Role of gut microbiota in liver diseases


  • Yasuhiro Miyake,

    Corresponding author
    • Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama, Japan
    Search for more papers by this author
  • Kazuhide Yamamoto

    1. Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama, Japan
    Search for more papers by this author

Correspondence: Dr Yasuhiro Miyake, Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry, and Pharmaceutical Sciences, 2-5-1 Shikata-cho, Kita-ku, Okayama 700-8558 Japan. Email:


The liver constantly encounters food-derived antigens and bacterial components such as lipopolysaccharide translocated from the gut into the portal vein. Bacterial components stimulate Toll-like receptors (TLR), which are expressed on Kupffer cells, biliary epithelial cells, hepatocytes, hepatic stellate cells, endothelial cells and dendritic cells and recognize specific pathogen-associated molecular patterns. The signaling of TLR to its main ligand triggers inflammation. Usually, in order to protect against hyperactivation of the immune system and to prevent organ failure by persistent inflammation, TLR tolerance to repeated stimuli is induced. In chronic liver diseases, a breakdown in TLR tolerance occurs. Furthermore, Kupffer cells, hepatic stellate cells and natural killer T cells are key components of innate immunity. Decreased numbers and impaired ability of these cells lead to failures in immune tolerance, resulting in persistent inflammation. Recently, the activation of inflammasome was revealed to control the secretion of pro-inflammatory cytokines such as interleukin-1β in response to bacterial pathogens. Innate immunity seems to be an important contributor to the pathogenesis of fatty liver disease and autoimmune liver disease. Recently, probiotics were reported to affect various liver diseases via shifts in gut microbiota and the stability of intestinal permeability. However, many unresolved questions remain. Further analysis will be needed to gain a more comprehensive understanding of the association of innate immunity with the pathogenesis of various liver diseases.


THE LIVER, THE largest organ in the body, weight 1200–1500 g and has a double blood supply.[1] The hepatic artery, coming from the celiac axis, supplies the liver with arterial blood and the portal vein brings venous blood from the intestines and spleen. Portal blood flow in humans is approximately 1000–1200 mL/min. Thus, the liver constantly confronts food-derived antigens and bacterial components such as lipopolysaccharide (LPS) translocated from the gut into the portal vein; however, the liver has the unique capacity to induce immune tolerance.

Previously, regulatory T cells (Treg), Kupffer cells, natural killer T (NKT) cells and hepatic stellate cells (HSC) were reported to contribute to immune tolerance in the liver. Interaction between Treg and Kupffer cells promotes the secretion of interleukin (IL)-10 from Treg, and the depletion of Treg breaks antigen-specific immune tolerance.[2] The depletion of liver NKT cells also exacerbates hepatic inflammation in carbon tetrachloride-induced liver injury.[3] HSC induce the apoptosis of conventional CD4+ T cells in a Fas/Fas ligand-dependent manner and increase Treg proliferation via cell–cell contact; moreover, HSC-expanded Treg express high levels of programmed cell death 1 and cytotoxic T-lymphocyte antigen 4, show enhanced production of IL-10, and cause the suppression of alloreactive CD4+ T-cell proliferation.[4]

Toll-like receptors (TLR), which comprise a highly conserved family of receptors that recognizes specific pathogen-associated molecular patterns (PAMP), play a key role in innate immunity by triggering inflammatory responses to the main ligands of TLR. Various TLR are expressed on liver cells (Table 1). The liver constantly encounters various antigens, and in order to prevent organ failure due to hyperactivation of the immune system, TLR tolerance to repeated stimuli is induced.[11] On the other hand, a breakdown in TLR tolerance results in persistent inflammation and contributes to the development of chronic liver diseases. Singh et al.[12] reported that bacterial translocation comparably occurs in both normal and diseased livers such as primary biliary cirrhosis (PBC) and non-alcoholic steatohepatitis (NASH) although the expression of TLR2 and TLR4 is enhanced in the diseased livers than normal. In normal biliary epithelial cells (BEC), repeated LPS-stimuli induced hyporeactivity to LPS.[13] However, BEC from PBC patients show hyperreactivity to LPS.[14]

Table 1. Intrahepatic expression of Toll-like receptors (TLR) on human liver cells
Kind of liver cellsExpressed TLR
HepatocyteTLR2,[5] TLR3,[6] TLR4[7]
Biliary epithelial cellTLR1–6, 9[8]
Hepatic stellate cellTLR1–9[9]
Sinusoidal endothelial cellTLR2[10]
Kupffer cellTLR2,[5] TLR3,[6] TLR4[5]

Herein, we review the association of gut microbiota with the pathogenesis of chronic liver diseases such as NASH, primary sclerosing cholangitis (PSC) and PBC.

