点击显示 收起
1 From the Department of Pediatrics, College of Medicine, University of Arkansas for Medical Sciences and Arkansas Children's Hospital Research Institute, Little Rock
2 The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the CDC or the NIH. 3 Supported by Cooperative Agreement no. U50/CCU613236-06 from the Centers for Disease Control and Prevention and by grants from the National Institute of Child Health and Human Development (5R01 HD39054-03) and the National Center for Research Resources (1C06 RR16517-01 and 3C06 RR16517-01S1). 4 Reprints not available. Address correspondence to CA Hobbs, Department of Pediatrics, College of Medicine, University of Arkansas for Medical Sciences, 11219 Financial Centre Parkway, Suite 250, Little Rock, AR 72211. E-mail: hobbscharlotte{at}uams.edu.
ABSTRACT
Background: It is well established that folic acid prevents neural tube defects. Although the mechanisms remain unclear, multivitamins containing folic acid may also protect against other birth defects, including congenital heart defects.
Objective: Our goal was to establish a maternal metabolic risk profile for nonsyndromic congenital heart defects that would enhance current preventive strategies.
Design: Using a case-control design, we measured biomarkers of the folate-dependent methionine and homocysteine pathway among a population-based sample of women whose pregnancies were affected by congenital heart defects (224 case subjects) or unaffected by any birth defect (90 control subjects). Plasma concentrations of folic acid, homocysteine, methionine, S-adenosylmethionine (SAM), S-adenosylhomocysteine (SAH), vitamin B-12, and adenosine were compared, with control for lifestyle and sociodemographic variables.
Results: After covariate adjustment, case subjects had higher mean concentrations of homocysteine (P < 0.001) and SAH (P < 0.001) and lower mean concentrations of methionine (P = 0.019) and SAM (P = 0.014) than did control subjects. Vitamin B-12, folic acid, and adenosine concentrations did not differ significantly between case and control subjects. Homocysteine, SAH, and methionine were identified as the most important biomarkers predictive of case or control status.
Conclusions: The basis for the observed abnormal metabolic profile among women whose pregnancies were affected by congenital heart defects cannot be defined without further analysis of relevant genetic and environmental factors. Nevertheless, a metabolic profile that is predictive of congenital heart defect risk would help to refine current nutritional intervention strategies to reduce risk and may provide mechanistic clues for further experimental studies.
Key Words: Birth defects congenital heart defects risk methionine homocysteine folate
INTRODUCTION
Congenital heart defects occur among 11 per 1000 live births in the United States (1) and are substantially more prevalent among stillbirths and miscarriages (2). Annually, heart defects account for 6000 total deaths and one-tenth of infant deaths (1). Only 15% of heart defects can be attributed to a known cause (1). The remaining cases are thought to result from complex interactions involving environmental exposures, maternal lifestyle factors, and genetic susceptibilities.
Folic acid is well established in preventing neural tube defects (NTDs). Evidence also suggests that multivitamins containing folic acid may protect against congenital heart defects (3-5) However, the mechanism or mechanisms underlying this effect remain unidentified. Several studies measured plasma biomarkers of folate metabolism among women with pregnancies affected by NTDs and orofacial clefts (6-10). In general, those studies demonstrated that increased plasma homocysteine and decreased folic acid are detrimental to the developing embryo.
Only 2 studies reported altered concentrations of biomarkers associated with homocysteine metabolism among women with pregnancies affected by heart defects. In one study, median fasting plasma homocysteine concentrations among 27 women who had affected pregnancies were significantly higher than 56 women with unaffected pregnancies (11). In the other study, fetal cells recovered from amniotic fluid of pregnancies affected by heart defects had elevated intracellular homocysteine relative to fetal cells of unaffected pregnancies (12).
Evidence demonstrates that homocysteine is elevated among women with a history of adverse pregnancy outcomes, including preeclampsia (13), recurrent early pregnancy losses (14), NTDs (6), orofacial clefts (15), and congenital heart defects (11). Elevated homocysteine is also associated with neurodegenerative disorders, cardiovascular disease, and cancer (16-19). Homocysteine is the product of the intracellular methionine cycle, in which methionine is initially activated by ATP to S-adenosylmethionine (SAM), the primary methyl donor for essential methyltransferase reactions (Figure 1). After methyl transfer, SAM is converted to S-adenosylhomocysteine (SAH). The sole source of homocysteine in the body is the hydrolysis of SAH. Interestingly, the equilibrium dynamics favor the reverse reaction, the synthesis rather than the hydrolysis of SAH. Thus, elevated homocysteine concentrations cause SAH to accumulate (20). Increased SAH is a potent product inhibitor of cellular methyltransferases, which during organogenesis can alter gene expression, cell differentiation, and apoptosis (21, 22).
