Sally P. Stabler * and Robert H. Allen. Metabolism. Clinical Chemistry 50: (2004) Endocrinology and

Transcription

1 Clinical Chemistry 50: (2004) Endocrinology and Metabolism Quantification of Serum and Urinary S-Adenosylmethionine and S-Adenosylhomocysteine by Stable-Isotope- Dilution Liquid Chromatography Mass Spectrometry Sally P. Stabler * and Robert H. Allen Division of Hematology, Department of Medicine, University of Colorado Health Sciences Center, Denver, CO *Address correspondence to this author at: 4200 East 9th Ave., Box B170, Denver, CO Fax ; Sally.Stabler@UCHSC.edu. Received August 28, 2003; accepted November 11, Previously published online at DOI: /clinchem Background: We have developed an assay that uses stable-isotope-dilution liquid chromatography mass spectrometry to assess S-adenosylmethionine (SAM) and S-adenosylhomocysteine (SAH) in body fluids to investigate the relationship of these metabolites to hyperhomocysteinemia. Methods: Commercially obtained SAM (D 3 methyl) and 13 C 5 -SAH uniformly labeled in the adenosyl moiety, which was synthesized using S-adenosylhomocysteine hydrolase, were added to samples followed by perchloric acid protein precipitation, C 18 chromatography, and analysis by liquid chromatography mass spectrometry with quantification by comparison of the areas of internal standard and endogenous peaks. Results: Estimates of intraassay imprecision (CV) were 5% and 17% for SAM and SAH, respectively (n 10). SAM decreased and SAH increased in serum and plasma samples at both 4 C and room temperature over 80 h. SAM and SAH were unstable in samples stored longer than 2 years at 20 C. In 48 volunteers, the estimated reference intervals [from mean (2 SD) of log-transformed data] for serum SAM and SAH were and 8 26 nmol/l, respectively. Fractional excretion of SAM was higher than that of SAH, and the urinary SAM:SAH ratio was much higher than the serum or erythrocyte SAM:SAH ratios. Conclusions: Stable-isotope-dilution liquid chromatography mass spectrometry can be used to quantify SAM and SAH in biological fluids and tissues. Sample handling and storage must be stringently controlled for any epidemiologic or clinical use of such assays American Association for Clinical Chemistry Serum total homocysteine (thcy) 1 is increased in both vitamin B 12 and folate deficiency and is corrected by treatment with the appropriate vitamin(s) (1, 2). Patients with renal failure also have increased thcy (3, 4), which is partially responsive to high-dose vitamin therapy (4). Increased thcy has been associated with thrombosis and other vascular diseases (5), mortality from vascular diseases and other causes (6), and cognitive disorders (7). It is not completely evident, however, whether these deleterious effects are attributable to hyperhomocysteinemia or to the underlying vitamin deficiency, with the increased thcy simply being a marker for the deficiency condition. Abnormal methylation of crucial neurotransmitters, phospholipids, myelin, and/or other proteins has been a long-standing postulated explanation for the neurologic and cognitive disorders frequently seen in vitamin B 12 deficiency (7). This hypothesis has been based on the folate and B 12 -dependent conversion of homocysteine to methionine, which is the precursor for S-adenosylmethionine (SAM). SAM is the most important physiologic methylator (8). The product of the methylation, S-adenosylhomocysteine (SAH) is an inhibitor of many of these reactions (9). The ratio of SAM to SAH has been termed the methylation index, and it could be abnormal because of a decrease in SAM, an increase in SAH, or both. SAH is cleaved to homocysteine and adenosine by SAH 1 Nonstandard abbreviations: thcy, total homocysteine; SAM, S-adenosylmethionine; SAH, S-adenosylhomocysteine; LC/MS, liquid chromatography mass spectrometry; SIM, selected-ion monitoring; and Cbl, cobalamin. 365

2 366 Stabler and Allen: Serum SAM and SAH hydrolase, a reaction that, however, is favored in the reverse direction (10). It is thought that when homocysteine or adenosine increases, significant amounts of SAH would be formed in vivo. Experimental vitamin B 12 deficiency induced by nitrous oxide was reported to change the methylation ratio in brain and spinal cord as the result of a build up of SAH (11). However, there have been few data in humans about the relationships of thcy to SAH and SAM until recently because of problems inherent in measuring these metabolites in readily accessible body fluids, such as plasma, where they are present in nanomolar quantities. A HPLC assay had been reported for erythrocyte SAM and SAH, and it has been shown that SAH is increased in the erythrocytes of patients with renal failure (12, 13). More sensitive assays using HPLC (14 16) or, most recently, stable-isotopedilution tandem mass spectrometry (17) have shown significant correlations between thcy and SAH (16, 18, 19) as well as between creatinine and these metabolites (18, 19). We have shown that serum SAH was increased, and then decreased with vitamin