the effect of maternal vitamin b status in combination with high … · 2018. 7. 18. · ii the...
TRANSCRIPT
i
The Effect of Maternal Vitamin B12 Status in Combination with High Folate Status on Gene-
Specific DNA Methylation in Cord Blood Mononuclear Cells
by:
Shahnaz Anne Vinta Fard
A thesis submitted in conformity with the requirements for the degree of Master of Science
Department of Nutritional Sciences
University of Toronto
© Copyright by Shahnaz Anne Vinta Fard 2018
ii
The Effect of Maternal Vitamin B12 Status in Combination with High Folate Status on Gene-Specific DNA Methylation in Cord Blood Mononuclear Cells
Shahnaz Anne Vinta Fard
Master of Science
Department of Nutritional Sciences University of Toronto
2018
Abstract
DNA methylation is an important epigenetic determinant in gene expression, with potential consequent
effects on disease susceptibility. Maternal folate and B12 has the potential to modulate DNA methylation via
the provision of S-adenosylmethionine. Using the Illumina Infinium MethylationEPIC array, we found
lower genome-wide (-0.0059±0.0147, p=0.013) and gene-specific (p=0.007) DNA methylation in genes
influenced by maternal low serum B12 and high RBC folate concentrations at early pregnancy. Analyses
using Ingenuity Pathway Analysis and ddPCR determined that B12 and folate-induced DNA methylation
changes altered expression of functionally relevant genes related to metabolism and development. Infants
born to mothers with a low B12 status are shown to have larger infant anthropometric measures. Our data
suggest that low B12 and high folate induced DNA methylation changes in the developing fetus. Given the
potential for these changes to mediate disease risk later in life, future large-scale studies are warranted to
confirm our exploratory findings.
iii
Acknowledgements “The noblest pleasure is the joy of understanding.” Although I do not have a double-blinded, placebo-
controlled, randomized control trial to support Leonardo da Vinci’s statement, what I do have are an
abundance of experiences that formed this unforgettable journey of learning that is my Master’s program. I
have gained invaluable lessons and have grown my understanding of scientific research and things concerning
life. This Master’s has granted me the opportunity to be curious, which has allowed me to produce this work.
I am truly blessed to have been part of a Master’s program at the University of Toronto’s Department of
Nutritional Sciences.
As this chapter of my life is nearing its end, I would like to recognize and thank the great people who have
joined me in this journey and helped make it more memorable. There is so much that I would like to say, but
with limited space, I will do my best to convey my thoughts.
To my supervisor, Dr. Young-In Kim, your belief in me has made this Master’s possible. Thank you for
selecting me, rooting for me to succeed and for pushing me to be better. I appreciate your guidance and I
value the lessons you have imparted. They will always be a part of me.
To my committee members, Dr. Debbie O’Connor and Dr. Rosanna Weksberg, thank you for showing me
the beauty in research and the significance of my topic. Dr. O’Connor – your perspective in the clinical and
real-world aspects added greater meaning to my study and gave me more reason to keep going. Dr. Weksberg
– your ability to take a complicated subject like epigenetics and explain it in a simple manner is such an
incredible skill. I endeavour to follow in your footsteps.
I would also like to specially thank Sanaa Choufani and Youliang Lou of the Weksberg lab. Sanaa – I simply
would not have been able to learn bioinformatics without you. Thank you for answering my many questions!
Youliang – thank you for always making a way to help me, even when it was something out of your breadth
you always went above and beyond.
To my biostatistician, Ruth Croxford – I am in awe of your statistical expertise. Thank you for your patience
when explaining statistics and teaching me SAS code. To Pamela Plant and Colleen Ding – thank you for your
sincere care and kindness and for making ddPCR possible.
iv
To the past and present members of the Kim lab: Kyoung-Jin Sohn, Lesley Plumptre, Eszter Pigott, Heajin
Cheon, Hannah Tateishi, David Im, Bajit Kaur and Siya Khanna, thank you for making the lab a warm and
welcoming place. Kyoung – I enjoyed our talks when you would give me lab and life advice. Thank you for
taking care of me and for feeding me yummy treats! Lesley – you set the bar high for the Kim lab. I admire
you and how you supported and helped me build this project. Eszter and Heajin – you were the lab’s supplier
of sweets! You truly made my experience sweeter. Hannah – thank you for your invaluable help with ddPCR.
I admire your efficiency and your calm disposition throughout the process. David – what would I have done
without you in the lab? You are the best lab partner I could have asked for. Thank you for understanding me,
and for repeatedly teaching me the calculations and wet-lab work in every way possible until I succeeded.
To the incredible individuals and friends at the university: Louisa Kung, Marie-Elssa Morency, Luke Johnston,
Neil Yang and Dana Katkhoda. Louisa – you were the one who first told me about Dr. Kim, so if it was not
for you, well, you would have received significantly less emails over the past two years! Thank you for
sincerely caring and going the extra mile for me. Marie-Elssa – your positivity is inspiring. Thank you for
your care and support through the highs and lows. Luke – thank you for the times that you have helped me
code SAS, even when you preferred R. Neil – thank you for checking up on me and making sure that I was
doing well. Dana – thank you for being there for me. You were someone I could lean on, and your pragmatism
has aided me to press through certain challenges.
To my dearest family, I love you. Mummy and Daddy – thank you for your unwavering love and support.
Thank you for believing in me and being proud of me from the very beginning. Kuya Michael and Joyce –
thank you for being beacons of hope and laughter. Your constant encouragement has helped me persevere.
Ate Christine and Victoria – the Lord knew to put you as my older sisters! You are extraordinary individuals
and I have always looked up to you. Thank you for always being by my side and having my back. To say the
least, I admire you. Ate Christine, thank you for lovingly realigning my mind and my heart to what is true
and good. Ate Victoria, thank you for your peaceful and steadfast reassurance that everything will not only
be okay, but will be greater than what I can imagine. Ate Alma, Ate Belen, Manang Linda and Ate Neneng –
maraming salamat sa pagmamahal at pagdadasal nyo saakin nung bata pa ako at hangang ngayon (thank you very
much for your love and prayers from when I was a child to now).
To God, the One who is my Source, who has held me together, and who made all these possible, I thank you.
You have been faithful until the end, showing me Your love, grace and super-abounding favour.
v
Table of Contents
Acknowledgements ................................................................................................................. iii
Table of Contents ....................................................................................................................... v
List of Tables .......................................................................................................................... viii
List of Figures ........................................................................................................................... ix
List of Appendices .................................................................................................................... x
Appendix A. Analysis of blood biomarkers ....................................................................... x
Appendix B. Final propensity score model in SAS code ................................................... x
Appendix C. Histograms of CpG island and non-CpG island methylation ....................... x
Appendix D. IPA functional analysis summary table of all four datasets .......................... x
Abbreviations .......................................................................................................................... xi
1- Introduction ...................................................................................................................... 1
2- Literature Review .............................................................................................................. 3
2.1 Folate/Folic Acid .......................................................................................................................... 3
2.1.1 Definition, Chemical Structure and Dietary Sources ................................................................. 3
2.1.2 Absorption and Metabolism .............................................................................................. 4
2.1.4 Folate Intake Requirements ............................................................................................ 10
2.1.5 Biomarkers of Folate ..................................................................................................... 12
2.1.6 Folate and Health ......................................................................................................... 17
2.2 Vitamin B12 .................................................................................................................................... 37
2.2.1 Definition, Chemical Structure and Dietary Sources ............................................................... 37
2.2.2 Absorption, Transport and Storage ................................................................................... 38
vi
2.2.3 Biochemical Functions ................................................................................................... 40
2.2.4 Vitamin B12 Intake Requirements ...................................................................................... 41
2.2.5 Biomarkers of Vitamin B12 .............................................................................................. 42
2.2.6 Vitamin B12 and Health .................................................................................................. 44
2.3 Epigenetics ................................................................................................................................... 52
2.3.1 Definition .................................................................................................................. 52
2.3.2 DNA Methylation ........................................................................................................ 52
2.3.3 Epigenetic Reprogramming ............................................................................................ 55
2.3.4 Analysis of DNA Methylation .......................................................................................... 56
2.3.5 Epigenetics and Health and Disease ................................................................................... 58
2.3.6 One-Carbon Nutrients and DNA Methylation ...................................................................... 60
2.3.7 Maternal One-Carbon Nutrients on DNA Methylation and Health Outcomes in Animal Offspring ...... 61
2.3.8 Maternal Folate and B12 status on DNA Methylation and Health Outcomes in Human Offspring ......... 67
2.4 The PREFORM Study ................................................................................................................... 72
3- Rationale, Research Questions, Hypothesis and Objectives ............................................... 74
3.1 Rationale ...................................................................................................................................... 74
3.2 Hypothesis ................................................................................................................................... 76
3.3 Objectives ..................................................................................................................................... 76
4- Effects of Maternal Vitamin B12 Status in Combination with High Folate Status on Gene-
Specific DNA Methylation in Cord Blood Mononuclear Cells ............................................... 77
4.1 Subjects and Methods .................................................................................................................. 77
4.1.1 Study Design and Sample Selection for the Present Study ........................................................ 77
4.1.2 Genome-Wide and Gene-Specific DNA Methylation and Functional Analyses ............................... 81
4.1.3 Functional Analysis ....................................................................................................... 84
4.1.4 Gene Expression Analysis ............................................................................................... 85
4.1.5 Statistic Analysis .......................................................................................................... 88
vii
4.2 Results .......................................................................................................................................... 89
4.2.1 Subject Characteristics ................................................................................................... 89
4.2.2 Genome-wide DNA Methylation ...................................................................................... 93
4.2.3 Gene-Specific Methylation .............................................................................................. 93
4.2.4 Gene Expression ........................................................................................................ 101
4.3 Discussion .................................................................................................................................. 107
4.3.1 Lower genome-wide DNA methylation in infants born to mothers with low B12 status ................... 107
4.3.2 Differentially regulated genes are more hypomethylated in mothers with low B12 status ................. 109
4.3.3 Ramifications of B12 and folate induced changes in DNA methylation on functional and clinical significance
of hypo- and hypermethylated genes ....................................................................................... 110
4.3.4 Upregulation of B12 and folate induced hyper- and hypomethylated genes ................................... 113
4.4 Conclusion ................................................................................................................................. 115
5- Overall Conclusions and Future Directions ...................................................................... 117
5.1 Summary and General Discussion .............................................................................................. 117
5.2 Strength and Limitations ........................................................................................................... 119
5.3 Future Directions ....................................................................................................................... 121
5.4 Final Conclusions ....................................................................................................................... 122
References .............................................................................................................................. 123
Appendix ............................................................................................................................... 141
viii
List of Tables Table 2.1.4. Dietary Reference Intakes (DRIs) for females at various life stages Table 2.1.5.2. Comparison of folate measurement assays Table 2.1.5.3. A summary of folate cutoffs Table 2.1.6.1. Health outcomes associated with inadequate and excess folate Table 2.1.6.3a. Summary of studies investigating maternal folate concentrations Table 2.1.6.3b. Summary of studies investigating cord folate and homocysteine concentrations Table 2.1.6.4. Protective and adverse effects of periconceptional folic acid supplementation on
pregnancy and birth outcomes. Table 2.2.2. Causes of B12 deficiency Table 2.2.4. Dietary Reference Intakes (DRIs) for females at various life stages Table 2.2.5 Comparison of B12 status biomarkers Table 2.2.6.2a. Summary of studies investigating B12 intakes and serum concentrations Table 2.2.6.2b. Adverse effects of periconceptional B12 supplementation on pregnancy and birth
outcomes Table 2.3.7. Summary of in utero and perinatal folate and B12 nutrient supplementation on DNA
methylation and related health and disease outcomes in the animal offspring Table 2.3.8. Summary of in utero and perinatal folate and B12 nutrient status on DNA
methylation and potentially related health outcomes in human offspring Table 2.4a. Maternal and cord blood concentrations of folate and B12 Table 2.4b. Genome-wide DNA methylation and hydroxymethylation and infant birth weight Table 4.1.1. Matching criteria for cases and controls Table 4.1.4.2. Genes-of-Interest and Housekeeping gene primer sequences for ddPCR Table 4.2.1.2a. Baseline maternal characteristics Table 4.2.1.2b. Maternal dietary intakes in early pregnancy Table 4.2.1.2c. Maternal blood concentrations in early pregnancy Table 4.2.1.2d. Newborn characteristics Table 4.2.2.1. Genome-wide methylation mean ß-value difference between cases and control for
the total, CpG island and non-CpG island based on a paired T-test and Wilcoxon Signed-Rank Sum test
Table 4.2.3.1a. Summary of the number of genes differentially methylated in cord blood MNCs relative to the control based on a T-test
Table 4.2.3.1b. Summary of the number of genes differentially methylated in cord blood MNCs relative to the control based on a Mann-Whitney U-test
Table 4.2.3.2a. The top molecular, cellular, physiological functions and disease processes for the 66 hyper- and hypomethylated genes
Table 4.2.3.2b. The top molecular, cellular, physiological functions and disease processes for the 219 hypomethylated and 162 hypermethylated genes
Table 4.2.4.1. Hypo- and hypermethylated genes selected for gene expression analysis Table 4.2.4.2. ddPCR gene expression analysis for candidate genes Table 4.4.4. Summary table of gene expression and functional analyses of hypo- and
hypermethylated genes
ix
List of Figures
Figure 2.1.1. Chemical structure of folate (a) and folic acid (b). Figure 2.1.2. Absorption and intracellular uptake of folates. Figure 2.1.3. Simplified diagram of the folate metabolism pathway. Figure 2.2.1. Chemical structures of vitamin B12. Figure 2.2.2. B12 absorption and transport Figure 2.2.3a. Biochemical functions of B12 (methylcobalamin) in folate/folic acid dependent one-
carbon metabolism. Figure 2.2.3b. Biochemical functions of B12 (5’deoxyadenosylcobalamin) in amino and fatty acid
metabolism. Figure 2.3.2 CpG island and gene body methylation in normal and tumor cells. Figure 2.3.2.1. DNA methylation regulation of transcription. Figure 2.3.3. Epigenetic programming. Figure 2.3.5 DOHaD. Figure 4.1.1a. Study Design. Figure 4.1.1b. Final sample selection and flowchart of the PREFORM participants. Figure 4.1.2.1. The HD Infinium assay for methylation. Figure 4.2.3.1a. Unsupervised hierarchical clustering and heat map of DNA methylation ß-value in
the case and control groups. Figure 4.2.3.1c. Principal component analysis plot of DNA methylation. Figure 4.2.3.2. Number of genes generated by T-test and MWU-test that are commonly
differentially methylated in cord blood MNCs. Figure 4.2.4.2a. Gene expression on candidate gene mean copy number. Figure 4.2.4.2b. Gene expression of high expressing hypo- and hypermethylated genes. Figure 4.2.4.2c. Gene expression of low expressing hypo- and hypermethylated genes. Figure 5.1. Summary of thesis conclusion.