FIGURE 1.. The folate-dependent methionine and homocysteine metabolic pathway. Biomarkers indicated by bold type were measured in the present study. B12, vitamin B-12 (cobalamin); MAT, methionine adenosyltransferase; mRNA, messenger RNA; MS, methionine synthase; MT, methyltransferase; SAH, S-adenosylhomocysteine; SAHH, SAH hydrolase; SAM, S-adenosylmethionine; THF, tetrahydrofolate; tRNA, transfer RNA.
The goal of our study was to determine whether biomarkers of the folic acid pathway, including homocysteine and methionine metabolism are altered among nonpregnant women who have had a pregnancy affected by congenital heart defects compared with women without such a history. This may provide new insights into mechanisms that confer an increased risk of having pregnancies affected by congenital heart defects.
SUBJECTS AND METHODS
The protocol and provisions for informed consent were reviewed and approved by the Institutional Review Board at the University of Arkansas for Medical Sciences. Case subjects were identified and ascertained through the Arkansas Reproductive Health Monitoring System, a statewide birth defects registry. Inclusion criteria for the case subjects were as follows: 1) resident of Arkansas at the time of the completion of the index pregnancy and at the time of enrollment in the study; 2) pregnancy outcome was a liveborn, stillborn, or elective termination; 3) pregnancy ended between February 1998 and September 2003; 4) a physician diagnosis of a nonsyndromic septal, conotruncal, or right- or left-sided obstructive heart defect that was confirmed by prenatal or postnatal echocardiogram, surgical, or autopsy report; 5) English or Spanish speaking; and 6) the case and control subjects had completed participation in the National Birth Defects Prevention Study (NBDPS) (23). Case subjects for whom the pregnancy was also affected by a known single gene disorder, chromosomal abnormality, or syndrome were excluded; only nonsyndromic congenital heart defects were included in this study.
Control subjects were Arkansas residents who had live births that were unaffected by any birth defect and were English or Spanish speaking. They were randomly chosen from all birth certificates registered at the Arkansas Department of Health, with birth dates between June 1998 and February 2003. Case or control subjects who were pregnant at the time of the blood draw or on antiepileptic medications were not eligible for the study. After determining eligibility for case or control subjects, a research nurse contacted the subjects by mail and telephone, describing this local study. During scheduled home visits, the nurse obtained written informed consent and a blood sample by routine venipuncture.
Covariates
The maternal educational level and household income were reported during a structured computer-assisted telephone interview that was specifically designed for an ongoing multisite case-control study, the NBDPS). Further details about the NBDPS were published previously (23). At the time of the home visit, a Block Abbreviated Food-Frequency Questionnaire (24) was administered, and information about the current use of multivitamins, cigarettes, and alcohol and about caffeine intake was obtained.
Sample preparation and biomarker measurement
Blood samples were collected into evacuated tubes containing EDTA, immediately chilled on ice, and centrifuged at 4000 x g for 10 min at 4 °C. Aliquots of the plasma layer were transferred into cryostat tubes and stored at 80 °C until analysis. Aliquots were thawed for extraction and HPLC quantification. Plasma folic acid and vitamin B-12 concentrations were measured by using Quantaphase II radioimmunoassay kit from Bio-Rad Laboratories (Hercules, CA).
To determine total homocysteine and methionine, 50 µL freshly prepared 1.43 mol/L sodium borohydride solution containing 1.5 µmol/L EDTA, 66 mmol/L NaOH, and 10 µL iso-amyl alcohol was added to 200 µL plasma to reduce sulfhydryl bonds. The samples were incubated at 40 °C on a shaker for 30 min. To precipitate proteins, 250 µL ice-cold 10% meta-phosphoric acid was added and mixed well, and the sample was incubated for 10 min on ice. After centrifugation at 18 000g for 15 min at 4 °C, the supernatant fluid was filtered through a 0.2-µm nylon membrane filter (PGC Scientific, Frederic, MD) and a 20-µL aliquot was injected into the HPLC system.