B 12 therapy, in a pilot study of patients with severe megaloblastic anemia attributable to vitamin B 12 deficiency (20). We describe here a stable-isotope-dilution liquid chromatography mass spectrometry (LC/MS) method used to determine reference intervals for serum, urine, and erythrocytes and to study relationships of other metabolites and clinical variables with SAH and SAM. Materials and Methods U- 13 C 5 -Adenosine (98%) was obtained from Cambridge Isotope Laboratories, Inc. D 3 -S-Adenosyl-l-methionine, tri(p-toluenesulfonate) salt (99 atom % D; 85% chemical purity; D 3 -SAM) was obtained from C/D/N Isotopes. d,l-homocysteine and S-adenosyl-l-homocysteine hydrolase from rabbit erythrocytes were obtained from Sigma, and Tris base was obtained from Sigma-Aldrich. C-18 YMC-gel, ODS-A (12 nm I-230/70 mesh) was obtained from YMC Co., Ltd. HPLC-grade methanol and other solvents were used throughout. Heptafluorobutyric acid (99%) was obtained from Aldrich Chemical Company. methods Synthesis of S-adenosyl-U- 13 C 5 -homocysteine ( 13 C 5 -SAH). SAH containing five 13 C molecules in the adenosyl moiety was synthesized by incubating 20 moles of U-( 13 C 5 )- adenosine with d,l-homocysteine (100 moles) in the presence of 20 nmoles of Tris base, 77 moles of dithiothreitol, and 2 U of SAH hydrolase in a total volume of ml at 35 C for 24 h. An additional 2 U of SAH hydrolase were added after 3 h. The reaction mixture was split and then applied to two individual columns ( cm) prepared with 20 m polyethylene frits and 200 mg of YMC-gel that had been prewashed with 3 ml of methanol and three 3-mL volumes of H 2 O. The loaded columns were washed four times with three 1-mL volumes of H 2 O, and the run-through was collected. The water wash was followed by three 1-mL washes of 750 ml/l methanol, and the run-through was collected. The aliquots were assayed for SAH as described below. The aliquots with the greatest concentrations of SAH (second water wash through first methanol wash) were pooled. The labeled SAH was quantified by assaying it with known quantities of SAH. The pooled material was aliquoted and stored frozen at 80 C. The recovery of SAH was 28% compared with the starting amount of labeled adenosine. A combination internal standard mixture was prepared by combining the commercially obtained D 3 - SAM and the synthesized 13 C 5 -SAH to give final concentrations of 40 mol/l D 3 -SAM and 20 mol/l 13 C 5 -SAH in 1 mmol/l HCl. The internal standard was used at this concentration for testing whole blood or tissue samples and was diluted 1:5 in H 2 O for serum, urine, or plasma. Sample preparation. Diluted internal standard containing 0.4 nmol of D 3 -SAM and 0.2 nmol of 13 C 5 -SAH was added to 0.5 ml of plasma, serum, or cerebrospinal fluid or to 50 L of urine; 1.0 ml of H 2 O (or 1.5 ml of tissue culture medium) and ml of 100 g/l perchloric acid solution in H 2 O were then added, followed by 5 L of 99% pure heptafluorobutyric acid. After mixing, the tubes were centrifuged for 30 min at g. The supernatant was loaded on YMC gel columns (200 mg of gel; cm; fitted with a frit) that were prewashed once with 3 ml of acetonitrile, four times with 3 ml of H 2 O, and once with 3mLofH 2 O containing 10 mmol/l heptafluorobutyric acid NH 4 OH, ph 3.2. The loaded columns were then washed four times with 3 ml of the heptafluorobutyric acid solution and were eluted with 1 ml of 750 ml/l acetonitrile in H 2 O. The eluates were taken to dryness in a Savant vacuum centrifuge, dissolved in 50 L ofh 2 O, and analyzed by LC/MS. Fresh or frozen whole blood was mixed thoroughly. We then added 50 L of the concentrated internal standard mixture and 1 ml of H 2 O, after which we added ml of 100 g/l perchloric acid with vortex-mixing. Samples of tissues such as liver or lung ( mg wet weight) were homogenized in 1 5 mlofh 2 O with 50 L of the concentrated internal standard in a VirTishear (VirTis Co.) at approximately half speed for 15 s, followed by homogenization twice with ml of 100 g/l perchloric acid for 15 s on high speed. Prepared samples were frozen at 80 C or assayed fresh by the method described above for plasma or serum after centrifugation for 30 min at g. Instrumentation and analysis. Analysis by LC/MS was performed with a Hewlett Packard Series 1100 MSD equipped with a Hewlett Packard Kayak XM Gap 600 Chem Station. Analysis was performed on an Aquasil C 18 column [ mm (i.d.); particle size, 5 m; Thermo Hypersil-Keystone; Keystone Scientific, Inc.]. Starting conditions for analysis included a column temperature of