x
List of Appendices
Appendix A. Analysis of blood biomarkers Appendix B. Final propensity score model in SAS code Appendix C. Histograms of CpG island and non-CpG island methylation Appendix D. IPA functional analysis summary table of all four datasets
xi
Abbreviations 5hmC, 5-hydroxymethylation 5mC, 5-methylcytosine 5-MTHF, 5-methyltetrahydrofolate AI, Adequate Intake APrON, Alberta Pregnancy Outcomes in Nutrition ChIP, chromatin immunoprecipitation CpG, cytosine-phosphate-guanine dinucleotide CRC, colorectal cancer ddPCR, digital droplet polymerase chain reaction DFE, dietary folate equivalent DHF, dihydrofolate DHFR, dihydrofolate reductase DMR, differentially methylated region DNMT, DNA methyltransferases DRI, dietary reference intake DOHaD, Developmental Origins of Health and Disease dUMP, deoxyuridine-5-monophosphate dUTP, deoxythmidine-5-monophosphate EAR, Estimated Average Requirement FFQ, food frequency questionnaire Hcy, homocysteine HDAC, histone deacetylase HoloTC, holotranscobalamin IUGR, intrauterine growth retardation LINE-1, long interspersed nucleotide element-1 MMA, methylmalonic acid MNC, mononuclear cell MS/MTR, methionine synthase / methyltransferase MSR/MTRR, methionine synthase reductase MTHFR, methylenetetrahydrofolate reductase NHANES, National Health and Nutrition Examination Survey NTD, neural tube defect PREFORM, PREnatal FOlic acid exposuRe on DNA Methylation in the newborn infant RBC, red blood cell RDA, Recommended Dietary Allowance SAH, S-adenosylhomocysteine SAM, S-adenosylmethionine THF, tetrahydrofolate TS, thymidylate synthase UMFA, unmetabolized folic acid UL, tolerable upper limit WCBA, women of childbearing age
1
1- Introduction
Folate and B12 play an important role in one-carbon metabolism which is involved in a variety of biological
methylation reactions including DNA methylation1,2. DNA methylation of cytosine-guanine (CpG) sites,
which can occur in clusters called ‘CpG islands’ located within the promoter regions of genes, inversely,
albeit with a few exceptions, often regulates gene transcription and expression. In contrast, CpG
methylation within the gene body also ensures genomic stability3. Aberrant patterns and dysregulation of
DNA methylation are mechanistically related to the pathogenesis of several human diseases including
cancer4. In vitro animal as well as human studies have shown that folate and B12 status can modulate DNA
methylation in a cell or tissue, and in a site and/or gene-specific manner through its effects on S-
adenosylmethionine (SAM), the universal methyl donor. However, results from these studies have not been
uniformly consistent5. DNA methylation is reprogrammed during the embryogenesis and is maintained
postnatally6. Metabolic and hormonal in utero environment has been shown to critically affect DNA
methylation programming of the fetus with consequent effects on disease susceptibility and health in
adulthood7.
Maternal diet during pregnancy has garnered significant attention due to its potential to modulate DNA
methylation4,8. Maternal status of folate and B12, either individually or in combination, have been shown to
be important in fetal development and growth. Suboptimal maternal status of these B vitamins during
pregnancy has been associated with adverse pregnancy and birth outcomes and disease susceptibility of the
offspring. Although optimal maternal folate and B12 status have been shown to be protective for certain
birth and pregnancy outcomes, excess folate and/or B12 status has been shown to be associated with adverse
health outcomes in the offspring 9.
2
The imbalance of folate and B12, that is a high folate/low B12 status, has been associated with more
pronounced adverse outcomes in the offspring. Since the levels of folate and B12 vary between studies, we
have set ≥1360 nmol/L as high RBC folate concentration and ≤150 pmol/L as a low B12 serum
concentration. The high folate cutoff is based on the 97th percentile of RBC folate concentrations using the
1999-2004 NHANES data10, and the low B12 cutoff is a commonly considered a deficiency cutoff11. Two
prospective observational studies in India have reported that high folate and low B12, intakes during
pregnancy are associated with small for gestational age babies12 and blood concentrations with insulin
resistance and adiposity in the offspring at six years of age13. However, mechanisms by which high folate and
low B12 status during pregnancy affect infant health outcomes have not been elucidated yet. We posit that
high folate and low B12 status during pregnancy modulates DNA methylation of critical genes involved in
growth and development and predispose the offspring to metabolic syndrome and obesity.
Hence, my overall research question is: can maternal folate and B12 status have an epigenetic effect in the
offspring, which may lead to functional, biological and clinical ramifications related to fetal growth and
development? I hypothesize that maternal folate and B12 nutrient exposure will directly affect DNA
methylation in genes associated with fetal growth and development.
3
2- Literature Review
2.1 Folate/Folic Acid
2.1.1 Definition, Chemical Structure and Dietary Sources
Folate is a water-soluble B vitamin (B9) in its naturally occurring form, while folic acid is synthetically
manufactured5. Folate encompasses naturally occurring dietary folates and all biochemically related
compounds including folic acid14.The common chemical structure of folate and folic acid are the 2-amino-4-
hydroxy-pteridine moiety (pterin ring) attached to p-aminobenzoic acid (PABA) by a methylene bridge and
glutamate residues attached by g-peptide bonds (Figure 2.1.1). Dietary folates (pteroylpolyglutamates) are
characterized as having polyglutamylated tails and reduced pterin ring, while folic acid
(pteroylmonoglutamic acid) contain monoglutamylated tails and an oxidized ring (Figure 2.1.1)5.
Figure 2.1.1. Chemical structure of folate (a) and folic acid (b). The common features entail a pterin (pteridine) ring attached to PABA by a methylene bridge with glutamate residues attached by g-peptide bonds. The pterin ring of dietary folate are reduced and contain more than one glutamate residue (polyglutamylated), while in folic acid the pterin ring is fully oxidized at the N5 and N10 positions and contain one glutamate residue (monoglutamylated). Reprinted with permission from Warzyszynska 20145.
4
The main naturally occurring sources of dietary folate include dark green vegetables, beans, legumes, citrus
fruits and organ meats1,14. While folate can be synthesized by colonic bacteria15 and absorbed across the
large intestine16,17, mammals cannot synthesize enough folate to maintain biological functions and must
therefore obtain additional folate from their diet18. Its bioavailability of naturally occurring folate in the diet
varies by food source and preparation14. Approximately 50-75% of naturally occurring folate is lost due to
food harvesting, storage, processing and preparation14. Furthermore, folate activity is lost at low pH levels
as it is oxidized1. The fully oxidized and monoglutamylated structure of folic acid is both stable and
bioavailable, making it suitable for supplements and fortified foods5. Bioavailability of folic acid when taken
with food and on an empty stomach is approximately 85% and 100%, respectively14.
2.1.2 Absorption and Metabolism
Folates are absorbed primarily in the duodenum and jejunum of the small intestine19 as well as across the
colon20,21 based on the presence of reduced folate carrier (RFC) and proton coupled folate transport
(PCFT). Active transport is required for monoglutamyl folate, but in concentrations of >10 µmol/L folate
is absorbed by passive diffusion in a nonsaturable manner. Additionally, folic acid supplementationcan
saturate and overwhelm the capacity of dihydrofolate reductase (DHFR) to reduce and methylate folic acid
to 5-methyltetrahydrofolate (5-MTHF)22. It is not uncommon to attain high levels of folic acid consumption
owing to supplementation and current fortification practices. Hence, appreciable amounts of unmetabolized
folic acid (UMFA) have been reported to appear in circulation23–26.
Folate in its polyglutamylated form cannot enter the enterocyte because it contains more than three
glutamate residues27,28. Thus polyglutamylated folate requires glutamate carboxypeptidase II (GCPII), an
exopeptidase anchored to the intestinal apical brush border membrane, to cleave glutamate chains until
folate is in its monoglutamylated form27,28.
5
Transportation of folate into the cell membrane is mediated by two main classes of systems: transmembrane
carriers and folate-binding protein-mediated systems. Types of transmembrane carriers are the RFC, PCFT,
a family of low-affinity membrane carrier multidrug resistance-associated proteins (MRPs), and
mitochondrial and lysosomal folate transport. Types of folate-binding protein-mediated systems include
three high-affinity folate receptors (FR) (a, b, and g) (Figure 2.1.2).
PCFT and RFC are highly expressed in the small intestine, and are the main transport carriers for
monoglutamyl folate. PCFT exhibits equal affinity for both reduced and oxidized forms of folate, and a
higher affinity for 5-MTHF at a low pH of <6.5 that is consistent with the microenvironment of the small
intestine19,29. RFC transports reduced folate at optimal physiological pH (pH > 6.5)29. It is ineffective for
transporting folic acid29. MRPs have low affinity for folate but have a high capacity for exporting folate out
of tissues29,30. FRs exhibit affinity for all forms of folate; however, their affinity is highest for folic acid29.
Both FR a and b are expressed in the gut mucosa of fetal tissue and a number of adult tissues in low levels.
FRs are found on the apical membrane of the enterocyte and thus function to transport folate into the cell.
Unlike transmembrane carriers like PCFT and RFC in the apical and basolateral membranes, FRs do not
transport folate in both directions31.
5-MTHF is the main form of folate found in portal circulation and is taken up by the liver. 5-MTHF is
metabolized to polyglutamylated derivatives in the liver where it is either retained for storage or converted
back to the monoglutamate form before being released into the blood or bile. Once monoglutamylated
folate enters the cell via the aforementioned transport systems, folate is converted back to its
polyglutamylated form via folypolyglutamate synthase (FPGS) in order to retain folate in the cell.
Polyglutamylated folates are better substrates for one-carbon metabolic reactions32. Before leaving the cell
6
via RFC and PCFT, terminal glutamate residues of folypolyglutamates are cleaved to create monoglutamyl
folate by g-glutamyl hydrolase (GGH), which contributes to the maintenance of intracellular folate
homeostasis.
Figure 2.1.2. Absorption and intracellular uptake of folates. Before entering the enterocyte of the small and large intestines, polyglutamated folates require cleavage by GCPII to one glutamate residue while monoglutamyl folate is actively transported by PCFT, RFC and FR. Once inside the cell, FPGS adds glutamate residues to retain polyglutamylated folate for intracellular one-carbon transfer reactions. For folate export out of the cell, GGH cleaves glutamate residues from polyglumatylated folate to a monoglutamylated form, which can then be transported out of the cell by RFC, PCFT and MRPs. Folate transporters have varying functions depending on the location and intraluminal and intracellular pH. [GCPII, glutamate carboxypeptidase II; PCFT, proton-coupled folate transporter; RFC, reduced folate carrier; FR, folate receptor; FPGS, folypolyglutamyl synthase; MRP, multidrug resistance-associated proteins.] Adapted from Kim 200728 and reprinted with permission from Plumptre 201633.
7
2.1.3 Biochemical Functions
Folate functions primarily to transfer one-carbon units involved in nucleotide biosynthesis, the
remethylation of homocysteine (Hcy) as part of the methionine cycle and biological methylation
reactions28,29. In de novo nucleotide biosynthesis, folate is a cofactor that is essential for DNA synthesis,
stability, integrity and repair29,34. In addition to methylation of DNA (for gene expression regulation and
gene stability), folate is also involved in methylation of proteins (for post-translational modifications) and
lipids (for synthesis)35.
Folate-mediated one-carbon metabolism is generally believed to be compartmentalized occurring at specific
areas intracellularly (cytoplasm, mitochondria, nucleus)36. Cytosolic folates such as tetrahydrofolate (THF)
and 5-MTHF are mainly associated with purine and thymidylate synthesis and biological methylation
reactions. Mitochondrial folates such as THF and 10-formylTHF are primarily related with the
serine/glycine cycle2. Nuclear folates are shown to mediate thymidylate synthesis, but are also involved in
the serine/glycine cycle36.
For naturally derived folates that have entered the cell, these folates can directly participate in one-carbon
metabolism34,37,38. This is because it exists mainly in the forms of 10-formlyTHF and more predominantly 5-
MTHF34. However, for folic acid, it must be reduced to dihydrofolate (DHF) then to THF by DHFR, and
methylated to 5-MTHF. This occurs primarily in the liver and to a lesser extent in the small intestine22. Due
to the limited enzymatic capacity of DHFR to reduce folate, high intakes of folic acid result in the
appearance of unaltered or UMFA in circulation39. This in turn has been shown to compete with active
forms of folate for folate transporters, binding sites, and folate-related enzymes34,37,38.
8
5-MTHF is critical in both DNA synthesis and DNA methylation (Figure 2.1.3). 5-MTHF is synthesized
from 5,10-methyleneTHF in a nonreversible reaction by methylenetretrahydrofolate reductase (MTHFR)40.
5,10-methyleneTHF donates methyl groups for de novo pyrimidine synthesis through a catalyzed reaction by
thymidylate synthase (TS) from deoxyuridine-5-monophosphate (dUMP) to deoxythmidine-5-
monophosphate (dUTP or thymidylate). As a result of the nonreversible methylation reaction by TS, 5,10-
methyleneTHF is oxidized to DHF and thymidylate can be used for DNA synthesis. Subsequently, THF is
produced from DHF by DHFR and can be converted to 5,10-methyleneTHF by vitamin B6-dependent
serine hydroxymethyltransferase (SHMT), and then to 5-MTHF by MTHFR. 5,10-methyleneTHF can also
be oxidized to 10-formylTHF for de novo purine synthesis (Figure 2.1.3).
For DNA methylation reactions, 5-MTHF transfers a single methyl group to Hcy catalyzed by vitamin B12-
dependent methionine synthase (MS) to generate methionine41,42. Methionine is converted to SAM through
the action of adenosine triphosphate (ATP) and methionine adenosyltransferases (MAT1A and MAT2A).
