The 5,10-methylenetetrahydrofolate reductase C677T polymorphism interacts with smoking to increase homocysteine

Atherosclerosis 174 (2004) 315–322

The 5,10-methylenetetrahydrofolate reductase C677T polymorphism interacts with smoking to increase homocysteine Karen S. Brown a , Leo A.J. Kluijtmans b , Ian S. Young c , Liam Murray c , Dorothy McMaster c , Jayne V. Woodside c , John W.G. Yarnell c , Colin A. Boreham d , Helene McNulty d , J.J. Strain d , Joseph McPartlin e , John M. Scott e , Laura E. Mitchell f , Alexander S. Whitehead a,∗ a

Department of Pharmacology and Center for Pharmacogenetics, University of Pennsylvania School of Medicine, 153 Johnson Pavilion, 3620 Hamilton Walk, Philadelphia, PA 19104-6084, USA b Laboratory of Pediatrics and Neurology, University Medical Center Nijmegen, Nijmegen, The Netherlands c Cardiovascular Research Centre, Queen’s University Belfast, Belfast, Northern Ireland, UK d Northern Ireland Centre for Diet and Health, University of Ulster, Coleraine, Northern Ireland, UK e Department of Clinical Medicine, Trinity College, Dublin, Ireland f Institute of Biosciences and Technology, Texas A&M University System Health Science Center, Houston, TX 77030, USA Received 18 August 2003; accepted 21 January 2004 Available online 15 April 2004

Abstract Elevated homocysteine is a risk marker for several human pathologies. Risk factors for elevated homocysteine include low folate and homozygosity for the T allele of the 5,10-methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism. Because nitric oxide may inhibit folate catabolism and endothelial nitric oxide synthase activity is reduced in smokers, we postulated that smoking status might modify the impact of the MTHFR C677T polymorphism on homocysteine (tHcy) concentrations. We tested this hypothesis in a healthy young adult population for which MTHFR C677T genotypes and tHcy concentrations were previously reported. The MTHFR 677TT genotype was significantly associated with elevated tHcy concentrations in smokers (P = 0.001) but not in non-smokers (P = 0.36). Among smokers, the MTHFR 677TT genotype was significantly associated with high tHcy in heavy smokers (P = 0.003) but not light smokers (P = 0.09), in men (P = 0.003) but not women (P = 0.11), and in subjects from the lowest serum folate quartile (P = 0.003) but not from folate quartiles 2–4 (P = 0.49). After adjustment for nutritional variables, interactions between MTHFR C677T genotype and NOS3 G894T genotype, and between MTHFR genotype, smoking status and gender were statistically significant. We propose that hyperhomocysteinemia in MTHFR 677TT homozygote smokers is the consequence of mild intracellular folate deficiency caused by a smoking-related reduction of NOS3 activity that is exacerbated when serum folate is low. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Hyperhomocysteinemia; Folate; 5,10-Methylenetetrahydrofolate reductase; Smoking; Nitric oxide; Nitric oxide synthase

1. Introduction An elevated concentration of the metabolic intermediate homocysteine (hyperhomocysteinemia) is a recognized risk marker for a number of important human pathologies. In both prospective and retrospective studies, even modest increases in homocysteine concentration have been shown to be a graded and independent risk marker for cardiovascular ∗ Corresponding author. Tel.: +1-215-898-2332; fax: +1-215-573-9135. E-mail address: [email protected] (A.S. Whitehead).

disease (CVD) (reviewed in [1]). Hyperhomocysteinemia is also a risk marker for neural tube defects (NTDs), pregnancy complications, inflammatory bowel disease (IBD), Alzheimer’s disease, and some cancers [2–7]. Plasma homocysteine (tHcy) concentrations are influenced by both environmental and genetic factors. Environmental risk factors for hyperhomocysteinemia include low dietary intake of folate and Vitamin B12 , and smoking [8]. The best-characterized genetic determinant of moderate hyperhomocysteinemia is the 5,10-methylenetetrahydrofolate reductase (MTHFR) C677T (Ala222Val) polymorphism [9]. MTHFR converts 5,10-methylenetetrahydrofolate to 5-methyltetrahydrofolate (5-mTHF); 5-mTHF is a co-substrate

