Society: AASLD

Background: The glucagon-like-receptor (GLP)-1 agonist semaglutide has demonstrated therapeutic efficacy on clinical endpoints in a recent phase 2 clinical trial in patients with non-alcoholic steatohepatitis (NASH)(Newsome et al. NEJM, 2021). The present study aimed to evaluate therapeutic efficacy of semaglutide on clinical endpoints and outcome in the GAN (Gubra-Amylin NASH) diet-induced obese (DIO) mouse model of NASH with advanced fibrosis and hepatocellular carcinoma (HCC).

Methods: Male C57BL/6J mice were fed the GAN diet high in fat, fructose, and cholesterol for extended 48 weeks prior to study start. Only animals with liver biopsy-confirmed NAFLD Activity Score (NAS ≥5) and advanced fibrosis (stage F3) were included and stratified into treatment groups. DIO-NASH-HCC mice received (SC, QD) vehicle (n=16) or semaglutide (30 nmol/kg, n=15) for 14 weeks. Vehicle-dosed chow-fed C57BL/6J mice (n=9) served as lean healthy controls. Untreated DIO-NASH-HCC mice (n=10) were terminated at baseline. Tumor histopathological classification was performed by an expert clinical pathologist. Pre-to-post liver biopsy histopathology was performed for within-subject evaluation of NAFLD Activity Score (NAS) and fibrosis stage. Additional endpoints included blood biochemistry and quantitative liver histology.

Results: Compared to baseline, DIO-NASH-HCC mice demonstrated progressive HCC burden over the 14-week study period. Tumors showed consistent architectural and cytologic features of HCC with a marked loss of reticulin-stained fibers. Notably, semaglutide completely prevented progression in HCC burden. Concurrently, semaglutide improved hallmarks of NASH, including transaminases, hepatomegaly and histopathological NAS (≥2 point) without improving fibrosis stage. In agreement, semaglutide reduced quantitative histological markers of steatosis (lipids, hepatocytes with lipid droplets), inflammation (number of inflammatory foci, galectin-3), fibrogenesis (α-SMA), proliferation (Ki67) and progenitor cell activation (CK19).

Conclusion: This is the first study to demonstrate that semaglutide improves both clinical histopathological endpoints for NAFLD Activity Score and HCC burden in a preclinical translational mouse model of NASH-driven HCC. This highlights the suitability of GAN DIO-NASH-HCC mice for profiling novel drug therapies targeting NASH with advanced fibrosis and HCC.

The formation of Cholesterol gallstone (CGS) disease is a complex process which is caused by supersaturation of biliary cholesterol due to either hyper-secretion of biliary cholesterol or decreased bile acids. Studies have found that dysbiosis of the gut microbiota is associated with cholesterol gallstones. The role and mechanism of probiotics in reducing cholesterol gallstones remains unknown. The proportion of lactobacilli in CGS was greatly reduced. However, whether supplementation with lactobacillus can reduce gallstones has not been well studied. In this study, L.acidophilus（LA）was administered to mice fed on a lithogenic diet (LD) for 8 weeks. LA significantly reduced LD-induced gallstones (P<0.05) and hepatic steatosis. At the same time, CSI of gallbladder bile decreased, the mucosal epitheliums of the gallbladders were thinner as the expression level of MUC2 and MUC5ac in gallbladder epithelial cells decreased. In the ileum of mice, the expression of NPC1L1 gene reduced therefore the intestinal absorption of cholesterol endocytosis from the brush margin of the small intestine was decreased. elevated levels of ABCG5 and ABCG8, which specifically expressed on the surface of intestinal cells, were also observed in the ileum by LA supplement and increases blood cholesterol excretion into the intestinal lumen. Further research also showed that LA small molecule metabolites regulated gene levels of cholesterol transporter NPC1L1 and ABCG5/G8. Whereas we did not find that LA was able to downregulate the intestinal FXR-FGF15 signaling pathway. Unlike the gut, the gene expression level of the ABCG5/G8 which is specifically expressed on hepatocytes, decreased therefore the excretion of cholesterol from the liver to the biliary tract reduced. LA could activate FXR-SHP signaling in livers and regulate the accumulation of liver lipids. In addition, LA increased the richness and α diversity of gut microbes which were dysregulated by high-cholesterol diet, improved the disorder of the ratio of Firmicutes to Bacteroidetes (F/B). LA in particular increased the abundance of Bifidobacterium, and Clostridiales containing BSH or 7α-dehydroxylase and altered the dysbiosis and composition of the mouse gut bile acid. There studies showed that LA could block CGS formation by regulating cholesterol transporter signaling pathways and remodeling intestinal microbiota. It may be a promising treatment strategy for treating CGS or preventing its recurrence.

