Potential conflict of interest: Nothing to report.
S-Adenosylmethionine (SAMe) has rapidly moved from being a methyl donor to a key metabolite that regulates hepatocyte growth, death, and differentiation. Biosynthesis of SAMe occurs in all mammalian cells as the first step in methionine catabolism in a reaction catalyzed by methionine adenosyltransferase (MAT). Decreased hepatic SAMe biosynthesis is a consequence of all forms of chronic liver injury. In an animal model of chronic liver SAMe deficiency, the liver is predisposed to further injury and develops spontaneous steatohepatitis and hepatocellular carcinoma. However, impaired SAMe metabolism, which occurs in patients with mutations of glycine N-methyltransferase (GNMT), can also lead to liver injury. This suggest that hepatic SAMe level needs to be maintained within a certain range, and deficiency or excess can both lead to abnormality. SAMe treatment in experimental animal models of liver injury shows hepatoprotective properties. Meta-analyses also show it is effective in patients with cholestatic liver diseases. Recent data show that exogenous SAMe can regulate hepatocyte growth and death, independent of its role as a methyl donor. This raises the question of its mechanism of action when used pharmacologically. Indeed, many of its actions can be recapitulated by methylthioadenosine (MTA), a by-product of SAMe that is not a methyl donor. A better understanding of why liver injury occurs when SAMe homeostasis is perturbed and mechanisms of action of pharmacologic doses of SAMe are essential in defining which patients will benefit from its use. (HEPATOLOGY 2007;45:1306–1312.)
Although S-adenosylmethionine (SAMe, also abbreviated as AdoMet and SAM)1 was discovered approximately 50 years ago, its story began in 1890 with Wilhelm His. When he fed pyridine to dogs, he was able to isolate N-methylpyridine from the urine and emphasized the need to demonstrate both the origin of the methyl group as well as the mechanism of its addition to the pyridine.2 Both questions were addressed by Vincent du Vigneaud, who, during the late 1930s, demonstrated that the sulfur atom of methionine was transferred to cysteine through the “transsulfuration” pathway, that is, the change of methyl groups from methionine, choline, betaine, and creatine. In 1951, Cantoni demonstrated that a liver homogenate supplemented with adenosine triphosphate (ATP) and methionine converted nicotinamide to N-methylnicotinamide. Two years later, he established that methionine and ATP reacted to form a product, which he originally named “active methionine,” capable of transferring its methyl group to nicotinamide or creatine in the absence of ATP. After determination of its structure, he called it SAMe (Fig. 1). Subsequently, Cantoni and his colleagues discovered methionine adenosyltransferase (MAT, the enzyme that synthesizes SAMe), S-adenosylhomocysteine (SAH, the product of the transmethylation reactions), and SAH-hydrolase [the enzyme that converts SAH to homocysteine (Hcy) and adenosine]. At about the same time, Benett discovered that folate and vitamin B12could replace choline as a source of methyl groups in rats maintained on diets containing Hcy in place of methionine, a finding that led to the discovery of methionine synthase (MS). In 1961, Tabor demonstrated that the propylamino moiety of SAMe is converted, via a series of enzymatic steps, to spermidine and spermine. In the biosynthesis of polyamines, 5′-deoxy-5-methylthioadenosine (MTA) was identified as an end product. Thus, by the end of the 1960s, Laster's group could finally provide an integrated view, similar to that depicted in Fig. 2, combining the transmethylation and transsulfuration pathway with polyamine synthesis and folate metabolism.3
Since then, SAMe-binding proteins have been identified with functions that vary from methylation of DNA, RNA, histones and other proteins, phospholipids, and small molecules such as arsenic to the synthesis of polyamines, radical formation, and binding to mRNA riboswitches.3 These reactions can affect a wide spectrum of biological processes ranging from metal detoxification and catecholamine metabolism to membrane fluidity, gene expression, cell growth, differentiation, and apoptosis, to establish what Cantoni called the “SAMe empire.”4
In mammals there are three distinct enzymes that synthesize SAMe: MATI, MATII, and MATIII. MATI and MATIII are the gene products of MAT1A, whereas MATII is the gene product of MAT2A.4 SAMe synthesis and utilization occur mainly in the liver via MATI/III and glycine N-methyltransferase (GNMT), the main enzymes responsible for the synthesis and removal of hepatic SAMe, respectively.4, 5
