The Role of Probiotics in Nonalcoholic Fatty Liver Disease: A New Insight into Therapeutic Strategies

Several lines of evidence indicate that intestinal flora composition and function play a paramount role in the development of obesity and NAFLD [37] and preclinical models are particularly worthwhile in the understanding of the implications of enteric dysbiosis and bacterial overgrowth in the pathophysiology of these diseases. Indeed, the unbalanced intestinal flora may exert a detrimental effect on the host nutrient metabolism thus facilitating overweight and fatty liver onset. In keeping with this finding, germ-free mice are protected against diet-induced obesity and display less severe fat accumulation into the liver upon Western diet, supporting the crucial role of intestinal flora in NAFLD pathogenesis [37,38]. In these mice, the protection against obesity and hepatic steatosis may be possibly explained by higher circulating levels of Angiopoietin-like 4 (ANGPTL4), a serum hormone that impaired fat storage in adipocytes, muscles and heart and by the ability of microbiota to cleave and ferments complex dietary plant polysaccharides [39]. In physiological conditions, indeed, SCFAs and monosaccharides, resulting from polysaccharides digestion, are absorbed by the host and delivered to the liver where they are converted to complex lipids. Therefore, the over-representation of specific classes of bacteria facilitates the catabolism of absorbed nutrients and through the regulation of the expression of host genes, it promotes energy harvest and storage in adipocytes favoring the progressive development of obesity and hepatic steatosis [25]. According to this notion, microbiota transplantation from mice with diet-induced obesity to lean germ-free recipients promoted greater fat deposition compared to mice transplanted intestinal flora isolated from lean donors, suggesting that obesity-associated gut microbiota is responsible for the transmission of the ability to promote fat storage [25]. Furthermore, maternal obesity exacerbated the risk of hepatic disorders onset in the offspring [40]. Indeed, germ-free mice colonized with stool microbes isolated from 2-week-old infants born to obese mothers displayed endothelial reticulum (ER) stress, activation of innate immunity and periportal inflammation, recapitulating the histological pattern of the childhood NAFLD [40]. The exposure of these mice to a Western diet promoted an excessive weight gain, further precipitating NAFLD onset [40]. Finally, germ-free mice were also reported to be resistant to high-fat diet (HFD)-induced IR, showing enhanced hepatic and adipose tissue insulin sensitivity and fecal lipid excretion [41], revealing that insulin resistance index (HOMA-IR) is transmissible [38].

Multiple preclinical models are currently exploited to study NAFLD development and progression to advanced stages of liver diseases, resembling human hepatic lesions [42] and recently, these models have been pointed out to deeply investigate the alterations of gut microbiota composition in the context of liver injury.

An unbalance of bacterial species was reported in Leptin deficient mice (Lep^ob/ob), a genetic model of obesity, IR and NAFLD, which had enhanced Firmicutes and lower Bacteroides levels, alterations that have been associated with obesity and subsequent chronic liver diseases [24,25]. Indeed, DNA sequencing of cecal microbiota of Lep^ob/ob indicated that the obesity-associated gut microbiome had an increased capacity for fermenting polysaccharides respect to the lean-associated one, due to the enhanced prevalence of Firmicutes [24].

As well as Lep^ob/ob mice, even HFD-fed mice carried a peculiar gut microbiota signature, which markedly impacted on obesity, IR and lipid metabolism in the liver [38]. Indeed, Le Roy and coworker demonstrated that gut microbiota play a causative role in the susceptibility to develop NAFLD features including hyperglycemia, IR and steatosis, in response to the HFD challenge and that the propensity to develop NAFLD is transmissible by means gut microbiota transplantation [38]. The authors revealed that alterations in taxonomic composition, such as decreased quantity of Bacteroidetes and increased levels of Firmicutes, were responsible of NAFLD development, similarly to what occurs in obesity [25]. Specifically, Barnesiella intestinihominis, which was previously related to increased hepatic steatosis and inflammation [43], was found increased in HFD mice, whereas Bacteroides vulgatus was reduced, as well as in patients affected by diabetes [44]. Nonetheless, the suppression of intestinal flora via chronic oral administration of antibiotics attenuated hepatic inflammation and fibrosis in HFD mice, as a result of the decrease in portal secondary bile acids, supporting the notion that a causal link between gut microbiota and liver damage exists [45].