Introduction: Lipodystrophy syndromes are characterized by congenital or acquired inheritance and generalized or partial distribution of adipose tissue loss. Inadequate fat storage leads to leptin deficiency and metabolic complications such as non-alcoholic steatohepatitis (NASH). NASH leads to early-onset fibrosis and cirrhosis in many patients with lipodystrophy, with some developing advanced fibrosis in childhood. Liver biopsy can diagnose and stage NASH; however, conventional non-invasive fibrosis markers may be inaccurate in patients with lipodystrophy due to incorporation of age and/or adiposity as predictors of fibrosis.

Aim: We aimed to 1) assess performance of existing predictive models in detecting advanced fibrosis and cirrhosis in patients with lipodystrophy, 2) create a novel predictive model in comparison with existing non-invasive serologic markers in literature, 3) validate the model in a cohort of patients.

Methods: An analysis of 82 patients with lipodystrophy with liver histology at the National Institutes of Health (NIH) during 1998-2019 was performed. Existing predictive scores (abbreviations in Table) including AAR, APRI, FIB-4, NFS, BAAT, and BARD score were calculated. Liver histology was scored by NASH Clinical Research Network method. Performance of existing predictive models was assessed by area under the receiver operating curve (AUROC) analysis. Predictive models were evaluated by logistic regression to create a novel scoring system and validated in a separate University of Michigan (UM) cohort of patients with lipodystrophy.

Results: 105 lipodystrophy patients (NIH=82 cohort, UM=23) were studied. NIH cohort was randomized into training and validation cohorts without statistical differences between groups. NIH cohort patients had median age 19 years (IQR 15-35), were mainly female (79%), with diabetes (87%), and with body mass index 21.9 (18.6-25.4) kg/m². Median ALT, AST, and leptin were 68 (30-112) IU/L, 40 (24-79) IU/L, and 1.68 (0.97-4.61) ng/mL respectively. 32% of patients had advanced fibrosis or cirrhosis. The UM cohort differed by age (median 42 years, IQR: 34-51), but not by sex (79% female) and diabetes (91%). 4% had advanced fibrosis. Amongst the existing predictive models, APRI performed best (AUROC: 0.74) and BAAT performed worst (AUROC: 0.51). Three novel predictive models (A, B, and C) used permutations of BMI, total bilirubin, triglycerides, platelets, and leptin. Model B exceeded performance of all other models, existing and novel, with an AUROC of 0.84.

Conclusion: Current predictive models in literature were not designed for hepatic fibrosis in patients with lipodystrophy. We created and validated a novel predictive model that outperforms existing models in detecting advanced liver disease in patients with lipodystrophy with better sensitivity, specificity, positive and negative predictive values.

