Understanding the Role of the Gut Microbiome and Microbial Metabolites in Non-Alcoholic Fatty Liver Disease: Current Evidence and Perspectives

SCFAs, mainly acetate, propionate, and butyrate are organic fatty acids synthesized from non-digestible proteins and fibers via anaerobic fermentation by the gut microbiota [66,67]. They are mainly produced in the distal colon, where they serve as a complementary factor to intestinal integrity and function. SCFAs are transferred to the liver by means of the portal circulation, thereby serving as precursors for gluconeogenesis and lipogenesis [68,69]. In fact, SCFAs are responsible for about 5–10% of the typical energy demands under normal conditions [70,71]. There are numerous studies demonstrating that SCFAs activate the G-protein-coupled receptors (GPCRs) GPR41 and GPR43 in the surface of the gut enteroendocrine L cells. In particular, activation of the GPCRs stimulates peptide YY (PYY) release, which results in the slowing down of gastric emptying and the promotion satiety [72,73]. In addition, activation of GPR41 and GPR43 on the surface of the L cells increases secretion of GLP-1, an incretin known to slow gastric emptying and induce satiety, thus decreasing food intake. Furthermore, GLP-1 enhances lipid oxidation in the liver, which contributes to diminished steatosis [73,74,75,76]. Besides their functionality as energy substrates, SCFAs have the potential to affect hepatic metabolism by functioning as signaling molecules. In particular, propionate and butyrate activate AMP-activated protein kinase (AMPK) to promote hepatic autophagy, a catabolic process which results in the hydrolysis of TG and the release of fatty acids for β-oxidation in the mitochondria [76,77,78,79,80]. The activation of AMPK by SCFAs has been related to increased Uncoupling Protein 2 (UCP2) levels and an increased AMP:ATP ratio [81]. Apart from the AMPK activation, SCFAs inhibit class I and II histone deacetylases and can thus alter gene transcription. Class I and II histone deacetylases are enzymes which catalyze the removal of acetyl groups from lysine residues on histones to reduce gene transcription. Butyrate and propionate have been shown to inhibit histone deacetylases in human colon carcinoma cells, whilst in macrophages the inhibitory effect of butyrate on histone deacetylases is likely to be responsible for its anti-inflammatory properties [76,77,78,79,80,81]. Loomba et al. have reported that patients with advanced fibrosis have increased acetate levels in their fecal samples, whereas patients with mild or moderate NAFLD presented increased levels of butyrate and propionate [20]. However, when circulating SCFAs were measured in cirrhosis patients the serum butyric acid levels inversely correlated to the inflammatory markers and serum endotoxin levels [76,77,78,79,80]. These differences may be attributed to a plethora of factors, such as variations in age, diet, environmental parameters, or even sample handling and processing. More specifically, the estimation of serum or fecal SCFA levels is troublesome per se as SCFAs are volatile substances, requiring immediate processing for an accurate measurement. However, SCFA supplementation in mouse models of NAFLD has shown beneficial effects. In High Fat Diet (HFD)-fed rodents, supplementation with butyrate resulted in a reduction in liver and adipose tissue inflammation. Besides, butyrate promoted alterations in the bacterial population of the gut microbiota. In particular, it enhanced SCFA-producing bacteria and reduced endotoxin-releasing bacteria [82]. Based on these animal model data, SCFA supplementation may have beneficial metabolic effects as well as decrease the severity of liver steatosis. Large-scale studies of SCFA supplementation among patients with NAFLD are lacking. It would be really intriguing to investigate SCFA supplementation in RCTs in humans.

