When the orphan nuclear hormone receptor farnesoid X receptor (FXR, NR1H4) was adopted in 1999 through identification of bile acids as its naturally activating ligands, nobody would have predicted the central role FXR is now recognized to play for a wide range of key metabolic and regenerative pathways in liver and extrahepatic tissues.1 Although it has been known for decades that bile acids can modulate bile secretory function, lipid and glucose metabolism, mucosal protection against bacteria in the gut, inflammation and immune function, as well as cellular proliferation and cancer formation,2 the underlying transcriptional pathways involving FXR have not been unravelled until recently.1 FXR, which is predominantly expressed in liver, intestine, kidney and adrenal glands is an intranuclear bile acid sensing receptor and binds as heterodimer with the common heterodimerization partner retinoid X receptor (RXR) to distinct FXR response elements in target gene promoter regions and initiates gene transcription.1 In addition to direct activation of gene transcription, FXR can also indirectly inhibit gene expression via induction of the transcriptional repressor short heterodimer partner (SHP).1
FXR protects liver integrity through control of bile acid homeostasis1 and hepatic regeneration.3 Mice lacking FXR (FXR−/−) show increased levels of inflammation and develop hepatocellular carcinomas (HCC), suggesting that FXR may also play a key role in the protection against hepatic inflammation and inflammation-driven carcinogenesis,4, 5 although the underlying molecular mechanisms have so far remained unresolved. Moreover, FXR may protect the liver against gut-derived toxins and inflammation by maintaining gut integrity through induction of antibacterial factors such as angiogenin, inducible nitric oxide synthase (iNOS) and interleukin (IL)-18.6 This may explain the detrimental effects of biliary obstruction with lack of bile acids in the gut, leading to intestinal bacterial overgrowth, mucosal injury followed by bacterial translocation across the injured mucosal barrier and a systemic inflammatory response.6
The work by Wang and coworkers in this issue of HEPATOLOGY7 now expands our understanding of the role of FXR for hepatic inflammation and development of inflammation-driven HCC by demonstrating that FXR is able to interact with nuclear factor-κB (NF-κB), thus adding FXR to the list of nuclear receptors reciprocally interacting with NF-κB. NF-κB is a key transcriptional regulator of the inflammatory response and cell proliferation, which is rapidly activated in response to proinflammatory stimuli such as bacterial lipopolysaccharide (LPS) and cytokines under various inflammatory conditions.8, 9 Interactions of NF-κB with other nuclear receptors such as the common nuclear receptor dimerization partner RXR and the xenobiotic receptor pregnane X receptor (PXR) may explain reduced cytochrome p450-mediated drug metabolism observed in inflammatory and infectious diseases.10–12 Conversely, activated PXR inhibits the NF-κB pathway which may contribute to immunosuppressive effects of drug metabolism-inducing xenobiotics/drugs, such as the antibiotic rifampicin.12 In addition, osteoporosis and periarticular bone loss in inflammatory arthritis have been linked to NF-κB mediated inhibition of the vitamin D receptor.13 With the current experimental work Wang et al. now demonstrate, that (1) livers from FXR−/− mice are more susceptible to inflammatory (e.g., LPS-induced) stress than wildtype animals, (2) FXR agonists, conversely, inhibit classical pro-inflammatory NF-κB target genes (such as tumor necrosis factor α (TNFα), iNOS and cyclooxygenase-2) in hepatocytes, (3) FXR activation inhibits hepatocellular NF-κB activation, and (4) inflammation-induced NF-κB activation reciprocally antagonizes FXR activity and FXR target gene expression.7 Interestingly, FXR has already been linked to the NF-κB pathway in cells other than hepatocytes. As such, FXR has previously been shown to inhibit NF-κB activation in endothelial cells and vascular smooth muscle cells, potentially via SHP-dependent mechanisms14 which could contribute to an anti-inflammatory net effect of FXR in liver. It is attractive to speculate, that SHP may also be involved in FXR-mediated inhibition of NF-κB in hepatocytes. Other possibilities such as physical protein–protein interactions and SUMOylation pathways, which were previously described for the inhibitory effects of the glucocorticoid receptor, the peroxisome-proliferator activating receptor γ and the liver X receptor on NF-κB activation, also remain to be explored.
Apart from the potential therapeutic relevance for FXR as novel anti-inflammatory therapeutic strategy, the findings of the current study7 may broaden our understanding of the multiple biological functions of bile acids — the natural ligands of FXR.2 Several in part conflicting studies have shown both pro- and anti-inflammatory properties of bile acids in hepatocytes and nonparenchymal cells of the liver. While previous studies reported pro-inflammatory effects of bile acids in macrophages, more convincing evidence now indicates that bile acids may in fact inhibit LPS-induced release of pro-inflammatory cytokines by macrophages,15 thus contributing to the well-known immunesuppressive effects of bile acid.5, 16, 17 In cells which do not express FXR (e.g., Kupffer cells/macrophages) these anti-inflammatory properties of bile acids may be mediated via the membrane-bound bile acid receptor TGR5.15 Notably, hepatocytes (expressing FXR) are not only targets of cytokine action, but can also produce proinflammatory cytokines18–20 making them active players in the inflammatory response. Hepatocellular accumulation of bile acids in cholestasis is clearly cytotoxic and thus contributes to inflammation.2 The current work by Wang et al.7 now raises the intriguing question whether lower (noncytotoxic) concentrations of bile acids could in fact exert anti-inflammatory effects in hepatocytes via FXR mediated inhibition of NF-κB.7 The interpretation of increased inflammation in FXR−/− mice is, however, complicated by the fact that these animals also have a severely disturbed hepatocellular bile acid homeostasis which per se could contribute to hepatocellular damage and thereby inflammation. Of note, the findings of this study by Wang and coworkers7 stand in some contrast to previous reports of bile acid-induced intracellular adhesion molecule-1 (ICAM-1) expression in human hepatocytes via FXR.21 From a therapeutic perspective, this Gordian knot may be overcome by pharmacologic FXR agonists lacking the potential cytotoxicity of endogenous bile acids.
