Intestinal microbiome and NAFLD: molecular insights and therapeutic perspectives

A number of studies have shown that the intestinal microbiota is a new environmental factor leading to obesity and NAFLD. First, Bäckhed and colleagues reported that the germ-free (GF) mice are protected against obesity produced by a Western-style, high-fat and high-sugar diet [29]. Moreover, the conventionalization of GF mice with a normal intestinal flora from conventionally raised animals increased the body fat content and insulin resistance within 14 days [30].

A direct involvement of gut microbes in the development of NAFLD was suggested by the finding that NAFLD could be delivered to GF mice by fecal microbiota transplantation (FMT) [31]. In this study, C57BL/6J mice fed with HFD usually showed hepatic steatosis and systemic inflammation (responders), but there were some other mice which were non-responders and failed to exhibit the symptom of metabolic disorders after HFD treatment. To explore the potential reasons for these inconsistent responses, the mouse gut microbes from responders or non-responders were transplanted to GF mice. Compared with the non-responder receiver group, the responder receiver group had a significant increase in fasting insulinemia and plasma aspartate aminotransferase (AST). Also, the responder receiver mice accumulated more hepatic triglycerides and subsequent liver steatosis [31].

Further studies suggest that the harmful metabolites produced by altered intestinal microbiome should be responsible for the development of NAFLD [32]. For example, Yamada and co-workers have reported the relationship between saturated fatty acids generated by intestinal microbiota and pathogenesis of NASH [33]. The authors found that feeding with STHD-01, a new class of HFD, led to the development of NASH in mice, accompanied by the dysbiosis and alterations in luminal metabolic profiles. This study revealed that the accumulation of saturated fatty acids, transformed from unsaturated fatty acids by gut microbiota, elicited inflammation by activating the migratory macrophages in liver [33]. More recently, Yuan et al. reported that the high alcohol-producing Klebsiella pneumoniae (HiAlc Kpn) was detected in 61% of individuals with NAFLD in a Chinese cohort [34]. To investigate the relationship between HiAlc bacteria and fatty liver disease, they fed specific-pathogen-free (SPF) mice with HiAlc Kpn. Strikingly, HiAlc Kpn feeding induced the chronic hepatic steatosis. Furthermore, the transfer of a HiAlc-Kpn-strain-containing microbiota from NASH patients by FMT into mice resulted in the occurrence of NAFLD. However, after selective depletion of the HiAlc Kpn strain by phage, there were no significant NAFLD symptoms in recipient mice. These results illustrated the contribution of a high level of alcohol-producing gut bacterial strain to pathogenesis of NAFLD [34]. Overall, the gut microbiota and their detrimental metabolites (ethanol, saturated fatty acids, polyamines, hydrogen sulfide, and so on) likely drive the damage to liver. And more studies are required to discover which intestinal microbes and/or their metabolites can promote the initiation and progression of NAFLD.

Excessive gut-derived endotoxin will induce the production of ROS in liver, while ROS can damage vulnerable hepatic cells and thus lead to the occurrence of NASH [35]. In 2001, Wigg et al. reported a higher prevalence of small intestinal bacterial overgrowth (SIBO) in NASH patients compared with healthy subjects [36]. In 2009, Miele et al. examined the incidence and potential mechanism of increased intestinal permeability in NAFLD patients [37]. They found that NAFLD patients had significantly increased intestinal permeability in comparison with healthy volunteers. And this abnormality was associated with an increased prevalence of SIBO. It was supported by Harte et al. who reported that a higher circulating level of endotoxin was detected in patients with NAFLD and NASH compared with healthy controls [38].

In addition to gut-derived endotoxin, multiple studies have been performed to examine the difference of gut bacterial compositions between healthy subjects and NAFLD patients. Boursier and co-workers investigated the association of disbalanced intestinal bacterial community with severe NAFLD lesions (i.e., NASH and fibrosis) [28]. Multivariate analysis shows that the enrichment of Bacteroides genus was independently associated with NASH, and the increased abundance in Ruminococcus was positively related to the deteriorated fibrosis [28]. In a pediatric study, Zhu et al. described the alteration of gut microbiomes in patients with NASH, suggesting that children with obesity or NAFLD featured higher abundances of Prevotella and Bacteroidetes as compared to healthy controls [39]. More recently, Loomba et al. provided a novel method based on gut microbiome for non-invasive detection of advanced fibrosis in NAFLD patients [40]. Given the association between specific microbiota and NASH, it is possible to develop a panel of gut microbiome-derived biomarkers to predict advanced fibrosis.

Taken together, the above studies support the view that intestinal microbiota dysbiosis is a key environmental factor contributing to the NAFLD development and its progression into NASH.

