一篇有关DNA甲基化的表观遗传学英文文献,与大家共勉

Epigenetics 4:6, 394-398; August 16, 2009; ? 2009 Landes Bioscience

REsEARch pApER

Introduction

Many health policies worldwide, including the US Department of Health and Human Services Public Health Service Centers for Disease Control,1 UK National Service Framework for Maternity Services 2 and National Institute for Health and Clinical Excellence guidance on Antenatal Care 3 highlight the critical importance of folic acid supplementation during pregnancy in the prevention of neural tube defects. However, supplementation with this vitamin has also been shown to have a range of other benefits. For example, it is implicated in reduction of the risk of low birth weight (with its associated sequelae)1-4 and has also been shown to be essential for normal development.5

A major effect of maternal folic acid intake duing pregnancy on the unborn child is thought to be through its role as an inter-mediate in the one carbon pathway.4,5 A reduction in folate sup-ply or in its conversion to one carbon methyl donors can lead to chromosomal strand breaks and altered methylation of spe-cific regions of the DNA (CpG islands).6 In monocytes, animal models, primary tumour cells and cell lines, the effect of folate and/or methyl group depletion on methylation-associated change has received considerable attention. In these studies, a near uni-versal finding of an overall decrease in genomic methylation has been identified, frequently accompanied by inappropriate gene-specific hypo- or hyper-methylation at promoter-associated CpG

islands.6-8 Thus, the bioavailability of folate and its one-carbon intermediaries to the fetus, particularly during the first trimes-ter of pregnancy, is a key component in “setting” the epigenetic programme through alteration of methylation patterns across the genome.

While these studies provide convincing data of the effect of folate on fetal genome-wide DNA methylation, much of the data is derived from animal or in vitro models. Furthermore, the mechanisms by which folate supplementation influences fetal methylation patterns and growth in humans are poorly under-stood. In this preliminary study, we examined the relationship between supplemental folic acid intake during pregnancy, blood folate parameters and genome-wide methylation levels (using methylation of long interspersed nucleotide element-1, LINE-1, sequences as a surrogate). Importantly, we present data on these relationships, for the first time, in human cord blood.

Results

Associations between folate-related parameters and LINE-1 methylation. In order to determine which of the folate-related parameters were the best predictors of LINE-1 methylation, we examined potential correlations between LINE-1 methylation and folic acid supplementation/blood folate indices. The selected indicator of genome-wide methylation, LINE-1, did not deviate

*Correspondence to: William E. Farrell; Email: w.e.farrell@https://www.360docs.net/doc/bb15787758.html, Submitted: 08/06/09; Accepted: 08/18/09

Previously published online: https://www.360docs.net/doc/bb15787758.html,/journals/epigenetics/article/9766LINE-1 DNA methylation is inversely correlated with

cord plasma homocysteine in man

A preliminary study

Anthony A. Fryer,1 Tamer M. Nafee,2 Khaled M.K. Ismail,2 William D. carroll,3 Richard D. Emes 1 and William E. Farrell 1,*

1

Institute of Science and Technology in Medicine; 2Academic Unit of Obstetrics and Gynaecology; Keele University Medical School; University Hospital of North

Staffordshire; Stoke-on-Trent, Staffordshire UK; 3Department of Paediatrics; Derbyshire Children’s Hospital; Derby, UK

Keywords: LINE-1, methylation, homocysteine, fetal, folic acid, cord blood

Abbreviations: LINE-1, long interspersed nucleotide element-1; HPLC, high performance liquid chromatography

Folic acid supplementation during pregnancy has known beneficial effects. It reduces risk of neural tube defects and low birth weight. Folate and other one-carbon intermediates might secure these clinical effects via DNA methylation. however, most data on the effects of folate on the epigenome is derived from animal or in vitro models. We examined the relationship be-tween cord blood methylation and maternal folic acid intake, cord blood folate and homocysteine using data from 24 pregnant women. Genome-wide methylation was determined by the level of methylation of LINE-1 repeats using pyrosequencing. We show that cord plasma homocysteine (p = 0.001, r = -0.688), but not serum folate or maternal folic acid intake, is inverse correlated with LINE-1 methylation. This remained significant after correction for potential confounders (p = 0.004). These data indicate that levels of folate-associated intermediates in cord blood during late pregnancy have significant consequences for the fetal epigenome.

