Nonalcoholic fatty liver disease (NAFLD) encompasses a disease spectrum ranging from relatively benign “pure” fatty liver to steatohepatitis [nonalcoholic steatohepatitis (NASH)] with inflammation and fibrosis, which can progress to cirrhosis and hepatocellular cancer. Because it is closely associated with metabolic syndrome and insulin resistance, NAFLD is already the leading cause of elevated liver function tests in Western countries, and we should be prepared to see more of it (including NASH at the more severe end of this spectrum) in the near future. Although considerable progress has been made with the use of insulin sensitizers such as metformin and glitazones, no generally accepted pharmacological treatment of NAFLD/NASH is available.2 As some of these therapeutic shortcomings have to be attributed to our limited understanding of the pathogenesis of NAFLD,3 studies such as the present article by Huang et al.,4 which shed more light on the pathogenesis and uncover novel potential therapeutic targets in NAFLD, are very welcome.
Ligand-activated nuclear receptors (NRs) serve as metabolic sensors that can integrate metabolic and hormonal signals to changes in gene expression.5 As such, NRs play a key role in the regulation of the hepatic metabolism of bile acids, lipids, hormones, and drugs,6 and this makes them attractive drug targets for the treatment and prevention of liver diseases. Intriguingly, NRs not only regulate metabolic processes but can also modulate inflammation and fibrosis, two key ingredients for the development of progressive liver disease. Small heterodimer partner (SHP; NR0B2) is an exceptional member of the NR family because it lacks a DNA binding domain and known ligands (i.e., it is still a true orphan receptor), but it directly interacts with a range of other conventional NRs and negatively regulates their transcriptional activity by acting as an inducible corepressor.5–7 It is expressed in several tissues, with relatively high expression in the liver, heart, and pancreas and also in skeletal muscle, white adipose tissue (WAT), and brown adipose tissue (BAT).7 In the liver, SHP suppresses bile acid uptake and synthesis, the latter by the inhibition of liver receptor homolog-1 (LRH-1; NR5A2) transactivation of the gene encoding the key bile acid synthetic enzyme, cytochrome P450, family 7, subfamily A, polypeptide 1 (Cyp7a1).6 Because bile acids themselves induce SHP through the farnesoid X receptor (FXR; NR1H4), this pathway provides a negative feedback loop, helping to prevent hepatic bile acid overload (Fig. 1). More recent studies have identified a broader role for SHP in the regulation of energy production in BAT,8 glucose homeostasis,9 and hepatic triglyceride metabolism.10 In order to fully appreciate the (patho)physiological consequences of SHP manipulation, it is important to consider its multiple hepatic and extrahepatic actions and the fact that in rare instances SHP can also act as a transcriptional activator.11
In this elegant study, Huang et al.4 shed more light on the role of SHP in the regulation of hepatic lipid metabolism and its potential role in the pathogenesis of NAFLD. Using transgenic and knockout mouse models for SHP, previous studies have already noted a role for SHP in hepatic lipid metabolism,8–10, 12, 13 although the underlying mechanisms have remained incompletely understood. The authors now were able to demonstrate that SHP expression was induced in genetic [leptin-deficient ob/ob (OB−/−) mice] and dietary (high-sucrose/high-fat diet) mouse models of NAFLD, providing a good rationale for targeting SHP. The underlying molecular mechanism or mechanisms of this interesting observation remain unclear but could involve metabolic and hormonal signals such as glucose and insulin, which modulate the expression of the nuclear bile acid receptor FXR,14 a key inducer of SHP expression. Apart from bile acid–induced FXR, the activation of the SHP promoter has also been reported through the estrogen-related receptor and the c-jun N-terminal kinase, although the administration of endotoxin and proinflammatory cytokines does not activate SHP expression in rodents.6 Surprisingly little information is available on the expression and function of other NRs in NAFLD. Peroxisome proliferator-activated receptor-γ (PPAR-γ; NR1C3) promotes the lipogenesis and deposition of fat and is expressed at very low levels in a normal liver (mainly in Kupffer cells) but is increased in animal models with insulin resistance and fatty liver.3 PPAR-α (NR1C1) is expressed robustly in a normal liver and stimulates the expression of enzymes involved in fatty acid (FA) oxidation.3 The key role of PPARs in the regulation of hepatic lipid metabolism is further underlined by the development of hepatic steatosis in PPAR-α−/− mice, whereas PPAR-γ−/− mice are protected against steatosis.3 The expression/function of other NRs in NAFLD, including the liver X receptor (LXR; NR1H3) and constitutive androstane receptor (NR1I3), remains to be further explored.