Derangement in gut microbiota composition was even described in mice fed methionine-choline deficient diet (MCD), a dietary model to study NAFLD/NASH in absence of obesity and IR, specifically, harboring a marked decrease in the abundance of Alistipes and the (Eubacterium) coprostanoligenes group and a parallel increase in Ruminococcaceae [46].

Gut dysbiosis has been also causally linked to the pathogenesis of cirrhosis and progression to end-stage liver disease [47]. Depletion of host microflora after gut sterilization could suppress tumor formation, reducing impressively size and number of nodules in diethylnitrosamine (DEN)-induced HCC [48]. In line with these findings, Dapito and colleagues reported that mice grown in germ-free conditions developed smaller and fewer HCC, and treatment with low dose of endotoxins reverted this situation [49]. These studies demonstrated that gut microbiota and Toll-like receptor 4 (TLR4) are required for the tumorigenesis promotion, mediating proliferation and prevention of apoptosis [48,49].

An increasing number of studies have demonstrated that bacterial overgrowth may adversely impact metabolic processes and immune responses, favoring obesity and obesity-related comorbidities, including NAFLD and IR [50]. However, the precise characterization of dysbiosis in the whole spectrum of NAFLD lesions has still been unexplored. In 35 consecutive patients with biopsy-proven NAFLD, Miele et al. demonstrated that NAFLD patients had a significantly increased gut permeability compared to healthy subjects and the prevalence of small intestinal bacterial overgrowth strictly correlated with the severity of steatosis, but not with lobular inflammation, ballooning and fibrosis [51]. Even more, patients affected by NASH displayed intestinal bacterial overgrowth, as assessed by the (14)C-D-xylose-lactulose breath test, increased endotoxins and inflammatory cytokines into the blood circulation [52]. Thus, the degree of NAFLD is correlated to dysbiosis and to modifications of metabolic properties of intestinal flora [53].

The main bacterial composition modifications observed in NAFLD patients are represented by an enrichment in Proteobacteria, Enterobacteriaceae, Lachnospiraceae, Escherichia and Bacteroidetes. However, there are several discrepancies in the proportion of the latters between the studies and the results are heterogeneous, mainly due to the presence of obesity and metabolic syndrome as confounders [54,55,56].

An unbalance in the ratio between Bacteroidetes and Firmicutes has been reported by Zhu and colleagues in fecal samples of obese and NASH children [57]. In particular, they assessed the composition of gut bacterial communities of 22 biopsy-proven NASH children, of 25 obese subjects and 16 healthy controls by 16S ribosomal RNA pyrosequencing and they revealed an enhanced abundance of Bacteroidetes and a decreased number of Firmicutes in fecal samples of obese and NASH children. Even the levels of Actinobacteria were reduced in NASH individuals, conversely the quantity of Proteobacteria rose progressively from healthy to obese to NASH patients [57]. The main finding of this research is the presence of elevated blood levels of alcohol and the highest activities of alcohol dehydrogenases (ADHs) only in NASH children, due to the increased concentration of ethanol-producing bacteria from carbohydrate catabolism such as Escherichia coli. In physiological conditions, indeed, endogenous alcohol is constantly produced by the intestinal microbiota and rapidly removed from portal blood by hepatic ADHs, catalases and microsomal ethanol-oxidizing system [58,59]. In NASH-induced dysbiosis, conversely, the over-representation of alcohol-producing bacteria determined an exaggerated release of ethanol into the blood flow, further corroborating liver inflammation, reactive oxygen species (ROS) production via the Cytochrome P450 2E1 (CYP2E1) [57,60] and intestinal hyperpermeability [57].

