Regulation by Hypoxia of Methionine Adenosyltransferase Activity and Gene Expression in Rat Hepatocytes

Transcription

1 GASTROENTEROLOGY 1998;114: Regulation by Hypoxia of Methionine Adenosyltransferase Activity and Gene Expression in Rat Hepatocytes MATIAS A. AVILA, M. VICTORIA CARRETERO, E. NELSON RODRIGUEZ, and JOSE M. MATO Instituto de Investigaciones Biomédicas, Consejo Superior de Investigaciones Científicas, Madrid, Spain See editorial on page 403. Background & Aims: Oxygen supply to the hepatic parenchyma is compromised by long- or short-term ethanol consumption and pathological conditions such as cirrhosis. Impairment in the production of S- adenosyl-l-methionine, the major methylating agent, occurs during hypoxia. In this study, the molecular mechanisms implicated in the regulation of S-adenosyl- L-methionine synthesis by oxygen levels were investigated. Methods: Rat hepatocytes were isolated and cultured under normoxic (21% O 2 ) or hypoxic (3% O 2 ) conditions for different periods. Methionine adenosyltransferase activity, messenger RNA levels, and nuclear transcription were evaluated. Results: Methionine adenosyltransferase was inactivated in hepatocytes kept under low oxygen levels. Hypoxia induced the expression of nitric oxide (NO) synthase, and the inactivation of methionine adenosyltransferase was prevented by the NO synthase inhibitor N G -monomethyl-l-arginine methyl ester. Methionine adenosyltransferase messenger RNA levels were down-regulated by hypoxia, through a mechanism that might involve a hemoprotein. Hypoxia dramatically reduced methionine adenosyltransferase gene transcription, and messenger stability was also decreased, although to a lesser extent. Conclusions: We have established the molecular basis for the regulation of methionine adenosyltransferase activity and gene expression by hypoxia. NO-mediated inactivation and transcriptional arrest seem to be the two major pathways by which oxygen levels control hepatic methionine adenosyltransferase, the enzyme necessary for methylation reactions and for the synthesis of polyamines and glutathione. Liver methionine metabolism starts with the formation of S-adenosyl-L-methionine (AdoMet). This reaction represents the preferred pathway for methionine catabolism and is catalyzed by the enzyme methionine adenosyltransferase (MAT) (EC ). 1 MAT catalytic protein subunits are encoded by two different genes; one is present only in liver, and the other is expressed in extrahepatic tissues and fetal liver. 2 4 The liver gene product can be found as two oligomeric forms, being either a dimer (MATIII) or a tetramer (MATI) of the same catalytic subunit. 4 AdoMet participates as the methyl donor for essentially all known methylation reactions, provides the propylamine group for the synthesis of polyamines, and acts as a precursor for the synthesis of reduced glutathione (GSH) through the transsulfuration pathway. 2,5 Methionine metabolism and the methylation cycle are altered in several liver disorders. AdoMet synthesis and utilization is impaired in conditions such as alcoholic and viral liver cirrhosis, 6,7 septic shock, 8 and several experimental models of liver damage (reviewed by Mato et al. 5 ) A situation common to many of these pathological conditions is a deficient supply of oxygen to the hepatic parenchyma. Short-term 9 and long-term 10 alcohol consumption as well as alcoholic liver necrosis 11 lead to a hypoxic status, strongly associated with liver injury. Rats exposed to hypoxia were recently shown to have reduced liver MAT activity and messenger RNA (mrna) levels. 12 In the present study, we directly evaluated the mechanisms underlying the effects of hypoxia on MAT activity and expression in a model of rat hepatocytes kept in primary culture. We also report on the role of nitric oxide (NO) induced by hypoxia in the regulation of MAT activity. Materials and Methods Preparation and Culture of Rat Hepatocytes Liver cells were isolated from male Wistar rats ( g) by collagenase perfusion as described previously. 13 Cells were plated onto 60-mm collagen-coated culture dishes at a density of cells per dish. Cultures were maintained in RPMI 1640 medium supplemented with 10% fetal calf serum, Abbreviations used in this paper: AdoHcy, S-adenosylhomocysteine; AdoMet, S-adenosyl-L-methionine; GSH, glutathione; inos, inducible NO synthase; L-NAME, N G -monomethyl-l-arginine methyl ester; MAT, methionine adenosyltransferase; SDS, sodium dodecyl sulfate; SOD, manganous superoxide dismutase by the American Gastroenterological Association /98/$3.00