To diminish the effects of microbial components and their metabolites on host health, antibiotics are usually used to reduce the number of intestinal flora. There are two types of antibiotics: absorbable antibiotics and non-absorbable ones. The former can effectively pass through the intestinal barrier to achieve therapeutic serum concentration, while the latter remains mainly within the gut milieu. Starting in the 1950s, antibiotics such as rifaximin, metronidazole and neomycin had been reported to treat patients with cirrhosis and hepatic encephalopathy [42, 43]. Also, the combined use of antibiotics (neomycin and polymyxin B) was proved to prevent fructose-induced hepatic lipid accumulation by decreasing the translocation of gut toxins [44].

Besides the suppression against local or systemic infection, antibiotics have regulatory effects on intestinal microbiota and are of benefit to NAFLD. For example, the treatment with cidomycin orally was found to promote small intestine transit rate and reduce serum levels of alanine aminotransferase (ALT), AST and TNF-α in NASH mouse model, indicating a potential of cidomycin in alleviating the severity of NASH by intestinal microbiota modulation [45]. Rifaximin, a largely water-insoluble and nonabsorbable (< 0.4%) drug, has been shown to exert antimicrobial activity against enteric bacteria such as Streptococcus, Bacteroides and Citrobacter [46]. Gangarapu et al. have demonstrated that a short-term administration of rifaximin (1200 mg/day for 28 days) improved the clinical status of patients with NAFLD/NASH, which was associated with reduced serum transaminases and circulating endotoxins [47]. Abdel-Razik et al. reported that after rifaximin therapy (1100 mg/day for 6 months), patients with NASH showed an evidently reduced levels of proinflammatory cytokines, ALT and NAFLD-liver fat score [48]. However, in an open-label clinical trial, rifaximin administration (800 mg/day for 6 weeks) was not effective for humans with NASH. The inconsistency may be due to the small sample size, the relatively low treatment dose or short duration in clinical study.

Specific antibiotics can act positively on intestinal microbiota and provided a so-called ‘eubiotic’ effect by promoting the growth of beneficial gut bacteria (e.g., Bifidobacteria and Lactobacilli), and this property may represent a therapeutic advantage in particular clinical practices [49]. On the other hand, although short-term treatment with antibiotics significantly improved NALFD, the long-term use of antibiotics should be careful in consideration of the possible side effects [50]. For example, antibiotics affect not only the harmful bacteria but also the healthy ones due to their wide ranges of action. Additionally, the use of antibiotics in immunocompromised or critically ill patients should be carefully evaluated to reduce the risk of infective endocarditis and bacteremia [51]. Besides, antibiotics may select for antibiotic-resistant strains in the human gut [52]. Collectively, the depletion or alteration of intestinal microbiota by antibiotics seems to alleviate the severity of NAFLD. However, considering the risk of antibiotics (side effects, resistance, etc.,), their clinical use must be cautious in treatment of NAFLD.

Probiotics are preparations or products containing viable, defined microorganisms in adequate amounts. Probiotics exert beneficial effects in host by altering the composition of microbial flora [53]. So far, the main commercialized probiotics in market include Lactobacilli, Streptococci, Bifidobacterial and Fungi [54]. Although most probiotics are from bacteria, the yeast strain such as Saccharomyces boulardii has also been proven to be an effective probiotic [55].

Probiotic treatment can protect the gut barrier from being damaged. The work by Karczewski et al. demonstrated that administration with L. plantarum strain WCFS1 enhanced the expression of Zonula occludens-1 (ZO-1) and Occludin close to the tight-junction structures [60]. MIYAIRI 588, a specific phenotype of the strain C. butyricum, has been used as a probiotic for treating colitis and antimicrobial-associated diarrhea [61, 62]. A recent study suggested that MIYAIRI 588 prevented the progression of steatosis to liver carcinogenesis in a rat NAFLD model [63]. Parallel studies from another group confirmed the same result, in which MIYAIRI 588 improved HFD-induced fatty liver in rats [64].

Collectively, these studies indicate that probiotics play a therapeutic role in NAFLD treatment. It seems that different probiotics may act on different target organs by changing the composition of intestinal microflora, producing antimicrobial peptides, reducing intestinal permeability or preventing the translocation of bacterial products [53]. Due to the multiple pathologic mechanisms of NAFLD, the combined administration of several probiotic strains may be more effective than a single one [65, 66]. Additionally, there are still some questions to be elucidated for the role of probiotics against NAFLD. For example, it is not clear how probiotics act on their specific target organs, and studies are required to clarify the crosstalk between probiotics and original bacterial inhabitants in the gut.

