Physiological Regulation of Phospholipid Methylation Alters Plasma Homocysteine in Mice*

Biological methylation reactions and homocysteine (Hcy) metabolism are intimately linked. In previous work, we have shown that phosphatidylethanolamine N -meth-yltransferase, an enzyme that methylates phosphatidylethanolamine to form phosphatidylcholine, plays a significant role in the regulation of plasma Hcy levels through an effect on methylation demand (Noga, A. A., Stead, L. M., Zhao, Y., Brosnan, M. E., Brosnan, J. T., and Vance, D. E. (2003) J. Biol. Chem. 278, 5952–5955). We have further investigated methylation demand and Hcy metabolism in liver-specific CTP:phosphocholine cytidylyl-transferase- (cid:1) (CT (cid:1) ) knockout mice, since flux through the phosphatidylethanolamine N -methyltransferase pathway is increased 2-fold to meet hepatic demand for phosphatidylcholine. Our data show that plasma Hcy is elevated by 20–40% in mice lacking hepatic CT (cid:1) . CT (cid:1)

mans, total plasma Hcy concentration normally ranges from 8 to 12 M. However, a small elevation (ϳ5 M) in plasma Hcy increases the risk of coronary artery disease by as much as 60% in men and 80% in women (3). Hyperhomocysteinema has also been correlated with smoking, obesity, diabetes, hypertension, and impaired B vitamin status (1,2). Furthermore, altered Hcy metabolism has been observed in Alzheimer's disease (4) and in the elderly with cognitive impairment (5).
Homocysteine is formed during the metabolism of methionine. Methionine is activated by methionine adenosyltransferase to form S-adenosylmethionine (AdoMet), an important biological methyl donor (6). Numerous methyltransferases catalyze the transfer of the methyl group from AdoMet to a methyl acceptor, producing a methylated product and S-adenosylhomocysteine (AdoHcy), which is subsequently hydrolyzed to form adenosine and Hcy. Homocysteine has several metabolic fates; it can be remethylated to methionine using either betaine, catalyzed by betaine:Hcy methyltransferase (BHMT), or N 5 -methyltetrahydrofolate, catalyzed by methionine synthase (MAT), both of which are methyl donors. The catabolism of Hcy is accomplished by the transsulfuration pathway, which consists of two enzymes, cystathionine ␤-synthase and cystathionine ␥-lyase. Finally, Hcy can be secreted from cells into the circulation.
It is clear that biological methylation and Hcy metabolism are closely related. However, the potential of specific methyltransferases to regulate plasma Hcy levels has been understudied. We have previously utilized the phosphatidylethanolamine N-methyltransferase (PEMT) knockout mouse to investigate the importance of phospholipid methylation in Hcy metabolism (7). PEMT is a liver-specific enzyme that methylates one membrane lipid, phosphatidylethanolamine (PE), into another membrane lipid, phosphatidylcholine (PC) (8). Three AdoMet molecules are consumed in the course of this reaction. Since PEMT accounts for 30% of hepatic PC synthesis (9 -11), we hypothesized that deletion of the gene would significantly alter Hcy metabolism. Indeed, plasma Hcy was reduced by 50% in the Pemt Ϫ/Ϫ mice compared with wild type controls (7). Moreover, hepatocytes isolated from Pemt Ϫ/Ϫ mice secrete less Hcy into the medium, providing clear evidence that Hcy production is reduced in the knockout mice. These studies illustrate the potential for PEMT to regulate plasma Hcy.
Since there is no PEMT activity in the Pemt Ϫ/Ϫ mouse (12), it is impossible to use this model to investigate how increased physiological demand for phospholipid methylation affects Hcy metabolism. As noted above, PEMT activity accounts for 30% of hepatic PC biosynthesis. The remaining 70% of hepatic PC is produced from choline by the enzymes of the Kennedy (CDPcholine) pathway (9 -11), and flux through this pathway is regulated by the activity of CTP:phosphocholine cytidylyltrans-ferase (CT) (13)(14)(15). Hepatic cells express two CT isoforms (16), CT␣ (major expression), and CT␤ (minor expression). We have generated a mouse model in which CT␣ is selectively deleted in the liver (17). In liver-specific CT␣ knockout mice, hepatic CT activity is 15% of normal, whereas PEMT activity is elevated 2-fold compared with controls. We hypothesized that induction of PEMT, while necessary to maintain close to normal hepatic PC levels, would increase hepatic Hcy production and result in elevated plasma Hcy. The results describe an important role for PEMT-derived PC biosynthesis in mice deficient in hepatic CT␣ and furthermore illustrate how physiological changes in PC metabolism influence Hcy metabolism.