2 February 1998 METHIONINE ADENOSYLTRANSFERASE IN HYPOXIA mmol/l glutamine, 50 mmol/l penicillin, and 50 mg/ml streptomycin sulfate. After a 2-hour incubation, the culture medium was removed, and cultures were refed the same medium with either 5% fetal calf serum or 1% fetal calf serum plus 1 mol/l triamcinolone. Triamcinolone was omitted when MAT activity or inducible NO synthase (inos) induction were measured, because glucocorticoids prevent the induction of inos. 14 Cells were maintained at 37 C in a humidified incubator containing 21% oxygen and 5% carbon dioxide in air (referred to as normoxic conditions). Hypoxic conditions were attained by exposure to 3% oxygen and 5% carbon dioxide with the balance as nitrogen in a humidified Billups-Rothenberg Modular Incubation Chamber (MIC-101); (Del Mar, CA) at 37 C. Cell viability was measured by trypan blue exclusion, and no significant differences were observed between normoxic and hypoxic cultures during the course of the experiments. When indicated, N G -monomethyl-l-arginine methyl ester (L-NAME) (Sigma Chemical Co., St. Louis, MO) and actinomycin D (Boehringer Mannheim, Mannheim, Germany) were used at 1-mmol/L and 5-mg/mL doses, respectively. Animals were treated humanely, and study protocols were in compliance with our institution s guidelines for the use of laboratory animals. Determination of Total MAT Activity Hepatocytes ( cells) were lysed by freeze thawing in 300 L of 10 mmol/l Tris-HCl (ph 7.5) containing 0.3 mol/l sucrose, 0.1% -mercaptoethanol, 1 mmol/l benzamidine, and 0.1 mmol/l phenylmethylsulfonyl fluoride. The homogenate was centrifuged for 30 minutes at 10,000g, and MAT activity was assayed in the supernatant as described previously 7 with minor modifications. Determination of AdoMet, S-Adenosylhomocysteine, GSH, and Adenosine Triphosphate For determination of AdoMet and S-adenosylhomocysteine (AdoHcy) concentrations in cultured hepatocytes, samples were deproteinized in 0.4 mol/l perchloric acid and centrifuged at 12,000g for 15 minutes at 4 C. Supernatants were filtered and analyzed by high-performance liquid chromatography as described previously. 15 For determination of adenosine triphosphate (ATP) concentration, samples were also deproteinized in 0.4 mol/l perchloric acid and then centrifuged at 3000g for 15 minutes at 4 C. Supernatants were neutralized with 3 mol/l K 2 CO 3 and 0.3 mol/l imidazole and centrifuged again under the same conditions. Aliquots of the supernatants (0.1 ml) were used for ATP determination as described previously. 16 GSH concentration was determined by the method of Hissin and Hill. 17 Protein concentration in samples was measured according to Bradford. 18 RNA Isolation and Northern Blot Analysis Total hepatocyte RNA was isolated by the guanidinium thiocyanate method. 19 Aliquots (20 g) of total RNA were size-fractionated by electrophoresis in a 1% agarose gel under denaturing conditions. RNAs were then blotted and fixed to Nytran membranes (Schleicher & Schuell, Keene, NH). Prehybridization and hybridization were performed as described previously. 20 MAT mrna levels were measured using a 2.2-kilobase (kb) EcoRI complementary DNA (cdna) fragment of rat liver MAT. 21 The levels of inos mrna were determined using a 0.8-kb EcoRI/HindIII fragment from the macrophage inos cdna. 22 Manganous superoxide dismutase (MnSOD) mrna levels were determined using a 0.8-kb EcoRI/HindIII fragment from rat liver cdna (gift of Dr. D. Massaro, Georgetown University, Washington, DC). Equal loading of the RNA gels was assessed by hybridization with a probe specific for -actin. The probes were labeled with [ - 32 P]deoxycytidine triphosphate (3000 Ci/mmol; Amersham, Little Chalfont, England) by random priming using the Megaprime DNA labeling system (Amersham). Specific activity was usually approximately cpm/ g of DNA. Quantitation was performed by scanning densitometry of the x-ray films. Isolation of Nuclei and Run-on Transcription Assay Nuclei from normoxic or hypoxic cells were prepared as described previously. 