0021-9150/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2004.01.023

316

K.S. Brown et al. / Atherosclerosis 174 (2004) 315–322

Fig. 1. Folate metabolic pathway. CBS: cystathionine ␤-synthase; DHF: dihydrofolate; DHFR: dihydrofolate reductase; dTMP: deoxythymidine monophosphate; dUMP: deoxyuridine monophosphate; Hcy: homocysteine; MTHFR: 5,10-methylenetetrahydrofolate reductase; MTR: methionine synthase; MTRR: methionine synthase reductase; NO: nitric oxide; SAH: S-adenosylhomocysteine; SAM: S-adenosylmethionine; THF: tetrahydrofolate; TS: thymidylate synthase.

for the synthesis of methionine from homocysteine (Fig. 1), as well as the primary circulating form of folate [1]. The MTHFR 677T variant encodes an enzyme that is thermolabile and mildly dysfunctional in vivo; MTHFR 677TT homozygotes with low folate status exhibit a phenotype encompassing mildly increased homocysteine concentrations [10], lower circulating folate concentrations [10], and altered distribution of red blood cell (RBC) folate derivatives [11]. The relationships, if any, between the MTHFR C677T polymorphism and diseases associated with hyperhomocysteinemia have also been extensively investigated and supported in some studies of subjects with CVD, NTDs, IBD, and cancer, but not others [12]. A recent meta-analysis, which assessed the relationship between MTHFR 677TT genotype and CVD, found an increased risk of approximately 16% that was largely confined to individuals with low serum folate [13]. We recently reported that the endothelial nitric oxide synthase (NOS3) G894T polymorphism is a genetic determinant of tHcy concentrations in healthy non-smokers with low serum folate [14], and proposed that the underlying genotype-dependent mechanism is an indirect, nitric oxide (NO)-mediated modulation of folate catabolism. Since NOS3 activity is markedly reduced in smokers, we subsequently postulated that smoking, by increasing folate catabolism, could induce a state of mild intracellular folate deficiency and thus make smokers more susceptible to the homocysteine-raising effects of the MTHFR 677TT genotype. We have tested this hypothesis in a population of healthy young subjects. The data presented below demonstrate that the MTHFR 677TT genotype is a risk factor for

hyperhomocysteinemia only in smokers with low serum folate, despite 677TT smokers and non-smokers having comparable serum folate concentrations.

2. Materials and methods 2.1. Subjects The study population was drawn from the Young Hearts Project, an ongoing prospective study designed to monitor the prevalence of coronary disease risk factors over time in young subjects from Northern Ireland [15–17]. Ethical approvals were granted by the Research Ethics Committee, Queen’s University Belfast, and all subjects provided written informed consent. The biochemical data used for the analyses reported here were acquired at the third screening visit, at which time subjects were between 20 and 26 years old. At this visit, subjects were also classified as current cigarette smokers or non-smokers and the number of cigarettes smoked per day was recorded There were 36 previous smokers in the population who were classified as current non-smokers. The analyses reported here include the 407 subjects for whom MTHFR C677T genotype, gender and tHcy data were available. 2.2. Determination of homocysteine, folate and Vitamin B12 Blood samples were collected from fasted subjects for determination of biochemical parameters and DNA

K.S. Brown et al. / Atherosclerosis 174 (2004) 315–322

extraction. tHcy concentrations were measured using an established high performance liquid chromatography method [18]. Serum folate and Vitamin B12 concentrations were determined by time-resolved immunofluorescence on an AutoDelfia analyzer (Wallac, UK) [17]. RBC folate concentrations were determined by microbiological assay as previously reported [19]. 2.3. Genetic analysis MTHFR C677T and NOS3 G894T genotypes in this study population and the respective genotyping methods used have previously been reported [14,17].

317

the model only if they remained significant, given the base model and all previously entered interaction terms. The Pvalues for this latter test are given in the text for significant interaction terms. The total amount of variation in tHcy explained by the base model and the final model containing the significant interaction terms were estimated by Efron’s pseudo-R2 statistic [20]. All tests were two-tailed and P-values of <0.05 (or Bonferroni-adjusted values where appropriate) were considered statistically significant. All statistical analyses were carried out using SAS version 8.2.