REsEARch pApER

BRIEF REpoRT

Discussion

These data show that cord plasma homocysteine concentration is significantly inversely correlated with genome-wide methylation, as assessed using mean LINE-1 methylation, in humans. To our

knowledge, this is a novel and previously unreported but highly

significant finding.

In this study, we used LINE-1 methylation by Pyrosequencing

as a marker of genome-wide methylation. Over 500,000 LINE-1

sequences exist,9 comprising about 5% of the human genome.

Yang et al.9 and others have shown that LINE-1 methylation measured by Pyrosequencing is reproducible and correlates well with global DNA methylation level. In particular, Pyrosequencing technology can detect subtle differences in mean LINE-1 methylation levels. Indeed, our data illustrate the view of Ogino et al.10 that the use of an accurate and precise method to measure methylation such as LINE-1 is essential to demonstrate the subtle differences in methyla-tion patterns expected in response to changes in folate exposure.

The relationship between folate and methylation patterns is complex. Folate has a range of effects that are critical to fetal development. Dietary

folate deficiency leads to genome-wide hypomethylation 11 and hyper-homocysteinaemia.12 However, folate is also critical for DNA synthesis and

is a substrate for a number of reactions that influence the metab-olism of several amino acids. While the contribution of folate or its intermediaries to these individual pathways is difficult to dissect, our data suggests that homocysteine may reflect a better indicator of methyl group supply and therefore more accurately reflect ‘functional’ folate availability and utilization in genome-wide methylation. This view is supported by Ingrosso et al.13 who showed that hyper-homocysteinaemia was associated with DNA hypo-methylation in uraemic patients.Our data also suggests that factors affecting the one carbon pathway in late gestation may have an important role in influenc-ing epigenetic patterns. This is particularly important given the focus on supplementation during the first trimester in the preven-tion of neural tube defects 1-3 and raises the question as to whether further research should focus on changes in the one-carbon path-way throughout gestation.The potential clinical implications of our findings are worthy of further investigation. Interestingly, in our preliminary studies, we also identified an association between LINE-1 methylation and birth weight centile (p = 0.014, data not shown), support-ing the concept that factors associated with changes in methyla-tion during pregnancy may be important in fetal and potentially

longer term outcomes, as descibed in the Barker hypothesis.14

from normality (p = 0.854), allowing the generation of correla-tion coefficients. Our data (Table 1) showed that, in univariate analyses, neither cord serum folate levels (p = 0.338) nor reported total folic acid supplemental intake during pregnancy (p = 0.152) were significantly associated with mean LINE-1 methylation. The association between prescribed folic acid dose per day and mean LINE-1 methylation approached statistical significance (p = 0.080). However, as shown in Figure 1, fetal plasma homo-cysteine concentration was highly significantly inversely cor-related with mean LINE-1 methylation (p = 0.001, r = -0.688; Table 1).Multivariate assessment. In order to examine the inter-relationship between the predictors and identify potential con-founding factors, we then used a multivariate stepwise regression model to determine the best predictors of mean LINE-1 methy-lation. In this model, we included folate-related parameters listed in Table 1 together with fetal gender, gestational age, maternal age, maternal parity and maternal body mass index. In this analysis, plasma homocysteine levels remained the only significant predictor (p = 0.004).Table 1. Associations between folate-related parameters and cord blood LINE-1 methylation

Variable

n

p

adjusted R 2

r

precribed folic acid dose per day

(mg)

24

0.080

0.093

0.364

Reported gestational supplemen-tal folic acid intake (mg)230.1520.0520.308cord serum folate (μmol/L)230.338-0.0020.209cord plasma homocysteine (μmol/L)210.0010.446

-0.688Figure 1. Association of mean LINE-1 methylation with cord plasma homocysteine concentration. Error bars represent standard error of the mean (sEM) of at least three replicates for each of the 4 LINE-1 cpG sites.

in humans. Undoubtedly, replication of our findings in larger independent populations, and providing answers for the above questions, would have major implications on current policies for folic acid supplementation during pregnancy and of general foodstuffs.