13 Transcription reactions were performed for 30 minutes at 26 C with 100 L of the isolated nuclei in a final volume of 150 L in the presence of 100 Ci of [ - 32 P]uridine triphosphate (800 Ci/mmol; Amersham) and 0.25 mmol/l unlabeled nucleotides. Elongated RNA transcripts (at equivalent amounts of radioactivity, i.e., cpm) were denatured and hybridized to 5 g of linearized plasmid DNAs immobilized on Nytran filters in hybridization buffer (0.2 mol/l NaPO 4 [ph 7.2], 1 mmol/l ethylenediaminetetraacetic acid, 7% sodium dodecyl sulfate, 45% formamide, and 250 g/ml Escherichia coli transfer RNA as carrier) for 48 hours at 42 C. The plasmids used were pssrl, 22 a rat -actin cdna, 23 and puc18 for background control. After washing, the filters were exposed to x-ray films. Data were normalized to transcription of the -actin gene. Immunoblot Analysis For determination of MAT protein levels, hepatocytes were homogenized as described above for the measurement of enzyme activity. Equal amounts of protein (20 g) were subjected to 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis. Proteins were electrophoretically transferred to nitrocellulose membranes. Immunodetection of MAT was performed using a rabbit anti-rat MAT antiserum 24 and a horseradish peroxidase conjugated secondary antibody. Blots were developed by enhanced chemoluminescence according to manufacturer s instructions (Dupont, Boston, MA). For determination of MATI/MATIII ratio, the hepatocytes were homogenized as described above and cytosolic fractions were loaded onto phenyl Sepharose CL-4B columns (Pharmacia, Uppsala, Sweden). MATI was present in the flowthrough fraction, whereas MATIII was eluted in the presence of 50% dimethyl-

3 366 AVILA ET AL. GASTROENTEROLOGY Vol. 114, No. 2 sulfoxide as described previously. 24 The relative content of MATI and MATIII in normoxic and hypoxic hepatocytes was then assessed by immunoblotting as mentioned above. Statistics The data are the means SEM of at least four independent experiments. Statistical significance was estimated with Student s t test. A P value of 0.05 was considered significant. Results MAT specific activity was evaluated in rat hepatocytes cultured under normoxic (21% oxygen) or hypoxic (3% oxygen) conditions for different periods. As shown in Figure 1A, MAT activity was reduced when cells were exposed to a low oxygen concentration; the values were significantly different after 6 hours of hypoxia. Cellular viability at the end of the hypoxic treatments was never below 88.5% 3.1% of the viability in normoxic cultures. Along with MAT activity, AdoMet levels were also determined under the same conditions. Culture of rat hepatocytes induced a time-dependent reduction in AdoMet and AdoHcy levels (Table 1). In agreement with previous studies, 25 the decrease in AdoMet concentration Figure 1. (A) Specific activity of MAT in rat hepatocytes cultured for different periods under normoxic (21% oxygen) (N) or hypoxic (3% oxygen) (H) conditions ( * P 0.05). Data are means SEM (n 5 10 for each group). (B) Western blot analysis of MAT protein from hepatocytes cultured in normoxic or hypoxic conditions for up to 24 hours (a representative blot is shown). was greater when hepatocytes were exposed to hypoxia (Table 1). No significant differences were observed in AdoHcy levels between normoxic and hypoxic cultures (Table 1). This situation led to a decreased AdoMet/ AdoHcy ratio, which decreased after a 6-hour incubation from 4.2 in normoxic cells to 2.6 in hypoxic cultures. To determine if changes in MAT specific activity were the consequence of a reduction in MAT protein levels induced by hypoxia, Western blot analyses were performed on protein extracts of hepatocytes kept under normoxic or hypoxic conditions. As shown in Figure 1B, MAT protein levels remained constant at short periods of hypoxia (up to 8 hours), supporting the notion of MAT inactivation as the cause of the decline in specific activity. The MATI/MATIII ratio was also not affected after 6 hours of hypoxia (not shown). When hepatocytes were kept in culture for longer periods (24 hours), the amount of MAT protein was reduced in normoxic cells. However, this decrease was always more pronounced in hypoxic cultures (Figure 1B). Our observation that rat hepatocytes in culture show a decrease in MAT protein content (Figure 1B) in response to hypoxia confirms in vivo results 12 and indicates that this effect is a specific response of the liver parenchymal cells to the hypoxic condition. MAT activity is known to be regulated in vivo by GSH levels. 5 We have therefore measured GSH concentration in cells cultured under normoxic or hypoxic conditions. GSH levels were reduced after 6 hours of hypoxia to approximately 70% of the normoxic value (Table 1). This result is in agreement with the previously reported dependence of hepatocyte GSH synthesis on oxygen levels. 