3. Results 2.4. Statistical analyses Distributions for tHcy, folate and Vitamin B12 were positively skewed even after logarithmic transformation; therefore, all analyses were performed using untransformed ranked data. Genotype frequency differences and, for discrete variables, differences between groups, were assessed by the χ2 -test. For continuous variables, differences between groups were assessed using the Kruskal–Wallis test or the Wilcoxon Rank Sum test. Where appropriate, Pvalues for pairwise differences were corrected for multiple comparisons using the Bonferroni method. Odds ratios (ORs) for MTHFR 677TT homozygotes relative to 677CC homozygotes were estimated by logistic regression analysis after stratification by smoking status, degree of smoking, gender and folate quartile. Generalized linear regression was used to evaluate the significance of interactions between MTHFR C677T genotype, smoking status, gender, and NOS3 G894T genotype. A loglink function and gamma error structure were employed to account for the rightward skew of the data. The likelihood ratio test (LRT) was used to determine the significance of interaction terms after adjusting for the effects of additional variables. The distribution of this statistic approximately follows a χ2 distribution, with degrees of freedom equal to the difference in the number of parameters estimated by two models. A series of models were fitted to the data. The base model included known determinants of tHcy concentrations: MTHFR C677T genotype, serum folate, Vitamin B12 , gender, and a MTHFR C677T genotype/serum folate interaction term. Main effects for smoking and NOS3 G894T genotype as well as two-way interactions between MTHFR C677T genotype and smoking status, gender, or NOS3 G894T genotype, between gender or NOS3 genotype and smoking status, and a three way interaction between MTHFR C677T genotype, smoking status, and gender were tested to determine whether they significantly improved the base model. Significance was evaluated using the LRT to compare the model including a single main effect or interaction term to the base model. Terms that were associated with a significant LRT were then added to the base model one by one, starting with the most significant term, and were retained in

3.1. Population characteristics Population characteristics, including MTHFR C677T genotype frequencies, tHcy, serum folate, RBC folate and Vitamin B12 , are summarized for the population as a whole and after stratification by gender or smoking status in Table 1. There were no statistically significant differences in the proportions of MTHFR C677T genotypes between smokers and non-smokers. Median tHcy concentrations were somewhat higher in smokers than non-smokers, but the difference was not statistically significant (P = 0.09, Table 1). Serum folate, RBC folate, and Vitamin B12 concentrations were significantly lower in smokers than in non-smokers (Table 1). 3.2. Biochemical parameters by genotype Biochemical parameters are presented by MTHFR C677T genotype for the entire population and after stratification by smoking status in Table 2. In the total sample, the MTHFR C677T genotype was significantly associated with tHcy, serum folate and RBC folate but not Vitamin B12 concentrations. Among smokers, MTHFR C677T genotype was significantly associated with tHcy, serum folate and RBC folate concentrations. Compared to smokers with the 677CC genotype, 677TT smokers had significantly higher tHcy concentrations, significantly lower serum folate and RBC folate concentrations, and non-significantly lower Vitamin B12 concentrations. Among non-smokers, the MTHFR C677T genotype was significantly associated only with RBC folate. Among 677TT homozygotes, tHcy concentrations were significantly higher in smokers than in non-smokers (median tHcy of 11.9 ␮mol/l versus 10.0 ␮mol/l, P = 0.04). In contrast, median tHcy concentrations for individuals with the 677CC genotype were comparable in smokers and nonsmokers (Table 2). 3.3. Impact of smoking status on the association between MTHFR 677TT genotype and elevated tHcy The impact of the MTHFR 677TT genotype on the risk of having elevated tHcy was assessed for both smokers and

318

K.S. Brown et al. / Atherosclerosis 174 (2004) 315–322

Table 1 Population characteristics Characteristic

All

Smokers

Non-smokers

n (%) Male (%) Female (%)

407 (100) 220 (54) 187 (46)

155 (38) 82 (53) 73 (47)

250 (61) 136 (54) 114 (46)

MTHFR C677T genotype (n, %) CC CT TT

176 (43.2) 176 (43.2) 55 (13.5)

Homocysteine (␮mol/l)a Folate (nmol/l)a RBC folate (␮g/l RBC)a Vitamin B12 (pmol/l)a

8.9 (7.5–11.0) 407 12.8 (9.4–18.9) 346 284 (212–367) 362 267 (197–344) 347

66 (42.6) 64 (41.3) 25 (16.1)

109 (43.6) 112 (44.8) 29 (11.6)

9.4 (7.5–11.6) 155 11.1 (8.7–15.3) 130b 255 (204–343) 132c 238 (179–320) 131d

8.7 (7.5–10.6) 250 14.2 (9.9–20.7) 214 296 (226–381) 228 280 (203–357) 214

RBC: red blood cells. a Median (interquartile range), n. b P = 0.0004, smokers compared to non-smokers. c P = 0.007, smokers compared to non-smokers. d P = 0.01, smokers compared to non-smokers.

non-smokers. Smokers with the 677TT genotype had a 14fold increased risk of having a tHcy concentration in the top 10% of the population distribution (tHcy > 14.2 ␮mol/l, OR = 14.3, 95% confidence interval [CI] = 4.0 − 51.5) relative to 677CC smokers. The corresponding OR for 677TT non-smokers relative to 677CC non-smokers was also elevated but not statistically significant (OR = 3.0, 95% CI = 0.6–14.4). Smokers still had increased risk relative to non-smokers after adjustment for folate, Vitamin B12 , age, sex, NOS3 G894T genotype, and (in smokers) cigarettes smoked per day; adjusted ORs were 10.2 (95% CI = 1.5–67.8) for smokers and 1.6 (95% CI = 0.03–76.1) for non-smokers.

