Nonalcoholic steatohepatitis (NASH) represents an advanced stage of fatty liver disease developed in the absence of alcohol abuse. Its increasing prevalence in western countries, the diagnostic difficulties by noninvasive tests, and the possibility of progression to advanced fibrosis and even cirrhosis make NASH a challenge for hepatologists. NASH is frequently associated with type 2 diabetes and the metabolic syndrome, and several genetic and acquired factors are involved in its pathogenesis. Insulin resistance plays a central role in the development of a steatotic liver, which becomes vulnerable to additional injuries. Several cyclic mechanisms leading to self-enhancement of insulin resistance and hepatic accumulation of fat have been recently identified. Excess intracellular fatty acids, oxidant stress, tumor necrosis factor-α, and mitochondrial dysfunction are causes of hepatocellular injury, thereby leading to disease progression and to the establishment of NASH. Intestinal bacterial overgrowth also plays a role, by increasing production of endogenous ethanol and proinflammatory cytokines. Therapeutic strategies aimed at modulating insulin resistance, normalizing lipoprotein metabolism, and downregulating inflammatory mediators with probiotics have promising potential.
Nonalcoholic steatohepatitis (NASH) is a liver disease characterized by steatosis and periportal and lobular inflammation. In its initial phases, during which fat accumulates in the liver, no clinical symptoms are evident. In advanced stages, fibrosis (eventually progressing to established cirrhosis in some patients) is detectable histologically, along with a mixed inflammatory cell infiltrate, glycogen nuclei, and Mallory’s hyaline (1). Because its adequate diagnosis requires histological evaluation of the liver, the prevalence of NASH is probably underestimated. Nonalcoholic fatty liver disease has been suggested to be the most common cause of chronic liver disease in the U.S., with a suggested incidence of 10–24% in the general population and probably similar figures in Europe and Japan (2,3).
Two types of NASH exist: primary NASH (which is associated with metabolic syndrome–related conditions, such as obesity, type 2 diabetes, and hyperlipemia) and secondary NASH (which occurs after obesity-related intestinal surgery, rapid weight loss in the obese, total parenteral nutrition, treatment with drugs such as amiodarone or perhexiline maleate, lipodystrophy, or Wilson’s disease). Although many aspects of the disease are common to both presentations, this short review focuses mainly on the pathogenesis of primary NASH.
The actual prevalence of NASH in type 2 diabetes and obesity is unknown. It is estimated that 75% of type 2 diabetic patients present some form of nonalcoholic fatty liver of different degrees. An association of NASH with hyperinsulinemia, as well as with clinical features of insulin resistance, has frequently been reported (4–9). As far as obesity is regarded, steatosis has been reported in 70% of obese and 35% of lean patients and NASH in 18.5% of obese and 2.7% of lean patients in a consecutive study (10), although some authors have reported even higher figures (up to 95% in some studies ). The prevalence of simple steatosis in obese patients is ∼60%, whereas 20–25% present NASH and 2–3% present cirrhosis (7,12,13).
Criteria for diagnosis of NASH were established at the American Association for the Study of Liver Diseases single topic conference on NASH (1). In most cases, asymptomatic patients are referred to the hepatologist after abnormal liver enzyme levels obtained during routine evaluation or during antihyperlipidemic drug therapy. Because clinical signs and liver test values have a poor predictive value for making a specific diagnosis, histological evaluation of morphological changes in a liver biopsy may be required, in particular, to differentiate between simple steatosis and steatohepatitis. The presence of obesity or type 2 diabetes, high (at least two times that of normal) levels of alanine aminotransferase (ALT) and triglycerides, hypertension, and an aspartate aminotransferase/ALT ratio greater than unity may justify performing a biopsy, since prognostic information is greater in these subgroups of patients (12,14–16). A standardized scoring system for nonalcoholic fatty liver disease has been published (17).