25 We showed recently that NO is a strong inhibitor of rat liver MAT that causes inactivation of the enzyme both in vitro and in vivo. 8 The mechanism seems to be the direct interaction of NO with a cysteine residue of the liver enzyme located in position These observations led us now to evaluate whether the induction of hepatocyte inos was among the cellular responses to hypoxia. This indeed was the case; a time-dependent increase in inos transcript in cells grown under 3% oxygen was observed (Figure 2A). inos mrna began to accumulate after 2 hours of hypoxia, peaking at 8 hours (eightfold induction), and was still well above normoxic levels after 20 hours of culture (data not shown). To check if other genes related to free radical metabolism were also regulated in this model, we evaluated the expression of MnSOD, which is induced under different oxidative stress conditions. 26 Figure 2A shows a decline in MnSOD mrna levels after 2 hours (38%) and 8 hours (50%) of hypoxia, suggesting that the generation of reactive oxygen species was not increased in this condition. To

4 February 1998 METHIONINE ADENOSYLTRANSFERASE IN HYPOXIA 367 Table 1. AdoMet, AdoHcy, GSH, and ATP Levels in Hepatocytes Kept for Different Periods in Normoxia or Hypoxia 2h 4h 6h Normoxia Hypoxia Normoxia Hypoxia Normoxia Hypoxia AdoMet a AdoHcy GSH a ATP a NOTE. Data (nmol/mg of protein) are means SEM (n 5 for each group). a P 0.05 compared with the normoxic value. study the role of NO in the inactivation of MAT during hypoxia, hepatocytes were incubated in the presence or absence of the inos inhibitor L-NAME. Addition of 1 mmol/l of L-NAME to the culture medium resulted in a complete protection of MAT inactivation by hypoxia (Figure 2B). Addition of L-NAME had no effect on the reduction in GSH levels induced by hypoxia (data not shown). Incubation of hepatocytes in the presence of 3 mmol/l of N-acetyl-L-cysteine, a thiol reductant, partially prevented the inactivation of MAT during hypoxia (55.0% 11.2% vs. 36.6% 8.4% inactivation in the presence of N-acetyl-L-cysteine after 6 hours of hypoxia). Thus, NO production together with the decrease in GSH levels could explain the loss of MAT activity in hypoxia, supporting a role for NO and GSH in MAT regulation under different stress conditions. Given the prominent role that glucocorticoids play in vivo in the maintenance of high levels of expression of liver MAT, 27 we tested whether oxygen levels could modulate the effect of these hormones on MAT mrna induction. Hepatocytes were incubated for 12 hours in the presence of 1 mol/l triamcinolone, a glucocorticoid shown to increase MAT transcription in cultured rat hepatocytes, 27 and the cells were then further incubated in normoxia or hypoxia. MAT mrna levels were induced 2.3-fold by triamcinolone treatment (Figure 3A). When the cells were placed under hypoxic conditions, a rapid decrease in MAT mrna levels was observed. MAT mrna levels decreased to 40% of control (normoxic) values by 2 hours of hypoxia and continued to decrease down to 30% of the control value after 12 hours of hypoxia. It has been proposed that oxygen is sensed by means of a heme protein and that this mechanism is shared by most cells. 28 This hypothesis is based on the ability of certain transition metals, such as cobalt, to mimic hypoxia by substituting the iron atom in the heme moiety, rendering it unable to bind oxygen. When hepatocytes were incubated for 6 hours in the presence of 50 mol/l CoCl 2, a dose shown to be nontoxic for primary rat hepatocyte cultures, 29 a decrease (65% of control value) in MAT mrna levels was observed (Figure 3B). Stronger support for the role of a heme protein as oxygen sensor has been provided by the use of carbon monoxide. 30 When this gas was added to the hypoxic mixture (10% CO), the effect of low oxygen concentration exposure for 4 hours on MAT mrna levels was partially blunted (Figure 3B). Finally, we wanted to Figure 2. (A) Northern blot analysis of inos and MnSOD levels in hepatocytes cultured for different periods under normoxic (N) or hypoxic (H) conditions. -Actin hybridization was performed as loading control (representative blots are shown). (B) MAT specific activity in hepatocytes kept under normoxic or hypoxic conditions for 6 hours in the presence or absence of the inos inhibitor L-NAME. Data are means SEM (n 4 for each group). The asterisk indicates that, in the absence of L-NAME, differences in MAT activity between normoxic and hypoxic cells were statistically significant (P 0.05). No statistical differences were found between MAT activity in normoxia vs. hypoxia in the presence of L-NAME.