Inflammation is a hallmark feature of NASH, in which gut microbiota play a pivotal role [83]. Bacterial products stemming from gut microbiota, such as LPS, peptidoglycan, and bacterial DNA, may be transferred via the portal vein to activate the TLRs on Kupffer cells, leading to an inflammatory cascade which promotes the development of NASH. Elevated levels of LPS have been detected in NAFLD in rodent (rats and mice) and human studies [84,85,86]. Pathogen-associated molecular pattern (PAMP) receptors, such as TLRs, are deeply implicated in the pathogenesis of NASH by the activating of NF-κB, inducing the secretion of chemokines from macrophages, and the recruiting of Kupffer cells to the liver to promote the inflammation process [84,85,86,87,88,89,90]. In addition, Nod-like receptor protein (NLRP)-3 may stimulate immunity by means of forming an inflammasome with ACS (the adaptor molecule apoptosis-associated speck-like protein containing a CARD), an apoptosis-associated protein, in order to activate pro-caspase 1 [91]. Inflammasome dysfunction leads to exaggerated liver inflammatory response, liver fibrosis, and cell death [91]. This role of NLRP3 has been documented in HFD-fed rodents, which exhibited decreased liver steatosis by inhibition of the NLRP3 inflammasome pathway [92]. Several TLRs have been shown to be of key importance, with the most significant being TLR-4 and TLR-9. For example, rodents deficient in TLR4 and myeloid-differentiation factor-2 (MD2) are protected from methionine- and choline-deficient diet-induced liver inflammation and liver steatosis [93]. Furthermore, plasma from patients with NASH has been found to possess increased levels of mitochondrial DNA as a potent TLR-9 activator. Mice deficient in TLR9 have been documented to be protected from HFD-induced liver steatosis and inflammation, thus pointing out the importance of the TLR-9 pathway in modulating inflammation in NASH [91,92,93]. Lastly, TLR-5 may play a protective role in diet-induced NASH, as mice lacking TLR-5 on hepatocytes showed exacerbated disease after being fed with a methionine- and choline-deficient diet [94]. These examples help clarify how PAMPs may provoke inflammation in the liver and suggest an interplay of bacteria and gut microbiota in the pathogenesis of NASH.

Bile acids (BAs) are mainly produced by cholesterol in the liver. They are categorized as primary BAs, such as cholic acid (CA) and chenodeoxycholic acid (CDCA), and secondary BAs, such as deoxycholic and lithocholic acid [95,96]. Primary BAs are further conjugated with glycine or taurine and stored in the gallbladder before being released into the intestine after consumption of a meal. In the gut, BAs are implicated in the absorption of dietary fat, cholesterol, and fat-soluble vitamins [97,98]. The primary BAs are deconjugated and dehydroxylated by gut microbiota to the more hydrophobic secondary BAs, which are reabsorbed in the distal ileum and returned to the liver by means of the portal vein [97,98,99]. There are several studies in favor of the notion that there are specific BA profiles related to NASH [100,101]. For example, Yara et al. have analyzed serum BAs and have documented that the ratios of primary to secondary BAs, taurine-conjugated BAs to glycine-conjugated BAs, unconjugated BAs to total BAs, and secondary BAs to total BAs are reduced in NASH patients compared to those of healthy individuals [100]. Moreover, Chen et al. have documented that increased ratios of circulating conjugated Chenodeoxycholic acids (CDCAs) to muricholic acids in NASH patients are correlated to the histological severity of NASH and the grade of fibrosis [101]. In addition, a BA intermediate and marker for de novo BAs synthesis, 7α-hydroxy-4-cholesten-3-one (C4), has been shown to be increased in the serum of patients with NASH and has also been related to changes in the gut microbiota [102].

BAs, especially secondary ones, may serve as signaling molecules by binding to cellular receptors, such as the FXR and the G protein-coupled bile acid receptor 1 (also known as TGR5). Different BAs possess variable abilities to activate these receptors [103]. For example, secondary BAs are more potent than primary BAs in activating TGR5 [104]. TGR5 is ubiquitously expressed throughout the human body with increased levels of TGR5 mRNA detected in metabolically active organs, such as the small intestine, stomach, and liver. Activating TGR5 has been shown to increase the intestinal GLP-1 release in obese animal models [104,105]. TGR5 has also been documented to be expressed in monocytes, macrophages, and Kupffer cells, modulating immune responses [104,105,106]. Indeed, in isolated Kupffer cells, bile acids activated TGR5 and inhibited LPS-induced cytokine expression in a cAMP-dependent manner [104,105,106]. However, there is a scarcity of studies in animal models as well as in humans regarding the role of BAs in the pathogenesis of NAFLD in animal models as well as in humans.