DNA methyltransferases 1, 3a and 3b (DNMTs) then transfers a methyl group to DNA from SAM43,44.
SAM is a universal methyl donor for over 100 biological methylation reactions45. During this process, S-
adenosylhomocysteine (SAH) is produced and converted to Hcy by SAH hydrolase and allows the
methylation cycle to start again46. Another mechanism to remethylate Hcy is through a folate-independent
pathway involving betaine or choline. After the methyl group is donated from 5-MTHF, again THF is
converted to 5,10-methylene THF by SHMT. 5,10-methylene THF is necessary throughout the one-carbon
metabolism pathway and can be used to regenerate methionine or directed towards nucleotide biosynthesis
(Figure 2.1.3).
9
Figure 2.1.3. Simplified diagram of the folate metabolism pathway. GCPII, glutamate carboxypeptidase II; FR, folate receptor; PCFT, proton-couples folate transporter; RFC, reduced folate carrier; FPGS, folylpolyglutamate synthase; GGH, g-glutamyl hydrolase; DHF, dihydrofolate; THF, tetrahydrofolate; DHFR, dihydrofolate reductase; SHMT, serine hydroxymethyltransferase; TS, thymidylate synthase; BHMT betaine-homocysteine S-methyltransferase; MS, methionine synthase, MSR/MTRR, methionine synthase reductase; MAT 1a, 1b, methionine adenosyltransferase; SAHH, SAH hydrolase; MTHFR, methylenetetrahydrofolate reductase; DNMT1, 3a, 3b, DNA methyltransferases; MBD2, methyl-CpG binding domain protein 2; SAM, S-adenosyl methionine; SAH, S-adenosyl homocysteine; CßS, cystathionine-ß-syntase; CTH, g-cystathionase; CpG, cytosine-guanine dinucleotide sequence; dUMP, deoxyuridine-5-monophosphate; dTMP, deoxythymidine-5-monophosphate; FAD, flavin adenine dinucleotide; FADH2, 1,5-dihydro-flavin adenine dinucleotide. Each filled triangle represents a glutamate, which is linked via a peptide bone to form polyglutamated folate. Each filled circle represents reduced or oxidized folates or folic acid. Adapted and modified with permission from John Wiley and Sons44.
10
2.1.4 Folate Intake Requirements
Dietary folate intake is expressed as dietary folate equivalents (DFEs) in order to account for the lower
bioavailability of natural folate compared to folic acid. Specifically, 1 µg DFE = 1 µg of food folate = 0.6 µg
of folic acid from fortified foods or supplements taken with meals = 0.5 µg of a supplement taken on an
empty stomach47. Hence, DFEs can be calculated by the following equation14:
µg DFE = x µg dietary folate + 1.7 (x µg folic acid)
Dietary Reference Intakes (DRIs) established by the Institute of Medicine14 are scientifically evidence-based
reference values used for dietary intake assessment and health planning for populations and individuals. The
DRIs vary by sex, age and have separate requirements for particular populations such as pregnant and
lactating subpopulations. DRIs consist of four measures: Estimated Average Requirement (EAR), the
Recommended Dietary Allowance (RDA), the Tolerable Upper Limit (UL), and the Adequate Intake (AI).
Briefly, the EAR is the daily nutrient intake value that is estimated to meet the requirements of 50% of the
population. The RDA is the average daily intake required to meet the requirement of approximately 97-
98% of the population, and it can be calculated by using the EAR + 2 standard deviations. If the EAR is not
available due to insufficient data, the RDA can be calculated using the EAR x 1.2 as this assumes 10%
coefficient of variation for the EAR, and is therefore equal to 120% of the EAR. When not enough
information is available to calculate an EAR and RDA, the AI is used as an average daily recommendation
based on observed or experimental estimates of nutrient intakes from a healthy population48. The UL is the
highest level of daily nutrient intake for individuals without risk of adverse health effects. For most
nutrients, as the intake increases above the UL, the risk of adverse health effects increases14.
11
DRIs for folate are based on observational and experimental studies. The EAR and RDA for folate for both
males and females 19 years old and above are 320 and 400 µg/d DFEs, respectively. For pregnant and
lactating women, the EAR and RDA are increased to 520 µg/d and 600 µg/d DFE, respectively49 (Table
2.1.4). An additional 400 µg/d of folic acid either in the form of supplements or fortified foods is
recommended 2-3 months prior to conception and during pregnancy to reduce the risk of neural tube
defects1. For pregnant women, the increase in requirement is due to multiple fetal and maternal
physiological changes that increase metabolic demands for folate. Ensuring adequate folate intake is critical
for maintaining one-carbon transfer reactions including DNA synthesis and repair and biological
methylation reactions necessary for optimal maternal health and fetal growth and development. For
lactating women, the increase in requirement is necessary to replace the amount of folate secreted daily in
the breast milk, thereby ensuring optimal folate status50.
There is no UL for naturally occurring folate, but the UL is set at 1000 µg/d for folic acid for adults 19
years and older and pregnant and lactating women (Table 2.1.4). Although studies are limited, the UL for
folic acid was set at this level based on suggestive evidence that excessive amounts of folic acid may mask B12
deficiency, thereby allowing irreversible neurological damage to continue to progress. This phenomenon
where high folate masks B12 deficiency is called the methyl folate trap or ‘methyl trap’ 51. In this scenario,
folate is trapped in the form of 5-MTHF due to the constant and irreversible conversion of 5,10-
methyleneTHF to 5-MTHF by MTHFR. With low B12, MS activity is reduced and subsequently leads to a
decrease in remethylation of methionine from Hcy. The remethylation of methionine works by the transfer
of a methyl group from 5-MTHF to homocysteine to form THF (Figure 2.1.3). During this process, the
regeneration of methionine from homocysteine by MS is also diminished, resulting in reduced methylation
activity52.In addition to masking B12 deficiency, excess folate has been implicated in resistance to antifolate
drugs used in cancer treatment and inflammatory disorders and reduced natural killer cell cytotoxicity5,53–55.
12
Table 2.1.4. Dietary Reference Intakes (DRIs) for females at various life stages
Age and life stage EAR (µg/d DFEs)
RDA (µg/d DFEs)
UL (µg/d folic acid)
Female, 19y+ 320 400 1000 Female, pregnant 520 600 1000 Female, lactating 450 500 1000
DFE, dietary folate equivalents; EAR, estimated average requirement; RDA, recommended dietary allowance; UL, tolerable upper limit. 2.1.5 Biomarkers of Folate
2.1.5.1 Folate Status
Serum or plasma is a short-term indicator of folate status as it represents recent dietary intake, while red
blood cell (RBC) folate concentrations reflect tissue stores56. Together, they are used to measure systemic
folate status. Plasma folates are monoglutamylated folates with variations in oxidation state and one-carbon
substitutions. 5-MTHF predominates in the serum/plasma, generally accounting for >80% of total plasma
folates57. Plasma also contains UMFA and non-5-methylated reduced folates usually in small concentrations.
RBC folates have higher folate concentrations than in the serum or plasma and are mostly 5-MTHF
polyglutamates, except in persons with MTHFR677T polymorphism where a proportion of 5-MTHF is
converted to formyl-folates46. Measurement is also more complex than with serum folate as polyglutamates
need conversion to monoglutamates before analysis. RBC folate measurement is not influenced by recent
folate intake as RBCs have a lifespan of approximately 120 days, and is therefore indicative of long term (~3
months) folate intake and tissue stores. It is for this reason that RBC folate measurement is considered the
gold standard for determining folate status46,58.
Plasma homocysteine is a functional indicator of folate status; however, it is not a specific marker as it is
also influenced by the status of vitamins B2, B6 and B1246. Although not a clinical biomarker of folate status,
the measurement of UMFA in serum or plasma is becoming increasingly important since the detection of
13
UMFA reflects high intake levels of folic acid. UMFA is found to be increasingly prevalent in blood samples
of many populations, and is not limited to countries with mandatory folic acid fortification59. Observational
studies have found detectable levels of UMFA in children, adolescents, and adults57,60, older adults61,
pregnant women and umbilical cord blood23 and in 4-day old infants62.
2.1.5.2 Analysis of Folate in Blood and Other Biological Fluids
There are three main methods for analysing folates in serum, plasma and RBCs, and other biological fluids.
They are accomplished by the microbiological, protein-binding and chromatographic assays as shown in
Table 2.1.5.2. It is well-known that these aforementioned folate assays produce varying concentrations
when measuring RBC folate46,58,63,64 and are often not readily comparable64.
Table 2.1.5.2. Comparison of folate measurement assays
Assay Assay Principle Strengths Weaknesses Microbiological Growth of bacteria
(L.rhamnosus) is proportional to amount of folate in sample
-Gold standard of overall folate status -Very sensitive -Inexpensive -Measures all biologically active forms equally
-Does not distinguish between individual folate forms -Low precision -Can be affected by antibiotics
Protein-binding Folate binding proteins “extract” folate from samples, can be radio- or nonradio-labelled
-Quick results -High-throughput analysis -Good precision -Low cost
-Narrow detection range -Binding proteins respond differently depending on forms of folate in sample
Chromatographic Individual forms of folate are separated and quantified based on their interaction with the adsorbent material. May be quantified by measuring the mass to charge ratio.
-Can measure individual folate forms -Highly selective and sensitive
-Extensive sample cleaning -Expensive specialized equipment -Trained personnel -Interconversions of folate forms needed for correct interpretation of data
Table modified from Plumptre, 2016 thesis33.
14
2.1.5.3 Common Cutoffs Used to Define Deficient and High Folate Status
The only formally accepted RBC cutoff is for folate deficiency, and it is defined as having a RBC folate
concentration less than 305 nmol/L (140 ng/mL), which is the level at which hematological (e.g.
macrocytic anemia) and metabolic (e.g. rise in homocysteine concentrations) abnormalities begin to
appear14,64. For serum folate, a concentration less than 7 nmol/L (3 ng/mL) is also an indicator of folate
deficiency; however, this is less commonly used as serum folate is sensitive to recent intakes14.
Due to increasing folate status arising from fortification and supplement use, several cutoffs for high folate
status have been suggested; however, no formally accepted cutoffs exist to date. A RBC folate
concentration of 1360 nmol/L using Bio-Rad immunoassay is considered high based on the 97th percentile
of RBC folate concentrations using the 1999-2004 National Health and Nutrition Examination Survey
(NHANES) data10. Using the 2007-2009 Canadian Health Measures Survey (CHMS), three upper RBC
folate concentrations cutoffs have been to facilitate future research examining the impact of high folate
status on health outcomes – 1450, 1800, and 2150 nmol/L65. These were estimated from post-fortification
2005-2010 NHANES data in increments of 305 nmol/L (adjusted from the Immulite 2000 immunoassay to
microbiologic assay). Based on the CHMS data, 16%, 6% and 2% of sampled Canadian population were at
or above the 1450, 1800, and 2150 nmol/L, respectively65.
I. Summary of Cutoffs
Low cutoffs have been more consistent as they are mainly based on deficiency levels (Table 2.1.5.3). A few
proposed high cutoffs have been proposed based on epidemiological studies. The ‘low’ cutoffs are 305, 906
and 1000 nmol/L. The 305 nmol/L is the most widely accepted cutoff for folate deficiency, below which
hematological (e.g. macrocytic anemia) and metabolic (e.g. an increase in plasma homocysteine
concentrations) indicators begin to appear14,64. For serum folate, a concentration less than 7 nmol/L (3
15
ng/ml) is also an indicator of folate deficiency; however, this is a less reliable marker as serum folate is
sensitive to recent folate intakes14. A cutoff of at least 906 nmol/L RBC folate concentration is
conventionally used as the level of maximal protection to prevent NTDs. A large Irish case-control study
(n=56,000 pregnant women) found that the risk for NTD decreased by eight-fold in those with RBC folate
concentrations >906 nmol/L, compared those with RBC folate concentrations <305 nmol/L66. This study
also reported a dose-response relationship between RBC folate concentrations and the risk of NTD-affected
pregnancy66, where the risk of NTD begins to decrease at 906 nmol/L and continues to do so until it
reaches a plateau at 1292 nmol/L66. Although the 906 nmol/L is the cutoff conventionally used, two large
studies in China (total n=249,045) demonstrated that a cutoff above 1000 nmol/L substantially attenuated
risk for NTD67. Applying this cutoff via logistic model regression to various large-scale population data like
the NHANES (2005 � 2010), the 1000 nmol/L cutoff further reduced the risk for NTD to a level below
moderate risk (<8 NTDs per 10,000 births) 67. Hence, a cutoff above 1000 nmol/L is proposed cutoff is
suggested to be an optimal RBC folate concentration for NTD prevention. The association of ‘high’ RBC
folate cutoffs with adverse health outcomes has not been established. Nevertheless, these cutoffs have been
proposed to define high folate status68 (Table 2.1.5.3). The 1090 nmol/L cutoff is based on RBC folate
concentrations from individuals consuming 2.1 mg DFEs or of the combined intake of 400 µg DFEs (RDA
excluding for pregnant and lactating women) and 1000 µg folic acid supplement69. The 1360 nmol/L cutoff
represents the 97th percentile according to the 1999-2004 National Health and Nutrition Examination
Survey (NHANES) data, using Bio-Rad immunoassay10. Based on a more recent 2005 - 2010 NHANES, the
1820, 2150, 2490 nmol/L cutoffs represent the 90th, 95th and 97.5th percentile of the 2005-2010 NHANES
cutoffs, adjusted to microbiologic assay70. In the 2007-2009 CHMS, three upper RBC folate concentrations
in increments of 305 nmol/L were selected and estimated from the 2005-2010 NHANES cutoffs � 1450,
1800, and 2150 nmol/L65.
16
Table 2.1.5.3. A summary of folate cutoffs
RBC folate cutoff (nmol/L) Rationale Citation
Low: 305 Considered for folate deficiency; hematological (e.g.
macrocytic anemia) and metabolic indicators (e.g. increasing plasma homocysteine concentrations).
IOM 1998
906 Conventionally used as the level of maximal protection to prevent NTDs based on a large Irish case-control
study.