Fecal dysbiosis and decreased quantity of Firmicutes has been also observed by Wong et al. [55] in 16 NASH patients compared to 22 controls. The authors showed the presence of lower fecal abundance of Faecalibacterium and Anaerosporobacter in these subjects but higher abundance of Parabacteroides and Allisonella [55]. Moreover, Sobhonslidsuk and coworkers highlighted an increase in the Bacteroidetes/Firmicutes ratio in 16 adult patients affected by NASH independently of age, body mass index (BMI), diabetes and medications [61]. In particular among the Bacteroidetes phylum, the Bacteroides and Prevotella genera are the most abundant in NASH subjects [61]. Conversely, Mouzaki and collaborators revealed a reduction in Bacteroidetes and higher levels of fecal Clostridium coccoides in 22 NASH subjects compared to 17 healthy subjects and 11 simple steatosis, thus facilitating the growth of other bacterial species and the override energy intake from dietary fat [24,25,62]. The variety in the proportion of Bacteroidetes and Firmicutes observed by Zhu and colleagues [57] and Mouzaki et al. [54] may reflect the diversity in age, BMI, environmental and dietary factors of the two study cohorts. Therefore, to rule out the impact of obesity on gut microbiota composition, Wang B. and colleagues sought to identify the variability in fecal microbiota composition between non-obese adult individuals with and without NAFLD (43 NAFLD vs. 83 healthy controls) [63]. They demonstrated that adult non-obese NAFLD patients harbored 20% more phylum Bacteroidetes and 24% less Firmicutes, showing a significant correlation of metabolic markers with the disturbed microbiota in NAFLD [63]. Hence, the prevalence of Firmicutes is considered a fingerprint of obesity-associated NAFLD, whereas the Bacteroides override is related to ‘lean’ NAFLD.

Overwhelming evidence provided by Loomba et al. suggested the presence of gut microbiota-derived signature, which predicts the presence of advanced fibrosis in NAFLD patients [64]. Through a whole-genome shotgun sequencing approach on stool samples, the authors analyzed the bacteria taxonomic composition of 86 biopsy-proven NAFLD of whom 72 had mild fibrosis and 14 had advanced fibrosis (stages 3/4). They identified 37 different bacterial species, which enabled us to distinguish mild and advanced fibrosis in NAFLD patients and they showed that advanced fibrosis is characterized by an exasperation of Proteobacteria and Escherichia coli along with a decrease in Firmicutes [64]. The same research group, in a very recent paper, identified the specific intestinal microflora profile of NAFLD cirrhotic patients, determining a panel of 27 fecal bacteria that may discriminate NAFLD cirrhosis using a random forest classifier model [65]. Besides, Boursier et al. determined the presence of rising quantity of Ruminococcus in NASH patients affected by advanced fibrosis, while Prevotella abundance was decreased [53] and higher counts of Escherichia coli and Staphylococcus have been found in stool samples of subjects with mild encephalopathy and cirrhosis [66]. Quantitative metagenomic analyses identified 75,245 genes that differ cirrhotic patients from healthy individuals [67], showing fewer Bacteroidetes but higher levels of Proteobacteria and Fusobacteria [67]. Schierwagen and colleagues focused on the assessment of the circulating microbiome in the portal vein of seven patients with decompensated cirrhosis, during the implantation of a transjugular intrahepatic portosystemic shunt and they demonstrated that 65 genera belonging to four phyla, predominantly Proteobacteria, were strictly correlated with cytokines secretion [68].

Finally, addressing to fecal microbiota diversity in HCC patients, Ren Z. and coworkers revealed an enrichment in the phylum Actinobacteria and in 13 genera, including Gemmiger and Parabacteroides in fecal samples of 75 HCC patients compared to 40 cirrhotic ones. In particular, butyrate-producing genera were decreased, while lipopolysaccharides (LPS)-producing genera were increased in early stage of HCC [69].

Collectively, these findings suggest that manipulation of intestinal microflora may be a strategy to prevent or treat NAFLD and metabolic syndrome features.

Emerging evidence has suggested that alterations in bile acids metabolism are associated with chronic liver diseases and comorbidities, i.e., development of cholestasis. The liver synthetizes primary bile acids, which first accumulate in the gallbladder and then are released in the duodenum lumen where they are converted in secondary bile acids by the gut microbiota and favor lipid solubilization, emulsification and absorption. In addition, bile acids act as signaling molecules (i.e., deoxycholic acid (DCA) and di-hydroxy chenodeoxycholic acid (CDCA)) as they activate the intestinal Farnesoid X Receptor (FXR), leading to the release of Fibroblast Growth Factor 19 (FGF19), which can modulate gut barrier integrity, and β-Klotho in the bloodstream [10]. Both FGF19 and β-Klotho can downregulate bile acids synthesis in the liver by inhibiting cholesterol 7-α hydroxylase 1 (Cyp7A1) [84]. The enhanced production of bile acids pool can also stimulate Takeda G-protein-coupled receptor 5 (TGR5) to activate the proinflammatory cascade on Kupffer cells surface [85]. Indeed, patients with NASH showed elevated levels of cytotoxic bile acids in the liver, serum, stool and urine, which may worsen liver damage to cirrhosis [86,87,88]. To date, many FXR agonists with hepatoprotective properties (i.e., obeticholic acids) have been proposed as they are able to reduce hepatic steatosis and necroinflammation [88,89]. However, the prolonged exposure to bile acids analogues is associated with severe side effects. Therefore, alternative approaches, such as ω-3 long-chain polyunsaturated fatty acids administration and/or combined to gut microbiota modulation, which may have the advantage to adjust bile acid pool, is still under investigations for the treatment of NAFLD [90,91].

