Background and aim: Widespread applicability of tissue-based mRNA expression screening for Barrett esophagus (BE) is likely to require (1) accurate methods for assaying archival formalin-fixed, paraffin-embedded (FFPE) histopathology specimens taken at endoscopy, and (2) validation studies of promising biomarkers in different patient cohorts.
Results: 30 genes were significantly differentially expressed by histopathology tissue type. The direction and magnitude of the differences were very similar to those found in previous studies using fresh frozen tissues. Novel upregulated genes were TSPAN8 (also known as CO-029), TSPAN24 (CD151), EGR1 and TCIRG1. Novel downregulated genes were DPYD, TSPAN29 (CD9) and Ets1. Strong associations between histopathology type and gene expression were observed with the overexpressed genes: cyclo-oxygenase-2 (COX-2), for which histopathology type explained 77.7% of the variation in expression, TSPAN1 (73.5%), TSPAN8 (62.9%), SPARC (62.1%), MMP7 (50.8%); and the under-expressed genes ADH7 (53.7%), APC (58.2%), RAR-gamma (51.3%).
Methods: mRNA was isolated from 54 FFPE small endoscopic biopsies from patients with Barrett intestinal metaplasia (BE), esophageal adenocarcinoma (EAC), or control patients with a normal squamous-lined esophagus. Multiplexed tandem PCR (MT-PCR) was used to quantitate 50 selected candidate genes for BE and control genes in duplicate. Principal component analysis (PCA) was conducted to explore the presence of global differences in gene expression profiles. One-way analysis of variance (ANOVA) of the transformed data was used to identify genes that were differentially expressed between different histological subtypes. Differentially expressed genes with a fold change of ≥2 (upregulated) or ≤-2 (downregulated) are reported with the p value for each comparison (EAC vs. normal, EAC vs. BE and BE vs. normal). The Benjamini-Hochberg method was used to control the false discovery rate at 0.01 for all comparisons.
Conclusions: Alterations in expression of select genes are strongly associated with BE or EAC or both. This study's findings for many highly differentially expressed genes are very similar to those previously reported, suggesting that these genes should be tested further in longitudinal studies for their potential role as biomarkers of progression to more advanced Barrett disease.