Daly et al. 1995
1000 A more recent measure based on two large and recent studies in China.
Crider et al. 2014
High: 1090 Based on the relationship between RBC folate and
intakes in DFEs. MacFarlane et al. 2011
1360 97th percentile of the 1999 � 2004 NHANES Colapinto et al. 2011 1820, 2140, 2490 90th, 95th and 97.5th percentile of the 2005 � 2010
NHANES Pfeiffer et al. 2012
1450, 1800, 2150 Estimated from 2005 � 2010 NHANES in increments of 305 nmol/L
Colapinto et al. 2015
17
2.1.6 Folate and Health
2.1.6.1 Folate in Health and Disease
Folate has been shown to affect human health and disease through its role in purine and thymidylate
synthesis and biological methylation reactions. These one-carbon transfer reactions play a critical role in
nucleotide biosynthesis and repair, genomic stability, and regulation of gene expression5,14. Adverse health
outcomes associated with both inadequate and excess intakes of folate have been summarized in Table
2.1.6.1.
Table 2.1.6.1. Health outcomes associated with inadequate and excess folate
I. Inadequate folate II. Excess folate Convincing • Megaloblastic anemia • Masking B12 deficiency
Probable • Coronary heart disease
• Stroke • Cancer initiation • Cognitive impairment
• Cancer progression
Possible/Insufficient
• Neuropsychiatric disorders • Other congenital disorders
• Antifolate drug resistance • Decreased natural killer cell
cytotoxicity • Increased risk of obesity,
insulin resistance in offspring
Table modified from Plumptre 2016 thesis33.
I. Folate Deficiency
Evidence to support the most purported adverse health outcomes associated with folate deficiency except
for megaloblastic anemia and NTDs14 is not consistent and is tenuous in the literature5,14,45,68.
Megaloblastic Anemia
Megaloblastic anemia is a hematological indicator of folate deficiency14. It has been characterized by the
appearance of hypersegmented neutrophils and has been associated with low RBC folate concentrations.
18
Clinically, common manifestations of megaloblastic anemia start with pallor of the skin and can progress to
fatigue, weakness, shortness of breath and palpitations71. The evidence supporting the link between folate
deficiency and megaloblastic anemia has been consistent14, and has thus been used as a basis for establishing
national folate deficiency cutoffs in Canada and the US64.
Probable
Coronary Heart Disease and Stroke
Epidemiological and clinical studies have investigated the relationship between high homocysteine levels and
the risk of coronary heart disease and stroke; however, the evidence is inconsistent and equivocal72. The
Homocysteine Studies Collaboration that pooled 30 prospective and retrospective studies reported that for
every 25% reduction in homocysteine concentration (approximately 3µmol/L), an 11% decreased risk of
coronary heart disease and a 19% decreased risk of stroke were observed72,73. In contrast to the results from
observational epidemiologic studies, most placebo-controlled randomized clinical trials and their systematic
reviews and meta-analyses have reported a null effect of folic acid supplementation with or without other B
vitamins on coronary heart disease risk 4,74–78. However, these studies have suggested that folic acid
supplementation may exert a protective effect on stroke risk78–81.
Cancer Initiation and Progression
The role of folate in cancer development and progression has been highly controversial. A large body of
epidemiologic studies collectively suggest an inverse association between suboptimal folate status and the
risk of several human malignancies including cancer of the lungs, oropharynx, esophagus, stomach,
colorectum, pancreas, cervix, endometrium, ovary, bladder, prostate, skin, and breast as well as
neuroblastoma and leukaemia82–84 However, the precise nature and magnitude of this purported relationship
have not been unequivocally established and may be influence by the baseline folate status of the population
19
studied. The role of folate in carcinogenesis is best studied for colorectal cancer (CRC)85. Collectively
observational epidemiologic studies and their systematic reviews and meta-analyses have indicated a 20 –
40% reduction in the risk of CRC in subjects with the most adequate folate status (assessed by dietary folate
intake or blood measurement of folate biomarkers) compared to those with the lowest28,82,83,86. However,
several placebo-controlled randomized clinical trials have reported an increase87, decrease88, or null89,90
effect of folic acid supplementation on the risk of CRC or recurrent adenomas (a well-established
precursor).
Animal studies have suggested dual modulatory effects of folate on the initiation and progression of CRC
depending on the dose and stage of cell transformation at the time of folate intervention28. Folate deficiency
has been shown to increase, while modest levels folic acid supplementation reduces the risk of neoplastic
transformation in normal tissues. However, it appears that excessive amounts of folic acid supplementation
can paradoxically increase the risk of neoplastic transformation in normal tissues. Conversely, in
preneoplastic and neoplastic lesions, folate deficiency has been shown to suppress, whereas folic acid
supplementation promotes the progression of these lesions28. Although the mechanisms to explain folate’s
dual effects on CRC initiation and progression have yet to be clearly elucidated, several plausible
mechanisms relating to nucleotide biosynthesis and DNA methylation have been proposed and extensively
studied85.
Neuropsychiatric Disorders & Cognitive Impairment
Although not uniformly consistent, epidemiologic studies have suggested an inverse association between
folate status and the risk of cognitive impairment, dementia and depression91. Furthermore, randomized
clinical trials have suggested a possible protective effect of folic acid supplementation with and without
other B vitamins on cognitive decline92.
20
Other Congenital Disorders
In contrast to the well-established inverse relationship between maternal folate status and the risk of NTDs,
the relationship between maternal folate status and the risk of other congenital defects has been equivocal
90,92. Several randomized clinical trials90,93,94 and observational studies92 have found a decrease in risk of
congenital defects with multivitamin supplementation containing folic acid.
II. Excess Folate
Due to the increased folate status resulting from folic acid supplementation, in addition to voluntary and
mandatory folic acid fortification in many countries, there has been increasing interest in potential adverse
health outcomes associated with excess folate and folic acid.
B12 Masking
B12 masking is a scenario of excess folate and deficient B12 status. With low B12, MS activity is reduced
resulting in reduced remethylation of homocysteine to produce methionine95. Folic acid supplementation
negates the adverse hematological effects (i.e. megaloblastic anemia) of low B12, thus masking and delaying
diagnosis of B12 deficiencyf. As a result, there remains a risk for the progression of B12 deficiency-associated
neurological conditions (e.g., subacute combined degeneration). Furthermore, B12 has been important for
maintaining cognitive function in the aging population96,97, and B12 masking has been shown to increase the
risk of cognitive impairment in the elderly95,98,99.
Cancer Progression
The effect of excess folate on cancer progression has been generally inconsistent and conflicting. Animal
studies suggest a dual modulatory role of folate in cancer progression. Although folic acid supplementation
may protect against neoplastic transformation in normal tissues, it appears to promote the progression of
21
established (pre)neoplastic lesions28. Several biologically plausible mechanisms relating to the folate’s role in
one-carbon transfer reactions exist to explain the tumor promoting effect of folic acid supplementation.
Most likely, folic acid supplementation provides nucleotide precursors to rapidly replicating cells such as
neoplastic cells for accelerated progression. Gene silencing of tumor suppressor or mismatch repair genes
via CpG island promoter DNA methylation has been proposed as another plausible mechanism for the
tumor promoting effect associated folic acid supplementation43. In large-scale placebo-controlled
randomized clinical trials that investigate folic acid supplementation and risk of recurrent adenomas as the
primary endpoint; an increase87, decrease88, or a null effect was observed89,90,100 (121,129,130). The
Aspirin/Folate Polyp Prevention Study associated folic acid supplementation with a 67% increased risk of
recurrent advanced adenomas with a high malignant potential in individuals with prior colorectal
adenomas87. In the 12-year French SUVIMAX prospective study (n=12741), dietary folate intake and RBC
folate concentration was associated with an increased risk of overall skin cancer101. Clinical trials with
cancer incidence or mortality as a secondary endpoint have shown either a tumour promoting or null effect
on cancer outcomes with folic acid supplementation5,78,102. A Norwegian study that investigated the effect of
folic acid (800 µg/d) and B vitamins (B12 400 µg/d) supplementation on ischemic heart disease as the
primary outcome found an overall increase in cancer incidence by 21% and mortality by 38% over a 36-
month period102. In contrast, a meta-analysis of eight randomized control trials of folic acid
supplementation on cardiovascular disease outcomes as the primary endpoint found no significant effects on
overall cancer incidence and mortality over a 5-year period78. Similarly, a meta-analysis of 13 trials (studies
investigating folic acid supplementation with cardiovascular risk as the primary endpoint and CRC risk in
patients with previous adenomas) found no overall effect of folic acid supplementation on CRC risk.
Nevertheless, a non-significant trend toward an increased risk of CRC was observed in the 3 trials involving
patients with previous adenomas and in all 13 trials103. Two ecologic studies on the trend of CRC incidence
post-folic acid fortification era found increased CRC rates in the US, Canada and Chile, suggesting that folic
22
acid fortification may have been responsible for this observation104,105. On the other hand, two large US
prospective studies conducted in the postfortification era found that a moderate consumption of folate
reduced the risk of CRC, and no tumor promoting effect was associated with folic acid supplementation or
fortification106,107.
Resistant to Antifolate Drugs
Antifolate drugs are used against arthritis108,109, inflammatory conditions, and cancer110–112 as they disrupt
intracellular folate metabolism and related one-carbon metabolism. In an environment with excess folate,
the potency or efficacy of antifolate drugs may be diminished and a resistance to these drugs may
emerge45,110,111.
Immune System
Excess folate (mostly supplemental folic acid) has also been linked with diminished natural killer (NK) cell
cytotoxicity. NK cells are immune cells with important defence mechanisms against viral infections and
cancer cells113. A study in postmenopausal women found that a high folate diet and folic acid
supplementation were associated with reduced NK cell cytotoxicity in comparison to the control114.
Women with low combined dietary folate and folic acid supplementation were also reported to have 23%
lower NK cell cytotoxicity114. This inverse U-shaped curve was more pronounced among women over 60
years old. Furthermore, in the 78% of women with detected UMFA, their NK cell cytotoxicity was 6.2%
lower compared with women without plasma FA present114. Similarly, a recent non-controlled intervention
(n=30) in Brazil observed a reduction in cytotoxic capacity and number of NK cells with 5000 µg folic acid
supplementation and increased UMFA concentrations after 45 and 90 days55. In an animal model, female
mice were fed a 20 x RDA folic acid diet for 3 months were found to have lower NK cytotoxicity compared
to control mice (1 x RDA folic acid)54.
23
Insulin Resistance and Obesity
Several studies have investigated prenatal folic acid supplementation with the risk of adiposity and insulin
resistance in the offspring. Animal studies have shown more consistent effects on these outcomes in
offspring from mothers supplemented with folic acid alone, whereas findings are conflicting in studies
where folic acid supplementation was used in combination with other B vitamins, including B12 (See
Section 2.2.6.2). Two rodent studies have shown that pups born to folic acid supplemented dams have a
higher birth weight115,116 and are more insulin resistant116. In a sheep model, periconceptional folate, B12 and
methionine that was restricted in dams and ova from these mothers, were fertilized in vitro and transferred
to surrogate dams with normal methionine status117. The offspring, especially males, were found to be
obese and insulin resistant compared to males from ewes in the non-restricted diet117. In humans, the Pune
Maternal Nutrition Study from India was the first to find an increased risk of insulin resistance and obesity
in children at 6 years of age associated with high maternal folate and low B12 status13 at mid-pregnancy (18
weeks gestation). A study in Myosore, Southern India reported similar findings to the Pune study in that
higher maternal folate status was associated with higher offspring insulin resistance children at 9.5 and 13.5
years old118. However, unlike the Pune study, the Myosore study did not find an interaction between
maternal folate and B12 status. Another interesting finding of the Myosore study was that children at 5 years
old had fasting glucose concentrations associated positively with folate and negatively with B12 blood
concentrations118. Since the Myosore study sampled mothers in late gestation (30 weeks), the window for
detecting the effects of B12 on insulin resistance may have passed, and this may be a reason for the observed
lack of interaction between folate and B12118. Further analysis from the Pune study has revealed that late
gestation (28 weeks) folate concentrations and mid-gestation (18 weeks) B12 concentrations are the
strongest predictor of insulin resistance in the offspring13.
24
In summary, both folate deficiency and excess have been shown to affect human health. Due to the
significantly increased folate status resulting from folic acid fortification and supplemental practices and
virtually non-existent folate deficiency in the North American population, there has been a shift in research
focus on investigating the effects of a high folate status on health and disease outcomes.
2.1.6.2 Folic Acid Fortification
As mentioned previously in Section 2.1.6.1, there are several adverse health outcomes related to folate
deficiency including the increased risk of NTDs. A large body of evidence has consistently suggested a
decrease in NTD risk with periconceptional use of folic acid93,119–121. It is primarily for this reason that the
Canadian and US governments mandated folic acid fortification in 1998 and recommended women at
reproductive age to supplement with folic acid122,123. The addition of 150 µg folic acid/100 g white wheat
flour and 200 µg folic acid/100 g cornmeal and enriched pasta (140 µg folic acid/100 g for all three grains
in the US) led to an approximately 50% decrease in NTD prevalence in Canada and the US, deeming folic
acid fortification a public health policy success124–127.
Since then, several studies have suggested an over-fortification of the food supply beyond the recommended
100 – 200 µg128, and that even these high levels are thought to be an underestimate of the actual amount of
folic acid in the fortified products129. Epidemiological studies have suggested that mandatory folic acid
fortification has significantly contributed to the increase in serum and RBC folate concentrations evident in
the North American population post folic acid fortification126,128. Comparison of blood folate measures from
NHANES pre (1988-1994) and post (1999-2004) folic acid fortification indicate a decline in the prevalence
of folate deficiency by 21% as evaluated by serum folate ( <1% for serum folate <6.8 nmol/L) and 38% as
evaluated RBC folate (5% for RBC folate <305 nmol/L) concentrations10. An increase in serum and RBC
folate concentrations by 119-161% and 44-64%, respectively, was also observed. This translates to an
25
increase of RBC concentrations from 398 to 636 nmol/L 10. More recently, the Centre for Disease Control
(CDC) and Prevention’s Second Nutrition Report for 2003-2006 indicated that less than 1% of the US
population was folate deficient130. Canadian data illustrate similar trends with 96.3% of the population from
2007-2009 with adequate status evident by RBC folate concentrations. Interestingly, 40% of the Canadian
population had RBC folate concentrations above the high cutoff (1360 nmol/L) set as the 97.5th percentile
of the NHANES 1999-2004 data131. Applying the most recent proposed high cutoffs of 1450, 1800 and
2150 nmol/L recommended to evaluate the impact of high RBC folate on health outcomes, 16%, 6% and
2%, respectively, of participants in the 2007-2009 CHMS, had high RBC values 65. The prevalence of high
RBC folate concentrations was higher in females, those between 60-79 years of age, overweight or obese
individuals, and folic acid supplement users65.