MATERIALS AND METHODS
In Vivo Experiments-All procedures were approved by the University of Alberta's Institutional Animal Care Committee and were in accordance with guidelines of the Canadian Council on Animal Care. The liver-specific CT␣ knockout mouse was generated using the Cre/ Lox technology as previously described (17). All mice (12-20 weeks old) were fed ad libitum a chow diet (LabDiet, PICO laboratory Rodent Diet 20) and were exposed to a 12-h light/dark cycle starting at 8:00 a.m.
Primary Hepatocyte Cultures-Primary hepatocytes were isolated by collagenase perfusion (18) and plated on 60-mm collagen-coated dishes at a density of 1.0 ϫ 10 6 cells/dish in Dulbecco's modified Eagle's medium containing 15% fetal bovine serum. 2-4 h after plating, the cultures were rinsed twice in serum-free Dulbecco's modified Eagle's medium over a 1-h period and then incubated in serum-free Dulbecco's modified Eagle's medium (containing 30 M choline and 50 M methionine) for 10 h. Medium was collected and frozen for Hcy measurements. Hcy export from hepatocytes was also assessed by measuring the formation of [ 14 C]Hcy from [1-14 C]methionine. Labeled Hcy was separated from methionine by thin layer chromatography (NH 3 /propanol, 30:70). Following visualization with 0.2% ninhydrin (dissolved in 95% ethanol), the band corresponding to Hcy was scraped, and radioactivity was measured by liquid scintillation.
PC biosynthesis via PEMT was measured in hepatocytes that were incubated with either [ 3 H]ethanolamine, [ 3 H]methionine, or [ 14 C]betaine. Medium was collected, and cells were sonicated in phosphate-buffered saline (pH 7.4). Protein analysis was performed with the BCA protocol (Bio-Rad) using albumin as a standard. Lipids were extracted using the Folch method (19), and phospholipids were separated by thin layer chromatography (chloroform/methanol/acetic acid/water, 25:15:4:2). Following iodine visualization, bands corresponding to PE and PC were scraped, and radioactivity was measured by liquid scintillation counting.
In a separate set of experiments, the flux through the Kennedy (CDP-choline) pathway was measured by incubating cells with [ 3 H]choline for 2 h, after which the cells were sonicated in phosphate-buffered saline (pH 7.4). Lipids were extracted using the Folch method, and phospholipids were separated by thin layer chromatography (chloroform/methanol/acetic acid/water, 25:15:4:2). Following iodine visualization, bands corresponding to PC were scraped, and radioactivity was measured by liquid scintillation. Radioactivity in soluble choline-containing compounds (choline and phosphocholine) was measured in the aqueous phase from the lipid extraction. The aqueous phase was first dried under constant air flow. The samples were then resuspended in phosphate-buffered saline, and choline-containing compounds were separated by thin layer chromatography (CH 3 OH, 0.6% NaCl, NH 4 OH; 10:10:0.9). Following iodine visualization, bands corresponding to choline and phosphocholine were scraped, and radioactivity was measured by liquid scintillation.
Measurement of Amino Acid Concentrations and Enzyme Activities-Livers were removed from mice and flash-frozen. The protein content of the liver samples was determined by the Biuret method (20). Thawed liver samples were homogenized in 50 mM phosphate-buffered saline (pH 7.0). The homogenate was centrifuged at 18,000 ϫ g for 30 min at 4°C, and the following enzyme activities were measured in the supernatant: cystathionine ␤-synthase (21), methionine adenosyltransferase (22), methionine synthase (23), 5,10-methylenetetrahydrofolate reductase (24), and betaine:Hcy methyltransferase (25). To determine CT activity, membranes were isolated by centrifugation of the liver homogenates at 600 ϫ g for 10 min to pellet unbroken cells and nuclei. The supernatant was centrifuged at 100,000 ϫ g for 1 h. The membrane pellet, containing membrane-associated CT, was resuspended in the above homogenization buffer. The supernatant resulting from the 100,000 ϫ g centrifugation contained non-membrane-associated CT. CT activity was measured in the homogenate, as well as in soluble and membrane fractions in the presence of PC-oleate vesicles, as previously described (26).