In adults, MAT1A is expressed exclusively in the liver and pancreas, whereas MAT2A is expressed in all tissues, including the liver. In fetal rat liver, Mat1a expression increases progressively after birth, and reaches a peak at 10 days of age. Conversely, Mat2a expression decreases after birth, reaching a minimum in the adult liver (approximately 5% that of Mat1a).4 Because of differences in the regulatory and kinetic properties of the various MATs, MATII cannot maintain the same high levels of SAMe compared with the combination of MATI and MATIII.4 Consequently, in Mat1a (−/−) mice, despite a significant increase in Mat2a expression, the liver content of SAMe is reduced approximately 3-fold from birth, when the switch from Mat2a to Mat1a expression takes place.6
The most abundant methyltransferase in mammalian liver is GNMT, comprising approximately 1% of the soluble protein in rat liver.7 GNMT is also present in large amounts in the exocrine pancreas and the prostate. The importance of GNMT is to maintain a constant SAMe/SAH ratio, which is considered to be an indicator of the methylation capacity in the cell.8 Consequently, in Gnmt (−/−) mice, despite the existence of many other methyltransferases,9 the liver content of SAMe is increased approximately 25-fold.10 GNMT is also a major folate binding protein in rat liver,11 because 5-methyl-tetrahydrofolate (5-MTHF) is a potent inhibitor of GNMT activity.12, 13 This, together with the finding that SAMe is a potent inhibitor of MS activity, an inhibitor of betaine homocysteine methyltransferase (BHMT) activity and expression, and an activator and stabilizer of cystathionine β-synthase (CBS, the first step of the transsulfuration pathway)14 has established the role of GNMT as an important factor in the control of methionine metabolism.15
Hcy is formed on demethylation of SAMe and subsequent hydrolysis of SAH and lies at the junction of two intersecting pathways, the transsulfuration pathway, which converts the sulfur atom of methionine to cysteine and glutathione, and the remethylation pathway, which conserves Hcy as methionine and is coupled to the folate cycle (Fig. 2).16 Hcy belongs to a group of molecules known as cellular thiols. The most abundant cellular thiols are glutathione (the intracellular concentration is approximately 1-10 mM) and cysteine (plasma total cysteine ranges from 200 to 300 μM). These two thiols are considered the “good thiols,” because their functions include maintaining intracellular and extracellular redox homeostasis, facilitate the removal of toxic compounds, and are part of the cellular antioxidant defense system.16 Hcy is normally found at much lower concentrations than glutathione and cysteine. The concentration of Hcy within the cell is approximately 1 μM, and in plasma ranges from 5 to 15 μM. Different from glutathione and cysteine, Hcy is considered a “bad thiol” because its elevation in the plasma is correlated with complex diseases, such as cardiovascular and liver disease, end-stage renal disease, neural tube defects, and cognitive dysfunctions including Alzheimer disease.16
Folic acid, vitamin B12, and B6 are three B vitamins that regulate Hcy content and methionine metabolism (Fig. 2). Mice deficient in Folbp1 (−/−), a folate-binding protein that functions as a membrane receptor to mediate the high-affinity internalization and delivery of folate to the cytoplasm of the cell, have elevated plasma Hcy levels.17 Similarly, mice deficient in Mthfr (−/−), the enzyme that catalyzes the conversion of 5,10-methylenetetrahydrofolate to 5-MTHF, the primary circulating form of folate and the methyl donor for Hcy remethylation to methionine (Fig. 2), have elevated Hcy plasma levels.18 Vitamin B12 is a required co-factor for MS so that B12 deficiency can also lead to Hcy accumulation. Finally, vitamin B6 is a co-factor required for both CBS and cystathionine (Fig. 2). Mice deficient in Cbs (−/−), the enzyme that catalyzes the conversion of Hcy to cystathionine (Fig. 2), also have elevated Hcy plasma levels.19
Hcy is in equilibrium with SAH, but under normal conditions this reaction is sufficiently rapid to maintain the flux in the direction of hydrolysis. SAH is an inhibitor of many SAMe-dependent methyltransferases.8, 9 Because this inhibition is competitive, for any given methyltransferase, the effect on activity will depend on the Km for SAMe, the Ki for SAH, and the concentrations of these two metabolites. Thus, SAH hydrolysis serves physiologically not only to synthesize Hcy but also as a key regulator in biological methylations.