According to Day et al. (14,18), the pathogenesis of NASH comprises two steps. First, the healthy liver becomes steatotic. This is mainly a consequence of peripheral resistance to insulin, whereby the transport of fatty acids from adipose tissue to the liver is increased. Although some protective mechanisms are developed to survive this stress, the fatty liver is in most cases particularly fragile and vulnerable to additional insults, such as ethanol or bacterial lipopolysaccharide. Then, a second step elicited by oxidative stress and cytokines (basically, tumor necrosis factor [TNF]-α) occurs. This leads to exacerbation of insulin resistance, further oxidative stress, and organelle dysfunction within liver cells, resulting in an inflammatory process, hepatocellular degeneration, and fibrosis.
Insulin resistance and development of steatosis
A combination of genetic and acquired factors contribute to the first hit–originating liver steatosis through increased lipolysis and delivery of free fatty acids (FFAs) to the liver (Fig. 1). Insulin resistance plays a primary role: it is the most specific finding in NASH, and it can be observed in most patients, whether they present overweight or with type 2 diabetes (4,9). This primary resistance leads to hyperinsulinemia and is associated with the metabolic syndrome (4,9,19), of which NASH is considered to represent the hepatic expression (20). The severity of insulin resistance has been shown to parallel the severity of fatty liver disease, with clinically overt type 2 diabetes being most common in NASH patients with cirrhosis (21). In this regard, the presence of diabetes has been suggested to be useful for the identification of those elderly NASH patients who might have severe liver fibrosis (16). These findings further support the notion that insulin resistance plays a pathogenetic role in NASH.
Several mechanisms for the development of peripheral resistance to insulin have been described, including acquired and inherited factors (22–24) (Fig. 1). A breakthrough in this regard was the discovery that chronic stimulation of IκΒ kinase β (IKKβ), which promotes activation of nuclear factor (NF)-κB (a transcription factor involved in inflammatory cytokine production) also causes insulin resistance (25). This is caused by IKKβ-mediated changes in phosphorylation (in serine rather than in tyrosine) of insulin receptor substrate (IRS)-1, thus disrupting the intracellular signaling triggered by binding of insulin to its receptor (25).
Interestingly, IKKβ is activated by two main types of stimuli (Fig. 1). The first is increased hepatic oxidative stress. This may be caused, among other factors, by an increased mitochondrial, peroxisomal, and/or microsomal oxidation of FFAs as a result of either environmental factors (fat diets, lipopolysaccharide, ethanol, and drugs) or genetic factors (β-oxidative defects). The second type of stimuli is proinflammatory cytokines, particularly TNF-α. This permits perpetuation of the cycle, whereby TNF-α activates IKKβ, which in turn induces TNF-α production. TNF-α stimulation of IRS-1 serine phosphorylation may also be mediated by c-Jun NH2-terminal kinase and several protein kinase C isoforms (1). Data supporting the major role played by TNF-α in the development of insulin resistance come from studies with TNF-α knockout mice, which do not develop insulin resistance in response to obesity induction (26). Notably, TNF-α expression in adipose tissue and liver is augmented in NASH patients, whose serum TNF-α levels correlate with insulin resistance (27).
As mentioned, the diet may be responsible for steatosis and oxidative stress in some liver diseases (28). Evidence exists demonstrating that the diet of NASH patients is rich in unsaturated fat and cholesterol but poor in polyunsaturated fat, fiber, and vitamins E and C compared with that of healthy subjects. These levels of unsaturated fat in the diet correlate with a lower sensitivity to insulin, with high postprandial triglyceride levels in these patients, and with other aspects of the metabolic syndrome (29).
As a consequence of insulin resistance, the lipogenic effects of this hormone on adipose tissue are modulated, resulting in the degradation of triglycerides into FFAs, which are then released into the blood flow (30). This situation is similar to that observed in overweight patients undergoing a fast and disproportioned weight loss (10). There are indications that visceral and central adipose tissues play a more critical role than peripheral adipose tissue in FFA release and steatotic liver formation (4,19). The reason might be a more direct access to the liver through the portal system for fat derived from those tissues, as well as a lower secretion of leptin (31).