Choline is mainly obtained from dietary sources, such as red meat, eggs, cheese, and peanuts, although de novo choline synthesis may also take place in the liver [107]. Choline is a component of the cell membrane, necessary for the production of phosphatidylcholine and sphingomyelin, which are indispensable structural and functional membrane phospholipid components. In the liver, choline is also necessary for the production of VLDL. Therefore, choline deficiency may lead to a reduced production of VLDL, resulting in the accumulation of TG in the liver [108,109]. For this reason, choline-deficient diets have been used in animal models to induce NASH [110]. Choline is known to be converted to trimethylamine (TMA) by gut microbiota. TMA can be oxidized by hepatic monooxygenases to produce trimethylamine N-oxide (TMAO) in the liver, which is afterwards released in the systemic circulation [107]. In HFD-fed rodents, higher conversion of choline to TMA by microbiota resulted in lower bioavailability of choline [111]. TMAO may also act directly on the liver and contribute to impaired glucose tolerance and the development of NAFLD [7,112]. In particular, studies have demonstrated that serum levels of TMAO are higher in patients with NAFLD than in healthy controls, being also correlated with the severity of hepatic steatosis [113]. Another study reported that elevated serum TMAO levels are related to NASH in patients with T2DM [114]. TMAO levels have also been associated with atherosclerosis via the increased production of foam cells in animal models. In particular, TMAO has been demonstrated to promote macrophage migration and their transformation into foam cells, while endothelial dysfunction as well as platelets dysfunction and thrombus formation have been correlated to increased TMAO levels. Based on these mechanisms, high blood TMAO has been suggested as a contributor to the increased cardiovascular disease risk. However, it remains unknown whether serum TMAO levels may serve as a biomarker for NAFLD prognosis or other metabolic derangements or whether it may function as a potential therapeutic target for atherosclerosis [115].

Hyperammonemia is suggested to be a marker of the severity of liver disease [116]. During NASH, ornithine transcarbamylase and carbamoyl phosphate synthetase mRNA, protein, and activity have been shown to be decreased, leading to hyperammonemia [117]. Notably, ammonia itself has been suggested to exert direct effects on hepatic stellate cells by activating them in cell culture as well as in vivo [118]. The above-mentioned findings suggest that hyperammonemia during NASH and cirrhosis may itself promote fibrosis. Ammonia is also generated from amino acids in the gut by bacteria [119]. Therefore, the composition of the gut microbiota contributes to the circulating ammonia levels. However, the exact amount of ammonia produced by gut microbiota and their role in determining serum ammonia levels in NASH and cirrhosis are not well known.

