Nonalcoholic Fatty Liver Disease Is Exacerbated in High-Fat Diet-Fed Gnotobiotic Mice by Colonization with the Gut Microbiota from Patients with Nonalcoholic Steatohepatitis

1Diet-induced nonalcoholic fatty liver disease (NAFLD) is one of the most common chronic liver disorders worldwide due to the epidemic increase in the prevalence of obesity in affluent societies. The development and progression of NAFLD involve complex pathophysiological processes and are influenced by host factors such as genetic polymorphisms, diet, and, as reported more recently, the composition of the gut microbiota. NAFLD includes a wide spectrum of hepatic pathological characteristics ranging from simple steatosis to nonalcoholic steatohepatitis (NASH) with or without fibrosis, cirrhosis, and hepatocellular carcinoma. In a subset of the population (~10–20%), NAFLD progresses to NASH. This population has hepatic steatosis accompanied by inflammation and fibrosis. Although fibrosis is not a requirement for the diagnosis of NASH, it is often present in NASH patients. Moreover, in the liver samples of patients with NAFLD, inflammasome components are significantly upregulated in the patients with NASH when compared to those with non-NASH NAFLD.

2Recent studies have highlighted the role of the gut microbiota in the regulation of energy homeostasis and revealed its association with a number of disease states such as allergy, inflammatory bowel disease, cancer, and diabetes. Gabele and coworkers have shown that combined administration of a high-fat diet (HFD) and dextran sulfate sodium (DSS) leads to disturbances in the homeostasis of the relation between the gut microbiota and the host; this change may promote bacterial translocation from the gut into the portal circulation, further increasing liver damage. Additionally, Henao-Mejia et al. have provided evidence that modulation of the intestinal microbiota through multiple inflammasome components is a critical determinant of NAFLD and NASH progression. These studies indicate that the gut microbiota is deeply involved in NAFLD and obesity.

3Several recent studies have used germ-free mice that were colonized by a human gut microflora to investigate the disease progression of inflammatory bowel disease (IBD) colitis and intestinal physiology. Turnbaugh and collaborators transplanted fresh human fecal microbes into germ-free C57BL/6J mice to create an animal model with a defined and representative human gut ecosystem. The gut microbiota of these humanized gnotobiotic mice yielded responses similar to those of their human counterparts including microbial colonization dynamics, biogeographical distribution, and responses to dietary perturbations. Follow-up experiments involving human fecal samples from twins discordant for obesity (when transplanted into gnotobiotic mice) showed significant body composition differences. Based on these findings, we believe that the humanized mouse platform for studies on NAFLD pathogenesis is a powerful tool. Therefore, our aim was to determine the role of the gut microbiota in NAFLD by colonizing germ-free (GF) mice with fresh feces from NASH patients or from healthy individuals. We uncovered a significant correlation between gut microbiota composition and progression of NAFLD and elucidated how these differences may interfere with host metabolism and dietary energy absorption and storage.

4All the animals were housed in sterilized isolators at the temperature of 22 ± 1 °C, 55–65% relative humidity, in a 12-h light/12-h dark cycle. All the procedures were performed at an animal facility accredited by the Association for Assessment and Accreditation for Laboratory Animal Care, International, with the approval of the Institutional Animal Care and Use Committee at the National Laboratory Animal Center (approval number IACUC2011M16R01; approval date 10 September 2012). The methods in this study were carried out in accordance with the approved guidelines.

5A total of 40 germ free (GF) mice (10 mice per group) were used in this study. Three- to four-week-old male GF C57BL/6JNarl mice were orally inoculated with a fresh fecal mixture obtained from donors: either healthy humans (HL) (n = 10) or NASH patients (NASH) (n = 10). Disease status (HL or NASH) was first determined based on serum chemistry (alanine transaminase (ALT), aspartate transaminase (AST), triglycerides (TGs), cholesterol, and fasting glucose) and imaging by abdominal ultrasonography, and then a liver biopsy was performed only on NASH patients for confirmation. The donors of healthy human’s feces had normal body mass index (BMI) and clinical biochemistry, including ALT, TGs, cholesterol (T-CHO), and glucose (GLU) (Table S1).

