Author of

• 10497

  +

  P1.01 - Poster Session 1 - Cancer Biology (ID 143)

  • Event: WCLC 2013
  • Type: Poster Session
  • Track: Biology
  • Presentations: 1
  • Moderators:
  • Coordinates: 10/28/2013, 09:30 - 16:30, Exhibit Hall, Ground Level

  • 10500

    +

    P1.01-003 - Targeting EMT in lung cancer: An integrated analysis of Axl and other mesenchymal targets in The Cancer Genome Atlas (TCGA) (ID 1991)

    09:58 - 10:12  |  Author(s): R. Cardnell
    • Abstract

    Background
    We previously developed a 76-gene signature of epithelial-to-mesenchymal transition (EMT) that predicted resistance to EGFR and PI3K inhibition in non-small cell lung cancer (NSCLC). This analysis also identified Axl, a receptor tyrosine kinase, as a novel target for mesenchymal lung cancers. Here, we conducted an integrated molecular analysis of EMT in resected, treatment-naïve tumors from three clinical cohorts, including the Cancer Genome Atlas (TCGA) lung adenocarcinomas (LUAD) and squamous cell carcinomas (LUSC), with particular focus on Axl as a potential target in mesenchymal NSCLC.

    Methods
    Using our 76-gene EMT signature, TCGA patient tumors (230 LUAD, 178 LUSC) and a large MDACC cohort of resected tumors (n=279) were assigned an “EMT score.” Expression of >160 total and phosphoproteins were measured in the tumors by reverse phase protein array (RPPA). Proteomic profiles and other molecular markers (including mutation status, miRNA expression, and copy number) were correlated with EMT scores and Axl expression levels.

    Results
    The EMT score, derived from our EMT signature, identified NSCLC tumors with mesenchymal gene expression signatures (average 23% of tumors across all cohorts, range 14-34%). In both LUAD and LUSC, EMT scores were highly correlated with (1) expression levels of the miR200 family, a group of miRNAs previously known to regulate EMT (p-values <0.001 by Pearson correlation) and (2) levels of proteins central to EMT (e.g., E-cadherin, alpha-catenin, beta-catenin, claudin-7, fibronectin; p<0.001 for all). Mesenchymal tumors also had lower expression of TTF1 in LUAD (p=0.0002) and lower p63 in LUSC (p=0.003). Although pEGFR levels were higher in epithelial LUAD tumors (p=0.01), the frequency of EGFR mutations was not significantly higher in this group. EMT score was not associated with smoking status. Consistent with our previous findings in cell lines and patients with advanced NSCLC (BATTLE trial), protein expression of the receptor tyrosine kinase Axl was significantly higher in tumors with mesenchymal signatures (high EMT scores) and with low E-cadherin protein expression (p<0.005 for both). The inverse correlation between tumor E-cadherin and Axl expression was confirmed in an independent group of NSCLC cases by immunohistochemistry. Although a small number of Axl mutations were observed (<3% of tumors), few occurred in the kinase domain and their biological significance is unknown. Other potential therapeutic targets expressed at higher levels in mesenchymal lung cancers included PKC-alpha, NFKB, and FGFR1.

    Conclusion
    The EMT gene expression signature performed well in the TCGA LUAD, TCGA LUSC, and MDACC cohorts, correlating strongly with established markers of EMT on other data platforms (miRNA and protein). We observed strong protein expression of the receptor tyrosine kinase Axl (as well as other targets) among mesenchymal tumors, supporting further investigation of AXL as a potential EMT target and into the mechanism of its overexpression in NSCLC.