Role of SAMe in Liver Injury
It has long been recognized that methionine metabolism is altered in patients with liver disease. In 1948, Kinsell et al.20 demonstrated marked impairment in methionine clearance after a methionine load in patients with liver injury, thereby establishing the central role of the liver in methionine metabolism. Patients with liver cirrhosis have decreased SAMe biosynthesis because of decreased expression of MAT1A and lower hepatic MAT activity,21, 22 and treatment with SAMe increased hepatic GSH levels and increased survival in patients with less advanced alcoholic liver cirrhosis.4, 23 When rats and mice are fed a diet deficient in methyl groups (choline, methionine), the liver develops steatosis within a few weeks, and if the diet continues, the liver develops steatohepatitis, with some animals developing HCC.24 Accordingly, we have shown, using the Mat1a (−/−) mouse model, that chronic hepatic SAMe deficiency predisposes the organ to further injury, spontaneous development of steatohepatitis, and HCC.25–27 The molecular mechanisms identified so far to explain the phenotype include increased oxidative stress, upregulation of cytochrome P450 2E1, downregulation of prohibitin 1 (a mitochondrial chaperon protein), and impaired mitochondrial function.25–27 Taken together, these results suggest that the deficiency in MAT activity observed in human liver cirrhosis due to various causes including alcohol, hepatitis B, and hepatitis C, may contribute to the pathogenesis and progression of the disease as well as predisposition to HCC. Deciphering where in the progression from normal liver to steatosis, steatohepatitis, cirrhosis, and HCC that SAMe is most likely to be involved is difficult. Findings from Mat1a (−/−) mice suggest that SAMe is likely to be involved in all of these stages.
Whereas SAMe depletion predisposes the liver to further injury and possibly malignant degeneration, impaired SAMe metabolism, which occurs in Gnmt (−/−) mice10 and in patients with GNMT mutations,28 can lead to steatohepatitis, increased hepatocyte apoptosis, fibrosis, and HCC (Mato, Lu, Martínez-Chantar, Luka, Wagner, unpublished observations); and a GNMT polymorphism (1289 C → T) has been associated with HCC.29 These results indicate that liver needs the right amount of SAMe to provide an adequate supply of methyl groups, and that too much or too little SAMe induces liver injury. Because the number of SAMe binding proteins is large—for instance, the number of known SAMe-dependent methyltransferases in mammals is 399—it seems likely that the pathophysiology of liver injury attributable to altered hepatic SAMe content will be extremely complex.
SAMe Regulation of Hepatocyte Growth
In hepatocytes, SAMe levels are related to the differentiation status, being high in quiescent and low in proliferating hepatocytes.30 In rat liver, after partial hepatectomy (PH), SAMe levels are dramatically reduced shortly afterward, coinciding with the onset of DNA synthesis and the induction of early response genes.31 When this fall in SAMe after PH was prevented by exogenous SAMe administration, hepatocyte DNA synthesis was inhibited.24, 32 Additionally, exogenous SAMe prevents the development of HCC in rats treated with hepatocarcinogen,33, 34 inhibits the growth of hepatoma cells,30 and blocks the mitogenic effect of hepatocyte growth factor (HGF) in hepatocytes.35
Conversely, chronic SAMe depletion in Mat1a (−/−) mice is associated with increased proliferating cell nuclear antigen expression and the spontaneous development of HCC.26, 36 Despite having higher baseline hepatic staining for proliferating cell nuclear antigen, Mat1a (−/−) mice with low hepatic SAMe levels have impaired liver regeneration after PH because of an inability to upregulate cyclin D1.36 Upstream signaling pathways involved in cyclin D1 activation include nuclear factor kappaB, the c-Jun-N-terminal kinase, extracellular signal-regulated kinases (ERKs), and signal transducer and activator of transcription-3. At baseline, c-Jun-N-terminal kinase and ERK are more activated in Mat1a (−/−) mice, whereas nuclear factor kappaB and signal transducer and activator of transcription-3 are similar to wild-type mice. After PH, early activation of these pathways occurred, but although they remained increased in wild-type mice, c-jun and ERK phosphorylation fell progressively in Mat1a (−/−) mice. In culture, Mat1a (−/−) hepatocytes have higher baseline DNA synthesis but failed to respond to the mitogenic effect of hepatocyte growth factor (HGF).36 Overall, these results indicate that SAMe regulates hepatocyte growth, and that chronic SAMe depletion results in abnormal growth and impaired response to growth factors.