6The average anthropometric variables for the HL and the NASH donors are summarized in Table S1. For each group of participants (the HL and NASH), 1 g of a fresh fecal sample from each of 10 individual human donors was mixed with 9 mL of PBS (1:9), and then the samples were pooled at equal volumes. Fecal mixtures were vortexed for 3 min at room temperature, and 0.5 mL was orally inoculated into each GF recipient mouse by gavage. After 4 weeks of colonization, the animals were started on either a standard diet (STD) (Research Diets D12450B, the diet composed of 20%, 70%, and 10% calories from protein, carbohydrate, and fat respectively) or the HFD (Research Diet D12492, composed of 20%, 20%, and 60% calories from protein, carbohydrate, and fat, respectively) purchased from Research Diets Inc. (New Brunswick, NJ, USA). This experimental design resulted in four groups: the HL feces inoculated in B6 mice fed the standard diet (HL-STD), the HL feces inoculated into B6 mice fed the high-fat diet (HL-HFD), the NASH patients’ feces inoculated into B6 mice fed the standard diet (NASH-STD), and the NASH patients’ feces inoculated into B6 mice fed the high-fat diet (NASH-HFD). After switching to the assigned diet, the mice were weighed weekly for 16 weeks and then euthanized by asphyxiation with 95% CO2 without fasting. Livers were extracted and fixed in 10% neutral buffered formalin for 24 h and next subjected to histopathological analysis or immunohistochemical staining. For preparation of frozen sections, tissues were fixed with Tissue Embedding Medium Compound (Tissue-Tek O.C.T. Compound, Sakura Finetek, Torrance, CA, USA), and rapidly frozen over a dry ice and hexane slush (−73 °C) and stored at −80 °C until sectioning on a cryostat at −20 °C. For liver TG assay or gene expression analysis, tissues were stored in liquid nitrogen.

7Whole blood was collected by cardiac puncture and centrifuged at 2600× g for 10 min at 4 °C. Serum was immediately stored at −80 °C until analysis of AST, ALT, GLU, TG, T-CHO, high density lipoprotein-cholesterol (HDL-C), and nonesterified fatty acids (NEFAs) on an automated analyzer (HITACHI 7080, Hitachi, Tokyo, Japan).

8Livers were fixed in 10% neutral buffered formalin for 1 day, dehydrated, embedded in paraffin, cut into 4-μm slices, and stained with hematoxylin and eosin (H&E) for histological examination.

9Fresh liver tissue was embedded in Tissue-Tek 4583 OCT compound (Sakura Finetek, Torrance, CA, USA). Tissue was sectioned at 4 μm on a Universal Microtome Cryostat (Leica CM3050S, Leica Microsystems, Nussloch GmbH, Nussloch, Germany) and processed for examination of fat accumulation by Oil Red O staining. The liver tissues were fixed with 10% neutral buffered formalin for 10 min, stained with isopropanol for 5 min, then stained with fresh 60% Oil Red O working solution for 7 min, followed by staining with 85% isopropanol for 3 min and counterstaining with hematoxylin for 2 min.

10The liver samples weighed ~50 mg and were homogeneous. The liver TG concentration of each sample was determined by means of the Triglyceride Colorimetric Assay Kit (Cayman, Ann Arbor, MI, USA).

11To assay endotoxin levels, serum samples were collected aseptically, then diluted 1:5 with the Limulus Amebocyte Lysate (LAL) reagent solution in a nonpyrogenic tube by means of a commercially available Pyrochrome Limulus Amebocyte Lysate Kit (Associates of Cape Cod, Falmouth, MA, USA). E. coli O113:H10 served as the standard; a standard series ranging from 0.04 to 1.28 EU/mL was used to evaluate the concentration of endotoxin. Standards, samples, and a negative control were heated for 15 min at 75 °C, then added into wells of a 96-well plate with 50 μL of pyrochrome already present in each well, mixed for 30 s, and incubated at 37 °C for 25 min. To stop the reaction, 25 μL of 50% acetic acid was added. Optical density (OD) was determined at 405 nm on a Multiskan™ GO ELISA reader (Thermo Scientific, Waltham, MA, USA). Calculations were performed in endotoxin testing with concentrations expressed in EU/mL.