To assess whether the effect of smoking on tHcy in the context of MTHFR C677T genotype was influenced by the number of cigarettes smoked per day, smokers were subdivided post hoc into light smokers (those smoking 10 or fewer cigarettes per day, n = 97) and heavy smokers (those smoking more than 10 cigarettes per day, n = 58). MTHFR C677T genotype was significantly associated with tHcy concentration for heavy smokers but not light smokers (P = 0.003 and 0.09, respectively; Table 3). For 677TT homozygotes, median tHcy was much higher in heavy smokers than in light smokers (16.8 ␮mol/l versus 9.5 ␮mol/l), although this difference did not reach statistical significance (P = 0.41). In contrast, median tHcy was comparable between heavy

Table 2 Associations between MTHFR C677T genotype and biochemical parameters Biochemical parameter

Subset

Pa

Median (interquartile range) n CC

CT

TT

Homocysteine (␮mol/l)

All Smokers Non-smokers

8.7 (7.5–10.5) 176 8.8 (7.3–10.7) 66 8.8 (7.5–10.4) 109

8.7 (7.4–10.8) 176 9.4 (7.5–11.6) 64 8.6 (7.4–10.6) 112

10.3 (8.0–16.8) 55 11.9 (9.2–19.1) 25 10.0 (7.7–11.8) 29

0.0006b 0.001b 0.36

Folate (nmol/l)

All Smokers Non-smokers

14.0 (10.1–20.0) 152 13.0 (9.7–16.5) 55 15.5 (10.9–22.1) 96

12.1 (9.3–18.9) 148 10.9 (8.7–16.5) 54 12.7 (9.6–19.1) 94

11.0 (7.0–15.2) 46 9.6 (7.8–11.7) 21 11.8 (9.2–17.6) 24

0.004c 0.03c 0.17

RBC folate (␮g/l RBC)

All Smokers Non-smokers

307 (239–389) 159 270 (219–361) 58 322 (259–398) 100

275 (212–364) 156 282 (209–363) 52 267 (212–364) 104

223 (161–320) 47 195 (138–223) 22 283 (213–365) 24

0.0002d 0.001b 0.01e

Vitamin B12 (pmol/l)

All Smokers Non-smokers

261 (196–342) 152 227 (179–320) 55 276 (199–353) 96

276 (204–351) 148 269 (200–348) 54 280 (204–359) 94

226 (165–322) 47 202 (143–278) 22 283 (197–346) 24

0.16 0.09 0.93

RBC: red blood cells. a Kruskal–Wallis test. b Bonferroni-corrected c Bonferroni-corrected d Bonferroni-corrected e Bonferroni-corrected

P < 0.05 P < 0.05 P < 0.05 P < 0.05

for for for for

TT TT TT CT

vs. vs. vs. vs.

CC and TT vs. CT. CC. CT, TT vs. CC and CT vs. CC. CC.

K.S. Brown et al. / Atherosclerosis 174 (2004) 315–322

319

Table 3 Homocysteine concentrations by MTHFR C677T genotype and smoking level Subseta

Homocysteine (␮mol/l), median (interquartile range), n CC

CT

TT

Light smokers Heavy smokers

8.9 (6.8–10.7) 44 8.8 (7.6–11.1) 22

9.7 (8.0–11.1) 38 8.6 (7.4–11.9) 26

9.5 (8.9–21.6) 15 16.8 (11.9–18.2) 10

a b c

Pb

0.09 0.003c

Light smokers smoked ≤10 cigarettes per day; heavy smokers smoked >10 cigarettes per day. Kruskal–Wallis test. Bonferroni-corrected P < 0.05 for TT vs. CC and TT vs. CT.

