Liver cholesterol: is it playing possum in NASH?

SINCE NONALCOHOLIC STEATOHEPATITIS (NASH) became an accepted entity in 1980, its pathogenic mechanisms have remained obscure. Strong associations with central obesity, type 2 diabetes, and dyslipidemia have always suggested NASH is a metabolic disorder (2, 7, 19, 23), and by 2003 the connection to insulin resistance was securely evidence based (6, 21). It has remained unclear, however, why only 10–25% of all those with nonalcoholic fatty liver disease (NAFLD) (25–45% of American adults) have NASH (32). An early suggestion was that this required a separate (second) injury/proinflammatory process, additional to the metabolic factors linked to steatosis (8). While useful in its day to stimulate research, the “two-hit” concept does not explain the strong links between NASH and diabetes or metabolic syndrome, a nexus that infers that the more severe the “metabolic movers”, the more does NAFLD manifest as NASH (6, 19, 21). For this and other reasons recently reviewed (1, 2, 7, 19, 23), most now accept that hepatocyte accumulation of triglycerides (TG) leads to steatosis, but different lipid molecules mediate the pathogenesis of NASH. Such “toxic lipid species” cause hepatocellular injury and cell death (1, 7, 30), directly or indirectly inciting inflammatory and profibrotic responses. This set of pathophysiological processes is collectively termed lipotoxicity (2, 7, 19, 23). Lipotoxicity is now the accepted mechanism for pancreatic -cell injury in type 2 diabetes, intimal damage in atheroma, and cardiac toxicity in metabolic syndrome. Some have even suggested renaming NASH as “liver lipotoxicity” (7). To date, however, the identity of the lipid molecule(s) that causes liver lipotoxicity has remained veiled (2, 7, 23). Perhaps, like the proverbial possum, the lipotoxic assassin is lying low (31). Enter the possum! Or, more specifically, the North American opossum. Gray short-tailed opossums (Monodelphis domestica) have been used for decades for biomedical research (26). About 20 years ago, one partially inbred line was noted to develop impressive increases in serum lipoproteins in response to an “atherogenic” high-cholesterol, high-fat (HCHF) diet (26). Investigators termed this line “high responders”, and comparisons are made with otherwise similar animals, “low responders”, that do not develop diet-induced hypercholesterolemia. In this issue of American Journal of Physiology-Gastrointestinal and Liver Physiology, Chan and colleagues from the National Primate Research Center, Genetics and Pathology at UT San Antonio report three highly novel findings (5). First, the genetic basis for this type of dietary hypercholesterolemia is a mutation in ABCB4, a phospholipid and cholesterol transporter expressed on the canalicular domain of the hepatocyte plasma membrane, deficiency of which impairs biliary cholesterol excretion. Second, feeding ABCB4 mutant opossums a HCHF diet caused a 10-fold increase of hepatic total cholesterol, more impressive than the 5-fold increase in TG. Third, such hepatic cholesterol accumulation was associated with steatosis, an inflammatory cell infiltration and ballooning of hepatocytes, together with perisinusoidal fibrosis; these are collectively the key findings of NASH (ductular proliferation and extramedullary erythropoiesis were also observed, but are not features of human NASH). Liver histology was normal in low-responder opossums fed the same diet. So the most salient feature of this novel animal model of NASH is that it does not depend solely on genetic predisposition (like db/db and ob/ob mice) nor on diet alone (e.g., 2% cholesterol HCHF diets in rats or mice). Rather, like humans with NASH (10, 19), it depends on two factors: a genetically predisposed host, who becomes exposed to one or more environmental factors that were not the norm before 1980. In this case, the HCHF diet contained 0.7% cholesterol, considerably less than the 2% cholesterol diets (often with cholic acid) that are toxic to livers of rats and mice (2, 11, 16, 20, 22, 23), not surprisingly as the amount of dietary cholesterol is roughly equivalent to 20 kg/day for humans ( 100 hamburgers?) (11). These are not the first data that incriminate cholesterol as a potential mediator of lipotoxicity in NASH (1, 4, 11, 22, 25). In fact, the few existing human lipidomic data are entirely consistent with one or more cholesterol fractions [free cholesterol (FC), oxysterols, cholesterol esters] associating with the NASH phenotype of NAFLD, in fact more consistently so than for any other lipid class (4, 25). Thus Puri and colleagues (25) found that total cholesterol was higher in NASH livers than in NAFLD without NASH; free fatty acids (FFA) and other lipid fractions were not. Using nonquantitative fillipin fluorescence, Caballero et al. (4) also confirmed that FC is present in abundance in human NASH livers. Until recently, the favored “lipotoxic mediator” in NASH has been either FFA, or lysophosphatidylcholine, or other phospholipid metabolites (12); for example, exposing hepatocytes to palmitic acid causes cell death via formation of lysophosphatidylcholine (15). This made sense because the dominant source of fatty acids in livers of obese patients with NASH is from lipolysis of peripheral tissues, such as adipose (7, 9), whereas saturated FFA cause lipotoxic injury to many cell types, including those of hepatic lineage (15, 24). However, an “inconvenient truth” is that there is no difference in hepatic FFA between NASH and not-NASH NAFLD in the few studies in which such data are available (1, 25). There is other evidence that cholesterol could be implicated in NASH pathogenesis, other than 2% cholesterol high-fat (HF) diet studies, which may, like methionineand cholinedeficient mice that we popularized 16 yr ago (10), really be a model of “toxic steatohepatitis”. Llacuna et al. (20) recently demonstrated that 2% dietary cholesterol causes mitochondrial GSH depletion, rendering mice highly susceptible to ischemiareperfusion injury. Furthermore, Dutch workers from the labAddress for reprint requests and other correspondence: G. Farrell, Gastroenterology and Hepatology Unit, The Canberra Hospital, PO Box 100, Woden, ACT 2605, Australia (e-mail: [email protected]). Am J Physiol Gastrointest Liver Physiol 303: G9–G11, 2012; doi:10.1152/ajpgi.00008.2012. Editorial Focus