12Mouse Serum Adipokine LINCOplex Kit (Millipore, Billerica, MA, USA) was applied to quantify serum tumor necrosis factor alpha (TNF-α), interleukin 6 (IL-6), monocyte chemotactic protein 1 (Mcp1), and insulin. The assays were conducted using the Luminex technology in accordance with the manufacturer’s protocol.

13RNA from the liver was isolated with the RNeasy Mini Kit (Qiagen, MA, USA). mRNA expression levels of Tnf-α, Il6, Mcp1, Toll-like receptor 2 (Tlr2), Toll-like receptor 4 (Tlr4), peroxisome proliferator-activated receptor gamma (Ppar-γ), and β-actin (Actb) were measured by qRT-PCR (for primer sequences used in the study see Table S2.) qRT-PCRs were carried out on LightCycler 1.5 (Roche, Mannheim, Germany) with the Light Cycler TaqMan Master Kit (Roche, Mannheim, Germany). Preincubation was performed at 94 °C for 10 min, followed by 40 cycles of denaturation at 94 °C for 10 s, then annealing at 60 °C for 30 s, and extension at 72 °C for 1 s. LightCycler® Software 3.5 (Roche, Mannheim, Germany) was used to perform relative quantification, with β-actin as internal control.

14To investigate the composition of the gut microbiota of humanized gnotobiotic mice (groups HL-STD, HL-HFD, NASH-STD, and NASH-HFD), fresh fecal samples from three individuals were collected at 4 weeks after oral inoculation and 16 weeks on the designated diets (indicated as Week 0 and Week 16, respectively). Each fresh fecal sample was first normalized to 20 mg/mL under anaerobic conditions, 10-fold dilution was performed under anaerobic conditions, and 0.05-mL samples were inoculated onto each of 5 nonselective and 13 selective agar media. Isolated bacteria were identified by analyzing colony and cell morphology, aerobic growth, and spore formation as well as by Gram staining. The bacterial counts per sample for each bacterial group were calculated and expressed as a log10 (colony-forming units (CFU)/g).

15All the data except the histopathologic score are presented as mean ± standard deviation (SD). SPSS ver. 18 (PASW Statistics for Windows, ver. 18.0, SPSS, Chicago, IL, USA) was used to analyze differences among groups. To compare values obtained from these groups, one-way analysis of variance was carried out, followed by Tukey’s post hoc test. Data with p < 0.05 were considered statistically significant. A two-way analysis of variance (ANOVA) was employed to compare multiple groups with a least significant difference (LSD) test. The significance of differences between histopathology scores was evaluated by the two-tailed Mann–Whitney U test. Data with p < 0.05 were regarded as statistically significant.

16In the HFD group, after the 16-week dietary intervention, the absolute epididymal fat weight and relative epididymal fat weight (percent of terminal body weight) were significantly higher in the NASH-HFD group than in the HL-HFD group. In the STD groups, the average of absolute epididymal fat weight and relative epididymal fat weight were higher in group NASH-STD than in the HL-STD group. Nonetheless, no significant differences in the terminal body weight, liver weight, and relative liver weight (percent of terminal body weight) were observed between groups NASH-HFD and HL-HFD (Table 1).