The implication of the gut–liver axis in the susceptibility of NAFLD has been widely investigated in both preclinical and human studies, although it remains a matter of debate. Nevertheless, new insights in the pathogenesis and progression of NAFLD have been attributed to the permeable gut barrier as it constitutes a favorable microenvironment for bacteria overgrowth, promotes endotoxemia and contributes to chronic liver damage in response to endogenous or exogenous cofactors, such as dietary pattern and lifestyle. Indeed, fatty liver is highly prevalent in obese patients and the 20%–30% of pathologically obese individuals show histological signs of necroinflammation and fibrosis [92,93], possibly even for diet-induced derangement of barrier integrity.

It has been demonstrated that the interaction between HFD and enteric bacteria promote intestinal inflammation through TNF-α production and NF-kB activation, a mechanism that precedes IR development in mice [94]. In NAFLD rats feeding high sucrose and high fat (HSHF) diet, Zhou et al. showed that animals exhibited damaged villous of the intestinal epithelium and low-grade inflammatory status due to increasing gut-derived endotoxins and inflammatory cytokines that translocate into the circulation [95]. Interestingly, Brun et al. revealed that leptin-deficient (Lep^ob/ob) and hyperleptinemic (Lep^db/db) obese mice displayed a dysmorphic mucosal barrier as demonstrated by a dramatic ZO-1 and tight junctions redistribution, and a remarkable release of IL1, IL6, TNF-α and Interferon (IFN-γ) in the portal vein [96]. Notably, sodium butyrate supplementation to HFD fed mice improved gut mucosa restoring intestinal ZO-1 levels and favoring abundances of the beneficial bacteria Christensenellaceae, Blautia and Lactobacillus [97]. Butyrate further impacts on liver damage, strongly reducing hepatic fat accumulation as well as markers of inflammation and fibrosis [97]. Nonetheless, HSCs isolated from Lep^ob/ob and Lep^db/db livers chronically exposed to LPS showed a pro-inflammatory and pro-fibrotic phenotype compared to HSCs isolated from lean mice, supporting that increased intestinal permeability may participate to the development and progression of obesity-related NASH [96].

Chronic fructose intake has been associated with loss of tight junction proteins and lower SCFA-producing agents, which foster PAMPs translocation, increase numbers of macrophages in the liver and activate TLR1-9 and myeloid differentiation factor 88 (Myd88)-dependent proinflammatory pathways in mice [98]. Similarly, acute and chronic high fructose consumption exacerbated endotoxemia in pediatric NAFLD and even correlates with markers of IR and liver inflammation [99]. Noteworthy, the G protein-coupled chemokine receptor CX3CR1 protects mice from steatohepatitis induced with HFD or MCD diet as it maintains intestinal homeostasis and barrier integrity [100].

Finally, in a model of early NASH induced with high glucose/fructose diet (HFGFD), the authors observed that rats developed portal hypertension, a severe complication of liver cirrhosis. Hence, they investigated whether enteric dysbiosis could modulate endothelial and hepatic functions. Notably, HFGFD-fed mice were enriched in Firmicutes rather than Bacteroides strains and selectively activated intestinal FXR thus suggesting that changes in intestinal microbiota communities impair bile acid metabolism, which, in turn, may be a driver of NASH-related complications [101].