SHP deficiency in OB−/−/SHP−/− double knockout mice prevented fatty livers with pronounced reductions of the hepatic FA and triglyceride (TG) contents of 56% and 85%, respectively, in comparison with those of leptin-deficient OB−/−/mice. In line with this finding, the adenoviral overexpression of SHP in primary hepatocytes increased TG accumulation, whereas decreasing the SHP expression through RNA interference had the opposite effect. Whether SHP-induced alterations of the FA/TG contents also had an impact on the degree of lipotoxicity was not addressed in this study. In this context, it should be considered that fat stored as TG is biologically quite neutral and may even be viewed as protective against the (lipo)toxicity of FA, which may trigger a number of harmful events ultimately resulting in hepatocyte death and liver fibrosis.15
In order to further address the molecular mechanisms underlying the antisteatotic effects of SHP deficiency, the authors focused on the major pathways of FA/TG metabolism, all known to be altered in NAFLD (Fig. 1): FA flux from WAT to the liver, FA uptake, de novo FA synthesis from glucose, FA oxidation, TG storage as fat droplets, and TG export as very low density lipoprotein (VLDL) from the liver.3 SHP deletion increased serum TG levels in OB−/−/SHP−/− mice, and this was attributed to higher rates of hepatic VLDL-TG secretion. In line with this key finding, the expression of microsomal triglyceride transfer protein (MTP), required for the assembly of TG with apolipoprotein B in VLDL production,3 was also increased in OB−/−/SHP−/− mice but reduced in OB−/− mice. The adenoviral overexpression of SHP inhibited MTP activity and VLDL–apolipoprotein B secretion in hepatocytes, whereas RNA interference knockdown of SHP showed the opposite effects. Promoter studies identified SHP as a key repressor of LRH-1–mediated transactivation of the mouse MTP promoter. Previous studies have revealed hepatocyte nuclear factor-4 (HNF-4; NR2A1) as a key target for the repressive effects of SHP on the human MTP promoter,16 whereas this pathway does not appear to play a role for the mouse MTP promoter. Of note, SHP in the repression of various hepatic genes (for example, Cyp7a1 and Cyp8b1) occurs through the targeting of LRH-1 and/or HNF-4.6 The repression of MTP expression through induced SHP could play a key role in the pathogenesis of NAFLD because MTP mutations in abetalipoproteinemia17 or the inhibition of MTP by the hepatitis C virus or pharmacological inhibitors (initially designed as hypolipidemic agents) can also cause severe hepatic steatosis.18 Interestingly enough, NAFLD in abetalipoproteinemia/hypobetalipoproteinemia is notoriously benign, without the development of progressive liver disease despite severe steatosis; this supports the concept that fat stored as TG may be biologically quite inert as long it is not associated with increased release of biologically active FA.
In addition to increased TG export, OB−/−/SHP−/− mice also showed reduced expression of pathways/genes (and their regulatory factors) for FA uptake (for example, CD36/PPAR-γ) and de novo FA synthesis [for example, fatty acid synthase/sterol regulatory element binding protein-1c (SREBP-1c)] in comparison with OB−/− mice, which could contribute to protection against steatosis (Fig. 1), whereas no effects on FA oxidation (for example, aldehyde oxidase and carnitine palmitoyltransferase 1/PPAR-α) were observed. The reduced expression of SREBP-1c, a key lipogenic transcription factor that activates several genes required for FA synthesis3 (Fig. 1), in OB−/−/SHP−/− mice may come as a major surprise because previous studies have suggested a role for FXR-induced SHP in the suppression of LXR/LRH-1–mediated SREBP-1c expression (Fig. 1), resulting in the repression of lipogenic target genes and the prevention of hepatic TG accumulation, VLDL secretion, and hypertriglyceridemia by bile acid feeding in dietary mouse models of obesity and NAFLD.13 The findings of this study are also difficult to reconcile with known effects of bile acids (which are potent inducers of SHP expression through FXR) on serum TG levels in humans: individuals treated with bile acids (for example, chenodeoxycholic acid for gallstone dissolution) show a reduction of serum TG, whereas patients treated with bile acid–binding resins or undergoing ileal resection show a marked (20%-30%) increase.19 However, bile acid/FXR effects on serum TG may be explained not only by the repression of de novo lipogenesis but also by their effects on activators (for example, apolipoprotein C-II) and inhibitors (for example, apolipoprotein C-II) of lipoprotein lipase activity and VLDL clearance (Fig. 1).19 To add to this controversy, FXR−/− mice (which also have reduced SHP expression) show increased SREBP-1c expression and accumulate TG in the liver and serum,20 whereas the activation of FXR (and SHP) improves steatosis and hyperlipidemia in leptin-resistant db/db mice.21 Nevertheless, the findings of this study are in line with previous studies of SHP transgenic mice that showed increased expression of PPAR-γ, SREBP-1c, hepatic lipogenesis, and TG accumulation.10 Both SREBP-1c−/− and PPAR-γ−/− mice are (partially) protected against hepatic steatosis, and this emphasizes the importance of these 2 lipogenic transcription factors for NAFLD.3 Recent studies suggest a role for PPAR-γ as an inducer of SREBP-1c expression and vice versa; this could explain the parallel reduction of both genes in OB−/−/SHP−/− mice.22 In addition, SHP could also (indirectly) affect SREBP-1c expression and function by its action on cholesterol catabolic enzymes, thus changing the intracellular concentration of metabolic intermediaries that activate LXR and inhibit SREBP-1c.3, 10 SHP has been reported to enhance PPAR-γ transactivation,11 but its effects on PPAR-γ gene transcription are still unclear. As correctly pointed out by the authors, acute (for example, bile acid–mediated or pharmacological) induction of SHP through FXR may have effects opposite to those of long-term transgenic or adenoviral SHP overexpression despite comparable tissue expression levels. We have to keep this in mind when speculating about potential pharmacological manipulation of the expression and/or activity of SHP.