There is accumulating evidence that NAFLD is related to suboptimal liver function, even during the early stages of the disease [120,121,122,123]. Urea synthesis occurs exclusively in the liver, which is the primary location for waste nitrogen, i.e., ammonia elimination, by converting excess amino-nitrogen to urea [124]. This cycle is significantly impaired in patients with cirrhosis due to the loss of function of the hepatic cells, resulting in the accumulation of ammonia [124,125]. However, this cycle appears to be impaired even in the pre-cirrhotic stages of NAFLD. It has been documented—both in animal models as well as in humans with NAFLD—that there is a decrease in the ability for urea synthesis, together with the expression and function of urea cycle enzymes, even at the stage of simple steatosis without fibrosis (F0), which leads to a reduced ammonia elimination and thus hyperammonemia even at a non-cirrhotic stage [117,118,126,127]. Furthermore, dietary intervention resulted in the restoration of the normal urea cycle enzyme activity in animal models of NASH, which was further related to a significant reduction in liver fat content [116,128]. There is mounting evidence arguing that the presence of steatosis has a detrimental effect on mitochondrial liver function, including the urea cycle, while in vitro studies have demonstrated that the accumulation of lipids in hepatocytes leads to reduced expression of the urea cycle enzymes and thus hyperammonemia [116,118,129,130]. Therefore, steatosis in early NAFLD seems to be the cause of dysfunction in the urea cycle rather than just a coincidental finding [131]. Ammonia is a neurotoxic molecule that easily crosses the blood brain barrier and plays a key role in the pathogenesis of hepatic encephalopathy [132]. However, the pathogenesis of hepatic encephalopathy is complicated, and whilst ammonia undoubtedly plays a key role, it is not solely responsible for the neurocognitive dysfunction in hepatic encephalopathy [133]. Many studies have documented that hepatic encephalopathy manifestations may be exacerbated in an inflammatory milieu [134,135,136]. Therefore, systemic inflammation acts together with the dysfunctional nitrogen metabolism in patients with progressive liver dysfunction. It is widely accepted that hepatic encephalopathy represents a primary gliopathy, which results from astrocyte swelling and oxidative stress. Even in the absence of clinically overt hepatic encephalopathy, low-grade astrocyte swelling, which may be present in NAFLD, could impair the crosstalk between neurons and swollen astrocytes [137,138,139]. Neuroinflammation is now considered a well-established feature of hepatic encephalopathy [140,141,142,143,144,145]. It is suggested that neuroinflammation caused by hyperammonemia could be reversible by decreasing systemic ammonia levels or by anti-inflammatory treatment [146].

A low-fat, low-carbohydrate diet with a caloric restriction (500–1000 kcal/day deficit to provoke a weight reduction of 0.5–1.0 kg/week) is generally recommended. Steatosis may be ameliorated by a 5% body weight loss, while a 7–10% is required for improving inflammation and fibrosis [148].

High-fat and high-cholesterol diets have been linked to hepatic steatosis, inflammation, and liver fibrosis [149]. Chronic administration of HFD in rodents has been associated with relatively increased numbers of Firmicutes and decreased numbers of Bacterioidetes species, resulting in a higher F/B ratio [150]. On the other hand, a high-fiber diet has been protective against hepatic inflammation and has been linked to the increased presence of Akkermansia muciniphila [151].

Fructose has been associated with dyslipidemia and insulin resistance. A study comparing fructose and glucose consumption has documented increased visceral adipose tissue volume as well as hepatic de novo lipogenesis, which has been approximately 3-fold higher in patients with NAFLD [152,153,154]. Glycotoxin levels, also known as advanced glycation end-products (AGEs), are very high in baked and fried food, especially under high-temperature cooking [155,156]. Although there is a scarcity of studies regarding AGEs and NAFLD, an animal study has reported that a high-AGEs diet provoked liver steatosis and fibrosis [157]. On the contrary, caffeine consumption has been demonstrated to be protective against the development of NAFLD as well as the progression of the stage of liver fibrosis [158]. The potential mechanism of the action of caffeine is the decrease in fat accumulation, liver inflammation, and oxidative stress as caffeine may increase glutathione and serve as an ROS scavenger at the cellular level [159]. It is also involved in restoring the F/B ratio [160]. In addition, another plausible mechanism of caffeine action is the up-regulation of the expression of Aquaporin-8 in the proximal colon, which is associated with increased growth of the Bifidobacterium species [161]. Moreover, coffee consumption has been shown to correlate with an amelioration of liver enzymes and a decreased risk of cirrhosis and HCC, as well as mortality rates, in a dose-dependent manner [162,163,164]. Besides, increased coffee intake has been associated with alterations in the gut microbiota composition; in particular, increased levels of Bacteroides, Prevotella, and Porphyromonas have been reported among high consumers of coffee [162,163,164].

Green tea extract in conjunction with epigallocatechin-3-gallate has been associated with improvements in several obesity parameters, presumably due to the restoration of the F/B ratio and the Bacteroides-to-Prevotella ratio [165,166,167]. It is noteworthy that patients with NAFLD receiving green tea extract together with 2.5% caffeine have shown remarkable improvements in liver enzymes after 12 weeks of administration [168]. Another study has confirmed the abovementioned results by reporting improvements in liver enzymes together with a significant decrease in the proportion of body fat in NAFLD patients [169]. In addition, green tea liquid consumption has also been documented to alter the human gut microbiome in many studies [169].