5 368 AVILA ET AL. GASTROENTEROLOGY Vol. 114, No. 2 Figure 3. (A) Effect of hypoxia on the glucocorticoid-dependent induction of MAT expression in cultured rat hepatocytes. Cells were incubated in the presence or absence of 1 mol/l triamcinolone for 12 hours, and incubation was then continued for different periods of time under normoxia or hypoxia. MAT mrna levels were analyzed by Northern blotting (a representative blot is shown). (B) Hypoxia-sensing mechanism for MAT mrna regulation. Triamcinolone-treated hepatocytes were incubated for 4 hours in normoxia (N), hypoxia (H), hypoxia plus 10% CO (CO), or in normoxia in the presence of 2.5 mmol/l NaN 3 and 2.5 mmol/l KCN plus 40 mmol/l 2-deoxyglucose (KCN 2DG), 50 mol/l dinitrophenol, or 50 mol/l CoCl 2 (6-hour treatment) (a representative blot is shown). -Actin hybridization was performed as loading control. test whether MAT mrna changes under hypoxia were caused by alterations in ATP cellular content. Indeed, exposure to hypoxia for 4 hours led to a 35% reduction in ATP levels (Table 1). However, the reduction of ATP levels by incubating hepatocytes for 4 hours with 2.5 mmol/l NaN 3 (60% ATP reduction), 2.5 mmol/l KCN plus 40 mmol/l 2-deoxyglucose (53% ATP reduction), or 50 mol/l dinitrophenol (20% ATP reduction) did not affect MAT steady-state mrna levels (Figure 3B). Intracellular H 2 O 2 levels have been proposed also to participate in the signaling of environmental oxygen concentrations to the cell. 28 When normoxic cells were treated with H 2 O 2 (from 10 mol/l up to 2 mmol/l H 2 O 2 ) for 4 hours, no changes in MAT mrna could be detected (data not shown). MAT mrna levels also remained unchanged when hepatocytes were treated with the H 2 O 2 -degrading enzyme catalase (100 mg/ml) (data not shown). Hypoxia control of gene expression is accomplished at the transcriptional and postranscriptional level. 28 We tested whether MAT gene transcriptional activity was dependent on oxygen tension by performing nuclear run-on studies. As shown in Figure 4A, MAT mrna synthesis was reduced under low oxygen levels with respect to -actin gene transcription, which remained constant. Exposure of hepatocytes grown in the presence of triamcinolone (1 mol/l) to 3% oxygen resulted in a 75% decrease in MAT gene transcription compared with control cells. The specificity of each band was supported by the lack of hybridization to puc18 DNA. To determine if MAT mrna stability is also affected by oxygen levels, actinomycin D (5 g/ml of media) was used to inhibit total RNA synthesis. After 1-hour treatment with this inhibitor, cells were further incubated in normoxia or hypoxia for 4 hours and MAT mrna was assayed by Northern blotting. As shown in Figure 4B, MAT mrna decreased in hypoxia in the presence of a transcriptional inhibitor (50% decrease). These data suggest that MAT mrna stability could also be regulated by oxygen tension. Nevertheless, it seems that the transcriptional component of hypoxia effects on MAT expression is of greater importance. Discussion Liver damage in conditions such as short- and long-term ethanol consumption 9 11 has been associated with a reduced oxygen supply to this organ. Moreover, there is growing evidence for a role of metabolic and antioxidant mechanism alterations during hypoxia in conditioning liver injury on reoxygenation. 31,32 GSH is probably the most important cellular antioxidant, 33 and its synthesis is impaired in hepatocytes kept under low oxygen levels. 34 Production of GSH from methionine starts with the ATP-dependent synthesis of AdoMet by MAT (reviewed by Mato et al. 35 ) Liver MAT activity and mrna levels were recently shown to be reduced in rats exposed to hypoxia. 12,36 In the present study, we directly evaluated the role of oxygen tension on MAT activity and expression. The experimental model of rat hepatocytes in primary culture eliminates the influence of other organs and cells on liver responses. Using this model, we have observed a reduction in MAT activity during hypoxia, accompanied by a decrease in the AdoMet/AdoHcy ratio that was not caused by a reduction in MAT protein levels, at least for up to 6 hours. These observations suggested that the inactivation of the existing protein could be the cause of the decrease in specific activity during hypoxia.

