297O - Stratifying ovarian cancer based on AXL signaling for therapeutic targeting

Date 16 December 2016
Event ESMO Asia 2016 Congress
Session Gynaecological cancers
Topics Ovarian Cancer
Translational Research
Presenter Ruby Huang
Citation Annals of Oncology (2016) 27 (suppl_9): ix94-ix103. 10.1093/annonc/mdw585
Authors J. Antony1, T.Z. Tan1, J. Low2, M. Choolani3, J.P. Thiery1, R.Y. Huang1
  • 1Cancer Science Institute Of Singapore, Cancer Science Institute of Singapore, 117599 - Singapore/SG
  • 2Gynaecologic Oncology, National University Cancer Institute, Singapore, 119228 - Singapore/SG
  • 3Obstetrics & Gynaecology, National University of Singapore, 119228 - Singapore/SG



Ovarian cancer (OC) is a complex disease demonstrated by the heterogeneity in gene expression molecular subtypes (GEMS). Kinome enrichment analysis identifies a receptor tyrosine kinase (RTK) AXL as a top-ranking kinase in the EMT-driven, poor prognosis Mes GEMS. The Gas6/AXL pathway initiates a recurrent and sustained ERK response exclusively for Mes GEMS, which contributes to an increase in motility and signal amplification via extensive crosstalks with other RTKs. This inherent amplification of oncogenic signaling makes Mes GEMS more sensitive to AXL inhibition. Thus, identifying Mes GEMS OC patients by utilizing the AXL signaling network warrants further investigation.


A gene signature of top 30 genes correlating with AXL was derived from the gene expression microarray analysis of 3500+ OC data points. This gene signature was further correlated with GEMS, EMT score, and patient outcomes. This AXL gene signature was further explored with the TCGA reverse phase protein array (RPPA) data to identify relevant AXL signaling nodes correlated with GEMS. The reproducibility was further validated by an independent cohort.


The AXL signature was found to be significantly overexpressed in Mes GEMS. The AXL signature showed positive correlation with the ovarian-specific EMT score (Rho = +0.4148, p = 5.23e-65). This AXL signature correlated with overall survival (OS) with high AXL signature conferring a worse OS with a hazard ratio (HR) of 1.263 (p = 0.0096) and 1.58 (p = 0.003) when using the cut-off at median or lowest-highest quadrants, respectively. In addition, in OC, the AXL signature was also significantly overexpressed in the omental metastasis (p = 0.0078) and in platinum-resistant relapsed tumours (p = 0.0059) compared to their paired primary tumours. Analysing the TCGA RPPA data shows higher phosphorylation of MAPK in Mes GEMS (p = 0.0282). The increased pERK in Mes GEMS is validated by western blotting in an independent cohort of GEMS-stratified OC samples.


In conclusion, molecular stratification of OC by GEMS enables the identification of AXL and its related signaling network being a crucial player in Mes GEMS. Stratifying OC patients based on GEMS and AXL signaling for AXL inhibition is a rational option.

Clinical trial indentification

Not applicable.

Legal entity responsible for the study

National University of Singapore


National Research Foundation of Singapore


All authors have declared no conflicts of interest.