Total Hcy levels in plasma and media were measured by reversephase high performance liquid chromatography and fluorescence detection of ammonium 7-fluoro-2-oxa-1,3-diazole-4-sulfonate thiol adducts (27). The concentration of free plasma amino acids was measured in samples that had been deproteinized by treatment with 10% sulfosalicylic acid. The protein was removed by centrifugation, and the pH of the supernatant was adjusted to 2.2. The amino acids were analyzed on a Beckman 121 MB amino acid analyzer using Benson D-X 0.25 Cation Xchange Resin according to Beckman 121MB-TB-O17 application notes and quantitated using a Hewlett Packard Computing Integrator model 3395A following postcolumn derivatization with ninhydrin.
AdoMet and AdoHcy Determination-The amounts of AdoMet and AdoHcy were measured in the liver as described previously (28). Briefly, freeze-clamped liver was quickly added to 5 volumes of 8% cold trichloroacetic acid. Samples were homogenized with a Polytron homogenizer for 10 s and were placed on ice for 10 min. The samples were then centrifuged for 10 min at 13,000 ϫ g. The supernatant was filtered through a 0.45-m syringe filter and analyzed by reverse-phase high performance liquid chromatography using a Vydac C18 column (model 2187P54) that was equilibrated with 96% 50 mM NaH 2 PO 4 , 10 mM heptane sulfonic acid (adjusted to pH 3.2 with concentrated sulfuric acid) and 4% acetonitrile. A 15-min linear gradient from 4 to 20% acetonitrile was used to separate AdoMet and AdoHcy. Peaks were monitored by UV detection at 258 nm and quantified using a 3390A Hewlett Packard integrator.
Immunoblotting of Hepatic BHMT1 and MAT1-15 g of liver homogenate was boiled in SDS buffer, and proteins were separated on a 10% SDS-polyacrylamide gel. Proteins were transferred to nylon membranes and probed with anti-BHMT1 (dilution 1:3000), anti-MAT1 (1: 1200), or anti-protein-disulfide isomerase (1:5000) antibody. Immunoreactive bands were visualized by enhanced chemiluminescence according to the manufacturer's instructions (Amersham Biosciences). Quantification of the immunoreactive bands was performed using Image Gauge version 3.0 software by Fuji.
RNA Isolation and Real Time PCR-Total liver RNA was isolated using Trizol reagent (Invitrogen) according to the manufacturer's instructions. Total RNA was treated with DNase I (Invitrogen) to degrade any genomic DNA and was then reverse-transcribed using an oligo(dT) [12][13][14][15][16][17][18] primer and Superscript II reverse transcriptase (Invitrogen) according to the manufacturer's instructions. Bhmt1, Bhmt2, Mat1a, Mat2a, Pemt, and cyclophilin (Cyc) transcripts were detected by real time PCR using a Rotor-Gene 3000 instrument (Montreal Biotech). Reaction mixtures contained Platinum® Quantitation PCR supermix (Invitrogen), a 0.25 M concentration of each primer, and SYBR Green I (Molecular Probes, Inc., Eugene, OR) in a total volume of 25 l. Data analyses were performed using the Rotor-Gene 6.0.19 program (Montreal Biotech). The Pfaffl method (29) was used to compare the relative transcript levels using cyclophilin as the housekeeping gene.
PCR Primer Sequences-All primers were synthesized at the DNA core facility, University of Alberta: Statistical Analysis-Data are presented as means Ϯ S.D. unless otherwise noted. Each experimental group contained 3-10 samples. Student's unpaired t test was performed to compare means unless otherwise specified. A p value of Ͻ0.05 was considered a significant difference.