6 February 1998 METHIONINE ADENOSYLTRANSFERASE IN HYPOXIA 369 Figure 4. (A) A representative nuclear run-on assay showing the effects of oxygen levels on transcriptional activity of MAT and -actin genes in triamcinolone-treated hepatocytes after 4 hours in normoxic (N) or hypoxic (H) conditions. (B) Effect of hypoxia on MAT mrna stability in triamcinolone-treated hepatocytes, assayed by Northern blotting. Hepatocytes were treated with actinomycin D (5 g/ml) and were then further incubated in hypoxia or normoxia for another 4 hours (a representative blot is shown). -Actin hybridization was performed as loading control. MAT activity is dependent on intracellular GSH levels 5 ; its decrease could therefore be attributed partly to the reduction in GSH induced by hypoxia. This point was further supported by finding that the addition of N-acetyl- L-cysteine to the hepatocyte cultures partially prevented MAT inactivation. On the other hand, we recently reported that rat liver MAT is very sensitive to NO inactivation both in vivo and in vitro. 8 This observation, with the recent description of hepatocyte inos induction in mice exposed to hypoxia, 37 led us to examine whether NO played a role in the modulation of MAT in hypoxic cells. First we assessed the levels of inos mrna in our normoxic and hypoxic cultures, observing a sharp induction in inos mrna that coincided in time with the decrease in MAT activity. Second, addition of inos inhibitors such as L-NAME to the culture media prevented MAT inactivation by hypoxia. These results and our previous observations 8 point toward a pivotal role for NO and GSH in MAT modulation under different stress conditions including hypoxia. When hepatocytes were kept under low oxygen tension for long periods (24 hours), a down-regulation in MAT protein levels was found. This result, together with previous observations, 12,36 prompted us to study the mechanisms underlying the regulation of MAT gene expression by hypoxia. We had previously shown that glucocorticoids play an important role in the maintenance of liver MAT. 27 Given this prominent action of steroids, we have studied whether oxygen levels could affect triamcinolone-dependent expression of MAT. Our results show that hypoxia was able to counteract the stimulatory effect of glucocorticoids, thus demonstrating that oxygen can modulate the tight control exerted by these hormones on MAT expression. Sustained hypoxia such as that observed in liver cirrhosis 11 and long-term ethanol consumption 10 would induce a persistent down-regulation of MAT expression, leading to a persistent reduction in the AdoMet/AdoHcy ratio and GSH levels. This situation may promote irreversible tissue damage through the impairment of methylation reactions 36 and a limited GSH-dependent detoxification capacity. 38 Interestingly, the rat hepatoma cell line H35 proved to be resistant to hypoxia in terms of MAT mrna down-regulation (unpublished observation). This situation could be important for the selection of hypoxia-resistant subpopulations during solid tumor development. 39 The mechanisms implicated in oxygen sensing and signaling are a matter of much debate. 28 Nevertheless, cumulative evidence suggests that oxygen is sensed by means of a heme protein. 28 This hypothesis is based on observations such as the ability of certain transition metals (cobalt and nickel) to mimic the induction by hypoxia of erythropoietin gene expression 30 and on the ability of carbon monoxide to blunt cellular responses to hypoxia. 30 By following both experimental approaches, we conclude that MAT responses to hypoxia could be mediated by such a heme-based sensor. Changes in intracellular levels of hydrogen peroxide have been proposed to be involved in oxygen levels signaling in hepatocarcinoma cells. 40 However, in our hepatocyte cultures, we were unable to induce changes in MAT mrna by manipulating intracellular H 2 O 2 levels. Similarly, MAT mrna reduction in hypoxia was not a result of changes in cellular ATP levels. Treatment of normoxic hepatocytes with different ATP-reducing agents had no effect on MAT mrna, suggesting that the respiratory chain does not function as the oxygen sensor. Oxygen modulation of gene expression can be achieved at different levels. 28 Both gene transcription and messenger stability are mechanisms affected by changes in oxygen tension In the case of MAT, we have observed that hypoxia mediates its effect mainly through a drastic reduction in gene transcription. To a lesser

