A cohort study compared intestinal microbiota in children with celiac disease and healthy children. Alterations of intestinal microbiota among patients with celiac disease may due to a gluten-free diet or reflect the pathogenic mechanisms of the disease. In this issue of Gastroenterology, Zafeiropoulou et al examined fecal samples from 141 children with new-onset celiac disease, children with celiac disease treated with gluten-free diet, unaffected siblings, and healthy children. Stool samples were also collected from some children with new-onset celiac disease after 6 and 12 months on a gluten-free diet. Stool samples were analyzed by 16S rRNA sequencing. Microbiota α diversity did not differ among groups, and microbial dysbiosis was not observed in children with new-onset celiac disease. Most of the operational taxonomic units that differed between patients on a gluten-free diet and healthy children were associated with nutrient and food group intake. A microbe signature composed of 11 distinctive operational taxonomic units was found to be specific with high diagnostic probability for celiac disease. This study showed that, although alterations in intestinal microbiota among children with celiac disease are largely due to a gluten-free diet, some specific bacteria may be distinctive biomarkers of celiac disease. Further studies are needed to determine their roles in the pathogenesis of celiac disease. See page 2039. Results from phase III clinical studies examined the histologic, endoscopic, and clinical end points among patients with ulcerative colitis receiving ustekinumab. Ustekinumab can induce and maintain histologic remission in patients with ulcerative colitis. In this issue of Gastroenterology, Li et al used 2630 colonic biopsies in ulcerative colitis patients treated with ustekinumab in phase III clinical trials to evaluate the association between histologic improvement, endoscopic healing, and clinical end points. Histologic improvement, defined as the composite of <5% neutrophil infiltration, no crypt destruction, erosions, ulcerations, or granulation tissue, was associated with clinical remission at the end of induction (weeks 8 and 16) and maintenance periods (week 44). Ustekinumab induced and maintained higher rates of histologic improvement at week 8 and week 44 compared with placebo. Histologic and endoscopic improvement (Mayo endoscopy score of 0 or 1) after induction were associated with higher rates of mucosal healing, clinical remission, and steroid-free remission at the end of the maintenance periods. Patients with both histologic and endoscopic mucosal healing achieved a high rate (61%) of clinical remission compared with patients with endoscopic remission (39%) or histologic improvement (34%) alone. These results showed that achievement of histoendoscopic mucosal healing after induction is associated with lower disease activity at the end of maintenance therapy. See page 2052. Gastric injury-induced regenerative responses are initiated by epithelial IL-33–induced Group 2 innate lymphoid cell recruitment and subsequent type 2 immunity activation promoting tuft cell and foveolar hyperplasia. Damage to the gastric epithelium leads to coordinated and generally protective injury responses that promote barrier restitution and resolution of regenerative processes. Reprogrammed chief cells drive spasmolytic polypeptide-expressing metaplasia, as well as foveolar and tuft cell hyperplasia, which underlie this process and recruit cells for cooperative tissue repair. Damaged epithelium releases IL-33, which induces an array of type II cytokines. Failure to terminate these normal regenerative processes can promote dysplastic changes and increase the risk of gastric cancer. Group 2 innate lymphoid cells (ILC2) reside in the gastric mucosa, evoke antigen-independent immune responses, and are activated by IL-33. In this issue of Gastroenterology, Meyer et al define the relationships among gastric injury, IL-33, ILC2 biology, and metaplastic regenerative responses. Using L635 to induce gastric damage, they identified IL-33–dependent ILC2 infiltration and ILC2-dependent type 2 cytokine production. ILC2 depletion with anti-CD90.2 blocked metaplastic reprogramming as well as tuft cell and foveolar hyperplasia after injury, linking these regenerative phenotypes to ILC2 recruitment and activation. In terms of clinical relevance, spasmolytic polypeptide-expressing metaplasia samples demonstrate greater numbers of ILC2s. These elegant experiments identify and sequence cellular signaling events that promote gastric regenerative responses after injury (Figure 2), and present potential therapeutic targets in gastric neoplastic progression. See page 2077. `Cooperation between pairs of oncogenes and tumor suppressors in hepatocellular carcinoma mouse models recapitulate histologic, immune microenvironmental, and transcriptional phenotypic variation in human disease. Hepatocellular carcinoma (HCC), often detected at an advanced stage, has few therapeutic options with limited efficacy. An improved understanding of the molecular biology underlying disease onset, progression, and therapeutic response is needed to improve outcomes. This goal, in turn, requires the development of more accurate experimental models. Limitations of current HCC mouse models include difficulties in faithfully recapitulating disease progression and intratumoral heterogeneity. In this issue of Gastroenterology, Molina-Sanchez et al develop nine distinct HCC models that represent intratumoral genetic and phenotypic heterogeneity. They used hydrodynamic tail vein injection to deliver either MYC or β-catenin in combination with 11 other genes previously identified as HCC drivers. Either loss-of-function or gain-of-function effects were modeled with CRISPR-based approaches to assess cooperation between tumor suppressors and oncogenes (Figure 3). Nine of the 23 tumor pairs generated had >20% tumor penetrance within 6 months (MYC;sg-p53, MYC;sg-Kmt2b, MYC;sg-Kmt2c, MYC;sg-Pten, MYC;CTNNB1, MYC;sg-Axin1, CTNNB1;sg-Pten, CTNNB1;Tert, and MYC;Tert); of note, these driver combinations are observed in >40% of patients with HCC. Single gene delivery of MYC or CTNNB1 failed to yield tumors. MYC-driven models were more efficient than β-catenin; however, their combination produced the highest penetrance and shortest survival. Cooperation between driver pairs resulted in unique histologic features, immune microenvironment, and represented HCC transcriptional subclasses. Interestingly, phenotypic variation seemed to arise owing to differential driver expression, indicating that tumors with identical genetic profiles may have different phenotypes owing to changes in driver levels. In cell lines generated from these tumors, interdriver cooperation modified chemotherapeutic response. These new mouse models will serve as a terrific resource for further studying mechanisms of progression, genetic and environmental interactions, and therapeutic approaches in HCC. See page 2203.