What can we learn from this study by Wang et al.7 regarding the pathogenesis and treatment of inflammatory liver diseases and inflammation-associated cancer? NF-κB activation in hepatocytes and nonparenchymal liver cells is a central event in various forms of liver injury induced by alcohol, fatty liver, cholestasis, hepatitis B and C virus infections9, 22 Generally, activation of NF-κB is considered a protective response, which may limit cell loss by apoptosis and stimulates hepatocyte proliferation/regeneration by providing a cellular survival signal.9 In line with the protective role of NF-κB, total inhibition of NF-κB exclusively in hepatocytes leads to increased hepatocyte apoptosis, which triggers hepatic inflammation via NF-κB expressing Kupffer cells, and finally results in the development of steatohepatitis and HCC.23 On the other hand, persistently activated NF-κB in hepatocytes contributes to a chronic inflammatory condition19 and constitutive activation of NF-κB is considered as one of the early key events involved in liver carcinogenesis.22 In line with its disease aggravating role, partial inhibition of NF-κB activation results in reduced liver injury after cytokine challenge and ischemia/reperfusion injury.24 These studies therefore indicate that NF-κB has a multitude of functions in the diseased liver and that the cellular source and degree of NF-κB inhibition may critically determine the final outcome of liver damage. Of note, NF-κB regulated anti-apoptotic genes were not repressed by FXR activation in the current study by Wang et al.,7 suggesting that FXR selectively inhibits only the effects of NF-κB on inflammation, although the molecular mechanisms underlying this selectivity require further studies. Lack of FXR-mediated repression of NF-κB-dependent anti-apoptotic genes may have important implications for the potential therapeutic use and safety of FXR agonists in inflammatory liver disorders, since major side effects of total NF-κB blockage may derive from deregulated programmed cell death.23 Notably, FXR ligands rescue HepG2 cells from serum deprivation-induced apoptosis25 and the FXR antagonist guggulsterone enhances apoptosis in Barrett's esophagus cells26 raising the possibility that FXR may directly inhibit apoptosis under other conditions.
One of the leading causes for liver inflammation and HCC development in adults are chronic viral infections with hepatitis B and C virus and both viruses are known to activate hepatocellular NF-κB.22 One might speculate that treatment with FXR ligands in these diseases might reduce hepatic inflammation and, thus, the risk for developing liver cirrhosis and cancer. However, it has to be kept in mind, that FXR activation by bile acids increases virus replication (at least hepatitis C genotype 1) and high bile acid levels are a negative predictor for the therapeutic outcome.27 Nonalcoholic fatty liver disease (NAFLD) is the most common cause of chronic liver disease in the Western world and a rising cause of HCC. The key pathophysiological feature of NAFLD is insulin resistance (IR) and a hallmark in the progression from bland steatosis to clinical relevant inflammation in nonalcoholic steatohepatitis (NASH) is the activation of NF-κB.28 Since FXR is critically involved in insulin sensitivity, lipid and glucose homeostasis, it is also a potential key player in NAFLD.1 Recent reports have linked NF-κB to IR. Mice with constitutive hepatocellular overexpression of inhibitor of kappa B kinase 2 (IKK2), which activates NF-κB via the classical/canonical pathway develop chronic inflammation leading to profound hepatocyte and moderate systemic IR.19 Conversely, deletion of IKK2 reverses IR, making NF-κB inhibition via FXR agonists also a possible strategy for the treatment of NASH.18 Pilot studies using a pharmacological pan-IKK2 inhibitor in obese and IR-mice were able to block liver steatosis, IR, and NASH.29 Collectively, these studies suggest that the use of FXR agonists for the treatment of NASH could have a dual beneficial effect through direct metabolic effects of FXR target genes and inhibiting chronic inflammation via blocking the NF-κB pathway. However, one has to keep in mind that targeting NF-κB via FXR agonists will limit the effects to FXR-expressing cells (i.e., hepatocytes, cholangiocytes, sinusoidal endothelial cells) and will not affect other cells involved in inflammation and fibrosis (Kupffer cells, myofibroblasts, stellate cells).
Finally, FXR may not only be an active player in inflammation but is also a target of the inflammatory response itself. LPS, TNFα and IL1β decrease FXR expression, DNA binding and FXR target gene expression during the acute phase response.30 This could finally end up in a vicious circle where reduced FXR expression results in less NF-κB inhibition, which augments the initial inflammatory signal, inhibits apoptosis and leads to HCC formation. Moreover, the current study by Wang et al. now shows that NF-κB itself inhibits FXR, which would further contribute to FXR reduction in inflammatory conditions and would therefore enhance NF-κB activity.7 In humans, genetic defects or variants of FXR contributing to development of HCC were so far not observed, but FXR is reduced in chronic (cholestatic) liver diseases.31
In summary, the study by Wang et al.7 has certainly expanded our understanding of FXR as key regulator of hepatic (patho)physiology and potential pharmacologic target. Apart from liver, these findings may also be relevant for extrahepatic tissues and organs linked to liver disease (e.g., inflammatory bowel disease, atherosclerosis) (Fig. 1). The biggest challenge will be to translate these findings into clinical practice. Of note, inhibition of NF-κB in cells expressing FXR may explain some of the anti-inflammatory effects of FXR observed in experimental studies in the liver and beyond (e.g., intestine) and could also contribute to clinical effects of FXR agonists currently investigated in ongoing clinical trials.