3 Clinical Chemistry 50, No. 2, C, and sample tray temperature of 6 C, The instrument was operated in positive ion atmospheric pressure ionization-electrospray mode with flow at 0.3 ml/min. The following settings were used: capillary voltage, 3000 V; nitrogen drying gas temperature, 350 C; nitrogen drying gas flow, 10 L/min; nebulizer pressure set point, 50 psi. The instrument was set for high-resolution selected-ion monitoring (SIM) with gain at 4 and fragmenter at 70. The solvents for LC analysis were as follows: solvent A, 100% H 2 O; solvent B, 100 ml/l H 2 O 900 ml/l methanol. Both solvents A and B contained 10 mmol/l heptafluorobutyric acid NH 4 OH, ph 3.2. The injection volume was L. Initial conditions were 99.9% A 0.1% B for 1.5 min, then a linear gradient up to 100% B over 14 min and holding for 35 min. The column was equilibrated at the initial conditions for 20 min between injections. The MSD start time was 7 min, and stop time was 15 min. Quantification. The protonated molecular ions for SAM and SAH were monitored. Unlabeled SAH (endogenous) was monitored at m/z 385, and the internal standard was monitored at m/z 390. The unlabeled SAM was monitored at m/z 399 and the labeled SAM at m/z 402. The stableisotope-enriched adenosine used in the synthesis of SAH was only 98% pure; thus the labeled SAH contained ion peaks at both m/z 390 (98%) and m/z 385 (2%). The internal standard mixture was run as a blank test mixture in each assay, and the peak area monitored at m/z 385 in the internal standard was then used to adjust the values in the samples. No correction was needed for the SAM internal standard because there was 0.1% peak area of m/z 399 compared with the peak area at m/z 402. Quantification was achieved by comparing the ratio of the peak areas of the endogenous m/z 385 or m/z 399 peaks with the peak areas of the respective internal standards, m/z 390 and m/z 402, and multiplying by the known quantity of internal standard added with further adjustments for the volume of sample added. clinical samples The collection of samples by phlebotomy from healthy individuals was approved by the University of Colorado Institutional Review Board after informed consent. Samples were obtained from nonfasting seated individuals from antecubital veins. In general, venous blood samples were allowed to clot for 15 min, and then iced at 4 C and centrifuged to remove serum or plasma immediately. Serum and plasma were stored frozen at either 20 or 80 C. Urine was fresh frozen. Venous blood was drawn into tubes suitable for serum and into EDTA- or heparincontaining tubes for plasma preparation at the same phlebotomy session to test differences. Serum, plasma, and urine was stored in polypropylene tubes for 2, 4, 24, and 80 h at either room temperature (21 C) or 4 C ina cold room. Total homocysteine, cystathionine, methylmalonic acid, methionine, and cysteine were analyzed by previously developed stable-isotope-dilution capillary gas chromatography mass spectrometry methods (21, 22). The central 95% intervals for some of these metabolites were determined previously from samples from 60 blood donors and were as follows; homocysteine, mol/l; methylmalonic acid, nmol/l (23, 24). Serum and urinary creatinine were measured in the University of Colorado Hospital Clinical Laboratory. Serum cobalamin (Cbl) and serum folate were measured by the DPC, Dual Count Solid Phase No Boil Assay (Diagnostic Products Corp.). The hematocrit was determined by centrifugation of whole blood in microtubes. The following equation was used to calculate the fractional excretion of SAM or SAH: SAM or SAH creatinine clearance ratio [urinary SAM (nmol/l) 1/urinary creatinine mol/l]/[serum SAM (nmol/l) 1/serum creatinine ( mol/l)] 100. statistical evaluations Most of the measured variables had a distribution skewed toward higher values. Log-transformed values were therefore used to estimate the central 95% interval as the mean 2 SD. The Spearman correlation coefficient ( ) was used to test bivariate associations between variables. For multivariate analysis of factors predicting serum SAM, SAH, and thcy, a stepwise multiple linear regression model was used. The level of significance for all analysis was The differences in means for continuous variables across two categories were evaluated with a t-test and the Levene test for equality of variances. The analyses were performed with SPSS Base software, Ver (SPSS, Inc.). Results Shown in Fig. 1 are chromatograms of a sample of normal human serum that had been prepared as described in Materials and Methods and subjected to LC/MS as described above. When serum was prepared and analyzed without added internal standard, there were no significant peaks interfering with the internal standard quantification (data not shown). The sample preparation, including evaporative centrifugation, took 5 h, and LC/MS analysis took 1 h/sample; therefore, 40 samples were routinely quantified over a 48-h interval. Separated serum, plasma, or urine samples (n 2 5) were incubated for 2 80 h at 4 C and room temperature to test the stability of the metabolites, as shown in Table 1. After 80 h at room temperature, there was virtually no detectable SAM in serum, and the SAH had increased 350%. There was a 20% difference in SAM and SAH values in serum vs heparin- or EDTA-plasma that was drawn at the same phlebotomy session from the same individual. Freezing and thawing either the serum or heparin-plasma three times did not significantly change the SAM or SAH concentrations.