Omega-3 polyunsaturated fatty acids (PUFAs) are suggested to be another promising dietary agent in combating NAFLD. Omega-3 PUFAs regulate the peroxisome proliferator activated receptors (PPARs) and decrease proinflammatory cytokines as well as oxidative stress. In a recent meta-analysis, it has been suggested that omega-3 PUFAs could result in improvements in liver enzymes and the lipid profile [170]. Foods rich in omega-3 PUFAs are found in excess in the Mediterranean diet, which is well known for its beneficial effects in preventing obesity and cardiovascular diseases [170,171].

Exercise, both aerobic and resistance training, is one of the most promising lifestyle modifications and has been proven to prevent or even reverse NAFLD/NASH [172,173,174]. As physical inactivity has been associated with the progression of the disease, exercise is suggested to be beneficial not only because of the weight loss but, above and beyond this, also because of the metabolic benefits which accompany exercise training [175,176]. In particular, exercise has been implicated in the increased production of several angiogenic factors, thus favoring fatty acid utilization and leading to a decreased entry of FFA into the liver [175,176,177,178]. Furthermore, alterations in the composition of the gut microbiota have been reported in response to exercise training. More specifically, in animal models, exercise has been associated with a reduction in the presence of Parabacteroides, Flavobacterium, and Alkaliphilus [179]. Besides, exercise has been related to a relatively increased presence of Verrucomicrobia and a reduction in Proteobacteria in overweight women, as well as a relative reduction in the F/B ratio in patients with T2DM [180]. Alterations that have been associated with exercise also include a reduced presence of the mouse-associated Bacteroidales S24-7 and Rikenellaceae families, which have already been related to increased intrahepatic TG content in mouse models [181]. The significant effect of exercise in NAFLD patients could be partially explained by the interplay between the gut and the liver. For example, exercise has been proven to increase the abundance of SCFAs, mainly butyrate [182]. In addition, exercise exerts a beneficial effect on other gut-derived metabolites related to hepatic metabolism, such as BAs and choline. Furthermore, exercise has been documented to exert epigenetic changes by modulating the expression of various genes implicated in lipid metabolism, such as SREBP-1c, FAT/Cd36, and C/EBPa [175,176,177,178]. Exercise provides more beneficial effects than diet alone. Notably, in a study comparing exercise to calorie restriction in HFD-fed animal models, exercise has been shown to increase insulin sensitivity and result in greater LDL reduction, mainly on account of the exercise-induced microbiome alterations [180]. Such modifications in the gut microbiota have been related to improvements in serum LDL cholesterol, liver fat mass, and liver TG [180]. Notably, Barton et al. have documented an increase in the abundance of Akkermansia spp. among athletes with a more active exercise training regimen than among more sedentary individuals, while Allen et al. have reported an increase in the presence of Faecalibacterium spp. with exercise among lean individuals when compared to patients with obesity [183,184]. Interestingly, both Akkermansia and Faecalibacterium are known for their beneficial effects on health [185].