7 370 AVILA ET AL. GASTROENTEROLOGY Vol. 114, No. 2 extent, a reduction in messenger stability could contribute to the total effect of hypoxia on the steady-state levels of MAT mrna. These observations may explain, from a mechanistic point of view, the decrease in rat liver MAT mrna reported in animals kept under low oxygen levels in breathing air. 12,36 The results of this study may be important to the more general question of the metabolic zonation of the liver. Oxygen tension plays a major role among the factors that condition the functional heterogeneity of periportal and perivenous hepatocytes. 44 Metabolic reactions requiring energy in the form of ATP are located in the more aerobic periportal zone, linked to oxidative energy metabolism. 44 Such is the case of MAT, which consumes ATP for the production of AdoMet. Although there is no evidence for a periportal location of liver MAT, it is interesting to note that sulfate formation, a product of cysteine breakdown and of the transsulfuration pathway, predominates in the periportal area. 45 References 1. Cantoni GL. Biochemical methylations: selected aspects. Annu Rev Biochem 1975;44: Finkelstein JD. Methionine metabolism in mammals. J Nutr Biochem 1990;1: Kotb M, Geller AM. Methionine adenosyltransferase: structure and function. Pharmacol Ther 1993;59: Kotb M, Mudd SH, Mato JM, Geller AM, Kredich NM, Chou JY, Cantoni GL. Consensus nomenclature for the mammalian methionine adenosyltransferase genes and gene products. Trends Genet 1997;13: Mato JM, Alvarez L, Corrales FJ, Pajares MA. S-Adenosylmethionine and the liver. In: Arias IM, Boyer JL, Fausto N, Jakoby WB, Schachter DA, Shafritz DA. eds. The liver: biology and pathobiology. 3rd ed. New York: Raven, 1994: Martin-Duce A, Ortiz P, Cabrero C, Mato JM. S-Adenosyl-Lmethionine synthethase and phospholipid methyltransferase are inhibited in human cirrhosis. Hepatology 1988;8: Cabrero C, Martin-Duce A, Ortiz P, Alemany S, Mato JM. Specific loss of the high-molecular weight form of S-adenosylmethionine synthetase in human liver cirrhosis. Hepatology 1988;8: Avila MA, Mingorance J, Martinez-Chantar ML, Casado M, Martin- Sanz P, Bosca L, Mato JM. Regulation of rat liver S-adenosylmethionine synthetase during septic shock: role of nitric oxide. Hepatology 1997;25: Arteel GE, Raleigh JA, Bradford BV, Thurman RG. Acute alcohol produces hypoxia directly in rat liver tissue in vivo: role of Kupffer cells. Am J Physiol 1997;271:G494 G Ji S, Lemasters JJ, Christerson V, Thurman RG. Periportal and pericentral pyridine nucleotide fluorescence from the surface of the perfused liver: evaluation of the hypothesis that chronic treatment with ethanol produces pericentral hypoxia. Proc Natl Acad Sci USA 1982;79: Tsukamoto H, Xi XP. Incomplete compensation of enhanced hepatic oxygen consumption in rats with alcoholic centrilobular liver necrosis. Hepatology 1989;9: Chawla RK, Jones DP. Abnormal metabolism of S-adenosyl-Lmethionine in hypoxic rat liver. Similarities to its abnormal metabolism in alcoholic cirrhosis. Biochim Biophys Acta 1994; 1199: Alvarez L, Avila MA, Mato JM, Castaño JG, Varela-Nieto I. Insulininduced inositol phosphate-glycan regulates mrna expression in rat hepatocytes. Mol Endocrinol 1991;5: Knowles RG, Salter M, Brooks SL, Moncada S. Anti-inflammatory glucocorticoids inhibit the induction by endotoxin of nitric oxide synthase in the lung, liver and aorta of the rat. Biochem Biophys Res Commun 1990;172: Henning SM, McKeer W, Swendseid ME. Hepatic content of S-adenosylmethionine and glutathione in rats receiving treatments modulating methyl donor availability. J Nutr 1989;119: Lowry OH, Passonneau JV. A collection of metabolite assays. In: Lowry OH, Passonneau JV, eds. A flexible system of enzymatic analysis. New York: Academic, 1972: Hissin PJ, Hill R. A fluorimetric method for determination of oxidized and reduced glutathione in tissues. Anal Biochem 1976;74: Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein dye binding. Anal Biochem 1975;72: Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate phenol chloroform extraction. Anal Biochem 1987;162: Thomas PS. Hybridization of denatured RNA and small DNA fragments transferred to nitrocellulose. Proc Natl Acad Sci USA 1980;77: Alvarez L, Asuncion M, Corrales F, Pajares MA, Mato JM. Analysis of the 5 non-coding region of rat liver S-adenosylmethionine synthetase mrna and comparison of the Mr deduced from the cdna sequence and the purified enzyme. FEBS Lett 1991;290: Xie QW, Cho HJ, Calaycay J, Mumford RA, Swiderek KM, Lee TD, Ding A, Troso