The aforementioned mechanisms further influence liver damage progression, possibly leading to advanced fibrosis and cirrhosis, the end-stage of chronic liver disease and the leading cause of liver failure and HCC. Activated HSCs show high LPS-induced TLR4 responsiveness thus repeatedly enhancing pro-fibrotic processes that compromise immune defense and toxins clearance from hepatic tissue supporting cirrhosis development [102]. Moreover, LPS/TLR4 signaling plays a critical role to induce hepatocarcinogenesis by promoting a senescence-associated secretory phenotype (SASP) in HSCs and stimulating chemoattractant cytokines production from HSCs and monocytes. Bacterial translocation even participates to hemodynamic complications linked to cirrhosis, such as hepatic encephalopathy, variceal bleeding and portal hypertension. Recently, Sorribas et al. showed that cirrhotic mice reduced mucus thickness and lost Goblet cells as well as mucins expression, allowing bacterial overgrowth and the pathological translocation of Escherichia coli [103]. Assimakopoulos et al. firstly demonstrated that leaky gut exerts a pivotal role in human cirrhosis, highlighting the importance of occludin and claudin 1 downregulation in cirrhotic patients and even more in those with decompensated cirrhosis [104]. In addition, higher levels of IL6, nitric oxide (NO) and decreased transepithelial resistance have been associated with the presence of activated intestinal macrophages in cirrhotic patients [105].

Obesity-induced changes in gut microbiota composition may raise the amount of DCA that has been associated with SASP phenotype in HSCs, which enhance pro-inflammatory cytokines and tumor-promoting factors, as well as DNA damage and ROS production in the liver [106,107]. In an experimental model of HCC induced with DEN, alterations in gut permeability seems to be the primary hit leading to amplified tumorigenic response of the liver to LPS. Indeed, antibiotics regimen or TLR4 ablation mitigated tumor growth and multiplicity in mice [48]. Dapito et al. exposed C3H/HeJ and C3H/HeOuJ mice to a mixture of DEN and hepatotoxin carbon tetrachloride (CCL4), a model that resembles human microenvironment for HCC raise, and demonstrated that intestinal microbiota and TLRs promote liver cancer as a long-term consequence of chronic liver injury [49]. In a most recent study, Ponziani and collaborators investigated the gut microbiome profile and intestinal features of 21 NAFLD patients with both cirrhosis and HCC compared to 20 cirrhotic individuals without HCC and healthy controls. Although they found a similar degree of intestinal barrier dysfunction between HCC and cirrhotic subjects, systemic levels of IL8, IL13, Chemokine C-C motif Ligand 3 (CCL3), CCL4 and CCL5 significantly correlated with circulating activated monocytes in presence of HCC [108].

Numerous studies demonstrated that probiotics administration might attenuate NAFLD features in animal models [112,113,114,115,116]. The administration of VSL#3, a mixture of three genera of bacteria (a multistrain formulation composed by Streptococcus, Thermophilus and several species of Bifidobacterium and Lactobacillus) for 4 weeks to Lep^ob/ob mice improved insulin sensitivity, total fatty acid content, serum alanine aminotransferase (ALT) levels and the histological spectrum of liver damage [112]. These improvements were mainly due to the reduction of Jun N-terminal kinase (JNK) and NF-kB activation and to the decreased expression of Uncoupling protein (UCP)-2 in Lep^ob/ob mice exposed to VSL#3, sustaining the hypothesis that intestinal bacteria may regulate the activation of host signaling pathways interfering with hepatic insulin response and lipid metabolism [112].

Nonetheless, VSL#3 supplementation in HFD-challenged young rats, dampened the production of TNF-α, inducible nitric oxide synthase (iNOS), metalloproteinases (MMP) and Cyclooxygenase 2 (COX-2), as well as it improved lipid peroxidation markers, limiting oxidative and inflammatory damage in the liver [115]. These findings are even supported by Ma X. et al.’s study, in which VSL#3 probiotics exposure ameliorates IR, steatosis and pro-inflammatory cytokines secretion, hampering NF-kB activity in HFD-fed mice [113]. Furthermore, oral Bifidobacterium longum supplementation in HFD-fed rodents for 12 weeks reduced hepatic fat accumulation more than the administration of Lactobacillus acidophilus, irrespectively of intestinal permeability restoration [117]. In addition, Lactobacillus johnsonii BS15 protected HFD mice from hepatic steatosis and hepatocyte apoptosis, exhibiting a positive effect on lipid peroxidation, sustaining the antioxidant defense system and improving mitochondria abnormalities [118]. The restoration of Bifidobacteria or Akkermansia muciniphila along with oligofructose implementation reduced endotoxemia, hepatic fat deposits and metabolic syndrome hallmarks in HFD mice [119]. Moreover, diabetic rats administered with Akkermansia muciniphila improve liver function, reduce gluco/lipotoxicity, alleviate oxidative stress, suppress inflammation and normalize intestine microbiota thereby ameliorating type 2 diabetes mellitus [120]. Similarly, nano-selenium-enriched Bifidobacterium longum delay the onset of streptozotocin-induced diabetes [121].