17The clinical biochemistry profiles are presented in Table 2. Serum ALT and AST levels were significantly higher in the NASH-HFD group than in the HL-HFD group. Moreover, a similar trend was seen in the NASH-STD group compared with HL-STD group. Serum glucose was significantly higher in the NASH-HFD group compared with the other groups. The serum concentrations of T-CHO, TG, and HDL-C were higher in the HFD-fed mice than in STD-fed mice. The NASH-HFD group showed marked accumulation of hepatic TGs (1.5-fold, p = 0.0021) in the liver tissues, as compared with the HL-HFD group. Of note, the levels of liver TGs in the NASH-STD group were 4.6-fold higher than those in the HL-STD group (Figure 1).

18Histological grading and staging system for liver steatosis and necrosis were based on the study by Brunt et al.. The NASH-HFD group showed a more advanced stage of steatosis (macro- and micro-vesicular) and multifocal necrosis. Moreover, the mild to moderate steatosis but not necrosis was seen in the HL-HFD group as compared with in the NASH-HFD group. Only mild steatosis and moderate necrosis were observed in the NASH-STD group (Figure 2A–D). In contrast, there were no lipid droplets or multifocal necrosis in the HL-STD group (Figure 2E–H). The average scores for steatosis were 3.8 ± 0.1, 3.1 ± 0.1, 0.6 ± 0.3, and 0 ± 0 in the NASH-HFD, HL-HFD, NASH-STD, and HL-STD groups, respectively (Figure 2I). The mean necrosis score was 3.2 ± 0.6 for the NASH-HFD group, which was greater than the score of 0.1 ± 0.1 in the HL-HFD group (p < 0.01; Figure 2J). Consistent with the results on steatosis on H&E staining, the Oil Red O-stained sections revealed a higher degree of stained lipid droplets in the NASH-HFD mice than in the HL-HFD or NASH-STD mice. Furthermore, very few lipid droplets were observed in the HL-STD animals.

19The serum level of IL-6 was significantly higher in the NASH-HFD group (113.7 ± 10.8 pg/mL) compared to the NASH-STD group (58.4 ± 4.2 pg/mL); however, it was not detected in the serum of HL-mice, regardless of whether they were fed HFD or STD (Figure 3). The serum levels of Mcp1 of the NASH-HFD group were 2.1-fold higher than those of the HL-HFD group; however, there were no statistically significant differences between groups HL-STD and NASH-STD. Serum Tnf-α was not detected in any group. The serum endotoxin levels were 2.96 ± 0.59, 4.64 ± 0.39, 3.27 ± 0.78, and 6.05 ± 1.04 EU/mL in groups HL-STD, NASH-STD, HL-HFD, and NASH-HFD, respectively. Our data indicated that the serum levels of endotoxin were significantly elevated in NASH-HFD mice (Figure 4).

20Tlr2 expression levels were elevated in the NASH group compared to the HL- group regardless of the diet (Figure 5A). The expression level of Tlr4 was the highest in the NASH-HFD group (Figure 5B). The expression of Tnf-α was significantly higher in the NASH-HFD mice and slightly increased in the NASH-STD group, and showed no differences between the HL-STD and the HL-HFD groups (Figure 5C). The mRNA expression of Il6 was higher in NASH mice than in HL mice, for both STD- and HFD-fed groups (Figure 5D). As for the Mcp1 expression, the HL-STD group had significantly lower levels than did the other groups, with group NASH-HFD showing the highest levels among them (Figure 5E). Ppar-γ mRNA expression was higher in the NASH group, with the NASH-HFD group showing the highest expression (Figure 5F).

21The Firmicutes and Bacteroidetes were the two most abundant phyla in the gut microbiota. The relative abundance of Firmicutes (including Lactobacillales, Streptococcaceae, Eubacteriaceae, and Clostridiaceae) was significantly higher in groups NASH-HFD, HL-HFD, and NASH-STD than in the HL-STD group; however, the relative abundance of Bacteroidetes was the highest in the HL-STD group (Table 3, Figure 6). The relative abundance of the phylum Proteobacteria (Enterobacteriaceae) was significantly higher in the NASH-HFD group than in the HL-HFD group. When we analyzed gut microbiota composition at the family level, the relative abundance of Streptococcaceae was higher in the NASH-HFD than NASH-STD group with no significant changes observed in groups HL-STD and the HL-HFD. A significant increase in the number of Lactobacillaceae cells was found in HL-HFD mice compared with the other groups. No significant difference in Clostridiaceae was observed among groups HL-STD, NASH-STD, and NASH-HFD. Of note, Clostridiaceae were less abundant only in the HL-HFD group. Moreover, Enterobacteriaceae were upregulated only in the NASH-HFD group. Taken together, our data suggested that the pathological progression of NAFLD may be associated with the relative abundance of these intestinal microbes.