Bariatric surgery is a successful therapy for severe obesity and related metabolic disorders, leading to permanent weight loss and amelioration of metabolic and inflammatory comorbidities such as NAFLD [186,187]. The most widely performed types of bariatric surgery are laparoscopic Roux-en-Y gastric bypass (RYGB) and laparoscopic sleeve gastrectomy (SG). After bariatric surgery, the anatomy, physiology, and energy and substrate metabolism adapt to a new and dynamic status [188,189]. The specific mechanisms through which bariatric surgery affects the composition of the gut microbiota remain to be elucidated. However, several factors have been postulated as being involved in the alteration of the gut microbiota, such as changes in dietary habits, gastrointestinal anatomy, nutrients as well as gastrointestinal transit time and pH, and bile acid metabolism [189]. Studies have documented that after bariatric surgery there is a mild increase in microbial gene richness but not to the levels seen in lean humans [7,190]. There are numerous studies which have investigated the influence of RYGB on the gut microbiota. For example, in a study of 16 patients who underwent RYGB, the gut microbiome was analyzed before and 3 months following surgery. Before surgery, there was relative abundance in the phyla Firmicutes and Actinobacteria, whereas Verrucomicrobia were less abundant when compared to the lean controls. After RYGB, the abundance of these phyla was approximately similar to that in the healthy controls. Only Proteobacteria abundance was more prominent after RYGB and lower in lean subjects [191]. At the genus level, Blautia, Roseburia, Faecalibacterium (Firmicutes), and Bifidobacterium (Actinobacteria) were reduced. It was noteworthy that these genera were still more abundant when compared to lean subjects [184,185]. At the species level, Streptococcus spp., Akkermansia muciniphila (Verrucomicrobia), Roseburia feces, Roseburia hominis, and Enterococcus faecalis were in relative abundance, whereas Faecalibacterium prausnitzii was reduced after RYGB [192,193,194,195,196,197,198,199]. Other studies analyzing the composition of the gut microbiota after SG have documented an abundance of the species C. comes, D. longicatena, Clostridiales bacterium, Anaerotruncus colihominis, Akkermansia muciniphila, and B. thetaiotaomicron [191,192,193,194,195,196,197,198]. Moreover, the abundance of the species Akkermansia muciniphila, Roseburia spp., Bacteroidetes, and Bifidobacterium have been demonstrated to be increased. One year after SG, the abundance of the phylum Actinobacteria was increased compared to the baseline and three months postoperatively [199]. Overall, several studies have reported substantial alterations in the composition of the gut microbiota after bariatric surgery. Different interventions have resulted in differential microbial profiles, whilst only partial restoration towards the lean gut microbiota composition was noted. Different methods of bariatric surgery, patient characteristics, differences in sample sizes and methodologies as well as the existence of comorbidities, such as T2DM, could account for the abovementioned variable results [192,193,194,195,196,197,198,199].

Bariatric surgery leads to prolonged weight loss with subsequent reductions in hepatic fat mass, inflammation, and fibrosis [186,187,188]. Indeed, a meta-analysis of 43 studies with 2809 participants has reported a dose-response association with liver inflammation, ballooning, and NAFLD/NASH resolution [186,187,188,189]. There is ongoing interest and research on this issue and the results of the ongoing trials are eagerly anticipated.

As gut dysbiosis appears to contribute to the pathogenesis of NAFLD, probiotics have been tested in the prevention and treatment of NAFLD. The term “probiotics” first appeared in 1974 but has since evolved to its current definition as “live microorganisms that confer a health benefit when consumed in adequate amounts,” as proposed by the World Health Organization (WHO) in 2002 [200]. To date, probiotics have been shown to improve the lipid profile as well as liver function tests of patients with NAFLD; however, data regarding hepatic histologic alterations are inconclusive. Trials with the administration of probiotics have been confounded by dietary factors which affect intestinal microbiota composition, different formulas, and dosages.

