Increased Abundance of Lactococcus Lactis as a Fecal Biomarker Predicting Development of Non-Alcoholic Fatty Liver Disease after Fecal Microbiota Transplantation in Human Microbiota-Associated Rodents

Abstract

Introduction: Gut microbiota dysbiosis is increasingly linked to the development of non-alcoholic fatty liver disease (NAFLD). Whether this relationship is causal in nature has not been established. It is also uncertain whether certain gut microbiota signatures can predict the disease course of NAFLD. We aimed to address these issues via fecal microbiota transplantation (FMT) in a human microbiota-associated rodent model. Aims & Methods: We recruited Asian human FMT donors categorized as three groups (n=8 each): obese NAFLD patients with body-mass index (BMI) ≥30 kg/m2 ; non-obese NAFLD with BMI < 25 kg/m2 ; and non-obese, non-diabetic healthy controls. NAFLD was diagnosed using controlled attenuation parameter (CAP) measurements via transient elastography (Fibroscan, Echosens, Paris), with steatosis defined as CAP ≥248 dB/m. After antibiotic administration, collected human fecal samples were transplanted individually without pooling to the same number of C57BL/6J mice, categorized as FMT-obese, FMT-lean and FMT-healthy. We collected liver, adipose tissue and blood from mice after a 12-week high-fat diet for assessment of histology, lipid metabolism, hepatic inflammation and intestinal barrier function. We performed shotgun metagenomics sequencing using the Illumina Novaseq 6000 platform and analyzed microbiome from the collected fecal samples. Results: Obese NAFLD human donors, when compared to non-obese NAFLD and healthy controls, had a significantly higher mean CAP (355.9±9.6 dB/m vs. 299.3±8.9 dB/m vs. 186.3±11.2 dB/m; p=0.01) and mean BMI (34.4±1.6 kg/m2 vs. 23.7±0.4 kg/m2 vs. 20.7±0.8 kg/m2 ; p<0.001). Principal coordinates analysis of microbiota profile at the phylum level based on Bray-Curtis distance showed non-obese NAFLD donors had a significantly more distinct gut microbiota composition compared with healthy controls (p=0.0341) and obese NAFLD donors (p=0.0479). FMThealthy mice, when compared to FMT-obese mice, had significantly lower plasma triglyceride levels (110.0±7.115 mg/dl vs. 207.9±11.52mg/dl, p<0.0001), significantly lower intrahepatic triglyceride content (48.80±4.463 mg/g vs. 68.51±4.183 mg/g, p=0.0061); decreased hepatic lipid accumulation (Oil Red O stained lipid droplets: 9.709%±3.231% vs. 21.37%±4.312%, p=0.0483); and decreased adipocyte size (10914±1437µm2 vs. 18455±2267 µm2 , p=0.0045). Tight junction protein zonula occludens-1 (ZO-1), an intestinal barrier marker, was significantly upregulated in FMT-healthy when compared to FMT-obese (1.370±0.4163 vs. 0.3302±0.1272, p=0.0316). There was no statistical difference in the abovementioned results when comparing FMT-healthy vs. FMT-lean (p>0.05 for all). Linear Discriminant Analysis Effect Size plot of taxonomic biomarkers showed a greater abundance of Lactococcus lactis in FMT-healthy mice when compared to FMT-obese mice (43.50%±12.92% vs. 24.82% ±12.76%, FDR-p=0.044). Lactococcus lactis correlated positively with tricarboxylic acid cycle pathway, a major energy-producing pathway involved in triglyceride breakdown (rho=0.59, p=0.002), but correlated negatively with plasma triglyceride levels (rho=-0.45, p=0.027). Conclusion: In this human microbiota-associated rodent model, FMT from healthy human donors prevented the development of NAFLD in mice, signifying a potentially causal relationship. The fecal presence of Lactococcus lactis as a biomarker was associated with a reduced risk of NAFLD development. Future animal and human interventional studies will be required to determine the feasibility of Lactococcus lactis as a therapeutic option for NAFLD. References: Supported by the Guangdong Natural Science Fund, Guangdong Province Science and Technology Department, Guangdong Province, China.