients with low expression of AXL or c-ABL proteins. Molecular investigations indicated that AXL positively regulates c-ABL protein expression through increased cap-dependent protein translation involving phosphorylation of EIF4E in EAC cells. Next, we investigated the functional relationship between AXL and c-ABL in EAC cells. We demonstrated that the pro-survival activity of AXL requires c-ABL expression in response to serum deprivation. Conclusion: This study highlights the importance of the co-overexpression of AXL and c-ABL proteins as a valuable prognostic biomarker and targeting these proteins could be an effective therapeutic approach in EAC or other solid tumors expressing high levels of AXL and c-ABL proteins.