This situation is extreme in lipodystrophies—genetic diseases characterized by absence of or insensitivity to leptin. In these patients, adipose tissue is not developed, fat accumulates in organs such as the liver (thereby originating a secondary steatosis ), and peripheral resistance to insulin occurs (22). Leptin is a hepatocyte-derived 16-kDa peptidic hormone mainly involved in lipid metabolism regulation. It precludes fat accumulation in nonadipose tissues by inhibiting lipogenesis and potentiating lipolysis after excessive caloric intakes (33,34). Controversial data have been reported on the involvement of leptin in the pathogenesis of NASH. Uygun et al. (35) found increased serum leptin levels in NASH patients and suggested that leptin might represent an independent predictive factor of the intensity of liver steatosis in NASH, along with C-peptide (which reflects pancreatic insulin secretion) and age (36). However, these results might have been influenced by the fact that subjects in the control group had a lower BMI than NASH patients. The apparent paradox of the correlation between leptin levels and liver steatosis (in principle, leptin reduces fat accumulation in the liver) could be explained by the development of resistance to leptin actions or by a close relation between leptin levels and insulin resistance (4). In this respect, leptin inhibits insulin-induced IRS-1 phosphorylation (37). However, in a separate clinical study that used appropriately matched control subjects (BMI, percent body fat, and abdominal fat distribution), no association between serum leptin levels and NASH could be observed (NASH: 21 ± 13 ng leptin/ml [n = 26]; control group: 18 ± 11 ng leptin/ml [n = 20]; P = 0.5) (38). Furthermore, there was no correlation between serum leptin and hepatic histology, serum transaminases, fasting insulin levels, or a measure of insulin resistance. Although these data suggest that leptin may not be relevant in NASH, additional studies with larger cohorts of patients and an improved control for factors such as interindividual leptin levels and the absence of liver pathology in control subjects could help to clarify the involvement of this hormone in the pathogenesis of NASH.
Exaggerated levels of FFAs may be deleterious for the liver through a variety of mechanisms (1), including 1) de novo synthesis of ceramides, which may cause apoptosis; 2) resistance to insulin, by interfering with intracellular phosphorylation processes; and 3) lipid peroxidation. Under normal conditions, the liver is prepared to defend against FFA-induced toxicity through their sterification and triglyceride formation, their oxidation, or the synthesis and release of VLDLs. Nuclear peroxisome proliferator–activated receptor (PPAR)-α plays an important role in these processes, by sensing excess FFAs and upregulating the genetic program of their disposal (39). However, under hyperinsulinemic conditions, the following processes will take place: 1) the expected increase in FFA sterification and triglyceride formation, but also 2) an increase in glycolysis and fatty acid synthesis, 3) an inhibition of oxidation (similarly to amiodarone and perhexyline maleate) (40), and 4) a reduced release of triglycerides in the form of VLDL (41). Apolipoprotein B secretion, necessary to form VLDL, has been shown to be reduced in NASH (29,42). Besides, NASH patients show hypertriglyceridemia in both fasting and postprandial situations, resulting in a higher fat upload to the liver (29,42,43). These alterations in lipid metabolism also contribute to making a healthy liver become steatotic.
The steatotic liver is a particularly susceptible organ to become resistant to insulin (19). However, the mediating mechanisms are not fully understood: apart from some of those already mentioned (TNF-α, NF-κB–mediated FFA effects), PPAR-α–mediated inhibition by polyunsaturated fatty acids of insulin-induced lipogenic and glycolytic enzymes may play a role (44). Leptin might also regulate hepatic sensitivity to insulin (4,36,37).
Injury to the steatotic liver and progression of liver disease
Although in most patients the process does not progress, a minority of subjects undergo a second stage, characterized by inflammation and hepatocellular degeneration. Steatosis renders hepatocytes vulnerable against external aggressions, and apoptosis becomes a frequent mechanism of cell death (45,46). Recent results obtained in a cohort of 264 prospectively enrolled patients with nonalcoholic fatty liver disease indicate that insulin resistance is a major, independent risk factor for advanced fibrosis, thus acting both as the first and second hits (47).
The reasons for progression of the disease are not fully understood, although oxidative stress and cytokines are the main effectors of this second hit on the vulnerable liver (Fig. 2). Furthermore, a certain genetic predisposition has been suggested to play a role in the development of NASH.