PC Synthesis from the CDP-choline Pathway Is Impaired in
CT␣-deficient Hepatocytes-The Kennedy (CDP-choline) pathway has been shown to produce ϳ70% of PC in a normal liver (9 -11). Disruption of the hepatic CT␣ gene resulted in an 85% reduction in total CT activity (17). The active form of CT in cells is considered to be associated with membranes, whereas the soluble form of CT is thought to be an inactive reservoir (30,31). Movement of CT to and from the membrane is linked to cell requirements for PC and is tightly regulated (32)(33)(34)(35)(36). Interestingly, all hepatic CT activity was found in the membrane frac-tion in the knockout mice; however, the active form of CT was still only 15% of control levels (Fig. 1A). We sought to determine to what extent the decrease in CT activity in the knockout mice was reflected in flux through the CDP-choline pathway.
To that end, we incubated primary hepatocytes with [ 3 H]choline for 2 h and measured radiolabeled phosphocholine (a substrate for the CT reaction) and PC. We observed that the amount of [ 3 H]phosphocholine was 2-fold higher in the knockout hepatocytes than in control hepatocytes (Fig. 1B). Furthermore, the amount of newly formed PC was 65% less in hepatocytes deficient in CT␣ (Fig. 1B). These data provide clear evidence for impaired flux through the CDP-choline pathway in livers deficient in CT␣.
Increased PE Methylation and Homocysteine Secretion from CT␣-deficient Hepatocytes-Despite the significant reduction in PC biosynthesis by the CDP-choline pathway, hepatic PC levels were only slightly lower (10 -20%) in CT␣-deficient mice than in their wild-type counterparts (17), suggesting that PE methylation may be compensating for the decreased CT activity. Indeed, in our previous work, we reported a 2-fold higher PEMT activity in the knockout mice that might have resulted in increased PC biosynthesis (17 Fig. 2A). Therefore, the ratio of radiolabel in PC and PE, a measure of PEMT flux, was 80% higher in CT␣-deficient hepatocytes than in control hepatocytes. To confirm this result, we also incubated the hepatocytes with [methyl-3 H]methionine. The amount of radio-labeled PC was nearly 2-fold higher in knockout cells than in the control cells (Fig. 2B). Taken together, these experiments show that induction of PEMT in CT␣-deficient hepatocytes (17) results in enhanced PE methylation and provides a compensatory mechanism for the decrease in PC biosynthesis via the Kennedy pathway.
Our model provides the opportunity to examine whether or not the physiological increase in flux through the PEMT pathway, driven by the absence of CT␣, results in elevated Hcy secretion from hepatocytes. We therefore measured the secretion of Hcy from primary hepatocytes isolated from control and knockout mice. Following a 2-10-h incubation period, hepatocytes from knockout mice secreted 40% more Hcy into the media than did hepatocytes from control mice whether measured by radioactivity (Fig. 3A) or by mass (Fig. 3B). These data indicate that a physiological enhancement of methylation demand does indeed increase Hcy secretion from hepatocytes.
Plasma Levels of Hcy Are Increased in Hepatic CT␣-deficient Mice- Fig. 4 shows that mice deficient in hepatic CT␣ have higher levels of plasma Hcy than do control mice. We observed a 20% higher level of plasma Hcy in male knockout mice than in male control mice, whereas in female knockout mice, plasma Hcy was 40% higher than in their female littermates. The plasma concentration of methionine, cysteine, serine, glycine, and taurine were not significantly affected by genotype in either male or female mice (data not shown).
The increase in plasma Hcy, coupled with an increase in Hcy secretion from the knockout hepatocytes, suggests that CT␣ deficiency in the liver appreciably alters Hcy metabolism. Therefore, we assayed the major enzymes involved in the production (transmethylation) and removal (transsulfuration and remethylation) of Hcy in the liver. The activity of BHMT, a key enzyme in choline oxidation, was increased by 80% in the livers of knockout mice compared with gender-matched controls (Table I). Similarly, the activity of MAT, which synthesizes AdoMet, was 2-fold higher in both male and female knockout mice. CT␣ deficiency did not alter the activity of methionine synthase, methylenetetrahydrofolate reductase, or cystathionine ␤-synthase in the livers (Table I).
