Covering the Cover

Abstract

In 2 prospective cohorts, histologic grade of intestinal metaplasia before radiofrequency ablation for Barrett’s esophagus predicts recurrent neoplasia and can be used to determine surveillance intervals. Barrett’s esophagus, the only established precursor to esophageal adenocarcinoma, can be treated effectively and safely with radiofrequency ablation. Although most patients achieve durable response after complete eradication of intestinal metaplasia, ≥25% of patients have recurrence of intestinal metaplasia and ≥1% of patients have recurrence with invasive adenocarcinoma. Although surveillance endoscopy is usually performed to identify and treat recurrence, this practice varies widely and is based largely on expert opinion. In this issue of Gastroenterology, Cotton et al used 2 large registries of patients undergoing radiofrequency ablation for Barrett’s esophagus in the United States and the UK to build and validate models to predict risk of recurrence after complete ablation. The incidence of neoplastic recurrence was associated with the most severe histologic grade before ablation, age, sex, baseline Barrett’s esophagus length, and performance of endoscopic mucosal resection. A model based solely on the most severe histologic grade predicted neoplastic recurrence with c-statistics of 0.892 (95% confidence interval, 0.863-0.921) in the US cohort and 0.728 (95% confidence interval, 0.584-0.871) in the UK validation cohort. For patients with low-grade dysplasia before ablation, the authors propose surveillance endoscopy at 1 and 3 years after complete eradication. For patients with high-grade dysplasia or intramucosal adenocarcinoma, the authors propose surveillance intervals at 3 months, 6 months, 1 year, then annually thereafter. Although these findings warrant further corroboration, this study is among the first to provide evidence regarding optimizing surveillance endoscopy for Barrett’s esophagus. See page 316. In a French nationwide cohort, the combination of tumor necrosis factor antagonists and thiopurine therapies was associated with a greater risk of serious and opportunistic infection in patients with inflammatory bowel disease. Combinations of immunosuppressive therapies, including tumor necrosis factor antagonists and thiopurines, have demonstrated superior efficacy over monotherapy with either agent in the management of inflammatory bowel disease. Previous studies have shown an increased risk of serious and opportunistic infection in patients treated with monotherapy (Figure 1). However, whether combination immunosuppressive therapy further increases the risk of infection is less known. In this issue of Gastroenterology, Kirchgesner et al identified 190,694 patients with inflammatory bowel disease from 2009 to 2014 in a French administrative health database. Roughly 33% of patients had been exposed to immunosuppressive therapies. Combination therapy was associated with increased risk of serious infection (defined as a diagnosis of infection requiring hospitalization) compared with tumor necrosis factor antagonist (hazard ratio [HR], 1.23; 95% confidence interval [CI], 1.05-1.45) or thiopurine monotherapy (HR, 2.11; 95% CI, 1.80-2.48). Tumor necrosis factor antagonist monotherapy was associated with a higher risk of serious infection (HR, 1.71; 95% CI, 1.56-1.88) compared with thiopurine monotherapy. Among patients with serious infections, 3.9% died within 3 months after infection occurrence. Similarly, combination therapy was also associated with increased risk of opportunistic infection compared with tumor necrosis factor antagonist (HR, 1.96; 95% CI, 1.32-2.91) or thiopurine monotherapy (HR, 2.11; 95% CI, 1.45-3.08). This study provides real-world evidence regarding the risk of severe and opportunistic infections in patients treated with thiopurines, tumor necrosis factor antagonists, and their combination. These data will be important to incorporate into the clinical decision making for optimal management of inflammatory bowel disease. See page 337; editorial on page 262. Changes in the gut virome discriminate between patients with colorectal cancer and controls and may be used for stage specification and prognostication. Increasingly, researchers have recognized the importance of host–microbial interactions in the pathogenesis of colorectal cancer (CRC) with most efforts focusing on bacterial microbiome identification and characterization. Studies have established links between the CRC development and other disease processes such as obesity, inflammatory bowel disease, and components of the metabolic syndrome. Shifts in gut bacterial populations can also affect the distributions of viral and fungal constituents. However, changes