For determination of SAM, SAH, and adenosine, 100 µL 10% meta-phosphoric acid was added to 200 µL plasma to precipitate protein; the solution was mixed well and incubated on ice for 30 min. After centrifugation for 15 min at 18 000 x g at 4 °C, supernatant fluids containing SAM and SAH were passed through a 0.2-µm nylon membrane filter, and 20 µL was injected into the HPLC system. The methodologic details for metabolite elution and electrochemical detection were described previously (25, 26).
Statistical analysis
Sociodemographic and lifestyle characteristics of case and control subjects were compared with the Fisher exact test for categorical variables or with the Mann-Whitney U or Wilcoxon rank-sum test for continuous variables. All plasma biomarkers exhibited positively skewed distributions; therefore, to achieve normality, biomarker data were log-transformed (natural log) before analysis. Normality of the transformed data was verified by using the Anderson-Darling test (27). Outlier analysis was performed for each biomarker, stratified by case-control status. Among the case subjects, 4 questionable values that fell beyond 3 SDs from the mean (in both the log-transformed and untransformed data) were considered to be outliers. There were no outliers identified among the control subjects.
Mean log-transformed biomarker concentrations of case and control subjects were compared by using a Student t test, whereas linear regression was used to adjust these comparisons for age, race, educational level, household income, multivitamin supplement intake, number of cigarettes smoked daily, alcohol consumption, caffeine intake, and interval between the end of pregnancy and the blood draw. Crude and adjusted odds ratios and corresponding 95% CIs for the association between plasma biomarkers and case-control status were computed by using logistic regression. The most important predictors were identified by using a "best subset" approach (28). Analyses were performed with use of the SAS statistical package version 8.2 (SAS Institute, Cary, NC).
RESULTS
Study population
Two hundred fifty-six case subjects were contacted to participate in the study; 32 (12.5%) of the case subjects refused to participate. Similarly, one hundred nine control subjects were contacted to participate in the study; 19 (17.4%) of them refused to participate. Two case subjects had pregnancies that resulted in a stillbirth; the remaining women had affected pregnancies that ended in a live birth. Selected characteristics of the 224 case subjects and 90 control subjects that participated in this study are presented in Table 1. The sample consisted primarily of white women. At the time of the blood draw, 62.5% of the case subjects were younger than 30 y, and less than one-half reported alcohol consumption (45.1%). No significant difference was observed between the number of case subjects who reported regular multivitamin use (45.1%) and the number of control subjects who reported regular use (46.7%). Smoking was more prevalent among case subjects (28.1%) than control subjects (16.7%; P = 0.043). The mean caffeine intake did not vary significantly between case subjects (44.59 mg/d) and control subjects (44.16 mg/d; P = 0.628). The median time between the end of pregnancy and the blood draw was significantly longer for control subjects (24.5 mo; range = 4.849.3 mo) than for case subjects (14.9 mo; range = 0.152.2 mo; P < 0.01). Eighty-nine (98.9%) of the 90 control subjects and 183 (81.7%) of the 224 case subjects participated 6 mo after the end of pregnancy (data not shown).
View this table:
TABLE 1. Selected characteristics of cases and controls at the time of study participation
Biomarker analysis
Plasma concentrations for the 7 biomarkers and the SAM-to-SAH ratio are summarized in Table 2. All biomarker concentrations, except for vitamin B-12, were significantly different between case and control subjects. As compared with controls, case subjects had higher mean homocysteine and SAH concentrations and lower mean methionine and SAM concentrations. The mean folic acid concentration for case subjects was significantly lower (10.33 mg/L) than for control subjects (11.99 mg/L). These biomarkers, except for plasma folic acid and adenosine, remained significant at the 5% level after adjusting for age, race, educational level, household income, smoking, multivitamin intake, alcohol consumption, caffeine intake, and time between the end of pregnancy and blood draw. Similar results were obtained when the outliers were included in the analysis.
View this table:
TABLE 2. Plasma biomarker concentrations and crude and adjusted P values for the comparison of log-transformed plasma biomarker concentrations between cases and controls1
Multiple linear logistic regression models stratified by multivitamin use were fitted to identify the simultaneous effect of the biomarkers on case or control status (Table 3). All models were adjusted for the mother's level of education, number of cigarettes smoked daily, and interval between the end of pregnancy and the blood draw. The SAM:SAH was not included in the model because of collinearity with SAM and SAH. When all subjects, regardless of their multivitamin use, were analyzed together, plasma homocysteine, SAH, and, to a lesser extent, methionine were identified as the most important predictive biomarkers by the best-subset selection method. None of the biomarkers were significantly different in multivitamin users than in nonusers (data not shown).