Despite the reported increase in folate intakes and blood concentrations post folic acid fortification, certain
groups in the population including women of childbearing age (WCBA) may be susceptible to suboptimal
RBC folate concentrations. In the US, > 90% of WCBA (20-59 years old) still had RBC folate
concentrations below <906 nmol/L, a purported level below which the NTD risk begins to increase,
postfortification. Nevertheless, RBC mean concentrations increased from 505 to 587 nmol/L pre to post
fortification132 in WCBA. In Canada, 22% of WCBA (15- 45 years old) also had RBC concentrations below
the <906 nmol/L cutoff postfortification in the most recent CHMS133. However, it is unknown whether
these WCBA were planning a pregnancy and no data on dietary patterns and supplementation were
collected. The controversy in folic acid fortification and supplementation thus persists; while high levels of
folate intake and blood concentrations are evident postfortification, there is a need to continue to
supplement with folic acid before and during pregnancy as a protective measure against NTDs.
26
2.1.6.3 Folate in Pregnancy
The effects of folate deficiency on pregnancy is discussed in Section 2.1.6.1 and summarised in Table
2.1.6.1. Briefly, folate deficiency has been associated with an increased risk of NTDs and congenital
disorders.
I. Folic Acid Supplementation during Pregnancy
Studies conducted in Canada and the US report that the majority of women take supplements containing
folic acid prior to and during pregnancy49,134–137. Two recent studies in the US show that around 51% - 78%
of women report taking folic acid before and during pregnancy49,134. In Canada, a recent study reported a
60% folic acid supplement use before pregnancy, and a 93% and 89% reported use in early and late
pregnancy135. Compared to the amount of folic acid obtained from fortification (100 µg/day), most prenatal
vitamins available on the market and consumed contain 1000 µg/d folic acid. This difference alone suggests
that folic acid folic acid supplementation, more than fortification, may be driving the folate status in
women. A study has also shown that the observed high serum folate levels from the non-pregnant
participants of 2001-2004 NHANES were primarily due to supplemental sources of folic acid, and not from
fortification of enriched cereal-grain products or ready-to-eat cereals138.
II. High Maternal Folate Levels
There have been observational studies from a number of countries that investigated maternal blood folate
levels during pregnancy13,23,49,51,59,137,139–142 (Table 2.1.6.3a). Among other countries with mandatory folic
acid fortification, studies report high supplemental use before and during pregnancy in addition to
mandatory fortification in Canada and the US135,143,144. Elevated RBC folate concentrations during pregnancy
have thus been evident in these studies and particularly in two recent observational studies in Canada and
the US. The most recent Canadian study (the PREnatal FOlic acid exposuRe on DNA Methylation in the
27
newborn infant [PREFORM] study) reported high RBC folate concentrations in early and late pregnancy
and increased RBC folate concentrations from first to third trimester (2417 and 2793 nmol/L
respectively)143. Furthermore, UMFA was detectable (≥ 0.2 nmol/L, range: undetectable to 244 nmol/L)
in 97% of the women143. Similarly, in the US, the 1999-2006 NHANES reported high RBC folate
concentration during pregnancy, which increased from the first to the third trimester (1255, 1527 and 1773
nmol/L respectively)49. On the contrary, an Irish randomized control trial (RCT) published this year
investigated the effect of supplemental folic acid (400 µg) with fortification (100 µg folic acid/day) on
UMFA in maternal and cord blood and found no significant difference in UMFA blood concentrations
between the treatment and placebo groups145. Nevertheless, the same study reported higher maternal
plasma total folate concentrations in the folic-acid supplemented group compared with the placebo
group145. Overall, while folic acid supplementation has been shown to be beneficial during pregnancy and at
birth, caution on the dose of periconceptional folic acid for supplementation must be considered as this may
attribute to the elevated blood levels of folate in mothers in the current folic acid fortified environment.
28
Table 2.1.6.3a. Summary of studies investigating maternal folate concentrations
For folate and homocysteine concentrations, aMolloy et al. 2002, Hure et al. 2011 reported median(IQR). bObeid et al. 2005 reported geometric means±SD. cTakimoto et al. reported mean±SD. dYajnik et al. reported median (25th-75th percentile). eObeid et al. reported median (10th-90th percentile). f Fayyaz et al. 2014, Plumptre et al. 2015 reported geometric mean (95%CI). gPentieve et al 2016. Reported mean±STD of placebo group. Immuno, protein-binding immunoassay; MBA, microbiological assay; Hcy, homocysteine; LC-MS, liquid chromatography-mass spectrometry. Table modified from Plumptre 2016 thesis33.
Guerra-Shinohara et
al. 2002
Molloy et al. 2002a
Obeid et al. 2005b
Takimoto et al. 2007c
Yajnik et al. 2008d
Sweeney et al. 2009
Obeid et al. 2010e
Hure et al. 2011a
Branum et al. 2013
Fayyaz et al. 2014f
Plumptre et al. 2015f
Pentieva et al. 2016g
Location Brazil Ireland Germany Japan India Ireland Germany Australia USA Canada Canada Ireland FA fortification? Voluntary Voluntary No No No Voluntary Limited Voluntary Mandatory Mandatory Mandatory Voluntary
FA supplementation?
Not addressed
Yes Yes (17%) No 500µg starting at
18 wks
No Yes (29%) Yes (84%) Yes (60-89%) Yes (nearly all taking ≥600
µg/d)
Yes (90-93%, 83%≥ 1000
µg/d)
Yes
Serum/RBC folate method
immuno MBA immuno immuno immuno MBA LC-MS immuno Bio-Rad adjusted for
MBA
immuno (ion-capture)
immuno MBA RBC: MBA
UMFA method - - - - - HPLC LC-MS - - - LC-MS LC-tandem MS
Sample size 69 201 82 82 562 20 87 138 1296 122-520, varies by trimester
364 125
Time of blood draw childbirth childbirth childbirth 34-36 wks 28 wks childbirth childbirth 36 wks 1st, 2nd, 3rd trimester
1st, 2nd, 3rd trimester
1st (0-16wk), 3rd (23-27wk)
1st (14wk) 3rd (36wk)
Fasted samples? - - No No Yes Yes 1-12 hrs Yes Yes No No No
Serum folate (nmol/L)
13 26 27 ±20
23
±69
- 21 19 26 - 36
1st: 51(49, 54) 3rd: 39 (37, 41)
24.9
Median(range) (7-34) (14-41) (7-70) (6-47) (20) (35,36) 1st,2nd
trimester
±10.9
RBC folate (nmol/L)
Median(range)
580
(252-1635)
1627
(1044-2093)
- 813
±475
961
(736-1269)
945
(278-2180)
- 1185
(702)
1st: 1255 (1048, 1525),
2nd: 1527 (1449, 1630),
3rd: 1773 (1694, 2012)
1st: 1280 (1114, 1393), 2nd:1504 (1450, 1568),
3rd: 1462 (1421, 1529)
1st: 2417 (2362,2472)
3rd: 2793 (2721,2867)
1200
±639
Hcy (µmol/L) 7.1 8.3 5.6
±1.6
5.9
±1.4
8.6 - 5.6 5 (2.3) - 1st:5 (4.9,5.2) 3rd:6 (5.8,6.2)
6.3
Median(range) (3.3-21.0) ±2.9 (6.7-10.8) (4.3–8.2) (n=14) - ±1.4
UMFA (nmol/L) - - - - - 0.29 0.15 - - - 2.44 0.44 Median(range) (0-1.34) (0-0.8) (1.99,2.88)
% with detectable UMFA
- - - - - 90% 44% - - - >90% 42%
29
III. High Cord Blood Folate Levels
Although fewer studies have reported on cord blood folate concentrations, all studies have shown higher
folate concentrations in the cord blood than in maternal blood23,51,139,140,143,145 (Table 2.1.5.3b). Four studies
have reported the presence of UMFA in 55 – 100% of the cord blood samples23,61,62,143; however, one RCT
study reported only 20%145. Notably, in RCTs, the level of folic acid supplementation and fortification in
the studied population were at 400 µg/d and 100 µg/d, respectively. While in observational studies, an
average folic acid intake of 400 µg to 1000 µg/d was reported. Interestingly, among the four studies
reporting the occurrence of UMFA in cord blood, two were in populations without mandatory folic acid
fortification and supplementation61,62.
Table 2.1.6.3b. Summary of studies investigating cord folate and homocysteine concentrations
Guerra-Shinohara
et al. 2002
Molloy et al. 2002a
Obeid et al. 2005b
Sweeney et al. 2005c
Sweeney et al. 2009
Fryer et al. 2009d
Obeid et al. 2010e
Plumptre et al. 2015
Pentieva et al. 2016g
Location Brazil Ireland Germany Ireland Ireland UK Germany Canada Ireland FA fortification? Voluntar
y Voluntary No Voluntary Voluntary Voluntary Limited Mandatory Voluntary
FA supplementation? Not addressed
Yes Yes (17%) No No Yes Yes (29%) (90-93%, 83%≥ 1000
µg/d)
Yes
Sample size 69 201 82 9 20 23 29 364 125 Serum/RBC folate
method immuno MBA immuno - - immuno LC-MS immuno MBA
UMFA method - - - HPLC/MBA
HPLC - LC-MS LC-MS LC-tandem MS
Serum folate (nmol/L)
28 47 61±21 - - 15.8±3.5 40 64 50.6
Median(range) (15-38) (33-62) (9-79) (61,68) ±20.1 RBC folate (nmol/L) 1027 2142 - - - - - 2689
1900
Median(range) (305-2627)
(1645-2549)
(2614,2765) ±718
Hcy (µmol/L) 6 7.9±2.9 5.4±1.9 - - 10.8±3.8 5.3 4.8
-
Median(range) (2.3-15.3)
(3.5–7.5) (4.7,5.0)
UMFA (nmol/L) - - - 0.42 0.28 - 0.21 0.68 1.34 Median(range) (0-0.60) (0-0.51) (0.60,0.77) (0.68, 2.25)
% with detectable UMFA
- - - 100% 85% - 55% >90% 20%
Molloy et al. reported median(IQR) unless otherwise stated.. bObeid et al.2005 reported geometric mean±SD. cSweeney et al. 2005 reported mean values. dFryer et al. reported mean±SD. eObeid et al. reported median(10th-90th percentile). gPentieve et al 2016. Reported mean±STD of placebo group. Immuno, protein-binding immunoassay; MBA: microbiological assay, HPLC: high performance liquid chromatography, LC-MS: liquid chromatography-mass spectrometry. Table modified from Plumptre 2016 thesis33.
30
2.1.6.4 Effects of High Folate Exposure During Pregnancy
Given folate’s role in cellular development, it is not surprising to find a high demand for folate.
Furthermore, previous studies have found and confirmed higher circulating folate concentrations in cord
compared to maternal blood. In the past fifteen years, research has focused on low maternal intake of folate
to decrease risk of neural tube defects (NTDs) in their offspring (See Section 2.1.6.1). However, recently,
due to the drastic increase in intakes, and serum and blood concentrations of folate resulting from
mandatory folic acid fortification in the population among many countries; as well as an increase in
periconceptional folic acid supplementation, a large body of research has shifted its focus to health effects
and disease risk of excessive intakes of folate. In addition, this study has assessed the literature and ranked
the strength of evidence as ‘convincing’, ‘probable’ or ‘possible/insufficient’ (Table 2.1.6.4).
Table 2.1.6.4. Protective and adverse effects of periconceptional folic acid supplementation on pregnancy and birth outcomes Protective effects Adverse effects
I. Pregnancy & birth outcomes
Convincing • NTDs
Probable • Congenital heart defects
• Preeclampsia
Possible/Insufficient • Preterm birth • Low birth weight
• Small for gestational age babies (with low maternal vitamin B12)
II. Infant health outcomes
Convincing
Probable • Pediatric cancers • Language development
Possible/Insufficient • Autistic traits • Asthma/wheezing • Adolescent adiposity and
insulin resistance
31
I. Pregnancy & Birth Outcomes
Convincing
Neural Tube Defects
NTDs refer to a complex of multifactorial disorders where the brain and spinal cord of the developing fetus
fail to properly develop. This occurs during early embryogenesis, between conception and approximately
21-28 days146. The evidence linking folate deficiency with NTDs has been strong and consistent. The first
observed link was observed in the Leeds Pregnancy Nutrition Study in 1969120. Two decades later, two
large randomized double-blind clinical trials demonstrated the protective effect of periconceptional folic
acid supplementation on NTD risk in women. The Medical Research Council Vitamin Study Research
Group clinical trial involving 1195 women across seven countries found a 72% decreased risk of having a
NTD-affected pregnancy with 4000 µg folic acid periconceptionally among women who previously had an
NTD-affected pregnancy119. Another clinical trial (n=4753) conducted as part of the Hungarian Family
Program in 1984 reported a 27% reduction in the first NTD occurrence in response to 800 µg/d of folic
acid as part of a multivitamin supplement consumed during the periconceptional period.
Probable
Preeclampsia
Preeclampsia is characterised by hypertension and proteinuria during pregnancy. Few studies have
associated periconceptional use of folic acid with a decreased risk of preeclampsia. Earlier studies reported a
43-63% decrease in preeclampsia associated with multivitamin supplements containing folic acid147,148.
However, no specific micronutrient was attributed to this outcome. Furthermore, some studies did not find
a significant association between folic acid supplementation in the range of 400-500 µg and
preeclampsia149,150. Nevertheless, more recent studies conducted in China and Canada found a dose-
response inverse relationship between folic acid supplementation and preeclampsia151,152. In the Canadian
32
study, the protective effect of folic acid supplementation was more pronounced for mothers with a higher
risk of developing preeclampsia153. Although this dose-response was not seen in a US study, folic acid
supplementation in early pregnancy was associated with a protective effect on preeclampsia on lean
compared to obese mothers154.
Congenital Heart and Other Congenital Defects
Both clinical155 and observational studies in the US156,157, Netherlands158 and China159 have generally shown
an inverse association between maternal folic acid supplementation and congenital heart defects. The most
recent population-based case-control study found that approximately 40% of congenital heart defects may
be preventable with folic acid supplementation of 600 µg/d160. Although consistent and unequivocal
evidence for the protective effect of periconceptional folic acid supplementation on NTD risk exists,
evidence for protective effects of folic acid supplementation on other congenital defects have not been well
established161. This may be due to varying folate status of the populations studied146.