To address the possible mechanism(s) for the increased activity of MAT and BHMT, we measured the mRNA levels of their respective genes by real-time PCR (Fig. 5A). Bhmt1, Mat1a, and Pemt mRNA levels were increased by 2.2-, 1.7-, and 2.4-fold, respectively, as compared with controls. However, the mRNA levels of Bhmt2 and Mat2a were unaltered. Protein levels of BHMT1 and MAT1 were also elevated in knockout liver homogenate (Fig. 5B), which is consistent with the observed increase in enzyme activity.
These data suggest that betaine (choline) oxidation is increased in the knockout mice, possibly as a mechanism for furnishing a supply of methyl groups for the increased rate of phospholipid methylation. To investigate this possibility, we incubated primary hepatocytes with [ 3 H]choline in the presence of either 50 M or 1 mM methionine. If choline oxidation had been stimulated to provide methyl groups for PE-methylation, a high concentration of unlabeled methionine would be expected to prevent the incorporation of radiolabel into PC. The amount of radiolabeled PC in control hepatocytes after a 2-h incubation was unaltered by incubation with 1 mM compared with 50 M methionine (Fig. 6A). However, when knockout hepatocytes were incubated with 1 mM methionine, the amount of radiolabeled PC was 25% less than in hepatocytes incubated with 50 M methionine (Fig. 6A). To confirm this result, hepatocytes were incubated with [ 14 C]betaine, and radiolabeled PC was measured. [ 14 C]PC production was increased 1.7-fold in knockout hepatocytes as compared with controls (Fig. 6B). These results clearly highlight the importance of choline (betaine) oxidation in the supply of methyl groups in CT␣-deficient hepatocytes.
The hepatic levels of AdoMet and AdoHcy were also meas-ured. A deficiency of hepatic CT␣ did not change the amount of either metabolite (data not shown), suggesting that the induction of BHMT and MAT was successful in maintaining a normal hepatic ratio of AdoMet to AdoHcy in the knockout mice. These results also indicate that AdoMet is not limiting for Hcy formation and that the increased secretion of Hcy from CT␣deficient hepatocytes is attributable to an increased flux through the PEMT pathway.

DISCUSSION
Methylation Demand and Homocysteine-Studies on the role of PEMT in regulating Hcy production have been limited, in part, by the lack of an appropriate model. For example, modulation of dietary choline and methionine levels have been shown to regulate PEMT activity/expression (37)(38)(39). However, choline and methionine are important substrates for the methionine cycle and also modulate Hcy metabolism (40). Experiments using AdoMet hydrolase inhibitors have shown that PEMT is important in hepatic PC biosynthesis (41,42); this model is of limited value, since most methylation reactions are inhibited and Hcy formation cannot occur. Finally, flux through the PEMT pathway has been shown to be insensitive to dietary changes of substrates such as ethanolamine and phosphatidylethanolamine (43,44).