Peripheral resistance to insulin and high levels of leptin allow entrance to the mitochondria of those FFAs reaching the liver as a consequence of the previously inhibited oxidation (14). Although oxidation of long-chain and very-long-chain fatty acids is partly extramitochondrial (in microsomes and peroxisomes), free oxygen radical production occurs mainly in mitochondria (48,49). Massive FFA hepatic upload, and particularly acyl-CoA, lead to PPAR-α–mediated activation of the synthesis of enzymes responsible for oxidation, thereby increasing peroxide levels (14,50). Formation of free oxygen radicals in a fat-rich medium induces lipidic peroxidation. Increased oxidative stress and lipidic peroxidation induce damage in plasmatic membranes, intracellular organelles, mitochondrial DNA, and respiratory chain-related proteins (51). Additionally, the end products of oxidative stress activate NF-κB–mediated nitric oxide synthesis, leading to the formation of peroxynitrites (13). FFAs also increase the expression of microsome oxidases CYP4A and CYP2E1, responsible for the production of hydroxyethil radicals (52–54). Because CYP2E1 is inhibited by insulin, its expression levels are higher in case of peripheral resistance to this hormone (55).
Another consequence of the increased production of free oxygen radicals is the induction of Fas ligand expression in hepatocellular membranes (not expressed under normal conditions), since its promoter contains a binding site for NF-κB. Interaction of Fas ligand with Fas-expressing hepatocytes leads to their death through a process termed “fratricidal apoptosis” (49).
Mitochondrial dysfunction seems to be a critical event in the second hit of NASH. Whether induced by lipidic peroxidation products secondary to oxidative stress or directly by TNF-α, it leads to alterations in electron transfer along the respiratory chain, thus generating more free oxygen radicals (49). The expression of mitochondrial oxidative phosphorylation uncoupling protein 2 is increased. This protein reduces free oxygen radical synthesis, but it also decreases ATP levels, thus making the cell more sensitive to insults (46,56,57) and facilitating hepatocellular apoptosis and necrosis. Decreased ATP synthesis has been reported in NASH patients after perfusion with fructose (58). Nevertheless, other authors question the detrimental potential of uncoupling protein 2 (59). Electron microscopic evaluation of liver biopsies from NASH patients has shown para-crystaline inclusions in 10–30% of hepatocellular mitochondria (6). Although the significance of this finding is unknown, these inclusions are observed in patients with myopathies associated with mitochondrial DNA alterations that affect respiratory chain enzymes (60) and might therefore represent a genetic basis determining the evolution of fatty liver in NASH (6).
TNF-α also contributes to the second hit in NASH pathogenesis. An increased TNF-α synthesis by hepatocytes and Kupffer cells may be caused by 1) NF-κB–mediated FFA oxidation-induced oxidative stress (61) or 2) endotoxemia resulting from intestinal bacterial overgrowth (see below). TNF-α has several modes of action: 1) it induces resistance to insulin, thus producing increased levels of FFAs; 2) it uncouples mitochondrial respiration, thereby inducing oxygen radical formation, similarly to amiodarone and perhexyline maleate (40); and 3) it induces hepatocyte apoptosis and necrosis (62).
NASH has been suggested to have an inherited component, involving genes related to 1) determination of sensitivity to insulin (22,63); 2) hepatic lipid storage, oxidation, or release into the blood flow (64); 3) obesity and its distribution (65,66); 4) regulation of hepatic iron levels and oxidative stress generation (67–69); or 5) cytokine synthesis (8,70,71).
Iron overload may play a role in NASH pathogenesis. An association between iron overload and the metabolic syndrome (68,72–75) or advanced liver disease (76) has been demonstrated. Iron catalyzes the transformation of hydrogen peroxide into hydroxyl groups through the Fenton reaction. Nevertheless, the relevance of mutations in the HFE gene and serum iron levels in the pathogenesis of NASH is a matter of discussion. Some authors have found higher prevalence of mutations in the HFE gene along with elevated hepatic iron and more advanced stages of fibrosis in NASH patients (67,69), whereas others have failed to identify an excess of iron and its supposed fibrogenic action (16,47,68,74,77) in these patients.