In a model of inherited dyslipidemia (ApoE-/-mice), the modulation of gut microbiota through VSL#3 mixture rescued hepatic and adipose tissue IR and counteracted atherosclerosis and NAFLD onset [122]. Specifically, VSL#3 reversed IR, prevented development of histologic features of mesenteric adipose tissue inflammation, NASH and reduced the extent of aortic plaques, through the modulation of Peroxisome Proliferator-Activated Receptor-γ (PPAR-γ), FXR and vitamin D receptor [122].

Velayudham and collaborators demonstrated that VSL#3 may even influence fibrosis development in MCD fed mice, favoring the reduction of collagens, MMPs and α-smooth muscle actin (α-SMA), not accompanied by the attenuation of hepatic steatosis and inflammation [116]. In agreement with Velayudham’s study, even Nardone and collaborators revealed the protective effect of Lactobacillus paracasei F19 (LP-F19), in an experimental model of ischemia-riperfusion in rats fed MCD [114]. Likewise, heat killed Lactobacillus reuteri GMLN-263 rescued hepatic and heart fibrosis, resetting the expression of pro-fibrotic markers, such as the transforming growth factor β (TGF-β) in HFD hamsters [123]. Alongside LP-F19 and GMLN-263, many other strains of Lactobacillus, i.e., Lactobacillus reuteri GMLN-13, have been demonstrated to attenuate the harmful impact of gut microbiota derangement in the contest of NAFLD and NAFLD-related comorbidities, such as hypertension, obesity, glucose intolerance, hyperglycemia, hyperinsulinemia, dyslipidemia, adipose tissue inflammation and oxidative stress [124,125,126,127].

In rodents, daily oral VSL#3 administration was next reported to be related to the restoration of enterocyte architecture and of intestinal barrier integrity by inducing mucus secretion, Muc2 colonic expression and ileal occludin levels, and avoiding viable microorganisms and bacterial products translocation into the bloodstream [128]. The expression of tight junction proteins (i.e., occludins, claudins and ZO-1) was forced by the supplementation with Lactobacillus rhamnosus, Lactobacillus paracasei and Bifidobacterium adolescentis [109]. Intriguingly, Clostridium butyricum strain MIYAIRI 588, a butyrate-producing probiotic, prevented the entire pathological spectrum of NAFLD, from steatosis to HCC in a rodent model of choline-deficient/L-amino acid-defined (CDAA)-diet, re-establishing the intestinal barrier integrity [129]. Besides, MIYAIRI 588 positively affects the development of IR, excessive triglyceride storage and attenuates serum endotoxin concentration, hepatic inflammation and oxidative stress [129]. Finally, Lactobacillus johnsonii La1 in combination with antioxidants prevented bacterial translocation and endotoxemia in a rat model of CCl4-induced cirrhosis, enlightening the idea to use probiotics and antioxidants as alternative strategy to antibiotics in the prevention of bacterial infections in cirrhotic patients [130].

Complexively, probiotics supplementation may positively intervene on the liver injuries induction in rodent models of NAFLD although most of the studies are addressed to prevent rather than treat diet-induced liver disease.

Over the year, the experimental models of NAFLD have collected promising findings in the field of microbiome therapeutics thus providing greater awareness for the clinical evaluation of probiotics to overthrow NAFLD development and progression.

Vajro et al. conducted a double-blind study on 20 obese children with ultra-sonographic proved steatosis that were randomized for Lactobacillus rhamnosus GG or placebo for 8 weeks. Despite probiotic administration had no effect on adiposity and fatty liver, it has been suggested as treatment for hypertransaminasemia in obese children noncompliant to lifestyle interventions [131]. Similar findings of reduced ALT and aspartate aminotransferases (AST) have been observed in 30 adults affected by NASH exposed to Lactobacillus acidophilus compared to those receiving placebo [132].