Bifidobacterium and Lactobacillus strains are still the most widely used probiotics in functional foods and dietary supplements, but next-generation probiotics, such as Faecalibacterium prausnitzii, Akkermansia muciniphila, or Clostridia strains, have demonstrated promising results [201]. VSL#3 and modified VSL#3 are mixtures of probiotics of the genera Lactobacillus, Bifidobacterium, and Streptococcus or Lactobacillus alone. VSL#3 has been documented to have a protective effect against NAFLD by inhibiting inflammatory pathways, such as c-Jun N-terminal kinase (JNK) and NF-κB, and restoring the number of natural killer (NK) T cells in the liver, caused by high-fat feeding [202,203]. On the contrary, in another animal model study, whilst VSL#3 did not have any effects on methionine-choline-supplemented diet-induced hepatic steatosis or inflammation, it improved liver fibrosis by down-regulating TGF-β signaling [204]. Other studies in rodents have found that probiotics improved the gut microbiota composition by maintaining tight junctions, thus restoring the intestinal mucosal barrier and suppressing the serum LPS levels. Liver inflammatory markers as well as serum cytokines levels were reported to be decreased, in parallel with a reduction in the serum LPS and liver TLR-4 mRNA concentrations [204,205]. In addition, Lactobacillus plantarum NA136 has been shown to reduce the mass of fat tissues in HFD-fed and fructose-fed animal models of NAFLD and to decrease serum liver enzymes. Besides, L. plantarum NA136 reduced lipogenesis and enhanced fatty acid oxidation by stimulating the AMPK pathway to phosphorylate ACC and suppress SREBP-1/FAS signaling in a NASH animal model. Moreover, L. plantarum NA136 reduced oxidative stress in the liver by stimulating the AMPK/NF-E2-related factor 2 (Nrf2) signaling in a NAFLD animal model. These results point towards a promising role of L. plantarum NA136 in improving NAFLD [204,205]. Lactobacillus paracasei has been found to reduce the expression of TLR-4, CCL2, and TNF-α and improve liver steatosis. In particular, L. paracasei has been documented to reduce the proportion of M1 Kupffer cells and increase that of M2, leading to an M2-dominant shift in the liver in a NASH animal model, a shift similar to that ascertained in animal models of obesity [204,205]. The addition of a combination of three probiotics has resulted in an enhanced production of lactate concentrations, which in turn led to the activation of lactate-consuming bacteria growth, resulting in a significant enhancement in SCFAs production. Furthermore, the abovementioned combination of probiotics has shown immunomodulatory potential, including enhanced production of anti-inflammatory cytokines, i.e., IL-10 and IL-6, and reduction in the production of proinflammatory chemokines, i.e., IL-8, CXCL 10, and Monocyte Chemoattractant Protein 1 (MCP)-1 [204,205]. However, the data regarding combinations of probiotics are difficult to interpret, as different dosages and proportions of different probiotics are used, while there are other factors to be taken in account, such as the duration of treatment as well as the timing of data collection after the discontinuation of treatment.

An animal study has found that Faecalibacterium prausnitzii may reduce adipose tissue inflammation in rodents and improve the metabolic parameters of liver function [206,207]. Nevertheless, this bacterial species is very oxygen-sensitive and difficult in cultivation and preservation; more research to improve its preservation and viability with the addition of antioxidants, such as riboflavin, cysteine, and cryoprotectant inulin, is ongoing. Another potential probiotic is Akkermansia muciniphila, a mucin-degrading bacterium, which has been found to reduce fat mass and ameliorate dyslipidemia in animal models [208]. Several meta-analyses have also advocated the use of probiotics in NAFLD, documenting improvements in various metabolic and inflammatory parameters. The combined use of Akkermansia muciniphila with metformin has been shown to induce better results than monotherapy alone with regard to serum levels of liver enzymes and sonographic characteristics [208,209,210,211,212,213,214].

Prebiotics constitute non-digestible carbohydrates, which lead to significant changes in the composition or activity of the gut microbiota, resulting in beneficial effects on host health [200]. Most data refer mainly to two chemical substances, inulin-type fructans and galactooligosaccharides. These data have reported increases in the production of Bifidobacteria and Lactobacilli, thus conferring beneficial alterations in the composition and activity of the gut microbiota [200]. Prebiotics may confer benefits to both NAFLD and NASH. In rodents, prebiotics have been documented to induce changes in the composition of the gut microbiota, resulting in increased plasma glucagon-like peptide-2 (GLP-2) levels, thus ameliorating the function of the gut barrier. Furthermore, prebiotics have been demonstrated to decrease liver inflammation and improve metabolic parameters in obesity [210]. Besides, prebiotics, mainly inulin and oligofructose, have been found to better control the growth of Faecalibacterium prausnitzii and Bifidobacterium as well as decrease serum LPS concentrations, due to increased GLP-1 production and the GLP-2-mediated trophic effect on the integrity of the gut barrier. Among patients with obesity, supplementation with oligofructose induced weight loss by regulating appetite hormones such as ghrelin and peptide YY [215].

