IN BARRETT'S EPITHELIAL CELLS, APE1/REF-1 REDOX FUNCTION MEDIATES EPITHELIAL-MESENCHYMAL TRANSITION INDUCED BY ACIDIC BILE SALT SOLUTIONS: A NOVEL TARGET FOR PREVENTING SUB-SQUAMOUS INTESTINAL METAPLASIA DEVELOPMENT IN BARRETT'S ESOPHAGUS

Background: Barrett’s Esophagus (BE) develops after reflux-mediated injury to distal esophagus and is only known precursor of esophageal adenocarcinoma (EAC), a highly deadly and increasingly prevalent cancer. Inadequate understanding of the genetic determinants and causative mechanisms of BE / EAC limits the development of effective prevention and treatment strategies. We have previously shown that BE and EAC aggregate in a proportion of families, and that a family history is present in nearly 10% of patients with BE or EAC. Aim: To understand familial/inherited predisposition to BE / EAC, and identify the genetic determinants and causative mechanisms driving Barrett’s neoplasia. Methods: We performed exome sequencing of affected relatives from select FBE families followed by validation with sanger sequencing. Results: We discovered inherited deleterious germline mutations in xanthene dehydrogenase (XDH), a gene fundamental to cellular homeostasis, segregates with BE and EAC in 6% of FBE families and genome-wide association studies and our own familial linkage studies have indeed suggested Chromosome 2, which contains XDH, to harbor susceptibility alleles strongly associated with BE/EAC risk and identified 15 heterozygous rare/private germline variants in 18 of 301 FBE families that we accrued. The significance of the observed rare/private XDH variants was calculated using SORVA package. This weighted scheme is utilized to determine the p-value from a binomial cumulative distribution function. Overall, considering a list of 4 candidate genes being tested, the p-value that the variants identified in XDH would be by chance alone was 0.011. Notably, 10 of the 15 variants we identified in XDH were discovered in families or individuals who had BE with high-grade dysplasia or EAC, suggesting that genetic defects in XDH may underlie risk of more aggressive/advanced disease phenotypes. In addition to inherited defects, analysis of TCGA dataset showed XDH to harbor somatic mutations, in up to 4% of EACs. Additionally, our findings from qPCR analysis, showed significant downregulation of XDH in BE dysplasia and EAC cell lines, compared to BE-metaplasia and esophageal SQ cells (***P<0.0005). Further assessment of molecular dynamics by 3D organotypic culture and subsequent RNAScope XDH ISH assay, showed that XDH to be predominantly expressed in the suprabasal/superficial layers, as compared to basal cells, of the esophageal mucosa. Intriguingly, our functional studies on XDH germline variants showed >2-fold higher ROS activity in stably expressed EPC2 cells, upon exposure to acidic bile compared to WT. Conclusion: Taken together, loss/deficiency of XDH likely supports the development of Barrett’s neoplasia in both inherited and sporadic cases. Further molecular and phenotypic studies of this germline variant XDH will determine its association with BE/EAC risk