Because SHP also plays a key role in the regulation of energy and glucose homeostasis outside the liver, a key question is whether the observed protective antisteatotic effects can be at least in part due to extrahepatic effects (Fig. 1). Previous reports have demonstrated that SHP deficiency protects against diet-induced obesity, and these effects have been attributed to the loss of repression of the transcriptional coactivator peroxisome proliferator-activated receptor-γ coactivator 1α (PGC-1α) by SHP, resulting in increased expression of the mitochondrial uncoupling protein [uncoupling protein 1 (UCP-1)] required for increased energy expenditure.8 In line with these studies demonstrating a key role of SHP as a repressor of PGC-1α expression, UCP-1 was induced in BAT of OB−/−/SHP−/− mice in the current study. However, these BAT effects may be of limited relevance for adult humans, although the overexpression of PGC-1α in WAT confers several BAT-specific features, including UCP-1 expression and increased FA oxidation.8 More importantly, as confirmed by glucose and insulin tolerance tests and clamp studies in the current work, SHP deficiency also improves hepatic and peripheral insulin sensitivity, and these effects may be mediated through multiple transcriptional effects/signaling pathways in the pancreas, muscle, liver, and WAT9 (Fig. 1). Interestingly, hepatic and peripheral insulin sensitivity improved in OB−/−/SHP−/− mice, whereas the body weight was not reduced. Improved insulin sensitivity may be key for the beneficial effects of SHP deficiency because insulin resistance increases hormone-sensitive lipase (HSL) activity, resulting in increased rates of TG lipolysis and enhanced FA flux to the liver (Fig. 1). In this context, it is important to keep in mind that, in contrast to mice, in human NAFLD more than 60% of TG accumulating in the liver is derived from FFA flux from WAT (as a result of insulin resistance and increased lipolytic activity), whereas only 25% is derived from de novo lipogenesis and even less is derived from diet.23 Interestingly, the expression of HSL remained elevated in OB−/−/SHP−/− mice, but HSL activity is mainly regulated at a posttranscriptional level, and the HSL activity itself was not tested. Moreover, a recently identified adipose triglyceride lipase may be more relevant for the first step of TG hydrolysis in WAT than HSL.24
Like any interesting publication, the study by Huang et al.4 raises some controversies and leaves several questions for the future. It will be important to see whether SHP polymorphisms and abnormal expression/activity of SHP play a role in the pathogenesis and progression of human NAFLD. The protective effect of SHP deficiency against obesity and NAFLD in mice contrasts with the reported association of SHP haplo insufficiency with increased body weight in Japanese populations25 and a possible link between decreased SHP activity and body weight in European populations,26 although it does not appear to be a common cause of obesity. As hepatologists, we are primarily concerned with NASH, whereas pure fatty liver runs a rather benign course. It will be interesting to see whether SHP knockout has an impact not only on steatosis but also on features of steatohepatitis, including hepatocellular ballooning, inflammation, and fibrosis,1 in more aggressive NASH mouse models (for example, methionine/choline deficiency). The factors responsible for progression from fatty liver to more aggressive NASH are still under debate.3 NRs (including SHP), as key regulators and gatekeepers of lipid metabolism, inflammation, and fibrosis,27 could play a central role in this progression. We may have to consider that the cellular localization of SHP to different parenchymal and nonparenchymal liver cell compartments may critically determine its biological effects, and we tend to oversimplify things by viewing the liver as one metabolic unit (Fig. 1). Finally, it remains entirely open to question whether ligand binding is possible at all for targeting SHP pharmaceutically,7 and there may be some concerns with targeting such central and ubiquitous pathways therapeutically. Because NAFLD is associated with increased cardiovascular risk,28 the increased VLDL secretion after SHP knockdown/inhibition observed in the present study in mice could become a clinical concern in humans. Despite these many open questions, this important study has put SHP on the stage, and it will be exciting to see what this little orphan holds for the future of NAFLD.