Society: AASLD

LIVE STREAM SESSION

BACKGROUND: Non-alcoholic steatohepatitis (NASH), a leading cause of liver-related morbidity and mortality, is highly prevalent in people living with HIV (PLWH).¹ Microbial translocation (MT) associated with hepatic inflammation and fibrosis may play a critical role in NASH pathogenesis.² Although HIV is associated with increased gut permeability and MT,³ few studies have evaluated the role of the gut-liver axis in NASH development in PLWH. We aimed to determine the associations of HIV and circulating biomarkers of MT on the FibroScan-AST (FAST)⁵ score, a non-invasive surrogate for NASH with advanced fibrosis, in a large US cohort of women with HIV (WWH) and women without HIV.
METHODS: Vibration controlled transient elastography (VCTE) was conducted from 2013-2018 among 1237 women without viral hepatitis (883 WWH, 354 without HIV). Serum and plasma MT biomarkers kynurenine to tryptophan (KT) ratio (a marker of MT), intestinal fatty acid binding protein (I-FABP, a marker of gut epithelial integrity), and the immune activation markers soluble(s) CD14 and sCD163 were tested within 6 months of the VCTE visit. We used linear regression models to evaluate the associations of HIV, demographic, and metabolic covariates with log-transformed FAST score. Next, we individually added KT ratio, I-FABP, and sCD14 and sCD163 to the models. Finally, we determined the mediating effect of MT using the difference method on the association of HIV with FAST.
RESULTS: Median age was 49 years. Most (74.5%) were Non-Hispanic Black and over half were obese (median BMI 30.5 kg/m²). Among WWH, median CD4 count was 651 cells/mm³ and 91% reported taking antiretroviral therapy. MT biomarker levels were higher in WWH (p< 0.001 for each). On multivariable regression, HIV infection was associated with a 49.2% higher FAST score (95%CI 29,73%), as were age (2.1% increase per year, 95%CI 1,3%), and heavy alcohol use (52% increase over no drinking, 95%CI 17,98%). Black race was associated with a 22% lower FAST score (95%CI -38,-2%). Higher levels of MT biomarker were each significantly associated with higher FAST score (Table 1). After adjustment for MT biomarkers, the association of HIV with FAST score was attenuated, and MT mediated 11-28% of the association of HIV with higher FAST (Figure 1).
CONCLUSION: Compared to women without HIV, WWH have higher circulating biomarkers of MT and higher FAST scores. MT biomarkers are strongly associated with higher FAST score and mediate the association of HIV with higher FAST score. Our findings suggest that MT is an important mechanism by which HIV increases the risk of NASH with advanced fibrosis. This study lays the groundwork for future efforts evaluating the relationship between MT, microbial composition, and NASH in PLWH.

Abbreviations: CI, 95% confidence interval, FAST, FibroScan- aspartate aminotransferase Score; KT, Kynurenine -Tryptophan; I-FABP, intestinal fatty acid binding protein; sCD14, soluble CD14; sCD163, soluble CD163<br /> MT biomarker levels were log transformed<br /> MT biomarkers underwent standardization defined as the log transformed variable divided by IQR to compare effect sizes over varying units<br /> FAST linear models were adjusted for HIV status, age, BMI, race, insulin resistance (HOMA IR), alcohol use, tobacco, and menopause state. Each MT biomarker was entered in the model separately.

Abbreviations: CI, 95% confidence interval, FAST, FibroScan- aspartate aminotransferase Score; KT, Kynurenine -Tryptophan; I-FABP, intestinal fatty acid binding protein; sCD14, soluble CD14; sCD163, soluble CD163
MT biomarker levels were log transformed
MT biomarkers underwent standardization defined as the log transformed variable divided by IQR to compare effect sizes over varying units
FAST linear models were adjusted for HIV status, age, BMI, race, insulin resistance (HOMA IR), alcohol use, tobacco, and menopause state. Each MT biomarker was entered in the model separately.

Figure 1: Path analysis showing multivariable adjusted effects of human immunodeficiency virus (HIV) infection status on steatohepatitis (as determined by FAST score). Standardized β coefficients for each MT mediator are shown, as well as percentages of the total effect attributable to microbial translocation (MT, as demonstrated by each biomarker), and non-MT pathways (other direct and indirect mechanisms of infection status). Models are adjusted for HIV status, CD4 count, and HIV viral load, age, BMI, race, insulin resistance (HOMA IR), alcohol use, tobacco, and menopause state.