22Some studies indicate that the gut microbiota plays an important role in the regulation of host metabolism homeostasis, innate immunity, development of insulin resistance, and progression of obesity. Several studies have used GF rodents to elucidate the intricate and complex interactions among host metabolism, intestinal flora, and progression of obesity, diabetes, and cardiovascular diseases. In addition, the gut microbiota was recently shown as an environmental factor that may promote metabolic diseases such as NAFLD and NASH. In the present study, we employed a novel humanized gnotobiotic mouse model and demonstrated that colonization with fecal bacteria from NASH patients along with HFD feeding can aggravate disease progression of NAFLD. These phenotypic variations may correlate with an increase in the levels of inflammatory cytokines (IL-6, Mcp1, and Tnf-α), TLR expression enrichment (Tlr2 and Tlr4), insulin resistance (insulin and Ppar-γ), and an alteration of gut microbiota composition (a decrease in the relative abundance of Bacteroidaceae and an increase in the relative abundance of Streptococcaceae). In addition, gnotobiotic mice colonized with fecal bacteria from healthy humans and fed HFD showed only simple steatosis without necrosis in the liver. The hepatic morphological and functional changes may correlate with a decreased proportion (abundance) of Clostridiaceae and increased proportion of Lactobacillaceae.

23Our results show that there were no differences in the body weight and liver weight between groups HL-STD and NASH-STD or between groups HL-HFD and NASH-HFD. Of note, the weight of epididymal fat in mice inoculated with the NASH human fecal mixture, regardless of whether the mice were fed STD or HFD, was significantly higher than that of mice inoculated with the HL fecal mixture. In one study, Backhed and coworkers reported that the colonization of GF mice with microbiota from conventionally raised animals produces an increase in body fat. Turnbaugh and colleagues reported that the obesity-associated microbiota has an increased ability to harvest energy from the diet. Hence, colonization of GF mice with the “obesity-related microbiota” resulted in a significant increase in total body fat relative to GF mice colonized with a “thinness-related microbiota”. Our results confirmed that inoculation of the gut microbiota from NASH patients in conjunction with high dietary fat content can indeed increase the accumulation of body and epididymal fat and aggravate the severity of NAFLD in the humanized gnotobiotic mice. These data suggest that the microbiota by itself may directly or indirectly cause an increase in weight and fat gain. Furthermore, in our study, regardless of whether STD or HFD was administered, the gut microbiota of mice inoculated with HL feces consisted of a greater proportion of Bacteroidetes than that of mice inoculated with the feces of NASH patients. The opposite effect was observed for Firmicutes. Research has shown that the relative abundance of Bacteroidetes is lower and that of Firmicutes is higher in obese mice compared to their leaner counterparts. The level of Bacteroidetes is significantly lower in obese subjects as compared with normal-weight subjects. These studies suggest that interventions targeting the metabolism-associated activity of gut microbiota might be effective in treating obesity and the associated metabolic disorders such as NAFLD and NASH.

24Recent evidence from humans and from animal models has also linked gut microbiota to the pathological progression of NAFLD/NASH via the gut–liver axis. In the present study, ALT activity was significantly higher in mice inoculated with NASH feces than in those inoculated with HL feces. These results are not consistent with the cirrhosis patient severity, and might not correlate with the ALT levels according to recent reports.