and light smokers with the 677CC genotype (Table 3). The OR for being in the top 10% of the tHcy distribution for 677TT smokers relative to 677CC smokers was higher for heavy smokers (49.0, 95% CI = 4.4–550.7) than for light smokers (6.8, 95% CI = 1.4–33.5), although the confidence intervals were wide and overlapping due to small numbers of subjects in the indicated subsets. To determine whether the relationship between smoking and MTHFR C677T genotype is modified by gender, data from males and females were analyzed separately (Table 4). Among smokers, MTHFR C677T genotype was significantly associated with tHcy in males (P = 0.003) but not in females (P = 0.11). tHcy tended to be higher in male 677TT smokers than in female 677TT smokers (median 16.1 versus 10.3 ␮mol/l); however, the difference did not reach statistical significance (P = 0.18). The OR for being in the top 10% of the tHcy distribution for 677TT homozygotes as compared with 677CC homozygotes was higher for male smokers (56.7, 95% CI = 4.9–656.8) than for female smokers (6.5, 95% CI = 1.4–30.3), although the confidence intervals were again wide due to the small number of subjects. Since male smokers consumed more cigarettes per day than female smokers (P = 0.03), OR estimates were adjusted for the number of cigarettes smoked per day. The adjusted OR was still much higher for male 677TT smokers (70.0, 95% CI = 5.1–968.7) than for female 677TT smokers (9.7, 95% CI = 1.9–48.2). Among non-smokers, MTHFR C677T genotype was not significantly associated with tHcy concentration for either males (P = 0.59) or females (P = 0.37), although male 677TT homozygotes tended to have modestly higher tHcy concentrations than their 677CC and 677CT counterparts. The tHcy-raising effect of the MTHFR 677TT genotype is known to be largely restricted to subjects in the lowest

serum folate quartile (folate quartile 1) in this study population [17]; therefore, the association between MTHFR C677T genotype and tHcy was assessed in smokers and non-smokers from serum folate quartile 1. There was a highly significant association between MTHFR C677T genotype and tHcy in smokers (P = 0.003), but not in non-smokers (P = 0.26). In folate quartile 1, as in the population as a whole, the interaction between smoking status and genotype was confined primarily to 677TT individuals (median tHcy of 18.2 ␮mol/l in smokers compared with 13.7 ␮mol/l in non-smokers), although the difference did not reach statistical significance in this small subset (P = 0.27). Median tHcy concentrations were somewhat higher for smokers with the 677CC genotype than for nonsmokers (Table 5). Importantly, in folate quartile 1, median folate concentrations were similar between smokers and non-smokers (8.0 and 7.6 nmol/l, respectively; P = 0.88); this was also true in the subset of 677TT smokers and non-smokers (7.4 and 6.7 nmol/l, respectively; P = 0.71). In contrast to results from subjects with low folate status, there was no association between MTHFR C677T genotype and tHcy in either smokers (P = 0.49) or non-smokers (P = 0.99) in folate quartiles 2–4. Similar results were obtained when the above analyses were carried out in subjects stratified by RBC folate concentrations (data not shown). 3.4. Assessment of interactions General linear regression was used to evaluate the significance of smoking status, NOS3 G894T genotypes, and interactions between MTHFR C677T genotype and gender, smoking status and NOS3 G894T genotype in determining tHcy concentrations. The significance of each of these

Table 4 Homocysteine concentrations by MTHFR C677T genotype after stratification by smoking status and gender Smoking status

Subset

Pa

Homocysteine (␮mol/l), median (interquartile range), n CC

CT

TT

Smokers

Men Women

8.5 (7.6–10.3) 35 9.2 (6.6–11.3) 31

8.5 (7.4–11.1) 39 10.3 (8.5–11.7) 25

16.1 (10.3–27.5) 8 10.3 (8.9–18.2) 17

0.003b 0.11

Non-smokers

Men Women

9.2 (7.9–10.7) 63 7.8 (7.5–9.6) 46

9.4 (7.7–11.1) 58 8.4 (7.0–9.5) 54

10.4 (7.8–13.4) 15 8.8 (7.6–11.5) 14

0.59 0.37

a b

Kruskal–Wallis test. Bonferroni corrected P < 0.05 for TT vs. CC and TT vs. CT.