Figure 1: Path analysis showing multivariable adjusted effects of human immunodeficiency virus (HIV) infection status on steatohepatitis (as determined by FAST score). Standardized β coefficients for each MT mediator are shown, as well as percentages of the total effect attributable to microbial translocation (MT, as demonstrated by each biomarker), and non-MT pathways (other direct and indirect mechanisms of infection status). Models are adjusted for HIV status, CD4 count, and HIV viral load, age, BMI, race, insulin resistance (HOMA IR), alcohol use, tobacco, and menopause state.

Background:
Non-alcoholic fatty liver disease (NAFLD) is highly prevalent, rising globally, and one of the primary contributors to chronic liver disease and liver-related mortality. Although prior studies suggest a link between the gut microbiome and NAFLD, investigations explaining these dysbioses in terms of genomic and molecular function are lacking.
Methods:
To investigate the role of gut microbial communities in NAFLD, we constructed integrated host-microbiome-molecular profiles, including decades of clinical metadata, stool metagenomes/metatranscriptomes, and metabolomes among 211 participants with NAFLD and 502 healthy controls from the Nurses’ Health Study II.
Results:
We found that NAFLD case status explained a small, but significant amount of variance in gut taxonomic (R²=0.003, P=0.001) and metabolic (R²=0.004, P=0.01) profiles. Per-feature testing using linear models adjusting for age, body mass index, physical activity, diabetes, and diet identified 27 differentially abundant species between NAFLD and healthy controls. Veillonella atypica, V. parvula, V. dispar, Rothia dentocariosa, and Clostridium lavalense were more abundant in NAFLD (notably enriched for oral-typical microbes), while Coprococcus comes, Fusicatenibacter saccharivorans, Faecalibacterium prausnitzii, Oscillibacter spp., and other Firmicutes (generally gut-typical fastidious anaerobes) were depleted in NAFLD (Fig. A, B). These alterations corresponded to other feature-level changes in microbial biochemical potential (i.e., MetaCyc pathways) and their related metabolites (e.g, fatty acids, acylcarnitines, and secondary bile acids, such as isoallolithocholic acid). We observed phylogenetic heterogeneity in the expansion of oral-typical microbes when comparing lean vs. non-lean NAFLD, with Streptococcus spp. relatively enriched in non-lean NAFLD, while other oral species (particularly Veillonella) were comparatively increased in lean NAFLD (Fig. C, D). In a pilot sample of 47 individuals with metatranscriptomic data (n=15 with NAFLD, n=32 controls), we observed altered expression of metabolic pathways involved in arginine biosynthesis (L-arginine biosynthesis I and III), which has previously been linked to attenuated host responses to hepatotoxic injuries, and methionine biosynthesis (HSERMETANA-PWY: L-methionine biosynthesis III), which is an important mediator of hepatic lipid accumulation.
Conclusion:
Our work-in-progress represents among the most dense and complete evaluations of NAFLD gut ecology to-date, incorporating functional metagenomics, metatranscriptomics, and metabolomics. We anticipate a significant expansion in the size of our pilot metatranscriptional cohort, empowering new insights into NAFLD gut microbial activity. Further study is needed to elucidate the biological importance of alterations in gut microbial biochemistry to NAFLD risk and severity.

<b>Figure. Alterations in gut microbial communities in the NAFLD gut microbiome.</b> (a) Phylogenetic differences drive segregation of NAFLD gut microbial communities from those of a control population. (b) Relative expansion and depletion of more than 27 different microbes in NAFLD compared to controls. Effect sizes from multivariable linear modeling adjusting for age, body mass index, physical activity, diabetes, and diet (represented by the Alternative Healthy Eating Index) plotted against FDR-corrected p-value. (c) Expansion of oral typical species was broadly observed in NAFLD. (d) Heterogeneity in gut community structure between lean vs. non-lean NAFLD, with increases in <i>Streptococcus</i> observed more frequently in non-lean NAFLD. Beta coefficient of non-lean NAFLD (versus controls) plotted against beta coefficient of lean NAFLD (versus controls), only among oral taxa.

Figure. Alterations in gut microbial communities in the NAFLD gut microbiome. (a) Phylogenetic differences drive segregation of NAFLD gut microbial communities from those of a control population. (b) Relative expansion and depletion of more than 27 different microbes in NAFLD compared to controls. Effect sizes from multivariable linear modeling adjusting for age, body mass index, physical activity, diabetes, and diet (represented by the Alternative Healthy Eating Index) plotted against FDR-corrected p-value. (c) Expansion of oral typical species was broadly observed in NAFLD. (d) Heterogeneity in gut community structure between lean vs. non-lean NAFLD, with increases in Streptococcus observed more frequently in non-lean NAFLD. Beta coefficient of non-lean NAFLD (versus controls) plotted against beta coefficient of lean NAFLD (versus controls), only among oral taxa.

Paul Nelson

Appearances