Additional mechanisms have been described to potentially contribute to fibrogenesis in NASH. A marked expression of connective tissue growth factor, which correlates with fibrosis (78), has been shown. Interestingly, glucose and insulin potentiate its synthesis. Hepatic stellate cell–derived leptin might stimulate fibrogenesis through an autocrine action and through the induction of transforming growth factor (TGF)-β1 by Kupffer and endothelial cells (79,80). However, serum leptin levels are not associated to more advanced stages of fibrosis (36). Finally, the expression of the TGF-β1 receptor endoglin is increased in sinusoidal endothelial cells and probably in hepatic stellate cells of NASH patients (13).
Role of intestinal bacterial overgrowth in the pathogenesis of NASH
A clear link between intestinal bacterial overgrowth and liver damage during NASH has recently been established (81). Bacterial overgrowth has been detected in NASH patients with breath tests with lactulose and d-xilose (82), as well as in some forms of secondary NASH, such as that associated with obesity-related intestinal surgery (83).
Studies about the pathogenesis of alcoholic fatty liver disease demonstrated protection of the liver from ethanol when intestinal bacterial overgrowth was inhibited (84,85). The reason was the reduced hepatic exposure to bacterial lipopolysaccharide (LPS). Intestinal bacteria may increase hepatic oxidative stress by at least two mechanisms (81): 1) increased endogenous ethanol production and 2) release of LPS. Both ethanol and LPS stimulate inflammatory cytokine production through an IKKβ-mediated mechanism (86), with Kupffer cells (a cell type that plays a critical role in NASH) as the main source of TNF-α. Because TNF-α is a central mediator in the pathogenesis of NASH, inhibition of bacterial overgrowth was also shown to result in protective mechanisms in this disorder (49). Further support for these effects is provided by evidence that Kupffer cells are oversensitive in NASH, as well as in obesity, probably because of leptin actions (87–89).
Ethanol synthesized by intestinal bacteria might be involved in NASH pathogenesis, since obese female NASH patients present higher levels of breath ethanol (90). This is supported by data obtained from leptin-deficient obese mice (ob/ob mice), regarded as a good experimental model for human NASH (91). These mice show a reduced endogenous ethanol production after treatment with neomycin to eliminate bacterial overgrowth (92). Other authors, however, propose a protective role for low doses of alcohol, by reducing resistance to insulin and inhibiting TNF-α synthesis by monocytes (7,93).
THERAPEUTIC ALTERNATIVES FOR NASH
At present, there is no widely accepted approach to treat NASH. The various therapeutic alternatives, as follows, are aimed at interfering with the risk factors considered to be involved in the etiology of NASH.
Correction of obesity with hypocaloric diets and physical exercise (94–97). Rapid weight loss and long-lasting fasting periods should be avoided, since they lead to an increase in the flow of FFAs to the liver. A gradual weight reduction has been associated with an improvement of hepatic lesions, including fibrosis (98).
Control of hyperglycemia with diet, insulin, or oral antidiabetic agents. Simultaneous treatment of overweight in these patients is of paramount importance.
Withdrawal from treatment with amiodarone, perhexiline maleate, tamoxifen, or other drugs to which NASH development has been attributed. Likewise, exposure to hepatotoxic environmental agents, including alcohol, should be avoided, particularly where fibrosis is histologically detected in a biopsy.
Control of hyperlipemia with diet, or, when indicated, with hypolipemic drugs. Nevertheless, the efficacy of this measure is controversial, according to a study in hypertriglyceridemic patients who received clofibrate (2 g/day) for 1 year, showing no biochemical or histological improvement (99). Gemfibrozil (600 mg/day) or bezafibrate showed more favorable results in terms of biochemical parameters and development of steatosis (100,101). Orlistat, an inhibitor of lipoprotein lipase, has been recently proven to be beneficial for NASH patients, inducing normalization of transaminases and reduction in liver steatosis and inflammatory activity (102).