320

K.S. Brown et al. / Atherosclerosis 174 (2004) 315–322

Table 5 Homocysteine concentrations by MTHFR C677T genotype stratified by smoking status in subjects with low and high serum folate Pa

Quartile folate

Smoking status

Homocysteine (␮mol/l), median (interquartile range), n CC

CT

TT

1

Smokers Non-smokers

11.1 (8.2–12.5) 11 10.1 (8.8–11.6) 17

10.9 (8.5–15.2) 19 10.7 (9.1–14.2) 20

18.2 (15.4–29.6) 10 13.7 (9.5–21.2) 6

2–4

Smokers Non-smokers

8.5 (7.0–10.5) 44 8.3 (7.5–10.0) 79

8.1 (6.6–9.9) 35 8.4 (7.3–10.1) 74

9.2 (7.9–9.5) 11 7.9 (6.9–10.6) 18

a b

0.003b 0.26 0.49 0.99

Kruskal–Wallis test. Bonferroni-corrected P < 0.05 for TT vs. CC and TT vs. CT.

terms was assessed relative to an initial model including MTHFR C677T genotype, serum folate, Vitamin B12 , gender, and a MTHFR C677T genotype/serum folate interaction term, which explained 35% of the total variation in tHcy. The model was significantly improved by the addition of an MTHFR C677T genotype/NOS3 G894T genotype interaction term (P < 0.0001) and an MTHFR C677T genotype/smoking status/gender interaction term (P = 0.001), but not by addition of the other terms. The final model including both interaction terms explained approximately 52% of the total variation in tHcy, 17% more than that explained by the model lacking these terms.

4. Discussion We have tested the hypothesis that smoking modifies the relationship between MTHFR C677T genotype and tHcy concentrations. In a population of healthy young adults, the MTHFR 677TT genotype was associated with markedly elevated tHcy concentrations in smokers (P = 0.001, Table 2) but not in non-smokers (P = 0.36). Interestingly, tHcy was significantly elevated only in 677TT heavy smokers (P = 0.003), although there was a trend towards increased tHcy in 677TT light smokers (P = 0.09, Table 3), suggesting a dose-dependent effect. The well-documented association between smoking and low serum folate may be the result of lower dietary intake, altered absorption, or increased oxidative catabolism [21]. Serum folate correlates inversely with tHcy concentration, and low serum folate potentiates the tHcy-raising effects of the 677TT genotype. Therefore, one possible interpretation of our data is that the apparent interaction between smoking and the MTHFR 677TT genotype to raise tHcy concentrations is largely attributable to smoking-mandated low serum folate. However, when we assessed the impact of MTHFR genotype on tHcy in smokers and non-smokers from folate quartile 1, the 677TT genotype was still significantly associated with high tHcy concentrations only in smokers, despite comparable serum folate concentrations in 677TT smokers and non-smokers in this subset of subjects (Table 5). Thus, the interaction between MTHFR C677T genotype and smoking appears to be independent of the impact of smoking on serum folate.

Another consequence of smoking that may have implications for homocysteine metabolism is the reduction of NOS3 activity in the vascular endothelium [22,23]. There is evidence that NO modulates tHcy concentrations through an effect on folate catabolism [14]. The iron storage protein ferritin has recently been shown to specifically promote the oxidative cleavage of folate [24]. Since NO inhibits ferritin expression [25], a decrease in the NO-generating capacity of NOS3 could lead to increased ferritin synthesis and hence increased ferritin-mediated folate catabolism. In addition, NO neutralizes many cellular oxidants, including superoxide anion and hydrogen peroxide [26]. A decrease in NO could therefore lead to increased concentrations of such non-enzymatic oxidizing agents and the increased nonspecific oxidative catabolism of folate. This may be exacerbated in MTHFR 677TT smokers, since 5-mTHF, through a poorly-defined mechanism, interacts with the NOS3 cofactor tetrahydrobiopterin (BH4 ) to reduce NOS3 “uncoupling” [27]. Uncoupled NOS3 has a decreased affinity for arginine together with an increased affinity for O2 which results in increased production of superoxide radicals [28], potent oxidants which may contribute directly to the non-enzymatic oxidation of folate. Thus, when circulating folate concentrations are low, impaired NOS3-dependent NO production in smokers might result in a state of intracellular folate deficiency due to increased folate catabolism. In the context of the above mechanism, a genotype-mandated deficiency in 5mTHF levels would promote NOS3 uncoupling to amplify its impact on folate catabolism and drive tHcy concentrations up. Several additional observations support our interpretation of the data outlined above. The tHcy-raising effect of the MTHFR 677TT genotype is largely confined to smokers in the lowest quartile of serum folate. In 677TT homozygotes with higher circulating folate concentrations, even if intracellular folate is somewhat depleted, it may still be high enough to stabilize the mildly dysfunctional MTHFR enzyme encoded by the T allele. Furthermore, we have demonstrated statistically significant interactions between the MTHFR C677T and NOS3 G894T genotypes, and between MTHFR C677T genotype, smoking status and gender, that together explain approximately 17% of the total variation in tHcy concentrations in this population.