View this table:
TABLE 3. Adjusted odds ratios (ORs) and 95% CIs for the association between case or control status and log-transformed plasma biomarker concentrations in all participants1
The concentrations of homocysteine, methionine, and SAH were compared between case and control subjects by computing crude odds ratios at various cutoffs on the basis of biomarker percentiles within controls (Table 4). Of the 224 case subjects, 137 (61.2%) had homocysteine concentrations above the 70th percentile of control subjects, and 31.7% had concentrations above the 95th percentile of control subjects, indicating a shift to the right (higher values) in the plasma homocysteine distribution of case subjects relative to control subjects. The distribution of plasma SAH values in case subjects was also shifted to the right with respect to control subjects: 103 case subjects (46.2%) had plasma SAH concentrations above the 70th percentile of SAH concentrations in control subjects, and 13.5% had concentrations above the 95th percentile of control subjects. Conversely, the distribution of plasma methionine in case subjects was shifted to the left (lower values) with respect to control subjects. Eighty-six cases (38.6%) had methionine concentrations below the 20th percentile of methionine concentrations in control subjects, and 17.9% had concentrations below the 5th percentile of control subjects. These distributional differences remained important after covariate adjustment. Similar results were obtained when the outliers were retained in the analysis.
View this table:
TABLE 4. Odds ratios (ORs) and 95% CIs for the association between case and control status at different cutoffs of plasma homocysteine, S-adenosylhomocysteine (SAH), and methionine
DISCUSSION
After controlling for covariates, plasma concentrations of homocysteine and SAH were elevated, and concentrations of SAM, methionine, and SAM:SAH were reduced among case subjects. There were no significant differences in folate, vitamin B-12, or adenosine between case and control subjects. The 95th percentile homocysteine concentration among women aged 2039 y in the third National Health and Nutrition Examination Survey sample was 10.4 µg/L. In our sample, 34.4% of case subjects and 5.6% of control subjects had greater homocysteine concentrations (29). The factors contributing to this abnormal maternal metabolic profile are unclear, as are the effects on cardiac development.
Possible causes for the altered biomarkers include deficiencies in folate or vitamin B-12 or both, genetic polymorphisms encoding enzymes in the methionine cycle, or exogenous factors. Mean plasma folic acid was not significantly lower among case subjects than control subjects, as would be expected if folate deficiency was strongly contributing. Most pregnancies in this study began after January 1, 1998, when folic acid fortification of grain products was mandated (30). The mean folic acid concentration among our control subjects was 11.99 mg/L (31), which is consistent with the mean postfortification concentrations among women of childbearing age in other studies (32).
The similarity in folic acid concentrations between case and control subjects suggest that case subjects were either more genetically susceptible to having increases in homocysteine and SAH or were more likely to have reduced activity of relevant enzymes. For example, an alteration in the function of, or genes encoding, methionine synthase may reduce the conversion of homocysteine to methionine (Figure 1). Multiple case-control studies have evaluated the association between NTDs and common genetic variants in the methionine synthase gene with inconsistent results (33-36). To our knowledge, the association between methionine synthase activity and congenital heart defects is as yet undetermined. Thus, further studies are needed of genetic variants associated with congenital heart defects that would cause elevated homocysteine.
Our findings of elevated homocysteine among case subjects are consistent with other studies comparing vitamin-dependent homocysteine metabolism among mothers of offspring with orofacial clefts (15) and NTDs (6) with those of mothers with nonmalformed offspring (15). Studies have also found elevated homocysteine concentrations in women who have had preeclampsia (13) and recurrent early miscarriages (14). In those studies, maternal vitamin and homocysteine concentrations were measured up to 6 y after the relevant embryonic period (15). Thus, it was not possible to determine whether alterations in biomarkers among nonpregnant women having adverse pregnancy outcomes reflect the biomarker concentrations during organogenesis or early pregnancy loss. However, it is known that temporal changes in homocysteine concentrations are minimal over a 12-y time frame (37). Also, we are unable to determine the temporal relation between the alteration in biomarkers and the adverse pregnancy from our findings and those of others. The evidence suggests that these alterations, if not causally related, are associated with an increased risk of having adverse pregnancy outcomes, including congenital heart defects.