Possible/Insufficient
Preterm Birth
Most studies did not find significant associations between folic acid supplementation and preterm
birth154,162,163. However, some studies have reported that a higher plasma folate concentration during late
pregnancy is associated with longer gestation and lower risk for preterm births164,165. Furthermore, a recent
study conducted in China found an 8% lower risk of preterm birth with periconceptional folic acid use152.
Low Birth Weight and Small for Gestational Age (SGA)
Among the previously mentioned pregnancy outcomes, low birth weight and SGA are considered major
influencers of infant long term health166. Numerous studies have reported associations between one-carbon
33
nutrients during pregnancy and the risk for low birth weight and SGA newborns. A study in Greece
reported that supplementation with high folic acid (5000 µg/d) in early to mid-pregnancy was associated
with a 60% decrease in the risk of a low birth weight newborn and a 66% decrease in risk of a SGA
newborn165. Studies in the Netherlands and Pakistan also found that low maternal folate concentrations
were associated with lower birth weight, increased risk for SGA167, and intrauterine growth restriction168.
More recently, a study in China found a 19% decreased risk of SGA with 400 µg/d of folic acid152.
However, other studies did not find significant associations between maternal folate status and SGA154,164 or
low birth weight154. Nevertheless, a review of observational studies reported that maternal folate status has
been positively associated with fetal growth parameters169. The most consistent positive association was
observed for RBC folate concentration and fetal growth parameters169. Comparably, animal studies have
shown positive associations with maternal methyl-supplemented diets and offspring birth weight. A study in
rats fed a low-methionine diet induced higher homocysteine concentrations and lower fetal and maternal
weight170. Another study that fed rats a methyl-supplemented (including folic acid) diet reported decreased
offspring leptin concentrations and impaired post-natal growth in both sexes171. Increased long-term body
weight gain was observed in males171. Similar increases in body weight gain and impaired fetal development
was observed in mice with combined fetal exposure to arsenic and maternal folic acid supplementation172.
II. Infant Health Outcomes
Probable
Pediatric Cancers
A number of studies have associated periconceptional folic acid supplementation with a reduction in risk of
developing pediatric cancers. An Australian case-control study associated folic acid supplementation with a
decreased risk of childhood brain tumours173. Similarly, as part of the Childhood Leukemia International
Consortium, 12 case-control studies were analysed in order to investigate the association between prenatal
34
use of folic acid and acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML). This study
found a reduced risk of ALL associated with maternal folic acid supplementation regardless of period of
intervention (preconception or during pregnancy); however, only a weak inverse association was reported
between prenatal folic acid use174 and AML. Another meta-analysis found a reduction in the risk of pediatric
leukemia, brain tumours and neuroblastoma associated with the use of multivitamins containing folic acid;
however, the study could not attribute the protective effective to a specific vitamin(s)175. In contrast, some
studies have reported that periconeptional folic acid supplementation may be associated with an increased
risk of 176,177 or a null178 effect on risk of pediatric cancers. Animal studies have shown that maternal folic
acid supplementation decreases colon cancer risk179, while it increases the risk of mammary tumours180 in
the offspring. However, no human studies have examined the effect of periconceptional folic acid
supplementation on colon and breast cancer risk in the offspring.
Language Development
Several studies have investigated a possible link between language development and periconceptional folic
acid supplementation and have generally found an inverse association. The Rhea Study in Greece reported
improvements in vocabulary, communication skills and verbal comprehension in 18-month-old children
associated with high (≥5000 µg/d) folic acid supplementation during pregnancy181. Similarly, a Norwegian
Mother and Child Cohort Study (MoBA) found that folic acid supplementation from preconception to
eight-week gestation decreased the incidence of severe language delay in children at 3 years of age182.
Possible/Insufficient
Autistic Traits
Studies investigating the relationship between autistic spectrum disorders or autistic-behaviours and
maternal folic acid supplementation have reported equivocal findings. The MoBA study in Norway found a
35
protective effect on autism spectrum disorders in 3-year-old children born to mothers who were
supplemented with folic acid for 4-8 weeks preconception compared with those born to mothers who did
not receive preconceptional folic acid supplementation183. Other studies have also provided evidence that
prenatal folic acid use is associated with decreased childhood autistic-behaviours184,185, although other
studies have reputed the purported protective effect of maternal folic acid supplementation on autistic
traits186.
Asthma/ Wheeze
The relationship between maternal folic acid supplementation and respiratory health in the offspring (i.e.
asthma and wheezing) has been studied with conflicting results9. Two meta-analysis of 14 studies (10
cohorts, 3 nested case-controls, and 1 case-control) and another more recent analysis looking at 10 large
prospective cohorts did not find an association between childhood asthma and prenatal folic acid
supplementation186,187. However, another meta-analysis of 26 studies (16 cohorts, 7 case-controls, and 3
cross-sectional) found that early pregnancy folic acid supplementation was associated with an increase in the
risk of wheeze188. Interestingly, irrespective of folic acid supplementation, children homozygous for the TT
genotype of the MTHFR677T polymorphism showed a higher risk for developing asthma in this meta-
analysis188. Other studies indicate that folic acid supplementation later in pregnancy (30-34 weeks gestation)
may be protective against asthma in children at three and a half years of age189.
36
Adolescent Adiposity & Insulin Resistance
See II. Excess Folate, Insulin Resistance and Adiposity in Section 2.1.6.1.
In summary, high prenatal supplementation of folic acid and folate status has been shown to have both
protective and adverse effects on mothers and their offspring. Although the evidence may still be equivocal
for the most part, the risk of high folic acid intakes (mostly from supplemental sources) and especially in
combination with low B12, may be potentially harmful to offspring’s health. Therefore, more research is
needed to elucidate the effect of periconceptional folate status and supplementation on pregnancy and birth
outcomes and infant health.
37
2.2 Vitamin B12
2.2.1 Definition, Chemical Structure and Dietary Sources
Vitamin B12 (cobalamin) is a water-soluble B vitamin that is essential for RBC production, normal
neurological function and metabolism of carbohydrates to glucose14. There are several different B12 forms
depending on the moiety attached to cobalt in the centre of this molecule; including methylcobalamin,
cyanocobalamin, hydroxycobalamin, sulfitocobalamin and 5'-deoxyadenosylcobalamin (Figure 2.2.1). The
two metabolically active forms are methylcobalamin and deoxyadenosylcobalamin95. Cyanocobalamin is the
most stable form and is found in supplements and fortified foods. Cyanocobalamin is also the form that is
readily converted to methylcobalamin and 5’-deoxyadenosynlcobalamin, where methylcobalamin is a
coenzyme for methionine synthase and 5’-deoxyadenosynlcobalamin is a coenzyme for amino and fatty acid
metabolism190.
Figure 2.2.1. Chemical structures of vitamin B12. These partial structures of B12 indicate the changes in the random group (L) where; (1), 5’-deoxyadenosylcobalamin; (2), hydroxycobalamin; (3), methylcobalamin; (4), sulfitocobalamin; (5), cynocobalamin. Reprinted with permission from Watanabe 2007190. Animal sources are major sources of dietary B12; plants generally don’t provide much if any B12 except
edible algae. Main animal sources are meats, eggs, fish and shellfish14. Some examples of fortified foods
with B12 generally involve meat replacements such as soy and tofu or plant-based beverages190.
38
2.2.2 Absorption, Transport and Storage
Before absorption of dietary B12 in the small intestine, B12 must first be released from the protein it is
bound to in food via hydrochloric acid and pepsin in the stomach. B12 present in fortified foods, however, is
unbound and can be absorbed immediately. Free B12 attaches to R proteins called haptocorrins (produced
in the salivary glands) to aid transport to the duodenum. In the duodenum, pancreatic proteases cleave and
release B12 bound to haptocorrins, which then bind to intrinsic factor (IF), a glycoprotein secreted by the
parietal cells of the stomach forming the IF-B12 complex. IF-B12 travels to the terminal ileum and is
absorbed by cubilin, cubam and amnionless receptors by calcium-dependent endocytosis (Figure 2.2.2).
These receptors can be saturated at 1.5 – 2.0 µg of B12, which limit absorption. The alternative pathway,
whereby ~ 1% of B12 is absorbed through passive diffusion, is relatively inefficient95. Upon entry into the
enterocyte, IF-B12 complex dissociates and B12 binds to plasma binding protein transcobalamin II. Upon
exit into circulation, there are three plasma binding proteins (transcobalamin I, II or III) that bind B12.
Transcobalamin I binds the majority (~80%) of circulating B12 and stores it in the liver. Transcobalamin III
has a similar binding role but this is still unclear. Transcobalamin II transports B12 to the tissues through
transcobalamin II receptor (TCII-R). Transport of B12 involves two main transport proteins: haptocorrin
and transcobalamin II. Most bind to haptocorrin and around 20-30% bind to transcobalamin II to form
holotranscobalamin (holoTC)95 (Figure 2.2.2).
39
Figure 2.2.2. B12 absorption and transport. P, dietary protein from food; HCl, hydrochloric acid; R; R proteins; IF, intrinsic factor; IF-B12; intrinsic factor- B12 complex; TC, transcobalamin; and HC, haptocorrin. Reprinted from Hughes et al. 201395.
HoloTC is the metabolically active fraction of B12 as it represents the fraction of circulating B12 for cellular
uptake11,95. The body stores around 500 µg B12 in the liver and excretes 1 – 10 µg B12/d into bile, where
90% is reabsorbed and 10% is excreted in feces. For this reason, B12 deficiency intakes years to develop in
healthy individuals. Nevertheless, B12 deficiency can be caused by either inadequate intake and/or impaired
absorption (Table 2.2.2). Healthy adults with normal digestive systems are believed to absorb 50% of
dietary B12190. However, gastrointestinal function is believed to decrease with age and is suggested to be
due to gastrointestinal atrophy and reduced gastric acidity. Malabsorption of B12 can be attributed to the
lack of IF due to loss of gastric parietal cells, resulting in pernicious anemia (Table 2.2.2). Hence, a higher
prevalence of B12 deficiency is seen in the elderly14,69,95. Nevertheless, in light of this, older adults have been
shown to increase their supplementation of B12, especially in older women. Hence, despite the possibility
of decreased intakes of animal products and malabsorption, no significant differences were observed in B12
deficiency across the lifecycle69.
40
Table 2.2.2. Causes of B12 deficiency
Cause Example Dietary inadequacy Diet low in meat and dairy products
Alcohol consumption
Malabsorption Lack of IF Pancreatic insufficiency Parasitism Transcobalamin deficiency Gastric bypass or other bariatric surgery
2.2.3 Biochemical Functions
The two metabolically active forms of vitamin B12 are methylcobalamin and 5’-deoxyadenosylcobalamin.
Methylcobalamin serves as a coenzyme for MS in the one-carbon metabolism pathway. Methylcobalamin
with MS transfers a single methyl from 5-MTHF to homocysteine to form methionine and regenerate THF.
This process is important in combination with folate/folic acid as THF can be converted to 5, 10-
methylene THF via B6-dependent SHMT for de novo DNA synthesis (Figure 2.2.3a).
Figure 2.2.3a. Biochemical functions of B12 (methylcobalamin) in folate/folic acid dependent one-carbon metabolism. Enzymes are shown in bold. MS, methionine synthase; BHMT, betaine-homocysteine methyltransferase; DMG, dimethylglycine; Met, methionine; MTHFR, methylenetetrahydrofolate reductase; SHMT, serine hydroxymethyltransferase; tHcy, homocysteine; THF, tetrahydrofolate. Reprinted with permission from Masih 2013191.
41
5’- deoxyadenosylcobalamin serves as the coenzyme for L-methylmalonyl-CoA mutase, which is the
enzyme that catalyzes the isomerazation of L-methylmalonyl-CoA to succinyl-CoA. This enzymatic
reaction is involved in amino and fatty acid metabolism190 (Figure 2.2.3b).
Figure 2.2.3b. Biochemical functions of B12 (5’deoxyadenosylcobalamin) in amino and fatty acid metabolism. Enzymes are shown in bold. MCM, methylmalonyl-CoA mutase; MMA, methylmalonic acid. Reprinted with permission from Masih 2013191.
2.2.4 Vitamin B12 Intake Requirements
For healthy adults between 19 – 50 years old, the EAR and RDA are set at 2.0 and 2.4 µg/d. The EAR is
based on the minimum serum B12 concentration needed to prevent adverse hematological effects. Not
enough information is available to determine the RDA; hence it is set by assuming the RDA is equal to the
EAR plus twice the coefficient of variation of 10% of the standard deviation of the B12 requirement.
Therefore, the RDA is approximately 120% of the EAR. The EAR and RDA are increased for pregnant
women to 2.2 and 2.6 µg/d, respectively. This increase in requirement is suggested to be due to an
estimated B12 fetal deposition of 0.1 to 0.2 µg/d. For lactating women, due to a loss of approximately 0.25
- 0.33 µg/d of B12 in milk, the EAR and RDA are increased to 2.4 to 2.8 µg/d. A UL has not determined
as current evidence suggest no adverse effects associated with excess B1214 (Table 2.2.4).
42
Table 2.2.4. Dietary Reference Intakes (DRIs) for females at various life stages
Age and life stage EAR (µg/d DFEs)
RDA (µg/d DFEs)
UL (µg/d folic acid)
Female, 19y+ 2.0 2.4 ND Female, pregnant 2.2 2.6 ND Female, lactating 2.4 2.8 ND
DFE, dietary folate equivalents; EAR, estimated average requirement; RDA, recommended dietary allowance; UL, tolerable upper limit; ND, not determined.
Data from the US generally report an adequate intake of B12 in the general population. In the 2003-2006
NHANES data, adult women consumed a median of 3.4 µg/d from a diet containing enriched cereal
grains192. However, when accounting for both enriched cereal grains and other supplemental sources, the
median increased to 16.5 µg/d192. Unlike in the US, intakes in Canada are reported to be less adequate
with up to 20% of females (≥14 years old) consuming dietary B12 (natural and fortified sources) at less than
the EAR193. However, inadequacy decreased to <5% when accounting for supplemental intakes.