In hepatocytes, SAMe inhibits HGF-induced cyclin D1 expression without affecting ERK phosphorylation, which indicates that the mitogen-activated protein kinase–signaling pathway is not the target of exogenous SAMe.35 Cyclin D1 mRNA levels are regulated by HuR, a RNA-binding protein that increases the half-life of a variety of cell cycle genes. In hepatocytes, HGF regulates cytoplasmic HuR content via phosphorylation and activation of the AMP-activated protein kinase (AMPK), and SAMe inhibits this process through a mechanism that involves the activation of protein phosphatases, probably protein phosphatase 2A.37 Similarly, in hepatocytes SAMe blocks 5-aminoimidazole-4-carboximide-riboside–induced phosphorylation and activation of AMPK, the transport of HuR from nucleus to cytosol, the induction of cyclin D1, and DNA synthesis.37 AMPK is a metabolic sensor that is activated by cellular stresses that deplete ATP. Once activated, AMPK switches on catabolic pathways that generate ATP, such as fatty acid oxidation, while switching off biosynthetic pathways that consume ATP, such as fatty acid biosynthesis.38 These results indicate that through the regulation of AMPK phosphorylation SAMe may modulate liver lipid and carbohydrate homeostasis and hepatocyte growth (Fig. 3).
SAMe Regulation of Hepatocyte Apoptosis
SAMe not only regulates hepatocyte growth response, it also regulates hepatocyte apoptotic response. Although SAMe protected against okadaic acid–induced apoptosis in normal hepatocytes, it induced apoptosis in liver cancer cell lines HepG2 and HuH-7 via the mitochondrial death pathway.39 MTA also recapitulated these actions. MTA can be derived from SAMe enzymatically and non-enzymatically. MTA is a product of SAMe metabolism in the polyamine pathway (Fig. 1). Exogenous SAMe can also undergo nonenzymatic hydrolysis into MTA.40 These findings agree with the reported chemopreventive action of SAMe and MTA in an in vivo model of chemical hepatocarcinogenesis in rats, which was accompanied by an increase of apoptotic bodies in atypical nodules and HCC foci in SAMe-treated animals.34
We have identified two mechanisms of SAMe's and MTA's differential effect on apoptosis in normal hepatocytes and liver cancer cells thus far.14, 41 Using microarray, we found SAMe treatment induced Bcl-x expression. Bcl-x is alternatively spliced to produce two distinct mRNAs and proteins, Bcl-xL and Bcl-xS. Bcl-xL is anti-apoptotic whereas Bcl-xS is pro-apoptotic. SAMe and MTA induced selectively Bcl-xS in HepG2 cells by increasing alternative splicing. Furthermore, inhibitors of protein phosphatase 1 (PP1) blocked SAMe's and MTA's ability to induce Bcl-xS. PP1 modulates alternative splicing by dephosphorylating SR proteins. SAMe and MTA increased PP1 catalytic subunit mRNA and protein levels and dephosphorylated SR proteins. Importantly, the effects of SAMe and MTA on Bcl-xS, PP1 catalytic subunit expression, and apoptosis were not seen in normal hepatocytes. Thus, in liver cancer cells, SAMe and MTA can affect the cellular phosphorylation state and alternative splicing of genes, resulting in the induction of Bcl-xS leading to apoptosis.
Another mechanism for SAMe and MTA to exert pro-apoptotic effect in liver cancer cells is their ability to inhibit the expression of BHMT.14 Both agents suppressed BHMT expression at the transcriptional level and, in the case of MTA, the inhibitory effect was more than 80%. Impairment in Hcy metabolism can result in endoplasmic reticulum stress, which also can lead to apoptosis.16 Indeed, MTA treatment led to increased expression of markers of endoplasmic reticulum stress.14 Importantly, SAMe and MTA have no effect on the expression of BHMT in normal human hepatocytes (Lu, unpublished observations).