Possible mechanisms by which homocysteine may have an embryotoxic effect include oxidative stress and secondary accumulation of SAH, which leads to product inhibition of DNA methyltransferase reactions, DNA hypomethylation, and altered gene expression (38). Lifestyle factors such as alcohol intake and cigarette smoking are associated with increased oxidative stress (39), which can decrease the functional activity of methionine synthase through limiting the availability of reduced vitamin B-12 (35, 40, 41). Of note in our sample is the greater number of smokers among case subjects than among control subjects. When homocysteine concentration was stratified by smoking status among case and control subjects, the differences were higher among smokers than nonsmokers in both the case and the control subjects, but the results were not significantly different (data not shown). Animal studies showed that oxidative stress during fetal organogenesis can result in damage to mitochondrial and nuclear DNA, to protein structure and function, or to membrane lipids and signal transduction pathways. Depending on the timing and duration of exposure, the functional consequences of oxidative injury are embryopathy and dysmorphogenesis. For example, exposure of rats to 20% oxygen during early neurulation significantly increased the incidence of NTDs relative to unexposed embryos (42). Hypomethylation also was shown to lead to embryonic lethality and multiple developmental malformations (43, 44). For example, maternal exposure to valproic acid was shown to induce DNA hypomethylation in NTD-affected pregnancies in mice (45).
Several methodologic limitations of our study should be considered. Despite efforts to enroll participants soon after the index pregnancy, the venipuncture occurred well after the index pregnancies had ended. Thus, our measurements may not represent the biomarker concentrations at the time of organogenesis. However, it is remarkable that, even far after delivery, the biomarkers measured are significantly different between case and control subjects.
Given the relatively rare prevalence of most birth defects, the most common epidemiologic design used to investigate etiologic factors is a case-control study after the completion of pregnancy. Substantially more resources would be required to launch a cohort study to measure biomarkers among women enrolled before conception and followed to the completion of pregnancy. However, confidence in our findings is supported by noting that although folate, vitamin B-12 (46), and homocysteine (47) gradually decrease during pregnancy, at 6 wk after delivery, the concentrations are similar to those before conception or early pregnancy (47). Further, we examined our data to determine whether concentrations of homocysteine and folate correlated with the time interval between the end of pregnancy and the blood draw in case and control subjects. No discernible trend was found. Therefore, it is likely that, on average, the metabolic patterns observed even more than a year after pregnancy reflect stable adult profiles.
We did not examine the relation between cardiac phenotypes and the biomarkers. Particular cardiac phenotypes may be more influenced by altered homocysteine and methionine metabolism by way of different developmental pathways. Another potential source of bias in our study could be the failure to identify cardiac lesions among control pregnancies. Although control subjects were excluded if the medical chart indicated any congenital malformation, echocardiograms were not required to rule out a cardiac lesion.
If further investigations replicate those of this study, this altered metabolic profile might provide a measure of preconceptional risk and, more importantly, new information for the refinement of nutritional intervention strategies designed to reduce risk. Currently, primary prevention efforts to reduce the occurrence of nonsyndromic congenital heart defects are limited to folic acid fortification and supplementation and to avoidance of specific teratogens. Clinical efforts are focused on secondary prevention by using fetal echocardiograms, thereby optimizing postnatal medical and surgical outcomes. Identification of the genetic, epigenetic, and metabolic factors that may increase the likelihood of having a pregnancy affected by congenital heart defects may enhance primary prevention. One might imagine that during routine gynecologic visits women are screened for cancer, sexually transmitted diseases, and an increased risk of adverse pregnancy outcomes. Such screening would only be helpful if effective interventions to modify the risk profile were available. Currently, a national effort is under way to identify biomarkers for cancer risk and prevention (48). Perhaps, it is now time to begin a similar effort for the primary prevention of birth defects.
ACKNOWLEDGMENTS
We appreciate Sadia Ghaffar for her diligent classification of cardiac lesions, and we are indebted to the Arkansas families whose participation made this study possible.
Author contributions were as follows: CAH (experimental design and analysis, manuscript writing), MAC (statistical data analysis design and manuscript writing), SM (laboratory biomarker analyses), WZ (statistical analyses), and SJJ (experimental analysis, manuscript writing). There were no potential conflicts of interests for the authors of this manuscript.
REFERENCES