25Blood endotoxin concentration and hepatic TG showed similar results. Hepatic steatosis, inflammation, and fibrosis are key factors in the pathogenesis of NASH. In this study, the results revealed that severity of micro- and macrosteatosis in the NASH-HFD group is higher as compared with the HL-HFD group. Moreover, mild to moderate necrosis was observed only in the NASH-HFD group. Histological characteristics of present mouse model, such as (macro- or microvesicular) steatosis and multifocal necrosis were similar to those of human NAFLD. However, other histological features were absent (e.g., pigmented macrophages, megamitochondria, and Mallory–Denk bodies). These findings are consistent with a previous report that indicated the above lesions would not appear in mice. In contrast, in the histological analyses, no liver fibrosis was observed in any of the animals. Recent studies indicate that the inflammasome may perform an important function in NAFLD and NASH progression and pointed out that the intestinal microflora can regulate activities of the inflammasome. Therefore, we suggest that inoculated NASH gut microbiota causing severe NAFLD might be related to inflammasome activation, and further experiments are needed for verification.

26Several reports point to strong correlations between the gut microbiota and inflammatory cytokines and the contribution to the progression of NAFLD. The TLR family of pattern recognition receptors is critical for host defense against invading pathogens. In the present work, the serum level of inflammatory cytokine IL-6 was higher in both the human NAFLD patients and the NAFLD feces-inoculated gnotobiotic animals regardless of dietary fat content. Recently, Mas and coworkers found that the severity of diet-induced NASH is lower in Il6-deficient mutant mice compared to controls. Our results are consistent with those in other reports, and the serum level of IL-6 may be associated with colonization by human NASH patients’ fecal bacteria. The serum Mcp1 protein levels and mRNA levels of Mcp1 were higher in the NASH-HFD group than in others. The Mcp1-deficient mice fed an HFD show reduced insulin resistance and hepatic steatosis. On the other hand, mice overexpressing Mcp1 in adipose tissues show increased insulin resistance and hepatic TG levels. Our findings are in agreement with other reports that Mcp1 may promote liver fat accumulation and the associated inflammation.

27Our data suggest that the mRNA expression levels of Tlr2 and Tlr4 are higher in the NASH-HFD group than in the other groups. Several animal models have revealed that TLR deficiency-associated changes in the gut microbiota composition are related to exacerbated hepatic steatosis and inflammation. Another study indicates that gut microbiota dysbiosis can cause an influx of Tlr4 and Tlr9 agonists into the portal circulation, leading to enhanced hepatic Tnf-α expression that drives NASH progression. Tlr4 is the receptor for endotoxin, which is an important mediator of liver inflammation associated with NAFLD and hepatocellular carcinoma. A close association between NAFLD and increased levels of circulating endotoxin has been observed in obese mice and humans. Furthermore, Ye and colleagues demonstrated that Tlr4 is essential for hepatic fat deposition and NASH development. Hence, gut microbiota colonization from NASH feces in conjunction with the HFD might drive NAFLD progression. Moreover, the inoculation with intestinal flora from healthy people and feeding the HFD did not induce an inflammatory reaction but steatosis could still be seen. Research suggests that immune modulation of the gut microbiota induces symptoms of metabolic syndrome in Tlr2 knockout mice. Conversely, we also found that expression of the Tlr2 gene was upregulated in the in the NASH-HFD group. We suspected that these alternations might due to changes of intestinal microbiota composition. These data indicate that inflammatory cytokines, Tlr2- and Tlr4-mediated pathways crucially contribute to the progression of NAFLD/NASH.