SAMe Treatment in Liver Disease
The importance of methyl groups in general, and SAMe in particular, to normal hepatic physiology, coupled with the convincing body of evidence linking SAMe depletion with the development of liver disease, led to the examination of the effect of SAMe supplementation in a variety of animal models of liver disease. SAMe administration to alcohol-fed mice, rats, and baboons ameliorated GSH depletion and liver damage.42 SAMe improved survival in animal models of galactosamine-, acetaminophen-, and thioacetamide-induced hepatotoxicity, and in ischemia–reperfusion-induced liver injury.43 SAMe treatment also diminished liver fibrosis in rats treated with carbon tetrachloride43 and reduced neoplastic hepatic nodules in animal models of HCC.44
Although studies in experimental animal models of liver injury support a therapeutic role for SAMe, more studies in patients with liver disease are still needed to better define its role. Analyses performed by the Agency for Healthcare Research and Quality Agency for Healthcare Research and Quality support the use of SAMe to relieve pruritus and decrease elevated serum bilirubin levels associated with cholestasis of pregnancy as well as due to other causes.45 However, the data on SAMe treatment of alcoholic liver disease are less clear. In 2001, the Cochrane Hepato-Biliary Group analyzed 8 clinical trials of SAMe treatment of alcoholic liver disease in 330 patients.46 This meta-analysis found SAMe decreased total mortality (OR 0.53, 95% CI 0.22-1.29) and liver-related mortality (OR 0.63, 95% CI 0.25-1.58). However, because many of the studies were small and the quality of the studies varied greatly, the Cochrane Group concluded, “SAMe should not be used for alcoholic liver disease outside randomized clinical trials.”46 The Agency for Healthcare Research and Quality reached a similar conclusion.45 The Cochrane Hepato-Biliary Group also concluded that only one trial of 123 patients with alcoholic cirrhosis used adequate methodology and reported clearly on mortality and liver transplantation. In this study.23 mortality decreased from 30% in the placebo group to 16% in the SAMe (1,200 mg/day orally) group (P = 0.077). When patients with more advanced cirrhosis (Child score C) were excluded from the analysis, the mortality was significantly less in the SAMe group (12%) as compared with the placebo group (25%, P = 0.025). Effectiveness of SAMe therapy in alcoholic hepatitis has not been adequately addressed.
Mechanisms of SAMe Action: Role of MTA
Various mechanisms have been proposed for SAMe's protective action in the liver. These include increased GSH levels, a change in DNA methylation, improved membrane fluidity, decreased tumor necrosis factor alpha expression, and inhibition of collagen I production by preventing transforming growth factor beta induction of the COL1A2 promoter.4, 42, 47 Whereas SAMe protects normal hepatocytes, it induces apoptosis in liver cancer cells.39, 41 This makes SAMe an attractive candidate for chemoprevention of liver cancer. Interestingly, SAMe's effect on growth, apoptosis, and tumor necrosis factor alpha expression48 can be mimicked by MTA. SAMe is highly unstable with a short half-life and converts to MTA rapidly in vitro (Lu and Mato, unpublished observation). In contrast, MTA is highly stable and is taken up rapidly. Whereas SAMe is a methyl donor and precursor of polyamines, MTA inhibits methylation and polyamine synthesis.40 Is it possible that at pharmacological doses of SAMe, the effect is actually mediated by MTA, particularly if MTA can recapitulate the same effects? This is a possibility recently raised40 and deserves further investigation. MTA has potent anti-inflammatory effects and has chemopreventive effects against liver cancer such as SAMe.32, 49 It also may be an attractive agent to examine in the treatment of liver disease.
We have come a long way in understanding SAMe's actions since its first discovery. However, a large knowledge gap remains in understanding its mechanisms of action and why abnormal SAMe metabolism results in liver injury. Future studies uncovering the molecular mechanisms of how impaired SAMe metabolism (deficiency or excess) leads to liver injury will help to identify novel targets of SAMe. Studies are also needed to identify mechanisms of action of pharmacologic doses of SAMe and which of these actions are mediated by MTA. Finally, the role of SAMe (and possibly MTA) in the chemoprevention of HCC deserves study. This includes defining the optimal dose and which patients will benefit from its use.