Subjects and Methods

Patients. Twenty-nine women attending the Maternity Unit of the University Hospital of North Staffordshire were recruited into the study between 2007–2008. The study was approved by the Local Research Ethics Committee and all women provided written informed consent. Demographic data on maternal age at delivery, parity, body mass index, fetal gender, anti-epileptic drug usage and folate supplementation during pregnancy were collected prospectively by the study research midwife and these were checked against the mother’s clinical record. Given the known anti-folate and teratogenic effects of carbamazepine,19 five women on this medication were excluded from subsequent analy-ses. Table 2 shows the demographics of the study population.Blood samples were collected at term from the umbilical cord into sample bottles containing ethylene diamine tetraacetic acid (for methylation analysis), heparin (for plasma homocysteine measurement) and without anti-coagulant (for serum folate measurement).

Measurement of blood folate parameters. Folate was measured in 23/24 serum samples using a competitive immunoassay-based assay employing direct chemiluminescent technology with an acridinium ester label on Siemens Advia Centaur equipment and reagents (Siemens Medical Solutions Diagnostics) as per manu-facturer’s instructions. Homocysteine was measured in 21/24 heparinised plasma samples using High Performance Liquid Chromatography (HPLC) at the Heart of England Foundation Hospital, Birmingham, UK as described by Martin et al.20 This method uses ion-paired reversed-phase HPLC of reduced sam-ples followed by oxidative mode coulometric electrochemical detection.

However, it is clear that the relationship between folate intake, levels of one carbon intermediates, DNA methylation and fetal growth is not a simple one. In this context, while maternal nutri-tion is a recognized predictor of birth weight, the precise role of folate remains a topic of debate.15,16

We recognize that there are limitations to this preliminary study. The relatively small sample size deterred us from under-taking multiple corrections for potentially confounding data such as maternal genotype, alcohol and coffee consumption and tobacco exposure that exposed the study to a risk of type II error. However, while a larger study is underway examining a range of folate parameters throughout gestation, we felt that these prelim-inary findings on late gestation effects were sufficiently compel-ling and novel as to warrant dissemination to allow reproduction elsewhere. Equally, there is also a small risk of type I error, though the signifcance of the relationship between LINE-1 methylation and cord plasma homocysteine is sufficiently strong to withstand Bonferroni correction. Furthermore, our analyses were hypothe-sis-driven and are consistent with findings in animal models.To our knowledge these data are unique in that they pres-ent, for the first time, measurements in human cord blood. The association between folic acid supplementation and the fetal epig-enome raises several intriguing questions: (1) Does maternal folic acid supplementation produce its beneficial effects through fetal epigenetic modifications? (2) If so, what is the best predictor of methyl donor supply and are these epigenetic changes dose- and time-related? (3) What are the long-term implications of these epigenetic modifications? (4) If the association between fetal birth weight and the fetal epigenome hold true, can this provide the mechanistic link between birth weight and risk of adult-hood disease later in life? Recent epidemiological data showing a robust association between folic acid supplementation and the risk of respiratory disease during childhood highlights the danger of assuming only benefits when considering the value of folic acid supplementation in public health programs.18

In summary, our data shows that folate indices during preg-nancy are associated with methylation of the fetal epigenome

Table 2. Demographic data on the study population

Maternal variables Age (years, mean ± sD), n = 2429.4 ± 7.0parity (median, range), n = 23 1 (0–3)BMI (median, IQR 1), n=23

26.2 (22.4–30.8)Daily folate dose (mg, median, range), n = 24400 (0–5000)Total supplemental folate intake (g, median, IQR), n = 23

3.2 (0–95)Neonatal variables

Gestation (weeks, mean ± sD), n = 2439.3 ± 1.9Birth weight (kg, mean ± sD), n = 24 3.33 ± 0.52Birth weight (percentile, median, IQR), n = 24

30.5 (12.5–58.4)