2.2.5 Biomarkers of Vitamin B12
There are four main biomarkers to measure B12 status; plasma/serum B12, holoTC, methylmalonic acid
(MMA) and homocysteine (Table 2.2.5). Serum B12 is primarily used as an assessment of clinical status as
it measures circulating B12. Similarly, holoTC measures the concentration of B12 available for cellular
uptake. MMA and homocysteine are inverse functional indicators of B12 status. With low B12, B12
(5’deoxyadenosylcobalamin)-dependent MCM activity decreases, and thus increases conversion of L-
methylmalonyl-CoA to MMA via D-methylmalonyl-CoA hydrolase (Figure 2.2.3b). Homocysteine is
increased with low B12 as MS activity is reduced; however, homocysteine is a less specific inverse functional
indicator to B12 as it is influenced by other B-vitamins such as B2, B6 and folate.
Cutoffs for deficiency have not unequivocally been established as a low serum B12 concentration does not
necessarily equate to deficiency, and similarly, a level above a ‘normal’ cutoff does not equate to
43
adequacy130. Most laboratories generally use a reference range to determine its own deficiency range
depending on the assay method. Over the past thirty years, the gold standard method has shifted to the
protein binding assay over the microbiologic assay27,95. A serum B12 concentration of 150 pmol/L is
commonly considered a deficiency cutoff11, 150 to 258 pmol/L for a marginal or subclinical194,195, and 220
pmol/L for maximal protection against NTDs196. As there is no set adequacy cutoff for serum B12, a value
above deficiency has been more commonly used130,197,198. For MMA, a cutoff >271 nmol/L is considered
a deficiency cutoff58. For homocysteine, a cutoff >13 µmol/L could indicate B12 deficiency69,199. For
holoTC, deficiency is considered to be <35 pmol/L137.
For more accurate assessment of B12 status, the NHANES roundtable summary suggested the need to
measure and report a combination of at least one biomarker of circulating B12 (serum B12 or holoTC) and
one functional biomarker (MMA or homocysteine)11.
Table 2.2.5. Comparison of B12 status biomarkers
Biomarker Type of biomarker
Strengths Weaknesses
Plasma/serum B12 Circulating measure -Measures total circulating vitamin B12
-Indicative of long-term low B12 intake
Needs confirmation with MMA or homocysteine
Holotranscobalamin (holoTC)
Circulating measure -Measures metabolically active portion of vitamin B12
-Sensitive to B12 deficiency status
Ongoing studies; holoTC concentrations are understudied
Methylmalonic acid (MMA)
Functional marker -Elevated MMA is more specific indicator for B12 deficiency than homocysteine -Sensitive to B12 deficiency status
MMA not a good functional marker of B12 status in certain cases (e.g. pregnancy)
Plasma homocysteine Functional marker -Elevated homocysteine can be indicator B12 deficiency
Influenced by other B-vitamins (folate, B2 and B6)
Table modified from Plumptre et al. 2016 thesis33.
44
2.2.6 Vitamin B12 and Health
2.2.6.1 Vitamin B12 in Health and Disease
B12 deficiency has been linked to neurological symptoms and sensory neuropathy14. These manifestations
include: megaloblastic anemia, myelopathy, neurodegeneration, depression and cognitive decline14,69,200.
Although B12 deficiency can occur without masking by folate, they are integrally linked. High folic acid
intake has been associated with the masking of B12 deficiency through the methyl trap and has been shown
to mask hematological manifestations, thereby delaying the diagnosis of B12 deficiency. Interestingly, a
study reported that the proportion of individuals with low B12 without macrocytosis (undiagnosed B12
deficiency) was higher in the post- than in the pre-folic acid fortification period201. Furthermore, while the
requirements for folate have been met by fortification and supplementation practices, the same cannot be
said for B12 as it remains untreated and inadequate (<111 pmol/L) among 35% of women of reproductive
age202.
2.2.6.2 Vitamin B12 in Pregnancy
I. Maternal B12 Intakes and Low B12 Levels
Low maternal concentrations of B12 have been reported despite adequate intakes. Most prenatal
supplements contain the amount of B12 equal to the RDA (2.6 µg/d). The PREFORM study found that
greater than 90% of women had intakes that met the RDA135. Serum holoTC concentrations were found to
be in the normal range at 35 to 140 pmol/L in 88 to 91% of pregnant Canadian women from the Alberta
Pregnancy Outcomes in Nutrition (APrON) 137. Nevertheless, the population data based on the 2007-2009
CHMS reported approximately 6% and 20% of the Canadian women between 20 to 79 years old with
serum B12 indicative of deficiency (<148 pmol/L) and marginal status (148-220 pmol/L), respectively69. A
study in the Netherlands showed a progressive decline reaching marginal and deficient (<150 pmol/L)
serum B12 by 28 – 32 weeks gestation203, suggesting that B12 concentrations decrease as pregnancy
progresses. Similar findings were reported by more recent studies from Canada. In the PREFORM study,
45
17% and 38% of women in early pregnancy and at delivery, respectively, had deficient (<138 pmol/L)
serum B12 concentrations204. A study from Vancouver found 23 and 35% of women had deficient (<148
pmol/L) or marginal (148-220 pmol/L) plasma B12 concentrations, respectively, in late pregnancy205.
Newborn B12 concentrations have been shown to be significantly higher than maternal concentrations. A
study reported cord serum B12 concentration (268 pmol/L) to be double that in maternal serum (188
pmol/L)206. Another study reported similar findings with serum B12 in cord blood as 321 pmol/L and in
maternal blood at delivery as 169 pmol/L204. At 27 weeks postpartum, the PREFORM study reported that
infants had a serum B12 concentration of 226 pmol/L where maternal B12 concentrations at 20 and 36
weeks gestation were 177 pmol/L and 149 pmol/L, respectively142. Other studies have also reported
decreasing B12 concentrations as pregnancy progresses despite adequate intakes137,204,207–209 (Table
2.2.6.2a). Animal studies show that B12 accumulates in the placenta of pregnant rats injected with B12 and
that it is actively transported to fetal tissues in a dose-dependent manner210. This implies that cord blood
and maternal B12 concentrations are closely correlated over the course of pregnancy. Hence, based on these
studies, maternal B12 status is important to ensure sufficient supply of B12 is transferred to the fetus206.
Reasons for this B12 reduction in the pregnant mother has been attributed primarily to physiological
changes occurring during pregnancy as well as to increased demands to ensure fetal growth and
development14,206. In the mother, some of these changes are related to hemodilution, hormonal changes,
decreased B12 absorption, alterations in B12 placental transport and binding protein activity14,206. In the
fetus, changes mainly occur in order to support rapid embryonic growth and development involving DNA
and protein synthesis 142.
46
Table 2.2.6.2a. Summary of studies investigating B12 intakes and serum concentrations.
aBaker at al. 2002 and Wu et al. 2013 reported mean±STD unless otherwise stated. bRay et al.2008 reported B12 deficiency as <125 pmol/L.
cFayyaz et al.2014 reported median 95% CI. dHerran et al. 2015 reported standard errors for serum B12, 95% CI for B12 deficiency, and used a cutoff <148 pmol/L. eVisentin et al. 2016 reported mean 95% CI. Serum B12 concentrations were from supplement users. B12 deficiency cutoff was <148 pmol/L and for marginal deficiency* 128-220pol/L serum B12.
Gstn, gestation; EP, early pregnancy (1-12 weeks gestation); LP, late pregnancy (12-16 weeks gestation) D, delivery (28-42 weeks gestation); NP, nonpregnant women.
Baker et al. 2002a
Ray et al. 2008b
MacFarlane et al. 2011
Wu et al. 2013 a Fayyaz et al. 2014c
Herran et al. 2015 d
Visentin et al. 2016e
Location US Canada Canada Canada Canada Colombia Canada Pregnant Yes Yes No
(WCBA) Yes Yes Yes Yes
Sample size 563 10,622 5,600 554
645 9,523 368
B12 Intake (µg/d)
12 - - - 1st:3.7 (3.3,4.0) 2nd:4.0 (3.7,4.2) 3rd:4.1 (3.8,4.4)
- EP:4.7±3.1 (0.9-26.1)
LP:4.6±2.6 (0.7-18.5)
Serum B12 (pmol/L)
1st:226±133
2nd:203±120
3rd:168±96
≤28 gstn:285 >28 gstn:249
<148 148-220
>220
2nd:287±126 3rd:224±96.2 NP: 413±157
-
221
(4)
LP:227 (216,237)
D:173 (165,182)
Median(range)
(86-446) (73-412) (64-323)
(277-293) (244-255)
HoloTC (pmol/L)
(<140pmol/L)
Median(range)
- - - - 1st:<35: 26 35-140:87 2nd:<35:31 35-140:81
- (81,94) (17,34) (77,83)
- -
MMA (µmol/L)
- - - - - - LP:108 (103,113)
D:137 (130,144)
Median(range) Plasma
homocysteine (µmol/L)
- - >13 <13
2nd:4.26±1.37 3rd:5.07±1.28 NP: 8.77±2.30
- - LP:5 (4.9,5.1) D:6 (5.8,6.2)
Median(range) B12 deficiency
prevalence (%)
1st:10 2nd:30
(no low B12
cutoff)
≤28 gstn:5.2 >28
gstn:10.1
5% 23% <1
1st:18.2(17,19) 2nd:41.0(38,44)
3rd:18.8(77,39.2)
LP:17,35* D:38,43*
47
II. Pregnancy and birth outcomes
The effects of B12 supplementation on health effects during pregnancy, birth and in the infant are shown in Table 2.2.6.2b. In addition, this study has assessed the literature and ranked the strength of evidence as ‘convincing’, ‘probable’ or ‘possible/insufficient’. Table 2.2.6.2b. Adverse effects of periconceptional B12 supplementation on pregnancy and birth and infant health outcomes.
Adverse effects I. Pregnancy & birth
outcomes Probable • NTDs
• Preeclampsia • Preterm birth • Low birth weight • Small for gestational age
babies (with high maternal folate)
II. Infant health outcomes
Probable • Cognitive development
Possible/Insufficient • Adolescent adiposity & insulin resistance
Probable
NTDs
The evidence linking maternal B12 and NTDs are few but generally points to an inverse association 211.
However, some studies did not find associations with NTD affected pregnancy and maternal B12 status197,212.
More recent case-control studies have reported an increased odds ratio of 2.5 to 5.4 for having a NTD
affected pregnancy associated with low serum B12 concentrations196,213–217.
Possible/Insufficient
Preeclampsia
Associations between maternal B12 status and the risk of preeclampsia have been inconsistent due largely to
variable B12 biomarkers used in these studies. Several studies have shown higher homocysteine
concentrations in women with preeclampsia than in normotensive women; however, serum B12
concentrations showed no significant difference between the two groups218–220.
48
Preterm Birth
High homocysteine concentrations, as a proxy of B12 insufficiency, have been associated with an increased
risk of preterm birth221,222. However, a case-control study in India reported that higher concentrations of
B12 was associated with preterm rather than term deliveries223.
Low Birth Weight and Small for Gestational Age
Studies investigating an association between maternal B12 status and birth weight and gestational age have
reported conflicting findings, depending on the location of the study population. Studies based in Western
populations such as in Australia224 and Netherlands167 reported no associations between maternal serum B12
concentrations during pregnancy and adverse pregnancy and birth outcomes (premature birth, intrauterine
growth restriction, small for gestational age and preeclampsia). In contrast, studies in Eastern populations
such as in India225,226 found that low serum B12 concentrations in the first and third trimesters and in cord
blood were associated with lower birth weights226. A similar study from Pakistan found an increased risk of
intrauterine growth restriction with lower maternal serum B12225 and high maternal homocysteine
concentrations168.
With regards to both folate and B12, a prospective analysis in Bangalore (n=1838) found that prenatal folic
acid supplementation in combination with low B12 intake was associated with an increased risk of SGA12. In
a subgroup analysis, an additional risk of SGA was observed in the presence of low B12 and high folate
intakes in the second trimester12. A similar study in Pune, India found that increasing folate:B12 ratio was
correlated with an increase in maternal plasma homocysteine concentrations and with a decreased in birth
weight, birth length, head circumference and chest circumference in the offspring202. Although a recent
study did not find a significant association between maternal B12 status and the risk for SGA and low birth
weight infants, an adjusted risk of fetal macrosomia was significantly higher in those born to mothers with
the lowest serum B12 and highest serum folate quartile than those born to mothers with the highest quartile
49
of serum B12 and lowest quartile of serum folate227. These mothers had no evidence of gestational
diabetes227.
Two studies investigated the association between maternal homocysteine concentrations based on the fetal
MTHFR C677T genotype and offspring birth weight. A French study found that the MTHFR 677TT genotype
had lower birth weight and birth length than those with the CC or CT genotypes. These differences
persisted until 1 year of age228. A more recent study found that the CT/TT genotypes rather than the wild-
type CC predicted higher maternal homocysteine concentration and lower offspring birth weight. Since
homocysteine is a functional inverse indicator of both folate and B12 status, these data suggest that low
folate and/or B12 status may be associated with low birth weight229. Other studies have also found similar
results in that high maternal homocysteine concentrations were associated with lower birth weight167. In a
Japanese study, a 0.1 µmol/L increase in maternal homocysteine concentration during the third trimester
was associated with 151 g lower infant birth weight 141 .
III. Infant health Outcomes
Probable
Cognitive Development
Low maternal B12 status has been associated with an increased risk of impaired mental and social
development in infants. Studies that have investigated this relationship are mainly from India or South Asia,
as many women in these countries are highly susceptible to a low B12 status, given their vegetarian dietary
practice13,230–234. The most recent Indian study in 2012, in which 62% of women were B12 inadequate
(<150 pmol/L), reported a positive association between maternal B12 status and motor and mental
(psychomotor) development quotients in children at two years of age231. Similarly, decreased cognitive
function as measured by decreased or slower performance in short-term memory and sustained attention
50
was observed in nine-year-old children born to mothers at the lowest decile of B12 concentration (<77
pmol/L) in comparison to those born to mothers in the highest decile (>225 pmol/L)230.