Non-Alcoholic Steatohepatitis

NON-ALCOHOLIC FATTY LIVER disease (NAFLD) is recognized as a common liver disorder that represents the hepatic manifestation of metabolic syndrome, and encompasses a spectrum of hepatology, ranging from simple steatosis to cirrhosis.[15, 16] NASH is the progressive form of liver injury and characterized by steatosis, lobular inflammation, hepatocyte ballooning, Mallory's hyaline and fibrosis. The histological findings of NAFLD and NASH are similar to the lesions caused by alcoholic liver disease.

The “two-hit” model is a widely accepted theory of the pathogenesis of NASH.[17] According to this theory, the first hit is an imbalance in fatty acid metabolism leading to hepatic steatosis, and the secondary hits are oxidative stress/metabolic stress and dysregulated cytokine production. In NASH patients, hepatic TLR4 expression is increased.[18] TLR4 deficiency ameliorates hepatic steatosis induced by high-fat diets.[19] Activation of TLR4 takes a role in the first hit. Next, as components potentially involved in the secondary hits, the gut microbiota have been investigated. In patients with NAFLD, intestinal permeability and the prevalence of small intestinal bacterial overgrowth are increased.[20] In NAFLD models, the translocation of bacterial components promotes tumor necrosis factor (TNF)-α release from Kupffer cells and induces hepatic inflammation through TLR4 and TLR9 signaling.[21, 22] High-fat diets induce the deposition of toxic lipids such as diacylglycerol and sphingolipid in Kupffer cells and promote the secretion of TNF-α, interferon (IFN)-γ, IL-6 and IL-1β from Kupffer cells via LPS stimulation.[23] Furthermore, hepatic NKT cell numbers have been shown to be decreased.[24] High-fat diets reduce hepatic NKT cell numbers through hepatic IL-12 production, which results in increases in the hepatic production of pro-inflammatory cytokines such as TNF-α and IFN-γ and the exacerbation of inflammation in the liver.[25] Modification of gut microbiota with probiotics has been found to increase hepatic NKT cell numbers and reduce the hepatic expression of TNF-α and inflammation.[24, 26, 27] In NASH patients, 24-week treatment with Bifidobacterium longum and fructo-oligosaccharides improves insulin resistance and reduces histological NASH activity.[28] Various findings to date support an association of gut microbiota with the pathogenesis of NASH. A breakdown in TLR tolerance seems to be significantly associated with the progression of NASH. On the other hand, in NASH patients, hepatic NKT cell number has been reported to increase.[29] Thus, there may be partial differences in the pathogenesis between NASH patients and animal models. Further studies in NAFLD patients are required.

Recently, the contribution of inflammasomes to the pathogenesis of NAFLD was reported.[30] Inflammasomes are groups of protein complexes that recognize a diverse set of inflammation-inducing stimuli, including PAMP and damage-associated molecular patterns (DAMP), and that directly activate caspase-1, resulting in the production of important pro-inflammatory cytokines such as IL-1β and IL-18 and a type of cell death called “pyroptosis”.[31] Csak et al.[30] reported that saturated fatty acid, but not unsaturated fatty acid, increases the expression of PYD domain-containing protein 3 (NLRP3) in hepatocytes, and that activation of NLRP3 by LPS stimuli via TLR4 leads to IL-1β release from hepatocytes. Furthermore, hepatocytes exposed to saturated fatty acid release danger signals that activate macrophages by the upregulation of NLRP3. The activation of hepatic macrophages leads to an exacerbation of hepatic inflammation.