Gender (% female), n = 24

58.3%cord serum folate (μmol/L, mean ± sD), n = 2315.8 ± 3.5cord plasma homocysteine (μmol/L, mean ± sD), n = 2110.8 ± 3.8cord LINE-1 methylation (mean %, mean ± sD), n = 24

71.9 ± 3.0

1

IQR: Inter-quartile range.

at dispensation 18 comprises, in the pre-converted sequence, 5 T residues and a non-CpG C residue. Hence, in the event of poor conversion, the signal of the peaks following this would demon-strate lower peak heights than those prior to the poly-T signal. Hence, we were able to assess efficiency and reproducibility of conversion by examining the ratio of the peak heights for the G residues at dispensations 21 and 17. This gave consistent ratios of 0.997 ± 0.055. As an additional check, LINE-1 Pyrosequencing of in vitro methylated and bisulphite converted DNA was also used as verification of the efficiency of bisulphite conversion. Sham bisulphite-conversion served as negative control for the specificity of the LINE-1 primers.

Statistical analysis. All data were analysed using STATA sta-tistical software (College Station, TX, version 8).

Normality of data were tested formally using the Shapiro-Wilk W test. In accordance with our null hypotheses, associa-tions between LINE-1 methylation and maternal folate intake,

cord blood folate and homocysteine were sought using univariate linear regression analysis. For normally distributed data, degrees of association were expressed as correlation coefficients. To exam-ine the data for potential confounding factors, a stepwise multi-variate regression model was created as described in the Results

section.Acknowledgements We are grateful to Mrs. Angela Rooney for invaluable help in

recruiting the patients for this study, and to Dr. Craig Webster for help on the measurement of plasma homocysteine. Supported in part by Egyptian Cultural Bureau (to K.M.K.I. and T.M.N.), West Midlands Association of Clinical Biochemistry Robert Gaddie Memorial Fund (to A.A.F.), North Staffordshire Medical Institute (to W.D.C., A.A.F. and K.M.K.I.) and World

Cancer Research Fund (to W.E.F., A.A.F., R.D.E., W.D.C. and K.M.K.I.).

The authors’ responsibilities were as follows: A.A.F., K.M.K.I., W.D.C. and W.E.F. were responsible for the initial study design, K.M.K.I. oversaw patient recruitment, W.D.C. designed the demographic data collection tool, T.M.N. performed the LINE-1 analysis under the supervision of W.E.F. and A.A.F. A.A.F., W.D.C. and R.D.E. performed the data analysis and all authors were involved in completing the final manuscript. All authors had a significant role in this project and approved the final ver-sion of the manuscript.

DNA extraction and bisulphite conversion. DNA was extracted from cord blood using standard phenol/chloroform

procedures. Sodium bisulfite conversion was carried out as pre-viously described 21 using the EZ DNA Methylation Gold kit (Zymo Research, Orange County, CA, USA).

LINE-1 methylation. LINE-1 assessment of global methyla-tion was performed essentially as described by the manufacturer using Pyrosequencing with the LINE-1 primer set (PyroMark LINE-1 kit, Biotage AB, Uppsala, Sweden) and results were analyzed using the Q-CpG software (Biotage AB, Uppsala, Sweden).

The analysed sequence allowed quantitation of the methylation status of four CpG sites. An example of the pyrograms obtained

is shown in Figure 2. The level of methylation was expressed as the percentage of methylated cytosines over the sum of methy-lated and unmethylated cytosines. These data were combined to give a mean methylation value for the four sites. Experiments were repeated at least three times on each cord blood sample and mean values were calculated from these separate experiments. Non-CpG cytosine residues were used as built in controls to verify bisulphite conversion. In the sequence analysed (Fig. 2

), the poly-T signal Figure 2. Example of LINE-1 sequencing program. Ratios of peak

heights at G residues at dispensations 17 and 21 (excluding enzyme [E]

and substrate [s] addition, marked with *) were used as an internal

control for bisulphite conversion.

References

1.

US Department of Health and Human Services Public Health Service Centers for Disease Control. Recommendations for the use of folic acid to reduce the number of cases of spina bifida and other neural tube defects. Morbidity and Mortality Weekly Report 1992; 41.