4 368 Stabler and Allen: Serum SAM and SAH assay precision and detection limit One healthy control was phlebotomized four times over 9 months; the SAM concentration ranged from 83 to 100 nmol/l and SAH from 10 to 22 nmol/l without any trend. Serum from healthy controls was aliquoted and assayed in each day s assay. Table 2 shows the interassay CV determined for samples from three healthy controls assayed over time. The intraassay CV is also shown in Table 2. Samples stored at 20 C for 2 years had decreases in SAM of at least 50% and increases in SAH of %. The sensitivity of the assay was assessed by assaying volumes of serum from 25 to 500 L. The SAM result was only 1% different with 250 L of serum compared with 500 L, which is within the intraassay variability. However, the result with 100 L varied by 16%, 50 L by 2%, and 25 L by 12%, suggesting that values obtained from serum volumes 250 L were unreliable. The SAH value obtained by assaying 250 Lof serum was 31% lower than that obtained with 500 L of serum, and SAH was not quantifiable in smaller amounts of serum. Therefore, the limit of quantification for either SAM or SAH is 5 pmol in a serum sample. We recovered 96% and 94% of the SAH and SAM, respectively, when we added enough unlabeled material to increase the serum values by 40 and 75 nmol/l. We added unlabeled SAH and SAM to the autosampler vials immediately before injection on the LC/MS system or before vacuum evaporative centrifugation and compared the peak areas with those obtained with added material at the start of the assay. The peak areas representing SAH for the three above conditions were within 3% of each other. In contrast, the peak area for SAM added at the beginning of the procedure was only 55% and 56% of the peak areas obtained when the unlabeled SAM was added immediately before injection on the LC/MS system or before vacuum evaporative centrifugation, respectively. Fig. 1. Chromatograms for normal human serum prepared as described in Materials and Methods. The top two panels show the peaks for endogenous SAH monitored at m/z 385 by SIM and the internal standard peak monitored at m/z 390, eluting at 10.3 min. The bottom two panels show the endogenous SAM peak monitored at m/z 399 and internal standard peak monitored at m/z 402, eluting at 10.4 min. estimation of human reference intervals Serum, urine, and whole blood were collected from 48 healthy controls [mean (SD) age, 38.5 (11.2) years; range, years; 44% female]. There were 39 whites (6 white Hispanics), 1 Native American Hispanic, 3 African Americans and 5 Asians. The occupations of the controls were diverse. Multivitamins were consumed at least weekly by 40% of the group. The log-normalized mean and estimated reference intervals for SAM, SAH, and the SAM: SAH ratio and some of the other metabolites are shown in Table 3. The mean thcy, methylmalonic acid, and creatinine concentrations were within the appropriate reference intervals, as shown in Table 3. The mean (SD) serum Cbl was 412 (493) pmol/l, and serum folate was 65.3 (26.4) nmol/l. The mean hematocrit was 46%, a value consistent with the altitude of Denver, CO. There was no significant difference in age between the males and females in the healthy controls. As expected, serum creatinine, hematocrit, and thcy were significantly higher in males than in females. Mean serum SAM