Introduction: Sub-squamous intestinal metaplasia (SSIM), which can be found in most patients with Barrett’s esophagus (BE), might contribute to recurrent intestinal metaplasia after endoscopic eradication therapy. In earlier reports, we proposed that SSIM might develop in BE when reflux induces the metaplastic cells to undergo epithelial-mesenchymal transition (EMT), a process that enables the cells to migrate into the lamina propria under adjacent squamous epithelium. In human BE cells in vitro, we found that exposure to acidic bile salt solutions (A&B) induced EMT via activation of hypoxia inducible factor (HIF)-1α that increased expression of vascular endothelial growth factor (VEGF) and its signaling. Now, we have looked for evidence of EMT in vivo in BE patients who had acid reflux induced by stopping PPIs, and in vitro in BE spheroids treated with A&B. Methods: Patients with long-segment BE treated with PPIs BID for ≥1 month had baseline endoscopy with BE biopsy, after which PPIs were stopped and endoscopy with biopsy was repeated 1 and 2 weeks later; pH monitoring was performed at baseline and 2 weeks after stopping PPIs. BE biopsies were assessed for HIF-1α and VEGF mRNAs by qPCR. For 3 patients, RNA sequencing of BE biopsies was performed and assessed for enrichment signatures for HIF-1α, angiogenesis, and EMT pathways. Primary BE cells were suspended in 25 µl Matrigel for 5 days for development of spheroids, which then were dissociated, re-suspended in Collagen I, and treated with A&B 15 minutes/day for 2 days. We assessed spheroids for VEGF mRNA by qPCR, for HIF-1α by immunofluorescence (IF), and for EMT features (protrusions of cells into the matrix and CDH1 re-localization by IF). Results: In 15 male BE patients (mean age 63 years), mean esophageal acid exposure increased from 10.5% at baseline to 27.3% at 2 weeks off PPIs (p=0.01). At 1 week off PPIs, BE biopsies exhibited significant increases in HIF-1α (p<0.0001) and VEGF (p=0.047) mRNAs and significant enrichment for the HIF-1α, angiogenesis, and EMT pathways (Figure 1); at 2 weeks, enrichment for all pathways was less than at 1 week. Spheroids treated with A&B also exhibited significant increases in VEGF mRNA (p=0.001), and demonstrated protrusions extending into the matrix. Cells in the protrusions exhibited re-localization of CDH1 from membrane to cytoplasm and increased nuclear staining for HIF-1α (Figure 2). Conclusion: Biopsies of BE in patients with profound acid reflux induced by stopping PPIs exhibit significant increases in HIF-1α and VEGF mRNAs and increased signaling via HIF-1α, angiogenesis, and EMT pathways. Moreover, BE spheroids exposed to A&B exhibit increased nuclear levels of HIF-1α and VEGF, and develop features of EMT. These findings elucidate molecular pathways supporting our hypothesis that GERD-induced EMT might underlie the development of SSIM in BE.

Introduction: Epithelial-mesenchymal transition (EMT) is the process whereby epithelial cells acquire mesenchymal cell features including the ability to migrate. In earlier reports, we proposed that GERD-induced EMT might cause Barrett’s epithelial cells to migrate under esophageal squamous epithelium to form the sub-squamous intestinal metaplasia (SSIM) that can be found in most patients with Barrett’s esophagus (BE). This SSIM is a potential source of the intestinal metaplasia that recurs frequently after endoscopic eradication therapy. In human Barrett’s cells, we found that acidic bile salt solutions (A&B) triggered EMT via activation of hypoxia inducible factor (HIF)-1α, which increases expression of vascular endothelial growth factor (VEGF). APE1/Ref-1 (hereafter called simply APE1) protein has a redox function that maintains HIF-1α in its active, reduced state. In earlier studies in Barrett’s cells, we found that APE1 redox function mediates the increase in HIF transcriptional activity induced by A&B. In this study, we explored whether we could block A&B from inducing EMT in Barrett’s cells by inhibiting APE1 redox function.
Methods: Barrett’s epithelial cell lines (BAR-T and BAR-10T) were treated with a 15-minute exposure to A&B medium (pH 5.5) for 1 or 3 days. We assessed VEGF mRNA levels by qPCR, and cell migration by transwell assays. We used siRNA to knockdown endogenous APE1, which we confirmed by assessing APE1 mRNA; scrambled siRNA served as control. We then used constructs resistant to APE1 siRNA to overexpress: 1) wild-type APE1, 2) a redox-dead APE1 mutant that lacks redox function, or 3) empty vector; these constructs contained a hemagglutinin (HA) tag that we used to confirm protein overexpression by Western blot. We treated the cells either with the novel APE1 redox inhibitor APX2009 or its analog RN7-58, which does not inhibit APE1 redox activity (both at 7µM concentration).
Results: A&B significantly increased migration rates in empty vector-containing control BAR-T cells; this increased migration could be blocked by APE1 knockdown with siRNA (Figure 1). A&B also significantly increased migration rates in BAR-T cells that overexpressed wild-type APE1, but not in cells that overexpressed the APE1 redox-dead mutant (Figure 1). Exposure to A&B significantly increased VEGF mRNA expression and migration rates (Figure 2) in both BAR-T and BAR-10T cells treated with RN7-58, but not in those cells treated with APX2009.
Conclusion: In Barrett’s epithelial cells, A&B increase VEGF mRNA expression and cell migration through the redox function of APE1. These findings suggest that APE1 mediates the GERD-induced increases in VEGF production that trigger EMT in BE. Thus, targeting the redox function of APE1 with agents such as APX2009 might prevent the development of SSIM in Barrett’s patients.