Primary Sclerosing Cholangitis

PRIMARY SCLEROSING CHOLANGITIS (PSC) is a chronic cholestatic liver disease characterized by inflammation, obliteration and fibrosis of the intrahepatic and/or extrahepatic biliary ducts.[32] Although the etiology of PSC remains unknown, gut microbiota are considered to play an important role in the pathogenesis of PSC. Sumitran-Holgersson et al.[33, 34] reported that approximately 60% of PSC patients have serum antibodies to BEC (anti-BEC), and stimulation of BEC with the immunoglobulin (Ig)G of anti-BEC+ PSC patients induces the expression of TLR4, whereas unstimulated or normal IgG-stimulated BEC do not express TLR4. TLR4-expressing BEC produce high levels of IL-1β, IL-8 and IFN-γ when stimulated with LPS. Mueller et al.[35] reported that BEC from end-stage PSC liver show marked expression of TLR4, increased activation of the myeloid differentiation protein 88/IL-1 receptor-associated kinase signaling cascade, and a loss of immune tolerance to endotoxin after repeated endotoxin exposure. However, the expression of TLR4 in BEC from the early-stage PSC liver is similar to that of healthy liver. Increased expression of TLR4 and a loss of immune tolerance to endotoxin in BEC may coordinate autoimmunity in the progression of PSC.

Approximately 1–3% of patients with ulcerative colitis (UC) had concurrent PSC,[36-38] and approximately 68–75% of PSC patients had UC.[39-41] UC patients with PSC more frequently have total colonic involvement than UC patients without PSC (68–85% vs 44–45%).[36, 42] In UC patients, the extent of disease is positively correlated with plasma concentrations of endotoxin.[43] Thus, endotoxin concentrations in the portal vein are expected to be higher in UC patients with PSC than in UC patients without PSC. Furthermore, in the bile of PSC patients, enteric bacteria such as Escherichia coli are frequently detected.[44] Thus, in PSC patients, the liver constantly confronts abundant gut bacterial antigens such as endotoxin, and reinforced confrontation with these antigens is considered to be among the causes of PSC.

Serum perinuclear antineutrophil cytoplasmic antibodies (pANCA), which are frequently seen in patients with UC, have been detected in approximately 80% of PSC patients.[45, 46] By indirect immunofluorescence study, pANCA in PSC show a heterogeneous rim-like staining in the nuclear periphery (atypical pANCA),[47] unlike classical pANCA, which show peripheral rim-like staining without intranuclear staining in patients with systemic vasculitis. Recently, the autoantigen of this atypical pANCA has been reported to be β-tubulin isotype 5.[48] Furthermore, this atypical pANCA cross-reacts with FtsZ, which is present in almost all bacteria of the gut microbiota. These phenomena reflect abnormal immune responses to gut bacterial antigens in PSC.

Recently, caspase recruitment domain-containing protein 9 (CARD9), υ-rel reticuloendotheliosis viral pmcogene homolog (REL) and IL-2, which are associated with the susceptibility to UC,[49] have been reported as candidate genes for PSC.[50] Of these genes, CARD9 and REL are associated with innate immunity. Importantly, REL takes part in nuclear factor (NF)-κB functions. CARD9 is the adaptor molecule essential for the control of fungal infection. Gross et al.[51] reported that all CARD9-deficient mice died within 5 days after infection with Candida albicans, whereas more than 50% of control mice survived for more than 12 days. β-Glucan is initially recognized by dectin-1, a type II transmembrane protein expressed in various inflammatory cells such as macrophages, monocytes, dendritic cells, neutrophils, a subpopulation of T cells, B cells, mast cells and eosinophils. After the recognition of β-glucan by dectin-1, Syk signaling leads to the complex formation of CARD9, Bcl-10 and mucosa-associated lymphoid tissue translocation gene 1 and results in the release of IL-1β.[51-54] Candida is detected in the bile of approximately 10% of PSC patients, and a finding of Candida in the bile worsens the prognosis.[44] Polymorphisms of the CARD9 gene may influence innate immunity to Candida in PSC patients. In addition, the activation of inflammasomes such as NLRP3 is involved in the process of IL-1β production by dectin-1 signaling. Silencing of NLRP3 expression partially impairs the processing of pro-IL-1β. Inflammasomes may be associated with the pathogenesis of PSC and are worth investigating in order to reveal the pathogenesis of PSC.