5 Clinical Chemistry 50, No. 2, Table 1. Mean (SD) change from baseline a of SAM and SAH in samples incubated in various conditions before analysis. Time of incubation 2h b 4h c 24 h d 80 h e Sample Temperature SAM SAH SAM SAH SAM SAH SAM SAH Serum or plasma 4 C 0 (14)% 10 (9)% 3 (9)% 19 (6)% 4 (7)% 33 (26)% 16 (2)% 76 (69)% Room temperature 1 (7)% 27 (13)% 12 (9)% 32 (22)% 41 (16)% 185 (88)% 97 (8)% 350 (218)% Urine 4 C 0 (3)% 37 (55)% 4 (2)% 1 (21)% Room temperature 7 (20)% 10 (11)% 12 (34)% 25 (16)% a SAM and SAH are expressed as the percentages of the concentrations in the individual baseline samples, which were stored at 80 C until the assay and thawed in ice. A negative sign denotes loss in value. Initial concentrations were nmol/l for SAM and 8 22 nmol/l for SAH in serum or plasma. b n 3 for plasma and serum; n 2 for urine. c n 4 for plasma and serum; n 2 for urine. d n 5 for plasma and serum at room temperature; n 4 for plasma and serum at 4 C; n 2 for urine. e n 3 for plasma and serum at room temperature; n 2 for plasma and serum at 4 C. concentrations were 119 and 102 nmol/l in males and females (P 0.025), and serum SAH was also higher, 17 vs 13 nmol/l (P 0.001). The SAM:SAH ratio trended to be higher in females (8.3) than in males (7.2; P 0.056). The red cell SAM was higher in males, 3.9 vs 3.4 mol/l (P 0.001), although red cell SAH and the SAM:SAH ratio were not different. The fractional excretion of SAM and SAH and the serum Cbl and serum folate values were not different between males and females. The use of multivitamins did not produce any differences in the concentrations of SAM, SAH, vitamins, or the other methionine metabolites. We calculated the mean fractional excretion of SAM and SAH compared with that of creatinine and obtained values of 93% for SAM and 39% for SAH, which may explain the high ratio of SAM to SAH in urine. In contrast, the fractional excretion of thcy and methionine in this group of healthy controls was 0.37% and 0.21%, respectively. Samples were obtained from two individuals on regular hemodialysis before their dialysis session, and the serum SAM values were 177 and 286 nmol/l. The serum SAH values were 244 and 402 nmol/l in the same patients; thus, the SAM:SAH ratios were 0.73 and 0.71, respectively. Values for rat liver and rat lung are shown in Table 4. The rat lungs were perfused with saline to prevent blood contamination and had an extremely high SAM:SAH ratio of 84. Table 2. Inter- and intraassay CV for serum samples. Mean CV, % (nmol/l) Individuals n a Time b SAM SAH Interassay months 5.7 (102) 23 (19) months 11 (89) 16 (13) months 7.4 (108) 22 (11) Intraassay day 5.2 (109) 17 (9) a Number of assayed samples. b Time interval during which samples were assayed. correlations between variables The Spearman correlation coefficients and P values for SAM, SAH, and the related metabolites are shown in Table 5. Red blood cell SAM and SAH did not correlate with serum or urinary SAM and SAH and thus are not shown in Table 5. Serum creatinine was strongly correlated with serum SAH, trended with serum SAM, and was strongly inversely correlated with the serum SAM: SAH ratio. Serum creatinine was also strongly correlated with urinary SAH and inversely correlated with the urinary SAM:SAH ratio but did not correlate with the urinary SAM value. The serum SAM and SAH values were directly correlated with their concentrations in urine, and the serum SAM:SAH ratio was inversely correlated to the urinary SAH but not the urinary SAM concentration. The serum and urinary SAM:SAH ratios were directly correlated. Serum thcy was weakly correlated with serum SAH; inversely correlated with serum SAM:SAH ratio; and correlated with serum folate. Serum Table 3. Geometric mean and estimated central 95% values from 48 healthy controls. Mean Estimated central 95% a Serum SAM, nmol/l SAH, nmol/l SAM:SAH ratio Red blood cells SAM, mol/l SAH, mol/l SAM:SAH ratio Urine SAM, mol/l SAH, mol/l SAM:SAH ratio Serum thcy, mol/l Methylmalonic acid, nmol/l Creatinine, mol/l a Central 95% interval cutpoints were estimated as the mean 2SDofthe log-transformed results.