Synbiotics are a combination of prebiotics and probiotics, which may be used to replace a dysfunctional gut microbiota [200]. A meta-analysis of probiotics, prebiotics, and synbiotics supplementation in NAFLD has confirmed significant decreases in the body mass index as well as improvements in liver enzymes. Probiotics/synbiotics use has been related to a significant decrease in alanine aminotransferase (ALT) levels and in liver stiffness assessed by elastography [215].

Antibiotics have also been suggested as one of the therapeutic options in NAFLD. Nevertheless, there is conflicting evidence about the efficacy of antimicrobial treatments. The use of norfloxacin and neomycin has been shown to improve liver function in cirrhotic patients by causing changes in bacterial translocation and overgrowth [216]. On the contrary, another study has not shown any beneficial effects on liver function in NAFLD patients administered with norfloxacin. Gangarapu et al. also showed that rifaximin treatment significantly reduced proinflammatory cytokines, ALT, and the NAFLD-liver fat score. This improvement by antibiotics was attributed to alterations in the gut microbiota population and bile acid metabolism as well as to reduced FXR signaling and decreased ceramide levels in the liver [217]. Hence, there is no established role for antibiotics in NAFLD, although longitudinal large-scale studies are lacking. The risk for bacterial drug resistance and potential side effects should also not be overlooked [185].

FMT has not been very widely used, apart from special circumstances, such as the treatment of severe/recurrent Clostridioides difficile infection, and to a lesser extent in inflammatory bowel disease, metabolic syndrome, and hepatic encephalopathy [141,218]. There is a scarcity of studies regarding the use of FMT in NAFLD. Only recently, a study in animal models has documented a beneficial effect on lipid accumulation in the liver as well as liver histology after FMT [219]. Nowadays, FMT has been used successfully in cirrhotic patients with hepatic encephalopathy and severe alcoholic hepatitis [219,220]. However, the success rates of FMT are mainly dependent on the “donor” characteristics, in particular fecal microbiota richness, diversity, and compatibility [221,222,223]. Further studies are needed to assess the long-term efficacy together with the safety of FMT in patients with NAFLD.

Overall, weight loss of 7–10% of total body weight in patients with biopsy-proven NASH remains the cornerstone of the treatment of NAFLD/NASH. Diet, physical exercise, and bariatric surgery have yielded impressive results, while the use of other conventional measures, such as vitamin E supplementation, the addition of pioglitazone—which may cause weight gain but has been documented to decrease inflammation and liver fibrosis—or glucagon like peptide (GLP) analogs seem to exert hepato-protective properties [8].

As most bacteriophages and archaeal viruses are specific to bacterial and archaeal strains, respectively, they may be used to target dysbiotic parts of the microbiota in patients with metabolic disorders in the near future [224]. The therapeutic strategy of targeting a single strain, specifically cytolytisin-positive E. faecalis, with the advent of a bacteriophage has only recently been found to be effective in ethanol-induced liver disease experiments using humanized mice [224]. More specifically, patients with alcoholic hepatitis have increased numbers of E. faecalis in their feces. The presence of this cytolytic E. faecalis has been shown to relate to the severity of liver disease as well as mortality among these patients. Using humanized mice that were colonized with bacteria from the feces of patients with alcoholic hepatitis, Duan et al. embarked on investigating the therapeutic effects of bacteriophages that target cytolytic E. faecalis. They have documented that these bacteriophages decrease cytolysin in the liver and mitigate ethanol-induced liver disease in humanized mice. This method with the advent of bacteriophages raises the possibility of modulating the gut microbiota [225,226,227]. Clinical trials are further required to assess the relevance of the abovementioned findings in humans and to test whether this therapeutic approach is effective for patients with NAFLD/NASH.

Not applicable.