The effects of probiotics as a combination of multistrains compound have reached better outcomes in randomized trials. Famouri et al. carried out a triple-blind trial on 64 obese children with sonographic NAFLD. Adolescents receiving a probiotic capsule of Lactobacillus acidophilus ATCC B3208, Bifidobacterium lactis DSMZ 32,269, Bifidobacterium bifidum ATCC SD6576 and Lactobacillus rhamnosus DSMZ 21,690 for 12 weeks showed significant reduction in ALT, lipid profile and intrahepatic fat content evaluated with ultrasound compared to placebo group. In a randomized placebo-controlled trials (RCT), Kobyliak et al. assessed the efficacy of “Symbiter”, containing 14 alive probiotic strains of Lactobacillus + Lactococcus, Bifidobacterium, Propionibacterium and Acetobacter, in NAFLD patients. The multiprobiotic cocktail ameliorated hepatic steatosis, aminotransferase activity, TNF-α and IL6 levels in patients with NAFLD [133]. Similarly, patients with histology-proven NASH randomly receiving Lepicol probiotic formula for 6 months attenuated intrahepatic triglycerides and reduced serum AST levels [134]. In a meta-analysis conducted by Ma et al., including 134 NAFLD/NASH from four randomized trials, the authors underlined that probiotic therapy with Lactobacillus, Bifidobacterium and Streptococcus beneficially impacts on hepatic fat content, total cholesterol, levels of aminotransferases and HOMA-IR [135]. Most recently, Gao and collaborators evaluated the efficacy of probiotic treatment in both pediatric and adult NAFLD including nine RCT with a total of 535 NAFLD cases [136]. They showed that probiotics improve the clinical outcomes of NAFLD patients, influencing insulin sensitivity and reducing TNF-α. However, probiotics ameliorated dyslipidemia only in Italian and Spanish population, suggesting that the effects of these molecules on high-density lipoprotein (HDL), low-density lipoprotein (LDL) and triglycerides might depend on ethnical background [136].

It has been demonstrated that VSL#3, the most studied multistrains probiotic, protects intestinal barrier integrity, dampens endotoxemia and oxidative/nitrosative stress thus favoring improvement in liver pathology in patients with different chronic liver diseases, among which 20 with ALD and 22 with NAFLD [10,137]. In a double-blind RCT (NCT01650025↗), in which were enrolled 48 pediatric NAFLD, 4 months VSL#3 supplementation improves severity of NAFLD as a consequence of VSL#3-induced eubiosis [138]. Administration of Bifidobacterium longum along with the prebiotic fructo-oligosaccharides (FOS) strongly improved circulating metabolic and inflammatory markers and fibrosis scores in patients with biopsy-proven NASH [139]. Intriguingly, the presence of Bifidobacterium longum in the VSL#3 compound modulates gut microbiota to produce conjugated linoleic acid (CLA), which, in turn, affect fatty acid composition in the liver further corroborating that the interplay between the gut and the liver in NAFLD assumes a relevant role for the development of therapeutic interventions.

In a phase I study, consumption of Lactobacillus GG for 8 weeks induced changes in gut microbiome of cirrhotic patients with minimal hepatic encephalopathy (n = 30), which reduced the abundance of Enterobacteriaceae, endotoxemia and TNF-α and facilitated the intestinal growth of Clostridiales Incertae Sedis XIV and Lachnospiraceae [140]. Recently, Romàn et al. evaluated the effect of a mixture of eight strains for 12 weeks in 36 patients with cirrhosis and found that the multistrain probiotic improved cognitive function, lower the incidence of falls at the follow-up and decreased inflammatory responses [141]. Conversely, in a randomized double-blind RCT including 44 cirrhotic outpatients, the long-term probiotic supplementation consisting of Bifidobacterium bifidum W23, Bifidobacterium lactis W52, Lactobacillus acidophilus W37, Lactobacillus brevis W63, Lactobacillus casei W56, Lactobacillus salivarius W24, Lactococcus lactis W19 and Lactococcus lactis W58, significantly influenced neutrophil resting burst, but not circulating endotoxins, gut permeability or inflammatory markers [142]. Although many of the current studies on effectiveness of probiotics are disappointing, this preliminary data has suggested that they are a well-tolerated and safe in cirrhotic subjects.

Few studies regarding the use of probiotics as HCC therapy have been reported in humans. Remarkably, it has been observed that preoperatively and postoperatively probiotic supplementation in patients with HCC who underwent hepatic resection favored liver function recovery and showed reduced frequency of intraoperative and postoperative complications [143].