The generation of PEMT knockout mice (12) provided an important tool for investigating the specific relationship between phospholipid methylation and Hcy production and, in a broader sense, the role of methylation demand. The level of plasma Hcy is 50% lower in Pemt Ϫ/Ϫ mice compared with wild-type mice, suggesting that flux through the PEMT pathway is an important determinant of plasma homocysteine levels (7). Our current report builds substantially on this initial observation. PC is vital for the structural integrity of mammalian membranes and, in the liver, is secreted into the bile and plasma as a constituent of lipoproteins. Under normal circumstances, the CDP-choline pathway produces 70% of hepatic PC (9 -11). However, in liverspecific CT␣ knockout mice, PC synthesis via the hepatic CDPcholine pathway is reduced to 20% of control levels (Fig. 1B). In order to compensate for the lack of PC biosynthesis, PEMT mRNA (Fig. 5A), protein, and activity (17) are induced in the CT␣-deficient livers, resulting in an increased flux through the PEMT pathway in hepatocytes. Importantly, these changes occur without manipulation of dietary substrates (methionine or choline), which allowed us to study directly to what extent increased PE methylation modulates Hcy metabolism. We show, for the first time, that physiological stimulation of PEMT increases production of Hcy in the liver and, consequently, increases the level of plasma Hcy. These results are consistent with earlier experiments in cultured hepatocytes in which a direct relationship was shown between the level of PEMT expression and homocysteine export. 2 Our data in the PEMT and the liver-specific CT␣ knockout mice challenge the long held assertion that creatine synthesis is the major consumer of AdoMet molecules and, therefore, the main methylation reaction involved in the generation of Hcy (6). Other studies provide data to support our assertion. Reo et al. (10) have estimated a rate of synthesis of PEMT-derived PC in rats to be 0.15-0.21 mol/h/g liver. When this number is multiplied by 10 (10 g of liver in a 250-g rat) and by 24 (for a 24-h day), the rate of PC synthesis via the PEMT reaction is 36 -50 mol/250-g rat/day. Since three AdoMet molecules are consumed for each PC molecule synthesized, the amount of AdoMet utilized by PEMT would, therefore, be 108 -150 mol/ 250-g rat/day. Creatine biosynthesis begins with the synthesis of guanidinoacetate in the kidney. L-Arginine:glycine amidino-2 D. J. Shields and D. E. Vance, unpublished data. transferase catalyzes the transfer of the amidino group of arginine to glycine, yielding ornithine and guanidinoacetate. Guanidinoacetate is carried in the blood to the liver, where it is methylated by AdoMet via guanidinoacetate N-methyltransferase to form creatine, which is then exported from the liver to extrahepatic tissues. Creatine and creatine phosphate are converted nonenzymatically to creatinine, the excreted form. In rodents, creatinine excretion has been measured at a rate of 58 mol/250-g rat/day (45). If the rate of creatinine excretion equaled the rate of creatine synthesis (a reasonable assumption), PEMT would consume at least 2-3 times more AdoMet than would guanidinoacetate N-methyltransferase and thus be a quantitatively more important source of plasma Hcy. We have previously studied the contribution of creatine synthesis in regulating plasma Hcy in rats (46). We hypothesized that if creatine synthesis accounted for the majority of AdoMet consumed, the provision of creatine in the diet would decrease plasma Hcy. We observed a 90% reduction in arginine:glycine amidinotransferase activity (the enzyme that synthesizes guanidinoacetate) but only a ϳ20% reduction in plasma Hcy, less than would be anticipated if creatine synthesis accounted for 75% of AdoMet consumption (6). In contrast, deletion of PEMT in mice reduced the level of plasma Hcy by 50% (7), a much greater reduction than would have been expected if PE methylation consumed only 15% of AdoMet (6).
Importance of BHMT-There is a clear link between PE methylation and plasma Hcy levels. However, neither the hepatic secretion of Hcy nor the plasma level of Hcy can be used as a measure of flux through the PEMT pathway. Indeed, changes in either the remethylation or transsulfuration path-  5. Hepatic expression of Bhmt1, Mat1a, and Pemt is increased in liver-specific CT␣ knockout mice. A, the Pfaffl method was utilized to determine relative mRNA levels using cyclophilin as the housekeeping gene. Means Ϯ S.D. for three samples are shown. The asterisks signify differences versus control, p Ͻ 0.05. B, hepatic proteins were fractionated by SDS-polyacrylamide gel electrophoresis and immunoblotting was performed using antibodies raised against BHMT1, MAT1, and protein-disulfide isomerase (PDI). For each measurement, there were three samples from each test group. Intensities of band were quantitated using Image Gauge version 3.0 software by Fuji and the relative levels of BHMT1 and MAT1 were compared with PDI. way can reduce hepatic Hcy secretion, thereby masking Hcy generation by PEMT. We have now observed that BHMT activity was increased by 80% in the CT␣ knockout mice (Table I). Furthermore, we have provided evidence that choline (betaine) oxidation is important in providing methyl groups for the PEMT reaction in hepatocytes isolated from the knockout mice (Fig. 6). The ability of BHMT to regulate plasma Hcy has received much attention in recent years. Schwahn et al. (47) observed a negative correlation between plasma betaine and plasma Hcy concentrations in humans. Dietary betaine supplementation has been shown to reduce plasma Hcy in humans (47,48), and in wild-type, Mthfr Ϫ/Ϫ (47,49), and Cbs Ϫ/Ϫ (50) mice. Even without betaine supplementation, Mthfr Ϫ/Ϫ and Cbs Ϫ/Ϫ mice have elevated flux through the BHMT reaction compared with their wild-type counterparts (47,50). Moreover, in these mice, increased Hcy remethylation was found to be necessary to reduce intracellular Hcy levels and thus AdoHcy (an inhibitor of AdoMet-dependent methylation reactions) and to provide more methyl groups in the form of methionine. Although the increase in plasma Hcy in the CT␣ knockout mice is not as great as that in the Mthfr Ϫ/Ϫ and Cbs Ϫ/Ϫ mice, the same principle for the induction of BHMT apparently applies. Indeed, in order to compensate for the increased demand for methylation reactions to generate PC, hepatic levels of BHMT and MAT are increased in CT␣-deficient livers. Although this induction does not normalize plasma Hcy, it does maintain AdoMet and AdoHcy at control levels.