K.S. Brown et al. / Atherosclerosis 174 (2004) 315–322

The disparity in the magnitude of the effect of the 677TT genotype on tHcy concentrations in males and females is also consistent with the above hypothesis. The association between MTHFR C677T genotype and tHcy concentrations is significant in male (P = 0.003, Table 4) but not female (P = 0.11) smokers. These results are not attributable to heavier smoking by 677TT males, since the number of cigarettes smoked per day was not significantly different between 677TT males and females. One possible counterbalancing factor is estrogen. It has been suggested that estrogen can lower tHcy concentrations in pre-menopausal women (or in post-menopausal women undergoing hormone replacement therapy) [29], possibly via its documented capacity to increase NOS3 activity [30]. Our findings may also be considered in the context of a recently-published meta-analysis which reported that the MTHFR 677TT genotype, in subjects who were not stratified by either smoking status or gender, is associated with only a 16% increase in CVD risk [13]. Thus, any increase in CVD risk conferred by the MTHFR 677TT genotype in male smokers that exceeded 16% would likely have been diluted by the inclusion of MTHFR 677TT non-smokers and females. Interestingly, the authors noted that the European studies included in the meta-analysis generally had higher OR estimates for the 677TT genotype than the North American studies and that smoking was more prevalent in the European study subjects than in the North American study subjects. In conclusion, we have demonstrated a strong interaction between MTHFR 677TT genotype and smoking that greatly increases the risk of elevated homocysteine in healthy young adults with low serum folate. The World Health Organization estimates that worldwide approximately one third of the adult male population smokes [31]. The acquisition of a hyperhomocysteinemic phenotype by MTHFR 677TT smokers with low folate early in life suggests that such individuals may be embarking on a decades-long chronic exposure to elevated tHcy concentrations with possible negative health consequences that include premature CVD, IBD, pregnancy complications in females, cancer and Alzheimer’s disease. If so, the resulting excess morbidity and mortality may represent a substantial ongoing public health and social burden. The data presented in this paper underscore the importance of both smoking cessation and folic acid fortification programs, the effective implementation of which would be particularly beneficial for individuals with genetic susceptibility to a low folate/high homocysteine phenotype.

Acknowledgements This work was supported by NIH grant AR47633, and in part by NIH grants HD39195 and HD39081. Karen S. Brown is supported by a grant from The American Heart Association and Leo A.J. Kluijtmans is supported by grant

321

1999T023 from the Netherlands Heart Foundation. Support for the Young Hearts Project has been provided by the British Heart Foundation and the Wellcome Trust.