Primary Biliary Cirrhosis

PRIMARY BILIARY CIRRHOSIS is an autoimmune liver disease characterized by intrahepatic bile duct destruction, particularly chronic non-suppurative destructive cholangitis, cholestasis, and presence in the serum of antimitochondrial antibodies (AMA). AMA are detected in approximately 95% of PBC patients.[55] In particular, M2 antibodies (M2Ab) against E2 components of pyruvate dehydrogenase complex (PDC-E2) are specific to PBC and are detected in nearly 80% of patients.

Increased expression of TLR4 is shown in the liver of PBC. TLR4 expression levels in the BEC and periportal hepatocytes of PBC are augmented.[7] Especially, the BEC of PBC patients clearly express TLR4, regardless of disease stage. On the other hand, the role of TLR in the pathogenesis of PBC has been investigated also using PBMC obtained from PBC patients. Compared to those from healthy controls, the monocytes from PBC patients produce high amounts of pro-inflammatory cytokines, particularly IL-1β and IL-6, in response to bacterial components such as LPS, flagellin and CpG, but not in response to viral components such as polyinosinic–polycytidylic acid (polyI:C).[56] LPS stimulation increases the expression of both TLR4 and MyD88 in monocytes from PBC patients.[57] Similarly, CpG-stimulated memory B cells from PBC patients express TLR9 and produce high levels of IgM.[58] However, the surface expression levels of TLR4 and TLR9, respectively, in unstimulated monocytes and B cells are similar in PBC patients and healthy controls. These findings raise the question of how innate immunity participates in the pathogenesis of PBC in vivo. Shimoda et al.[59] recently reported that when in the presence of IFN-α from polyI:C-stimulated monocytes, LPS-stimulated natural killer (NK) cells destroy autologous BEC. The activation and cross-talk of monocytes with NK cells are suggested to contribute to the pathogenesis of PBC. The various findings to date generally support the contribution of mechanisms of innate immunity in the pathogenesis of PBC.

The concept of molecular mimicry has been proposed as the cause of PBC. AMA in PBC serum cross-react with bacterial components. AMA have been reported to react with proteins of E. coli isolated in stool specimens from PBC patients.[60] HRPA153–167 and MALE95–109 of E. coli share 80% and 73% sequential similarity, respectively, with human PDC-E2212–226, and M2Ab in approximately 30% of PBC patients cross-reacts with HRPA153–167 and/or MALE95–109 of E. coli.[61] In addition, approximately 50% of PBC patients harbor IgG3 antibodies that cross-react with β-galactosidase (BGAL) of Lactobacillus delbrueckii, a probiotic microorganism essential to starter cultures and yogurt production.[62] BGAL266–280 of L. delbrueckii shares 67% similarity with human PDC-E2212–226. In approximately 25% of PBC patients, the serum reacts in a highly directed and specific manner to proteins of Novosphingobium aromaticivorans from fecal specimens.[63]

Probiotics and the Liver

GUT MICROBIOTA SHIFTS influence hepatic inflammation. In a model of liver injury induced by ischemic reperfusion, intestinal Enterococcus spp. and Enterobacteriaceae increase, while Lactobacillus spp., Bifidobacter spp. and Bacterioides spp. decrease. Supplementation with Lactobacillus paracasei decreases Enterococcus spp. and Enterobacteriaceae and increases Lactobacillus spp., Bifidobacter spp. and Bacterioides spp., which result in reduced levels of expression of TNF-α, IL-1β and IL-6 and amelioration of necroinflammation in the liver.[64] In liver injury induced by chemical substances or alcohol, probiotic supplementation with species such as Lactobacillus spp. and Bifidobacterium spp. decreases bacterial translocation to the liver through decreased concentrations of aerobic bacteria such as E. coli as well as due to increased intestinal stability (i.e. reduced intestinal permeability), and reduces hepatic inflammation.[65-67] Furthermore, gut microbiota shifts influence hepatic metabolism (e.g. amino acid, fatty acid, organic acid and carbohydrate metabolism) by the modulation of hepatic gene expression, without direct contact with the liver.[68, 69] In cirrhotic patients with hepatic encephalopathy, intestinal E. coli and Staphylococcus spp. overgrow, and supplementation with symbiotic reduces blood ammonia levels and ameliorates hepatic encephalopathy by reducing levels of E. coli and increasing Lactobacillus spp.[70] In rats fed a high-cholesterol diet, Lactobacillus spp. supplementation decreases intestinal E. coli and increases Lactobacillus spp. and Bifidobacterium spp., which leads to reduced levels of hepatic cholesterol and triglyceride.[71] In general, gut microbiota shifts have been shown to exert a substantial impact on the liver.