In parenteral nutrition-associated NASH, modifying the composition of the infusion, replacing glucose with lipids. Glucose stimulates insulin secretion, thus inhibiting FFA oxidation and leading to their accumulation and synthesis in the liver. Supplementation with choline is indicated to increase the synthesis of lecitin, necessary for VLDL formation (98).
In patients undergoing surgery to treat obesity, reconstructing intestinal transit to help improve hepatic lesions. Metronidazol may prevent the development of NASH by preventing the absorption of bacterial overgrowth-derived endotoxin in excluded loops (1,98).
Progression of NASH histological lesions is not always precluded by measures that target the etiological factors of NASH, as those mentioned above. Therefore, alternative treatments directed against specific pathogenetic mechanisms are currently under investigation, including clinical trials with anti–TNF-α antibodies (103).
Because bacterial overgrowth–derived lipoproteins may be involved in the development of NASH, oral metronidazol (0.75–2 g/day for 3 months, followed by a similar period without treatment) may be efficacious in reverting steatosis and, in some cases, inflammation and fibrosis (104,105). Oral polymixin B may improve parenteral nutrition–associated NASH by reducing liver exposure to intestinal flora-derived endotoxin.
The knowledge of the role of bacterial overgrowth in the pathogenesis of NASH has led to the proposal of probiotics as a therapeutic strategy for this disorder. Probiotics may interfere with the development of NASH at various levels: 1) decreases in proinflammatory cytokines, such as TNF-α; 2) alteration of the inflammatory effects of pathogenic strains of intestinal bacteria, through changes in cytokine signaling; 3) replacement of pathogenic strains of bacteria; and 4) improved epithelial barrier function (thereby avoiding excessive exposure of the liver to LPS and bacterial ethanol). Evidence in experimental animal models of fatty liver disease (103), as well as clinical data on other gastrointestinal diseases, strongly suggest that probiotics might be beneficial in NASH (81). Data from an uncontrolled clinical trial with NASH patients show promising results, with improvement of liver enzymes in treated patients (106).
Cytoprotective agents and antioxidants
Several agents are included in this group, including ursodeoxycholic acid, vitamin E, lecitin, β-carotene, selenium, S-adenosyl-methionine, metadoxine, or silimarin.
Ursodeoxycholic acid has several mechanisms of action that justify its use in NASH: hydrophilic effect (resulting in the displacement of toxic hydrophobic biliary salts), and immunomodulatory and cytoprotective properties. An oral dose of 13–15 mg/day for 12 months was efficacious in improving liver biochemistry alterations and steatosis, although no favorable changes occurred in the rest of the histological lesions of NASH (99). However, recently reported results of a randomized multicenter study in which NASH patients received between 13 and 15 mg · kg−1 · day−1 of ursodeoxycholic acid for 2 years showed no improvement of liver disease with respect to placebo-treated patients (107).
The therapeutic indication of S-adenosyl-methionine in intrahepatic cholestasis and alcoholic liver disease is based on its anti-steatotic, anti-inflammatory, anti-oxidant, and anti-fibrotic properties. Oral treatment with 600 mg/day or intramuscular administration of 50–100 mg/day have shown efficacy in terms of biochemical, histological, and echographic parameters of liver steatosis, in the absence of adverse effects (108).
α-Tocoferol is effective in improving liver biochemistry and histological lesions of NASH because of its actions as an antioxidant agent and as an inhibitor of TGF-β, a cytokine involved in liver fibrogenesis (109).
Metadoxine has proved efficacious in the treatment of alcoholic liver steatosis, as shown by biochemical data and echographical signs (110). This drug restores hepatic glutathione concentrations and acts as an antifibrogenic agent. These therapeutic effects, along with the proven efficacy in steatosis, may justify its indication for the treatment of NASH.
Silymarin also possesses antioxidant and antifibrogenic properties, with beneficial effects in alcoholic liver disease (111), supporting its indication for the treatment of NASH. Betaine treatment has shown beneficial biochemical and histological effects in a pilot study of NASH patients (112). Additional drugs currently in assessment include ghrelin (113) and pentoxyphylline (114). Other promising, potentially useful antioxidant agents include vitamin E and N-acetylcysteine.