References [1] Durand P, Prost M, Loreau N, Lussier-Cacan S, Blache D. Impaired homocysteine metabolism and atherothrombotic disease. Lab Invest 2001;81:645–72. [2] Steegers-Theunissen RPM, Boers GJH, Trijbels JMF, et al. Maternal hyperhomocysteinemia: a risk factor for neural-tube defects? Metabolism 1994;43:1475–80. [3] Mahmud N, Molloy A, McPartlin J, et al. Increased prevalence of methylenetetrahydrofolate reductase C677T variant in patients with inflammatory bowel disease, and its clinical implications. Gut 1999;45:389–94. [4] Dekker GA, De Vries JIP, Doelitzsch PM, et al. Underlying disorders associated with severe early-onset pre-eclampsia. Am J Obstet Gynecol 1995;173:1042–8. [5] Vollset SE, Refsum H, Irgens LM, et al. Plasma total homocysteine, pregnancy complications, and adverse pregnancy outcomes: the Hordaland Homocysteine Study. Am J Clin Nutr 2000;71:962–8. [6] Clarke R, Smith AD, Jobst KA, et al. Folate, vitamin B12 , and serum homocysteine levels in confirmed Alzheimer’s disease. Arch Neurol 1998;55:1449–55. [7] Wu LL, Wu JT. Hyperhomocysteinemia is a risk factor for cancer and a new potential tumor marker. Clin Chim Acta 2002;322: 21–8. [8] Nygard O, Refsum H, Ueland PM, Vollset SE. Major lifestyle determinants of plasma total homocysteine distribution: the Hordaland Homocysteine Study. Am J Clin Nutr 1998;67:263–70. [9] Frosst P, Blom HJ, Milos R, et al. A candidate genetic risk factor for vascular disease: a common mutation in methylenetetrahydrofolate reductase. Nat Genet 1995;10:111–3. [10] Harmon DL, Woodside JV, Yarnell JWG, et al. The common ‘thermolabile’ variant of methylene tetrahydrofolate reductase is a major determinant of mild hyperhomocysteinaemia. Q J Med 1996;89:571–7. [11] Bagley PL, Selhub J. A common mutation in the methylenetetrahydrofolate reductase gene is associated with an accumulation of formylated tetrahydrofolates in red blood cells. Proc Natl Acad Sci USA 1998;95:13217–20. [12] Ueland PM, Hustad S, Schneede JO, Refsum H, Vollset SE. Biological and clinical implications of the MTHFR C677T polymorphism. Trends Pharmacol Sci 2001;22:195–201. [13] Klerk M, Verhoef P, Clarke R, et al. MTHFR 677C->T polymorphism and risk of coronary heart disease. JAMA 2002;288:2023–31. [14] Brown KS, Kluijtmans LAJ, Young IS, et al. Genetic evidence that nitric oxide modulates homocysteine: the NOS3 894TT genotype is a risk factor for hyperhomocysteinemia. Arterioscler Thromb Vasc Biol 2003;23:1014–20. [15] Boreham C, Twisk J, van Mechelen W, et al. Relationship between the development of biological risk factors for coronary heart disease and lifestyle parameters during adolescence: the Northern Ireland Young Hearts Project. Public Health 1999;113:7–12. [16] Boreham CA, Twisk J, Savage MJ, Cran GW, Strain JJ. Physical activity, sports participation, and risk factors in adolescents. Med Sci Sports Exerc 1997;29:788–93. [17] Kluijtmans LAJ, Young IS, Boreham CA, et al. Genetic and nutritional factors contributing to hyperhomocysteinemia in young adults. Blood 2003;101:2483–8. [18] Ubbink JB, Vermaak WJH, Bissbort S. Rapid high-performance liquid chromatographic assay for total homocysteine levels in human serum. J Chromatogr 1991;565:441–6.

322

K.S. Brown et al. / Atherosclerosis 174 (2004) 315–322

[19] Molloy AM, Scott JM. Microbiological assay for serum, plasma and red cell folate using cryopreserved, microtiter plate method. Methods Enzymol 1997;281:43–53. [20] Hardin J, Hilbe J. Generalized linear models and extensions. Stata Press; 2001. [21] Mansoor MA, Kristensen O, Hervig T, et al. Low concentrations of folate in serum and erythrocytes of smokers: methionine loading decreases folate concentrations in serum of smokers and nonsmokers. Clin Chem 1997;43:2192–4. [22] Higman DJ, Strachan AM, Buttery L, et al. Smoking impairs the activity of endothelial nitric oxide synthase in saphenous vein. Arterioscler Thromb Vasc Biol 1996;16:546–52. [23] Barua RS, Ambrose JA, Eales-Reynolds LJ, et al. Dysfunctional endothelial nitric oxide biosynthesis in healthy smokers with impaired endothelium-dependent vasodilatation. Circulation 2001;104:1905– 10. [24] Suh JR, Oppenheim EW, Girgis S, Stover PJ. Purification and properties of a folate-catabolizing enzyme. J Biol Chem 2000;275:35646– 55.

[25] Juckett MB, Weber M, Balla J, Jacob HS, Vercellotti GM. Nitric oxide donors modulate ferritin and protect endothelium from oxidative injury. Free Radic Biol Med 1996;20:63–73. [26] Benz D, Cadet P, Mantione K, Zhu W, Stefano GB. Tonal nitric oxide and health—a free radical and a scavenger of free radicals. Med Sci Monit 2002;8:RA1–4. [27] Stroes ESG, van Faassen EE, Yo M, et al. Folic acid reverts dysfunction of endothelial nitric oxide synthase. Circ Res 2000;86:1129– 34. [28] Katusic ZS. Vascular endothelial dysfunction: does tetrahydrobiopterin play a role? Am J Physiol Heart Circ Physiol 2001;281: H981–6. [29] Dimitrova KR, DeGroot K, Meyers AK, Kim YD. Estrogen and homocysteine. Cardiovasc Res 2002;53:577–88. [30] Wycoff MH, Chamblis KL, Mineo C, et al. Plasma membrane estrogen receptors are coupled to endothelial nitric-oxide synthase through G␣i . J Biol Chem 2001;276:27071–6. [31] World Health Organization Factsheet: smoking statistics. World Health Organization; 2002.