6 370 Stabler and Allen: Serum SAM and SAH Table 4. Mean (SE) values for SAM, SAH, and SAM:SAH ratio in rat liver and lung. SAM SAH SAM: SAH ratio Rat liver, nmol/g wet 72 (3) 20 (1) 3.6 (0.3) weight (n 3) Rat lung, a nmol/mg protein (n 3) 1443 (32) 17 (0.6) 84 (2.0) a Perfused to eliminate blood contamination of tissue and expressed per mg of protein because wet weight was altered by perfusion. Cbl and folate were directly correlated, but neither vitamin correlated with the serum SAM or SAH values, although the serum SAM:SAH ratio was directly correlated with serum folate. Serum creatinine was correlated with serum methionine, total cysteine, N-methylglycine, and N,N-dimethylglycine, but not serum cystathionine. Serum cystathionine (not on Table 5) directly correlated with SAH (Spearman 0.39; P 0.006), trended with serum SAM ( 0.28; P 0.056), and correlated with serum thcy ( 0.399, P 0.005). Correlations between the other sulfur-containing amino acids and SAM and SAH were corrected for serum creatinine because of the strong influence of serum creatinine on the values. After we controlled for creatinine, thcy was no longer correlated with SAH and was now inversely correlated with SAM. N,N-Dimethylglycine was positively correlated with both SAM and SAH, whereas N-methylglycine was not. Methionine showed a trend with both SAM and SAH, and cystathionine trended with SAH. N-Methylglycine remained directly correlated with methionine, and thcy correlated directly with cysteine and indirectly with folate after correction for serum creatinine. A linear regression analysis with stepwise entering of variables showed that both serum SAH and urinary SAM were predictors of serum SAM. Predictors of serum SAH were serum creatinine and serum SAM, and predictors of thcy were serum creatinine and serum folate. Discussion We have developed a sensitive and specific assay for SAH and SAM in body fluids and biological materials that uses LC/MS and stable-isotope-dilution quantification meth- Table 5. Spearman correlation between selected variables. Serum Urine Serum Cr a SAM SAH SAM:SAH SAM SAH SAM:SAH thcy MMA Cbl folate Serum Cr P Serum SAM P Serum SAH P Serum SAM:SAH P Urinary SAM P Urinary SAH P Urinary SAM:SAH P Serum thcy P Serum MMA P Serum Cbl 0.31 P a Cr, creatinine; MMA, methylmalonic acid.

7 Clinical Chemistry 50, No. 2, ods. We used a negatively charged ion, heptafluorobutyric acid, on the C 18 column during sample clean-up and in the solutions for HPLC as a counter ion to retain SAM because it is volatile and well suited to LC/MS. The advantages of our method are that quantification is achieved with stable-isotope-labeled SAM and SAH as internal standards, which corrects for variations in recovery from sample to sample. Our assay is sensitive enough to quantify SAH in 0.5 ml of serum, but because normal serum SAH is on the order of nmol/l, our assay is close to the detection limit and the CV for SAH is greater than that for SAM. Several methods that are suitable for determining values in plasma, serum, or cerebrospinal fluid have been reported, including HPLC with fluorescence detection (15), HPLC with electrochemical detection (16), and a method that uses stable-isotope-dilution tandem mass spectrometry (17). Our values for serum SAM in controls were very similar to the values given for the previously reported stable-isotope-dilution method (17) and a corrected value for a HPLC method with coulometric electrochemical detection (16). Other HPLC methods gave higher (15) or lower (14) values than ours. Our mean SAH concentration in controls was lower than any of the other methods; thus our SAM:SAH ratio was higher than in the previous reports. Urinary SAM and SAH concentrations were roughly 100- and 30-fold higher than in serum, values similar to those in a previous report (23). We calculated a fractional excretion of SAM compared with creatinine, which was close to 100%, and also found that the SAM:SAH ratio in urine is much higher than in serum or red blood cells. These findings suggest that most SAM released into the bloodstream is probably excreted in the urine. The fractional excretion of SAH is less ( 40%), but it would be expected that renal insufficiency would have a major impact on excretion of both of these metabolites. The relationships between renal status and SAM and SAH are also complicated because the formation of creatine (precursor of creatinine) requires a SAM-dependent methylation, which is an important user of SAM (8). The SAM and SAH in serum vs urine should also be