Reduction of peripheral resistance to insulin
The thiazolidinediones are compounds that improve insulin sensitivity by binding the PPAR-γ class of nuclear transcription factors. Troglitazone led to a decrease in the transaminase levels in obese NASH patients (115), although no weight reduction was observed. Nevertheless, troglitazone was later discarded because of its risk for severe liver toxicity. Second-generation agents such as rosiglitazone have also shown some efficacy in improving liver enzyme levels and histology (116). In a pilot study in which 30 NASH patients were treated with rosiglitazone, 4 mg twice daily for 48 weeks, reductions in liver fat content and in mean ALT (from 86 to 37 units/l) were already observed by 24 weeks of treatment, along with the expected improvement (although not complete normalization) in insulin sensitivity and an acceptable tolerance profile (117). Whether the observed liver effects were caused by improved insulin sensitivity or by the anti-inflammatory actions of this class of compounds remains to be elucidated. Similarly, pioglitazone has also been tested in a pilot study with 18 nondiabetic NASH patients (118). Administration of a daily dose of 30 mg for 48 weeks resulted in normalization of ALT levels in 72% of patients. Hepatic fat content and size, as well as glucose and FFA sensitivity to insulin, were consistently improved, as well as histological signs of steatosis. Treatment side effects were manageable. Metformin has also been tested as a therapy for NASH. In addition to improving hyperinsulinemia and insulin sensitivity in animals and humans (119), metformin inhibits hepatic TNF-α and several TNF-inducible responses, which, as stated above, are likely to promote hepatic steatosis and necrosis. When administered to insulin-resistant obese ob/ob leptin-deficient mice, metformin reduced both hepatic steatosis and hepatic TNF-α expression (120). In humans, treatment of 14 NASH patients with 500 mg t.i.d. metformin for 4 months resulted in normalization of transaminase levels, improvement of insulin sensitivity, and a decrease in liver volume in 50% of patients (121). In a recent clinical study, 17 NASH patients were given 850 mg b.i.d. metformin plus dietary treatment and were compared with a similar group with calorie-restricted dietary treatment alone. Metformin statistically significantly improved serum alanine/aspartate aminotransferase levels as well as insulin resistance, whereas it decreased insulin and C-peptide levels (122). However, no significant differences between groups were observed in terms of necro-inflammatory activity or fibrosis. Although larger studies are needed, these data suggest that metformin could be a promising agent for the treatment of NASH patients.
Reduction of liver iron content
Because iron deposits are associated with a higher intensity of liver damage in NASH patients, repeated phlebotomies might prevent the development of lesions, delay their evolution, or even reduce or eliminate those already established (69).
Liver transplantation is indicated for NASH patients with decompensated cirrhosis (123). However, some cases of disease recurrence have been reported in female NASH patients and in cases of jejunum-ileum derivation in which the intestinal transit was not restored by eliminating the dysfunctional loop, simultaneously to transplantation.
In summary, the pathogenesis of NASH is multifactorial: the development of steatosis seems to be mainly determined by insulin resistance, whereas the causes of hepatocellular injury include factors such as excess intracellular fatty acids, mitochondrial dysfunction, oxidant stress, cytokines, iron overload, bacterial overgrowth, and genetic predisposition. At present, therapeutic strategies targeting the various etiologic and pathogenetic mechanisms are being investigated, including behavioral and pharmacological approaches to insulin resistance, cytoprotective agents, antioxidant agents, iron reduction therapy, and antihyperlipidemic agents.
Financial support was obtained from a grant (C03/02) from the Instituto de Salud Carlos III, Spain, SAF 2001-1414 from Ministerio de Ciencia y Tecnología, Spain (to R.M.O.), and a grant (02/3015) from Fondo de Investigaciones Sanitarias, Spain (to J.M.).
The authors are grateful to Brenda Ashley for assistance with English.
J.M. and L.I.F.-S. contributed equally to this work.
A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.