Diagnostic accuracy of the established and novel predictive models

Diagnostic accuracy of the established and novel predictive models

As one of most common chronic liver diseases, non-alcoholic fatty liver disease (NAFLD) is characterized by excessive fat deposition in hepatocytes (steatosis) and progresses to non-alcoholic steatohepatitis (NASH), the advanced form of NAFLD, when the liver displays overt inflammatory damage. Previous research demonstrated that stimulator of interferon genes (STING), an essential regulator that mediates DNA sensing and triggers innate anti-viral immunity, is positively associated with the degree of liver inflammation in human patients with NAFLD/NASH. In addition, disruption of STING in hematopoietic cells attenuates hepatic steatosis and inflammation whereas myeloid cell-specific deficiency of 6-phosphofructo-2-kinase/fructose-2,6-biphoshphatase 3 (PFKFB3, a master regulatory gene of glycolysis) exacerbates hepatic steatosis and inflammation in mice fed a high-fat diet (HFD, 60% fat calories). Since PFKFB3 disruption is associated with mitochondrial dysfunction that is capable of increasing STING activation, the present study postulates STING as a downstream mediator of PFKFB3 disruption as it relates to macrophage proinflammatory activation, thereby NAFLD pathogenesis. As such, the present study aimed to determine the extent to which STING disruption in myeloid cells alleviates the deleterious effect of myeloid-specific PFKFB3 deficiency (Mye-PFKFB3-KO) on hepatic steatosis and inflammation in mice fed an HFD. The mice in which PFKFB3 and STING were deleted only in myeloid cells (Mye-DKO) and Mye-PFKFB3-KO mice, as well as their control mice, were fed an HFD for 12 weeks and examined for liver steatosis and inflammation. Compared with control mice (whose PFKFB3 and STING were intact), Mye-PFKFB3-KO mice displayed significantly increased severities in diet-induced NAFLD phenotypes, evidenced by increased hepatic fat deposition and proinflammatory signaling through NFkB p65 and JNK p46. However, the severities of diet-induced hepatic steatosis and liver inflammation in Mye-DKO mice were significantly lower than their respective levels in Mye-PFKFB3-KO mice. Consistently, the severities of diet-induced glucose intolerance and systemic insulin resistance in Mye-DKO mice were significantly lower than those in Mye-PFKFB3-KO mice. Taken together, these results indicate that myeloid STING disruption alleviates the deleterious effect of myeloid cell-specific PFKFB3 deficiency on diet-induced NAFLD aspects. As such, STING may be a viable target for managing obesity-associated fatty liver disease.

Background: Glucagon-like peptide 1 receptor agonists (GLP-1RAs) are approved for treatment of diabetes and overweight, and are under investigation for treatment of non-alcoholic fatty liver disease (NAFLD). While many genetic variants influence risk of NAFLD and cirrhosis, whether these variants modify the effects of GLP-1RAs on liver enzyme levels is not known. We aimed to characterize trends in alanine aminotransferase (ALT) before vs. after GLP-1RA initiation depending on genetic risk factors for NAFLD.

Methods: This was a single center study utilizing the Michigan Genomics Initiative, a cohort of patients seen at a large integrated healthcare system who underwent genotyping for research purposes. We included all patients initiated on a GLP-1RA from 2010-2021 with outpatient ALT measurements <3 months before (“baseline”) and 12-24 months after (“follow-up”) GLP-1RA initiation. We did not require that patients have baseline NAFLD or elevated ALT. The primary outcomes were change in ALT in follow-up vs. baseline, and proportion with normal ALT (<= 19/30 U/L for women/men, respectively) at baseline vs. follow-up. The primary predictors were genetic variants associated with NAFLD, namely PNPLA3-rs738409, TM6SF2-rs58542926, and HSD17B13-rs6834314 genotypes.

Results: We included 714 patients meeting inclusion criteria. Mean age was 56 years, 59% were female, mean body mass index was 38.4, and 76% had diabetes. Mean ALT at baseline and follow-up were 35.1 and 31.3 U/L, respectively (P<.0001). Prevalence of elevated liver enzymes decreased from 68.8% to 63.6% from baseline to follow-up (P<.0001). At baseline, abnormal liver enzymes were more common in PNPLA3-rs738409-GG carriers than CG or CC carriers: 85%, 68%, and 67%, respectively (P=.0095 for GG vs. CC/CG genotype; Figure 1). However, at follow-up, there was no difference in prevalence of abnormal liver enzymes: 75%, 61%, and 64% in PNPLA3-rs738409-GG, CG, and CC carriers, respectively (P=.19 for GG vs. CC/CG genotype). Similarly, mean decrease in ALT from baseline to follow-up was 2.4, 5.1, and 11.2 U/L in PNPLA3-rs738409-CC, CG, and GG carriers, respectively (P<.05 for difference from 0 for all), though heterogeneity only trended toward statistical significance (P_{heterogeneity}=.055). PNPLA3-rs738409-GG genotype remained associated with change in ALT after adjustment for follow-up BMI or change in BMI from baseline to follow-up. TM6SF2-rs58542926 or HSD17B13-rs6834314 genotypes were not associated with change in ALT between baseline and follow-up (P_{heterogeneity}>.05 for both).

Conclusions: In patients started on a GLP-1RA, PNPLA3 risk alleles were associated with greater decrease in ALT and higher rate of ALT normalization. These findings are proof of concept that PNPLA3 genotype may modify response to treatment with GLP-1RAs and should be considered in clinical trials.

Vincent Chen

University of Michigan Medical School

Appearances