studied in populations with renal insufficiency because the decreased excretion of SAM relative to SAH in renal failure might impact the serum SAM:SAH ratio. Although increases in thcy seem to be a risk factor for vascular disease, the underlying pathophysiology has not been adequately explained. There have been recent reports that increased SAH (18, 19, 24), associated in some reports with DNA hypomethylation (18, 25), is present in populations with vascular disease. However, the role of poor renal status in increasing SAH and decreasing the SAM:SAH ratio has not been adequately explored in these populations. Our findings demonstrate that case control studies should examine renal status and gender as potential confounders of the relationship of SAM and SAH to vascular disease or vitamin status. We found that the serum SAM:SAH ratio was directly correlated with serum folate in our population, in contrast to a recent investigation (26). However, the earlier report (26) studied an elderly population, who clearly would have poorer renal function than the current group of healthy volunteers. We have found increased SAH and decreased SAM:SAH ratio in Brazilian individuals with severe Cbl-deficient megaloblastic anemia. The increased SAH corrected with Cbl therapy except in one individual with renal insufficiency (20). We also found a strong influence of renal status on response of increased thcy and SAH to vitamin therapy in a study of seniors (27). Therefore, the diagnostic utility of increased SAH in detecting Cbl or folate deficiency may be restricted to individuals with normal renal excretion of these metabolites. In addition, it may not be possible to infer the hepatic or other tissue SAM:SAH ratios from the serum values. We studied healthy individuals in the US in the postfolate food fortification era, 40% of whom also took vitamin supplements. Therefore, one limitation of our study is that the reference intervals were obtained in persons with high folate status. It will be of interest to determine whether populations with poorer folate status have higher SAH concentrations and lower SAM:SAH ratios. Measurements of SAM and SAH are of great interest in populations with neurologic or vascular disease (19, 25); when assessing the possible effects of drugs, such as levodopa and high-dose niacin; or in nutritional studies. We therefore examined conditions of sample handling and storage because the ability to quantify these metabolites in archived samples would be an obvious advantage. There are marked changes in the concentrations of SAM and SAH after exposure of samples to room temperature or prolonged frozen storage, which would impact the utility of archived or transported samples. The conditions for thawing of samples and benchtop storage during their preparation will need to be uniform. One advantage of our isotope-dilution method is that the loss of the expected peak area of the internal standard may alert the investigator to problems developing during sample preparation. It is possible that some of the post vivo sampleholding conditions are responsible for the differences in normal SAM and SAH concentrations that are reported in the literature. It will be crucial to know the storage history of clinical samples before relying on the measured SAM: SAH ratio as a true measure of the methylation index in vivo. This research was supported in part by National Institute on Aging Grant AG (to S.P.S.). Expert technical assistance was provided by Bev Raab and Carla Ray. Expert assistance in preparing the manuscript was provided by Theresa Martinez.

8 372 Stabler and Allen: Serum SAM and SAH References 1. Stabler SP, Marcell PD, Podell ER, Savage DG, Lindenbaum J, Allen RH. Quantitation of total homocysteine, total cysteine, and methionine in normal serum using capillary gas chromatographymass spectrometry. Anal Biochem 1987;162: Savage DG, Lindenbaum J, Stabler SP, Allen RH. Sensitivity of serum methylmalonic acid and total homocysteine determinations for diagnosing cobalamin and folate deficiency. Am J Med 1994; 96: Soria C, Chadefaux B, Coude M Gaillard O, Kamoun P. Concentrations of total homocysteine in plasma in chronic renal failure [Technical Brief]. Clin Chem 1990;36: Bostom AG, Lathrop L. Hyperhomocysteinemia in end stage renal disease: prevalence, etiology, and potential relationship to arteriosclerotic outcomes. Kidney Int 1997;52: Ueland PM, Refsum H. Review article: plasma homocysteine, a risk factor for vascular disease, plasma levels in health, disease, and drug therapy. J