Noticeably, though probiotics have been recognized as a potential therapeutic tool for NAFLD, their beneficial effect should be further demonstrated by large randomized and controlled studies [67]. Until now, most of the studies that demonstrate the therapeutic effects of probiotics were mainly from Middle East countries. It is well known that the composition of gut microbiota is highly heterogeneous among the populations from different regions. Hence, the beneficial effects of probiotics on NAFLD/NASH need to be verified among people with ethnic difference, and the optimal formulations and dosages are also required to be determined in the development of commercialized probiotic products.

Prebiotics, which contain no living microorganisms, are nondigestible food ingredients that can selectively promote the proliferation and/or activity of one or several gut microbes [68]. Synbiotics are combination of prebiotics and probiotics [69]. It is estimated that the average intake of prebiotic fiber is 1–4 g/d in United States and 3–11 g/days in Europe, which is related to the dietary habits of local residents [70]. In animal models, prebiotics are usually used at doses of 5–20% by weight.

There are numerous in vitro and clinical studies demonstrating that prebiotics and synbiotics can be used to treat NAFLD (Table 1). Common prebiotics include oligofructose (OFS), lactulose, inulin and Synergy1^® [71]. The de novo lipogenesis pathway was found to be threefold higher in patients with NAFLD, indicating that the increased de novo lipogenesis is a key feature of this disease [72]. Prebiotic supplementation may improve NAFLD by reducing the fatty acid synthesis pathway as has been shown in animal experiments [71, 73]. Kok et al. reported that feeding with 10% of oligofructose, a nondigestible but fermentable oligomer of β-d-fructose, significantly alleviated fructose-induced hepatic triglyceride (TG) accumulation in rats [74]. The decreased lipogenic capacity was due to the reduced gene expression of enzymes which regulated hepatic lipogenesis, such as acetyl co-A carboxylase and fatty acid synthase (FAS) [75, 76]. Cani et al. proved that dietary supplementation with oligofructose improved body weight gain, reduced adipose development, and controlled HFD-induced inflammation [77]. The study also suggests that oligofructose might control the occurrence of metabolic diseases by modifying gut microbiota in favor of Bifidobacterial, which has been shown to improve mucosal barrier function and reduce the level of gut endotoxin [77–79]. In recent years, our group focused on the improvement of prebiotic supplement on metabolic syndrome in obese or HFD-fed mice [80, 81]. The results indicate that chitosan oligosaccharides (COS), oligomers of β-(1–4)-linked d-glucosamine, displayed dramatically suppressive effect on glucolipid metabolism disorder, including alleviation of insulin intolerance, and prevention of intestinal barrier damage [80]. COS treatment also reshaped the unbalanced gut microbiota in NAFLD mice by upregulating populations of Lachnospiraceae_UCG_001 and Akkermansia, and reducing abundances of Lachnospiraceae NK4A136 group, Alistipes, Helicobacter and Odoribacter [80]. In parallel studies, treatment with N-acetylated chitooligosaccharides (NACOS) prevented the occurrence of NAFLD in HFD-fed mice by suppressing hepatic inflammation and lipid accumulation and promoting the growth of beneficial gut bacteria [81].

Other preclinical and clinical trials to test the potential benefits of prebiotics on NAFLD patients are underway (Table 1). More recently, in a pilot clinical trial, 14 NASH patients were allocated to oligofructose or placebo intervention for 9 months [82]. Despite no changes in body mass index (BMI), oligofructose treatment significantly improved hepatic steatosis and NAS score. In addition, oligofructose supplementation increased the abundance of Bifidobacterium spp, which was inversely associated with obesity and plasma LPS [83, 84].

In the previous study, Eslamparast et al. found that the synbiotic plus lifestyle modification was superior to lifestyle modification alone for NAFLD treatment [58]. Similar result was reported by Mofidi et al. who provided evidence that synbiotic supplementation improved the main features of NAFLD in patients with normal and low BMI through reduction in inflammatory indices [85]. In both clinical trials, patients were assigned to consume the same synbiotic capsules, containing seven strains of bacteria and fructo-oligosaccharide. Although the data from trials have shown that synbiotics may alter the progression of NAFLD (Table 1), the exact mechanism of their therapeutic effects remains to be determined, and the future work should focus on the elucidation of host energy balance, regulators of metabolism, as well as reshaping of gut microbiota in NAFLD.

Based on the results from animal experiments and clinical trials, both prebiotics and synbiotics have been recognized as a potential therapeutic tool for NAFLD. However, given the limited investigation in this field, the generalization of prebiotics/synbiotics for treatment of NAFLD needs to be confirmed by high-quality clinical trials. Additionally, though prebiotics are constituents of natural foods, there were still studies showing that the consumption of prebiotics in excess of 30 g/days would cause adverse gastrointestinal effects such as flatulence [86]. In considering the vast differences in intestinal microbiota, dietary habits and host health status, the doses of therapeutic prebiotics/synbiotics should be individually customized in clinical application.

