PLoS ONE (Jan 2011)

Whole genome expression array profiling highlights differences in mucosal defense genes in Barrett's esophagus and esophageal adenocarcinoma.

  • Derek J Nancarrow,
  • Andrew D Clouston,
  • B Mark Smithers,
  • David C Gotley,
  • Paul A Drew,
  • David I Watson,
  • Sonika Tyagi,
  • Nicholas K Hayward,
  • David C Whiteman,
  • Australian Cancer Study,
  • Study of Digestive Health

DOI
https://doi.org/10.1371/journal.pone.0022513
Journal volume & issue
Vol. 6, no. 7
p. e22513

Abstract

Read online

Esophageal adenocarcinoma (EAC) has become a major concern in Western countries due to rapid rises in incidence coupled with very poor survival rates. One of the key risk factors for the development of this cancer is the presence of Barrett's esophagus (BE), which is believed to form in response to repeated gastro-esophageal reflux. In this study we performed comparative, genome-wide expression profiling (using Illumina whole-genome Beadarrays) on total RNA extracted from esophageal biopsy tissues from individuals with EAC, BE (in the absence of EAC) and those with normal squamous epithelium. We combined these data with publically accessible raw data from three similar studies to investigate key gene and ontology differences between these three tissue states. The results support the deduction that BE is a tissue with enhanced glycoprotein synthesis machinery (DPP4, ATP2A3, AGR2) designed to provide strong mucosal defenses aimed at resisting gastro-esophageal reflux. EAC exhibits the enhanced extracellular matrix remodeling (collagens, IGFBP7, PLAU) effects expected in an aggressive form of cancer, as well as evidence of reduced expression of genes associated with mucosal (MUC6, CA2, TFF1) and xenobiotic (AKR1C2, AKR1B10) defenses. When our results are compared to previous whole-genome expression profiling studies keratin, mucin, annexin and trefoil factor gene groups are the most frequently represented differentially expressed gene families. Eleven genes identified here are also represented in at least 3 other profiling studies. We used these genes to discriminate between squamous epithelium, BE and EAC within the two largest cohorts using a support vector machine leave one out cross validation (LOOCV) analysis. While this method was satisfactory for discriminating squamous epithelium and BE, it demonstrates the need for more detailed investigations into profiling changes between BE and EAC.