2. Department of Health: National Service Framework for Children, Young People and Maternity Services, 2004.

3.

National Institute for Health and Clinical Excellence, Clinical Guideline 62; Antenatal care: Routine care for the healthy pregnant woman, 2008.

4.

Nafee TM, Farrell WE, Carrol W , Fryer AA, Ismail KMK. Epigenetic control of fetal gene expression. Br J Gynaecol 2008; 115:158-68.

5.

Zeisel SH. Importance of methyl donors during repro-duction. Am J Clin Nutr 2009; 89:673-7.

6.

Kim YI. Folate, colorectal carcinogenesis and DNA methylation: lessons from animal studies. Environ Mol Mutagen 2004; 44:1-25.

7. Kim YI. Will mandatory folic acid fortification prevent

or promote cancer. Am J Clin Nutr 2004; 80:1123-8.8. Jones PA, Baylin SB. The fundamental role of epigenet-ic events in cancer. Nature Rev Genet 2002; 3:415-28.9. Yang AS, Estécio MR, Doshi K, Kondo Y, Tajara EH,

Issa JP . A simple method for estimating global DNA methylation using bisulfite PCR of repetitive DNA elements. Nucleic Acids Res 2004; 32:38.

10. Ogino S, Kawasaki T, Nosho K, Ohnishi M, Suemoto

Y, Kirkner GJ, Fuchs CS. LINE-1 hypomethylation is inversely associated with microsatellite instability and CpG island methylator phenotype in colorectal cancer. Int J Cancer 2008; 122:2767-73.

11. Jacob RA, Gretz DM, Taylor PC, James SJ, Pogribny

IP , Miller BJ, et al. Moderate folate depletion increases plasma homocysteine and decreases lymphocyte DNA methylation in postmenopausal women. J Nutr 1998; 128:1204-12.

12. Forges T, Monnier-Barbarino P , Alberto JM, Guéant-Rodriguez RM, Daval JL, Guéant JL. Impact of folate and homocysteine metabolism on human reproductive health. Hum Reprod Update 2007; 13:225-38.

13. Ingrosso D, Cimmino A, Perna AF , Masella L, De

Santo NG, De Bonis ML, et al. Folate treatment and unbalanced methylation and changes of allelic expres-sion induced by hyperhomocysteinaemia in patients with uraemia. Lancet 2003; 361:1693-9.

14. Barker DJ. The fetal and infant origins of adult disease.

Br Med J 1990; 301:1111.

15. Yajnik CS, Deshmukh US. Maternal nutrition, intra-

uterine programming and consequential risks in the offspring. Rev Endocr Metab Disord 2008; 9:203-11.

16. Cutfield WS, Hofman PL, Mitchell M, Morison IM.

Could epigenetics play a role in the developmental origins of health and disease? Pediatr Res 2007; 61:68-

75.

17. Scholl TO, Johnson WG. Folic acid: influence on the

outcome of pregnancy. Am J Clin Nutr 2000; 71:1295-

303.18. H?berg SE, London SJ, Stigum H, Nafstad P, Nystad

W. Folic acid supplements in pregnancy and early

childhood respiratory health. Arch Dis Child 2009;

94:180-4.

19. Kishi T, Fujita N, Eguchi T-A, Ueda K. Mechanism for

reduction of serum folate by antiepileptic drugs during

prolonged therapy. J Neurol Sci 1997; 145:109-12.

20. Martin SC, Tsakas-Ampatzis I, Bartlett WA, Jones AF.

Measurement of plasma total homocysteine by HPLC

with coulometric detection. Clin Chem 1999; 45:150-

2.

21. Simpson DJ, Hibberts NA, McNicol AM, Clayton

RN, Farrell WE. Loss of pRb expression in pituitary

adenomas is associated with methylation of the RB1

CpG island. Cancer Res 2000; 60:1211-6.

22. World Health Organisation, WHO Child Growth

Standards based on length/height, weight and age. Acta

Paediatr Suppl 2006; 450:76-85.