28Some literature data indicate that, in rats fed a fructose-rich diet, the development of metabolic syndrome directly correlates with variations of the gut concentration of specific bacterial taxa. In our study, the blood glucose level in the NASH-HFD group was significantly higher than that of the HL-HFD group; however, the opposite effect was observed for insulin. Several research groups demonstrated a correlation between type 2 diabetes and obesity. Membrez and coworkers used ob/ob mice to show that modulation of the gut microbiota via antibiotics reduces hepatic steatosis and improves glucose sensitivity. Furthermore, we observed that the insulin levels were lower in the NASH-HFD group than in the HL-HFD group. A similar result was obtained for the STD-fed group. Our findings indicate that the colonization with fecal bacteria from HL should have a greater relation with obesity or NAFLD (simple steatosis type) induction. In contrast, the feces from NASH may correlate with NAFLD (steatohepatitis type). Cryer reported that an increase in serum glucose concentration induces an increase in insulin secretion by β-cells of the pancreas; however, β-cell failure did not result in a decrease in β-cell insulin secretion. Furthermore, Satoch and coworkers demonstrated that hepatic steatosis is induced by an increase in incorporation of fatty acids into the liver via increased Ppar-γ expression in the liver of insulin-resistant mice. The results of the present study are consistent with these findings. In contrast, mice inoculated with fresh feces from a healthy person showed only glucose intolerance. Therefore, the imbalance of gut microbiota might correlate with insulin resistance and contributed to NAFLD severity.

29In recent years, experimental and clinical studies have shown a pathogenic association between the gut microbiota and NAFLD. Le Roy and coworkers demonstrated that gut microbiota contributes to the development of NAFLD independently of obesity. The different bacterial groups that we investigated in this study are affiliated with three phyla: Bacteroidetes, Firmicutes, and Proteobacteria, which are crucial for gut homeostasis and are believed to be involved in NAFLD. Our results indicate a higher abundance of the phylum Firmicutes and a lower abundance of the phylum Bacteroidetes, as well as Proteobacteria in the NASH-HFD group when compared with group HL-STD. These results are in agreement with those of Sefcikova and other researchers. Additionally, Firmicutes, particularly Streptococcaceae, have been noticed in the NASH-HFD group but not in the NASH-STD group. In contrast, they were not observed in HL groups. As compared to healthy subjects, patients with NAFLD also had an increased percentage of bacteria from the families Streptococcaceae and Enterobacteriaceae, both known to induce persistent inflammation in the intestinal mucosa and to be associated with IBD. A significant increase in the number of Lactobacillaceae was found in group HL-HFD. Other studies that compared obese patients with healthy controls uncovered an increase in cell counts of genus Lactobacillus and a decrease in the family Ruminococcaceae in obese patients. Regarding the association with Lactobacillus, it is surprising, because several species from this genus are frequently used as probiotics. Lactobacillus is lactic acid bacteria that can suppress pathogens, enhance epithelial barrier function, and modulate immune responses: actions that seem to protect from the pathogenesis of NAFLD and/or NASH. In contrast, Lactobacillus may be associated with the production of volatile organic compounds such as acetate and ethanol, which may be important for the pathogenesis of obesity and NAFLD. Patients with NASH were recently shown to have lower abundance of bacteria belonging to the phylum Bacteroidetes (family: Bacteroidaceae) compared to subjects with simple steatosis and healthy individuals as shown by qPCR. In addition, there are gender differences in the results on Western-diet-induced steatosis severity. Additionally, Jena and coworkers reported that the Western-diet-fed male Farnesoid X Receptor (FXR) knockout mice have severer steatohepatitis as compared to male wild-type mice. The antibiotic treatment might reduce hepatic inflammation in each mouse. Especially, Lactococcus, Lactobacillus, and Coprococcus decrease the liver inflammation. These results revealed the relationship between Western diet and gut microbiota might play important roles in NAFLD and NASH.

30The results presented here provided evidence that composition of the intestinal microbiota may interfere with NAFLD and/or NASH progression through multiple inflammation reactions. As far as we know, this is the first report of exacerbation of NAFLD in the HFD-fed humanized gnotobiotic mice after direct colonization with fecal bacteria from NASH patients, but not healthy people. In addition, we found complex and cooperative effects of Tls, inflammatory cytokines, and host gut microbiota on NAFLD parameters. Taken together, our findings may lead to novel strategies (such as colonization with fecal bacteria from NASH patients) to investigate NAFLD pathogenesis.