Lab Clin Med 1989;114: Vollset SE, Refsum H, Tverdal A, Nygard O, Nordrehaug JE, Tell GS, et al. Plasma total homocysteine and cardiovascular and noncardiovascular mortality: the Hordaland Homocysteine Study. Am J Clin Nutr 2001;74: Weir DG, Scott JM. The biochemical basis of the neuropathy in cobalamin deficiency. Baillieres Clin Haematol 1995;8: Mudd SH, Levy HL, Svovby F. Disorders of transsulfurations. In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds. The metabolic and molecular bases of inherited disease, 7th ed. New York: McGraw- Hill, 1995: Finkelstein JD, Kyle WE, Harris BJ. Methionine metabolism in mammals: regulatory effects of S-adenosylhomocysteine. Arch Biochem Biophys 1974;165: De la Haba G, Cantoni GL. The enzymatic synthesis of S-adenosyl- L-homocysteine from adenosine and homocysteine. J Biol Chem 1959;234: Molloy AM, Orsi B, Kennedy DG, Kennedy S, Weir DG, Scott JM. The relationship between the activity of methionine synthase and the ratio of S-adenosylmethionine to S-adenosylhomocysteine in the brain and other tissues of the pig. Biochem Pharmacol 1992;44: Perna AF, Ingrosso D, Zappia V, Galletti P, Capasso G, De Santo NG. Enzymatic methyl esterification of erythrocyte membrane proteins is impaired in chronic renal failure. Evidence for high levels of the natural inhibitor S-adenosylhomocysteine. J Clin Invest 1993;91: Fu W, Dudman NP, Perry MA, Young K, Wang XL. Interrelations between plasma homocysteine and intracellular S-adenosylhomocysteine. Biochem Biophys Res Commun 2000;271: Loehrer FM, Haefeli WE, Angst CP, Browne G, Frick G, Fowler B. Effect of methionine loading on 5-adenosylmethionine and S- adenosylhomocysteine in plasma of healthy humans. Clin Sci 1996;91: Capdevila A, Wagner C. Measurement of plasma S-adenosylmethionine and S-adenosylhomocysteine as their fluorescent isoindoles. Anal Biochem 1998;264: Melnyk S, Pogribna M, Pogribny IP, Yi P, James SJ. Measurement of plasma and intracellular S-adenosylmethionine and S-adenosylhomocysteine utilizing coulometric electrochemical detection: alterations with plasma homocysteine and pyridoxal 5 -phosphate concentrations [Erratum published in Clin Chem 2001;47:612]. Clin Chem 2000;46: Struys EA, Jansen EE, de Meer K, Jakobs C. Determination of S-adenosylmethionine and S-adenosylhomocysteine in plasma and cerebrospinal fluid by stable isotope dilution tandem mass spectrometry. Clin Chem 2000;46: Yi P, Melnyk S, Pogribna M, Pogribny IP, Hine RJ, James SJ. Increase in plasma homocysteine associated with parallel increases in plasma S-adenosylhomocysteine and lymphocyte DNA hypomethylation. J Biol Chem 2000;275: Kerins DM, Koury MJ, Capdevila A, Rana S, Wagner C. Plasma S-adenosylhomocysteine is a more sensitive indicator of cardiovascular disease than plasma homocysteine. Am J Clin Nutr 2001;74: Guerra-Shinohara EM, Morita OE, Pagliusi RA, Blaia-D Avila VL, Allen RH, Stabler SP. Elevation of serum S-adenosylhomocysteine in severe cobalamin deficient megaloblastic anemia [Abstract]. Blood 2001;98:13a. 21. Stabler SP, Lindenbaum J, Savage DG, Allen RH. Elevation of serum cystathionine levels in patients with cobalamin and folate deficiency. Blood 1993;81: Allen, RH, Stabler SP, Savage DG, Lindenbaum J. Elevation of 2-methycitric acid I and II levels in serum, urine, and cerebrospinal fluid of patients with cobalamin deficiency. Metabolism 1993;42: Delabar U, Kloor D, Luippold G, Muhlbauer B. Simultaneous determination of adenosine, S-adenosylhomocysteine and S-adenosylmethionine in biological samples using solid-phase extraction and high performance liquid chromatography. J Chromatogr B Biomed Sci Appl 1999;724: Loehrer FMT, Angst CP, Brunner FP, Haefeli WE, Fowler B. Evidence for disturbed S-adenosylmethionine: S-adenosylhomocysteine ratio in patients with end stage renal failure: a cause for disturbed methylation reactions? Nephrol Dial Transplant 1998; 13: Castro R, Rivera I, Struys EA, Jansen EEW, Ravasco P, Camilo ME, et al. Increased homocysteine and S-adenosylhomocysteine concentrations and DNA hypomethylation in vascular disease. Clin Chem 2003;49: Becker A, Smulders YM, Teerlink T, Struys EA, de Meer K, Kostense PJ, et al. S-Adenosylhomocysteine and the ratio of S-adenosylmethionine to S-adenosylhomocysteine are not related to folate, cobalamin and vitamin B 6 concentrations. Eur J Clin Invest 2003;33: Stabler SP, Allen RH, Dolce ET, Hawthorne NA, Johnson MA. Response of elevated serum S-adenosylhomocysteine and total homocysteine to high dose oral cobalamin [Abstract]. Blood 2002;100:227a.