MANY FINDINGS TO date support the contribution of bacterial components (e.g. endotoxins, unmethylated CpG containing DNA) to the pathogenesis of various liver diseases (Fig. 1). Innate immunity plays an important role in the hepatic response to these bacterial components, and TLR4 and TLR9 signaling has been widely investigated (Table 2). However, many questions remain regarding the relation of innate immunity to the pathogenesis of liver diseases. First, it remains unclear why TLR tolerance is disrupted in various liver diseases. Second, how do the roles of innate immunity in the pathogenesis differ between PSC and PBC? BEC are the main targets of injury in both diseases, although the histological features of PSC and PBC markedly differ. Third, the factors that control the protective or detrimental roles of NKT cells and Kupffer cells remain to be determined. Fourth, we still need to determine which probiotic will be most effective for treating which liver disease(s). Further analysis will be needed to more fully understand the association of innate immunity with disease pathogenesis in the case of each specific disease.

Figure 1.

Concept of gut–liver axis in liver diseases. TLR, Toll-like receptors; Treg, regulatory T cells.

Table 2. Intrahepatic condition of liver cells, TLR, cytokines and causative microbes in the chronic liver diseases
 TregNKTHSCKupffer cellTLRCytokineMicrobe
  1. HSC, hepatic stellate cells; IFN, interferon; IL, interleukin; NASH, non-alcoholic steatohepatitis; ND, no data; NKT, natural killer T cells; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis; TLR, Toll-like receptors; TNF, tumor necrosis factor; Treg, regulatory T cells.
NASH↑[72]↑[29]↑[73]↑[72]TLR4 ↑[18]TNF-α ↑,[74] IFN-γ ↑,[74] IL-1β ↑,[74] IL-6 ↑,[74] IL-17 ↑[75]Porphyromonas gingivalis[76]
PSC→[77]NDND↑[78]TLR4 ↑,[35] TLR9 ↑[34]TNF-α ↑,[35] IFN-γ ↑,[34] IL-6 ↑,[33] IL-8 ↑[35]Escherichia coli,[44] Helicobacter pylori,[79] Chlamydia spp.,[80] Candida[44]
PBC↑[77]↑[81]↑[82]↑[78]TLR3 ↑,[6] TLR4 ↑,[7] TLR7 ↑,[6] TLR9 ↑[6]IFN-α ↑,[6] IFN-β ↑,[6] IFN-γ ↑,[6] IL-5 ↑,[83] IL-6 ↑,[83] IL-8 ↑,[84] IL-12 ↑,[83] IL-17 ↑[75]Escherichia coli,[60] Lactobacillus delbrueckii,[62] Novosphingobium aromaticivorans,[63] Mycoplasma pneumonia,[85] Streptococcus intermedius,[86] Propionibacterium acnes[87]

Recently, stimuli by TLR have been indicated to activate inflammasomes, and activated inflammasomes induce the processing of pro-IL-1β and pro-IL-18 by the activation of caspase-1 (Fig. 2). The association of IL-1β with the pathogenesis of various liver diseases has been already reported;[23, 34, 56] however, investigation of the association of inflammasomes with liver disease is still in the early stages. Inflammasomes warrant further analysis, which may reveal the mechanisms of innate immunity in various liver diseases.

Figure 2.

Relation of Toll-like receptors and inflammasomes. DAMP, damage-associated molecular patterns; IL, interleukin; PAMP, pathogen-associated molecular patterns; TLR, Toll-like receptors.