in viral species in CRC remain underexplored; their identification could provide an additional diagnostic tool. This issue of Gastroenterology features Nakatsu et al’s metagenome-wide association study of the gut microbiome in CRC. The authors used control and CRC cases derived from patient samples from the research stool bank at the Prince of Wales Hospital at the Chinese University of Hong Kong, a discovery cohort, and an independent Chinese cohort for validation purposes. They observed consistent increases in bacteriophage diversity that were associated with reduced bacterial diversity in patients with CRC, defining 22 viral genera whose presence could discriminated between CRC and controls. Orthobunyavirus, Inovirus, and Tunalikevirus, the top 3 genera, demonstrated disease-specific enrichment in CRC (Figure 2A, B). Furthermore viral signatures could distinguish early from late-stage CRC and had independent prognostic value (Figure 3). The data revealed associations between changes in the gut virome and CRC, which suggests a potential pathologic role for multikingdom changes in enteric constituents in CRC. Validation of these observations in additional populations will help to determine the broader implications of this work.Figure 3Clinical stage–associated dysbiosis of the gut virome in CRC. Biplot summarizing Canberra distance–based partial redundancy analysis of virome species-level profiles, controlling for compositional effects of age, sex, obesity, diabetes mellitus, and use of oral prescription drugs. Viral species that correlate with both constrained axes at false discovery rates of 5% (Spearman ρ) or less are shown. Projection axes were assessed individually by Wilcoxon rank-sum tests. P values were adjusted for multiple-hypothesis testing by Benjamini-Hochberg step-up procedures. ∗∗∗∗Q < .0001.View Large Image Figure ViewerDownload Hi-res image Download (PPT) See page 529. The transcriptional repressor MAFG contributes to the pathobiology of cholestatic liver disease and hepatocellular carcinoma. Whereas treatment of hepatic cells with SAMe and UDCA reduces MAFG expression, treatment with the FXR agonist OCA increases proliferation and growth, raising concern that OCA could promote tumor growth. Chronic cholestasis is one of the major preconditions for cholangiocarcinoma, yet molecular factors contributing to chronic cholestatic injury remain unclear. A bZIP transcription factor activated by the farnesoid X receptor (FXR), V-Maf Avian Musculoaponeurotic Fibrosarcoma Oncogene Homolog G (MAFG) represses bile acid synthesis. It has the dual purpose of heterodimerizing with nuclear factor-erythroid 2-related factor 2 (NRF2) and activating transcription of genes with an antioxidant response element (ARE). Hence, MAFG protects from reactive oxygen species. Notwithstanding, MAFG homodimers repress ARE target genes when induced in cholestatic liver disease. The stoichiometry of MAFG:NRF2 heterodimers thus shifts to MAFG homodimers, which results in reduced glutathione expression. When mice were treated with S-adenosylmethionine (SAMe) or ursodeoxycholic acid (UDCA), MAFG induction was completely blocked, but the mechanism of therapeutic effect could not be clearly elucidated. MAFG was also up-regulated in cholangiocarcinoma (CCA), and levels correlated with CCA growth in mouse models. As an FXR target, MAFG serves as a transcriptional repressor of bile acid synthesis and metabolism. Liu et al’s article in this issue of Gastroenterology defines the molecular mechanisms that cause MAFG expression in cholestatic liver disease and elucidates the effect of various therapeutic agents on this process. The authors first used a combination of promoter-mapping constructs to determine that activator protein-1 (AP-1), nuclear factor-κB, and E-box elements are required for lithocholic acid (LCA)-induced MAFG transcriptional up-regulation. This occurs via upregulation of methionine adenosyltransferase 2A (MAT2A). Both MAT2A and MAFG were induced in niethylnitrosamine liver tumors in mice. Both SAMe and UDCA inhibited LCA up-regulation of MAFG through common and independent mechanisms. Obeticholic acid (OCA), a potent FXR agonist, induced MAT2A, MAFG, and c-MYC; increased proliferation of liver cancer cell lines; and augmented xenograft tumor growth (Figure 4). High MAFG levels were associated with poor prognostic indices in both human hepatocellular carcinoma and CCA. This study elucidated the biochemical signaling pathways regulating MAFG expression and established its role in cholestatic injury and hepatocellular carcinoma. Results point to potential protumorigenic effects from OCA treatment. Further studies will help to clarify the implications of these observations. See page 557.