T, Nathan C. Cloning and characterization of inducible nitric oxide synthase from mouse macrophages. Science 1992;256: Levi A, Elridge JD, Patterson BM. Molecular cloning of a gene sequence regulated by nerve growth factor. Science 1985;229: Mingorance J, Alvarez L, Sanchez-Gongora E, Mato JM, Pajares MA. Site-directed mutagenesis of rat liver S-adenosylmethionine synthetase. Identification of a cysteine residue critical for the oligomeric state. Biochem J 1996;315: Shan X, Aw TY, Shapira R, Jones DP. Oxygen dependence of glutathione synthesis in hepatocytes. Toxicol Appl Pharmacol 1989;101: Wong GH, Goeddel DV. Induction of manganous superoxide dismutase by tumor necrosis factor: possible protective mechanism. Science 1988;242: Gil B, Pajares MA, Mato JM, Alvarez L. Glucocorticoid regulation of hepatic S-adenosylmethionine synthetase gene expression. Endocrinology 1997;138: Bunn HF, Poyton RO. Oxygen sensing and molecular adaptation to hypoxia. Physiol Rev 1996;76: Kietzmann T, Schmidt H, Probst I, Jungermann K. Modulation of the glucagon-dependent activation of the phosphoenol-pyruvate carboxykinase gene by oxygen in rat hepatocyte cultures. FEBS Lett 1992;311: Goldberg GMC, Gaut CC, Bunn F. Regulation of the erythropoietin gene: evidence that the oxygen sensor is a heme protein. Science 1988;242: Gonzalez-Flecha B, Cutrin JC, Boveris A. Time course and mechanism of oxidative stress and tissue damage in rat liver subjected to in vivo ischemia-reperfusion. Am J Physiol 1994;266:G799 G Caraceni P, Rosenblum ER, Van Thiel DH, Borle AB. Reoxygenation injury in isolated rat hepatocytes: relation to oxygen free

8 February 1998 METHIONINE ADENOSYLTRANSFERASE IN HYPOXIA 371 radicals and lipid peroxidation. Am J Physiol 1994;266:G799 G Meister A. Glutathione-ascorbic acid antioxidant system in animals. J Biol Chem 1994;269: Shan X, Aw TY, Shapira R, Jones DP. Oxygen dependence of glutathione synthesis in hepatocytes. Toxicol App Pharmacol 1989;101: Mato JM, Alvarez L, Ortiz P, Pajares MA. S-Adenosylmethionine synthesis: molecular mechanisms and clinical implications. Pharmacol Ther 1997;73: Chawla RK, Watson WH, Jones DP. Effect of hypoxia on hepatic DNA methylation and trna methyltransferase in rat similarities to effects on methyl-deficient diets. J Cell Biochem 1997;61: Osei SY, Ahima RS, Fabry ME, Nagel RL, Bank N. Immunohistochemical localization of hepatic nitric oxide synthase in normal and transgenic sickle cell mice: the effect of hypoxia. Blood 1996;88: Shan X, Aw TY, Smith ER, Ingelman-Sundberg M, Mannervik B, Iyanagi T, Jones DP. Effect of chronic hypoxia on detoxication enzymes in rat liver. Biochem Pharmacol 1992;43: Graeber TG, Osmanian C, Jacks T, Housman DE, Koch CJ, Lowe SW, Giaccia AJ. Hypoxia-mediated selection of cells with diminished apoptotic potential in solid tumors. Nature 1996;379: Fandrey J, Frede S, Jelkmann W. Role of hydrogen peroxide in hypoxia-induced erythropoietin production. Biochem J 1994;303: Czyzyk-Krzeska M, Paulding WR, Beresh JE, Kroll S. Posttranscriptional regulation of tyrosine hydroxylase gene expression by oxygen in PC12 cells. Kidney Int 1997;51: McQuillan LP, Leung GK, Marsden PA, Kostyk SK, Kourembanas S. Hypoxia inhibits expression of enos via transcriptional and posttranscriptional mechanisms. Am J Physiol 1994;267:H1921 H Liao JK, Zulueta JJ, Yu FS, Peng HB, Cote CG, Hassoun PM. Regulation of bovine endothelial constitutive nitric oxide synthase by oxygen. J Clin Invest 1995;96: Jungermann K, Kietzmann T. Zonation of parenchimal and nonparenchimal metabolism in liver. Annu Rev Nutr 1996;16: Thurman RG, Kauffman FC. Sublobular compartmentation of pharmacologic events (SCOPE): measurement of metabolic fluxes in periportal and pericentral regions of the liver lobule with microfluorometric and micropolarographic techniques. Hepatology 1986;5: Received June 3, Accepted October 23, Address requests for reprints to: Jose M. Mato, Ph.D., Departamento de Medicina Interna, Facultad de Medicina, Universidad de Navarra, Pamplona, Spain. jmmato@unav.es; fax: (34) Supported in part by Fondo de Investigaciones Sanitarias 94/0231, Plan Nacional de Investigación y Desarrollo Salud y Farmacia , Europharma, and Knoll. Dr. Rodriguez was supported by Instituto de Cooperación Internacional and Dr. Carretero by Instituto Reina Sofia de Investigación Renal. The authors thank Estefanía Fernández (supported by Finnova, Comunidad de Madrid) for technical assistance and Drs. F. J. Corrales and I. Varela-Nieto for the critical reading of the manuscript.