Considering the effect of FXR and TGR5 signaling on glucose and lipid metabolism, it is believable that the molecules which can modulate both receptors (i.e., FXR agonists) or regulate endogenous levels of bile acids might have therapeutic effects on NAFLD/NASH (summarized in Fig. 3).

Gastrointestinal tract is the largest endocrine organ in human body and secretes more than 20 different intestinal hormones [125]. Several G protein-coupled receptors (e.g., GPR41, TGR5) exist on the membrane of enteroendocrine L cells. The nutritional elements and some hormonal factors (insulin and leptin) can stimulate these receptors to induce the secretion of intestinal hormones such as ghrelin, Peptide YY and glucagon-like peptide 1/2 (GLP-1/2) [126, 127]. After being transported into systemic circulation, the intestinal hormones act on their target organs such as liver, adipose tissue and intestinal tract, thereby regulate NAFLD-related metabolic indices including glucose metabolism, insulin resistance and metabolic inflammation [128, 129]. Besides, intestinal hormones can be affected by the metabolites of gut microbiome, i.e., SCFAs [130]. A number of studies have suggested that intestinal hormones are critical to the pathogenesis of NAFLD and may become attractive targets for treatment of such disease [131, 132].

GLP-1, an incretin hormone secreted by L cells, can enhance glucose-induced insulin and inhibit the release of glucagon. The GLP-1 receptor agonist liraglutide has been recognized as a promising option for obesity treatment [133]. Wang et al. investigated the mechanism about the effect of liraglutide on weight control and reported that liraglutide could change the overall structure of intestinal microbiota in mice, leading to more lean-related phylotypes [134]. More recently, Moreira and co-workers investigated the effect of liraglutide on NAFLD in murine models of obesity [135]. It was found that liraglutide treatment led to obvious weight loss in obese mice, accompanied by the attenuated hepatic lipid accumulation. Moreover, microbiota analysis illustrated that liraglutide modified the diversity of gut microbiota by reducing the population of Proteobacteria and increasing the content of Akkermansia muciniphila, which were correlated with improved symptoms of NAFLD. Several clinical studies also demonstrated the efficacy and safety of liraglutide in patients with NASH or NAFLD [129, 136, 137]. Nevertheless, these clinical trials have not confirmed whether liraglutide exerts the efficacy by regulation of gut microbiota, and further investigations deserve to be performed to explore the potential relationship between therapeutic effect of liraglutide and its modulation on gut microbiota in patients with NAFLD or NASH.

Endotoxaemia is also implicated in the pathogenesis of NAFLD [36]. By using adsorbents (a series of highly adsorptive materials), the toxins and bacterial products in gut were bound and thereby their flow into the liver and systemic circulation were suppressed [138, 139]. Since adsorbents fail to be adsorbed or degraded in gastrointestinal tract, they were mainly discharged in the way of stool. For example, AST-120 is a spherical carbon adsorbent (0.2–0.4 mm in diameter) with broad non-specific binding surface area (> 1600 m²/g), the application of which effectively lowered absorption of gut-derived ammonia into the body circulation in rats with chronic liver failure [140]. Recently, a synthetic activated carbon, i.e.,Yaq-001 (Yaqrit Ltd. UK), has been developed with optimized pore size in both macro- and micro-porous range [141]. Yaq-001 can selectively absorb intestinal-derived toxins such as cytokines, hydrophobic bile acid and bacterial products. Oral Yaq-001 therapy revealed a significant reduction in ALT and hepatic TLR-4 expression in rodents with NAFLD [142]. Currently, the clinical trials to assess therapeutic effect of Yaq-001 are under investigation as a part of the European Commission Horizon 2020 program (carbalive.eu) [141].

Fecal microbiota transplantation (FMT) is a new approach to clinical treatment, in which gut microbes are transferred from healthy donor to diseased recipient. By this way, a ‘healthy’ gastrointestinal microbiota may be reconstructed. Zhou and co-workers found that FMT intervention remarkably increased the concentration of butyrate in fecal contents and improved the tight junction of small intestine. This study further proved that FMT attenuated steatohepatitis in mice by a beneficial regulation of gut microbiota [143]. A growing interesting in FMT and its potential in liver diseases has been reflected in ongoing trials such as NAFLD/NASH (NCT03803540↗, NCT02469272↗ and NCT02721264↗) and liver cirrhosis (NCT02862249↗). In the future, further high-quality clinical data are needed to determine the efficacy and safety of FMT. And standardized protocols should be formed such as sample preparation, archiving, formulations and dosages.