Strikingly, the PEMT-dependent increase in plasma Hcy is dependent on the gender of the mouse. Plasma Hcy was increased by 40% in female hepatic CT␣ knockout mice compared with that in gender-matched controls but only by 20% in the male knockout mice. These results are consistent with a study that showed that female rats produce a larger fraction of their hepatic PC from PE methylation compared with male rats (51)(52)(53)(54). Differential rates of Hcy remethylation might also explain the gender difference in plasma Hcy levels. Indeed, several studies have suggested that flux through the remethylation pathway is greater in male than in female rats (6,47,55). We note that in our study, the gender of the mice did not affect flux through the PEMT pathway or the rate of Hcy secretion from primary hepatocytes.
Homocysteine, PEMT, and Atherosclerosis-It is clear that the hepatic PEMT reaction is a major producer of Hcy, an independent risk factor for cardiovascular disease (1,2). The explanation for the correlation between plasma Hcy levels and cardiovascular disease remains elusive, but studies have demonstrated that elevated plasma Hcy is associated with reduced endothelium-dependent vasodilatation (56) and recruitment of monocytes to atherosclerotic lesions (57). Although there is an abundance of epidemiological data linking Hcy to the development of atherosclerotic lesions, it has not been demonstrated that Hcy is a cause of atherosclerosis. Recent studies showed that patients with mild hyperhomocysteinemia and without any other risk factor for cardiovascular disease have no increased risk for the disease (58). Recently, ApoE Ϫ/Ϫ mice, which develop atherosclerotic lesions similar to those in humans, have been used to investigate the effects of altered homocysteine metabolism on cardiovascular disease. Dietary manipulation, either with supplementation with methionine or Hcy or removal of B vitamins, has been shown to promote atherosclerosis in these mice (59 -62). Zhou et al. (63) have suggested that multiple cellular stress pathways, including endoplasmic reticulum stress, are associated with atherosclerotic lesion development in ApoE Ϫ/Ϫ mice. The low density lipoprotein receptor knockout mice also develop atherosclerosis (64 -66). Preliminary data have shown that deletion of PEMT reduces athero-sclerotic plaque formation in the Ldlr Ϫ/Ϫ (low density lipoprotein receptor) knockout mice by lowering the hepatic secretion rate of very low density lipoprotein (67) and thus the level of plasma cholesterol. 3 Hepatic Hcy secretion is also reduced in Pemt Ϫ/Ϫ compared with Pemt ϩ/ϩ mice (7), and it is possible that the resulting decrease in plasma Hcy plays a role in the reduction of atherosclerotic lesions. Alternatively, the level of Hcy might represent a marker of atherosclerosis. In our future work, we shall determine if elimination of CT␣ in the liver affects the development of atherosclerosis, since these mice have a reduction in plasma cholesterol (17) but an increase in plasma Hcy levels.
Conclusions-Induction of PEMT in the liver of hepatic CT␣deficient mice results in a 20 -40% increase in plasma Hcy. We have previously shown that deletion of PEMT results in a 50% reduction in plasma Hcy (7). Taken together, it is now clear that regulation of hepatic PEMT is not only important in producing PC but is also a major source of plasma Hcy.