Adolescent Adiposity & Insulin Resistance
As mentioned previously in Section 2.1.6.4, limited studies have investigated maternal B12 status with the
risk of adiposity and insulin resistance in children. In a study from the UK including 91 mother-infant pairs,
low serum maternal B12 concentrations (<191 ng/L) were inversely associated with offspring’s
Homeostasis Model Assessment 2-Insulin Resistance (HOMA-IR) and triglycerides and positively
associated with HDL-cholesterol after adjusting for age and BMI235. After adjusting for multiple
confounders, however, significance was lost for triglycerides and HOMA-IR235. In the Randomized
Control Trial of Low GI diet (ROLO) study, maternal intake of micronutrients, including vitamins D, E,
and B12, magnesium and selenium, was positively associated with neonatal anthropometry236. Third
trimester B12 intake was positively associated with birth weight. At a follow up in infants at 6 months-of-
age, maternal first trimester intakes of B12 was positively associated with a risk of overweight/obese
infants; a B12 intake greater than 1 µg was associated with a 2.5 fold increased risk237. Another study found
that children from mothers deficient in B12 (<148 pmol/L) during early pregnancy had 26.7% higher
HOMA-IR at 6-8 years of age than those from mothers who were B12 sufficient during early pregnancy238.
This study did not find any significant association of HOMA-IR in children with maternal folic acid or other
micronutrient supplementation (iron, zinc and multiple micronutrient supplementation).
With regards to the combined effects of folate and B12, the Pune Maternal Nutrition Study12,13 found that
low maternal serum B12 concentrations (<150 pmol/L) in combination with high RBC folate
concentrations were associated with higher central adiposity and insulin resistance in infants and 6 year old
children. Furthermore, mothers in the extremes (highest folate and lowest B12 concentrations) were
reported to have children with the most insulin resistant. A recent animal study that attempted to replicate
51
the Pune study found conflicting results239. Female mice were fed one of three diets containing folic acid (2
mg/kg) with B12 (50 µg/kg diet), high folic acid (10 mg/kg) with B12, or a high folic acid without B12
starting 6 weeks before mating and lasting through during mating, pregnancy and weaning. Male offspring
were weaned onto a control or western diet. Amongst the control-fed offspring, males born to dams with
the high folic acid and B12 were heavier, in contrast to the Pune study, and male offspring from the high
folic acid without B12 showed reduced insulin concentrations. Glucose tolerance and baseline glucose
concentrations in the offspring fed control or western diets were not associated with any of the three
maternal diets239. Despite this contradictory finding, the intervention may have been too short to elicit the
full effects of reduced insulin concentrations on the offspring from the dams fed a high folic acid without
B12 diet. This study also did not measure maternal RBC folate concentrations, making it difficult to assess
how high folate concentrations were. Nevertheless, in support of other studies mentioned in Section
2.1.6.4, this animal study did find an association with maternal folic acid supplementation and increased
weight gain in the offspring.
In summary, low maternal B12 has been primarily associated with adverse health effects for children. Of
interest to this study is the combined effect of maternal high folate and low B12 status on adiposity and
insulin resistance in the offspring.
52
2.3 Epigenetics
2.3.1 Definition
Epigenetics differs from genetics in that it refers to heritable changes in gene expression that are
independent of changes in the DNA sequence itself5. Most of these heritable epigenetic changes are
established during cell differentiation and maintained even after cell division, allowing cells to have distinct
identities and functions while containing the same genetic code240 Types of epigenetic mechanisms are
DNA methylation, covalent histone modification, chromatin remodelling and RNA interference. Histone
modification, chromatin remodelling and DNA methylation involve the regulation i.e. coiling and
uncoiling of the local structure of chromatin. Chromatin is DNA wrapped with histones, specifically an
octamer of four core histone proteins (H3, H4, H2A and H2B), to form repeating units of nucleosomes240.
Nucleosomes function to package DNA within a cell, and can alter accessibility of regulatory DNA
sequences to transcription factors and thus regulate gene expression241. RNA interference includes non-
coding microRNAs (miRNAs) that regulate gene expression by posttranscriptional gene
silencing242.Epigenetic mechanisms such as DNA methylation are generated and in parallel to histone
modifications243,244. Altogether, these epigenetic mechanisms interact to regulate the genome. Epigenetic
changes are potentially reversible by dietary and pharmacologic manipulations43,245.
2.3.2 DNA Methylation
DNA methylation generally refers to the methylation of specific cytosines of cytosine-guanine (CpG)
sequences in the genome. DNA methylation also occurs at sites other than CpGs and are called non-CpG
DNA methylation246. DNA methylation is species-, tissue- and cell type-specific and plays an important
role in gene expression, in the maintenance of DNA integrity and stability, and in chromatin
conformation247. Up to 80% of all CpG sites in mammalian DNA are methylated and most CpG
methylation occurs in low density CpG areas such as in exons, noncoding regions and repeat DNA sites248
(Figure 2.3.2). High density areas are termed ‘CpG islands’ that span the 5’end promoter, untranslated
53
region and exon 1 of many transcribed human genes (Figure 2.3.2). The characteristics of a CpG island
commonly include a minimum of 200 to 500 base pairs, an observed: expected CpG ratio >0.6 to 0.64,
and a GC content at 50 to 55%249. Genome-wide methylation refers to the total cytosine methylation in
the genome whereas gene-specific methylation refers to DNA methylation at specific genes.
In most cases, CpG promoter methylation is an inverse determinant of gene expression207. That is in
normal cells, CpG islands are unmethylated and allow for transcription; while in disease states such as
cancer, CpG island methylation leads to stable heritable transcriptional silencing or inactivation250 (Figure
2.3.2). DNA methylation occurring in non-CpG islands is relatively rare in most mammalian cell-types,
however, it is most abundant in oocytes, pluripotent embryonic stem cells, and mature neurons250. The
function of mammalian non-CpG methylation remains unclear250. Aberrant or dysregulation of DNA
methylation can affect developmental and cellular processes such as embryonic development, genomic
imprinting, transcription, X chromosome inactivation, chromosome structure, and stability251. This in turn
can contribute to the development and progression of chronic diseases252, including asthma253,
cancer4,247,252, cardiovascular disease254, and metabolic disorders170,255.
DNA methylation of cytosine (5-methylcytosine; 5mC) has been thought of as stable, irreversible
epigenetic marks until the discovery ten-eleven translocation (TET) proteins that oxidize 5mC to 5-
hydroxymethylcytosine (5-hmC), which is a marker for active DNA demethylation250. 5-hmC has also been
linked to regulating transcription by attracting or repelling specific DNA-binding proteins250. This suggests
that DNA methylation may be more susceptible to environmental factors than previously thought256.
Biological functions of 5-hmC are not fully understood, but some studies have indicated its importance in
embryonic and germ cell development and brain cancer250.
54
Figure 2.3.2 CpG island and gene body methylation in normal and tumor cells. CpG island methylation occurs near the promoter region, exon 1 (box 1), and untranslated regions. Methylation (black filled circles) in the remaining area is referred to as gene body methylation, such as in exons, noncoding regions and repeat DNA sites. In normal cells, a) CpG sites near promoter regions are usually umethylated and allow for transcription to occur and b) CpG low dense areas are usually methylated. In tumor cells, loss of methylation in CpG depleted regions and a gain of methylation in promoter regions of CpG islands are generally observed. Methylated sites are filled with black circles and unmethylated are white circles. Adapted from the publisher John Wiley and Sons28.
2.3.2.1 Regulation of DNA Methylation
DNA methylation is a dynamic process between active DNA methylation, and active and/or passive DNA
demethylation. CpG DNMT1, 3a, 3b251 using SAM are involved in active DNA methylation (Figure
2.1.3). DNMTs facilitate conversion of SAM to SAH, and are inhibited by an accumulation of SAH257. For
active demethylation, methyl DNA binding domain protein-2 (MBD2) removes methyl groups from 5mC
residues, and can occur independently of DNA replication. Passive demethylation occurs when DNMT1
are inactive during the cell cycle following DNA replication, and thus sites can remain unmethylated in the
newly synthesized DNA strand258.
As mentioned previously, DNA methylation of promoter region CpG islands usually results in
transcriptional silencing, although a there are a few exceptions259. Transcription silencing involves
55
regulation by numerous proteins that bind and alter the conformation of chromatin. For instance, when
CpG islands are hypermethylated, histone deacetylases (HDACs) suppress transcription in a complex with
a transcriptional repressor, methyl-CpG binding protein-2 (MeCP2), and DNMTs260. This complex blocks
the transcriptional machinery from accessing the promoter region (Figure 2.3.2.1). In contrast, when
CpG islands are un/hypomethylated, transcription can occur as the nucleosomes are in a transcriptionally
active state due to histone acetyltransferases that maintain an open chromatin conformation260–262.
Figure 2.3.2.1. DNA methylation regulation of transcription. Hypermethylated CpGs (red circles) attract protein complexes, including HDACs, MeCP2, and DNMTs to bind and inhibit transcription by altering chromatin conformation to a closed state. Hypomethylated CpGs allow for transcription to occur as chromatin is in an open conformation. Adapted with permission from Bird 2002260.
2.3.3 Epigenetic Reprogramming
Epigenetic reprogramming during embryogenesis starts with active and passive genome-wide
demethylation, respectively, of the paternal and maternal genomes. This is followed by de novo methylation
soon after implantation6. Demethylation takes place during fertilization at the zygote stage, and then in the
primordial germ cell, which develop to become a sperm or oocyte263. DNA methylation content is the
lowest at the blastocyst stage (Figure 2.3.3). Establishment of de novo methylation patterns by DNMT3a
and 3b do not require pre-existing methylation and are thus able to establish a new pattern; in contrast,
56
maintenance of methylation patterns depends on DNMT1 to identify hemimethylated sites to ensure
replication of DNA methylation patterns264. The new DNA methylation pattern is considered stable
postnatally but it can be modified by factors such as aging, cancer or the environment including diet.
Figure 2.3.3. Epigenetic programming. After fertilization, active demethylation of the paternal and passive demethylation of the maternal genome occurs. 5mC content is lowest at the blastocyst stage. Formation of de novo methylation patterns rise after implantation. ESCs, embryonic stem cells; ICM, inner cell mass cells (forms the embryo); TE, trophoectoderm cells (forms the placenta). Modified from Plumptre 2016 thesis33. Reprinted with permission from Wu and Zhang 2014250.
2.3.4 Analysis of DNA Methylation
Epigenetic analyses have progressed through many years of innovation. Sequencing technology for DNA
methylation analyses, previously restricted to specific loci, are now performed on a genome-wide scale at
single-base-pair resolution249. Similarly, for DNA methylation profiling, current techniques utilize
microarray hybridization instead of two-dimensional gel electrophoresis249,265.
57
DNA methylation measurements can be attained either as a pattern of methylated target sequences along a
section of DNA or as an average methylation level at a single genomic locus across many molecules266. The
complexity in measurement comes from the varying DNA methylation patterns associated with each cell
type, DNA sample and location in the genome. Since hybridization techniques do not accurately distinguish
between methylated and unmethylated cytosines, and DNA methyltransferases are not present during
PCR, DNA methylation patterns are erased during amplification. Hence, pretreatments are required in
order to reveal the presence or absence of the methyl group at cytosines of CpG sites. Once completed,
amplification or hybridization can proceed249,266. There are three primary methods for a methylation-
dependent pretreatment of DNA: (1) endonuclease digestion, (2) affinity enrichment, and (3) bisulphite
conversion. After this step, various analytical steps including probe hybridization and sequencing proceed
in order to determine the locations and patterns of 5mC residues249.
The first of three approaches for methylation-dependent pretreatment of DNA is sequence-specific
restriction endonuclease, which works in a complex with DNMTs to methylate bases in the recognition
site as protection from the restriction defense system. The first locus-specific DNA methylation analyses
uses methylation-sensitive restriction endonucleases followed by gel electrophoresis and hybridization on
Southern blots; a method that is still used today for some locus-specific studies. Also currently in use is
PCR across the restriction site as an alternative subsequent step to methylation-sensitive restriction
digestion. This is a very sensitive method despite some risk of false-positive results caused by incomplete
digestion249,266.
The second approach involves chromatin immunoprecipitation (ChIP), microarray hybdridization (ChIP-
chip) or next generation sequencing (ChIP-seq) powerful techniques used primarily for genome-wide
studies of histone modifications and DNA methylation. These methods identify the location of DNA-
binding proteins and epigenetic marks in the genome. Another method of affinity enrichment involves
58
either 5mC-specific antibodies (in the context of denatured DNA) or methyl-binding protein with an
affinity for methylated native genomic DNA. This tool is particularly useful for DNA methylation profiling
in complex genomes. Despite the advantages of this approach, affinity-based methods do not yield
information on individual CpG dinucleotides and require substantial experimental or bioinformatics
adjustment for the non-uniform distribution of CpG sites at different regions of the genome249.
The third approach of treating denatured DNA with sodium bisulphite depends on the conversion of only
unmethylated cytosines to uracil (read as thymine when sequenced), while keeping methylated cytosines
intact. Illumina has adapted the bisulphite conversion method to their Infinium platforms (e.g. the Illumina
Infinium MethylationEPIC array267) used in this study. Bisulphite conversion is followed by whole-genome
amplification, then fragmentation, and hybridization of DNA samples to methylation-specific DNA
oligomers that are linked to individual bead types. Each bead type matches to a specific CpG site and
methylation state to provide the DNA methylation status of individual CpG sequences249.
The choice of method for DNA methylation analysis is complex and highly dependent on various factors
and needs of the study. For instance, the number of samples, quality and quantity of DNA, potential
sources of bias, desired coverage, and resolution are some considerations. Additionally, accounting for the
type of organism being studied is important; an array-based analysis would require the presence of a
suitable array for the species of interest, whereas, a sequence-based analysis would be applicable to all
species as long as a reference genome exists249.
2.3.5 Epigenetics and Health and Disease
The Developmental Origins of Health and Disease (DOHaD) is a theory that explains the fetal origins of
adult disorders through the involvement of epigenetics268. This theory proposes epigenetics as a possible
mechanism to explain the association between in utero environmental factors and infant/adult disease
59
outcomes. It suggests that metabolic, hormonal and/or nutritional changes in the intrauterine environment
cause epigenetic changes, either genome-wide or gene-specific DNA methylation patterns, which in turn
influence disease susceptibility and health outcomes in the offspring later in life269 (Figure 2.3.5).
Figure 2.3.5. DOHaD. Environmental factors can influence hormonal, metabolic or epigenetic changes in the mother altering gene expression and stability, which may influence chronic disease of offspring later in life.
Early observational studies have supported the DOHaD hypothesis, an example being the Dutch Winter
Famine Cohort where undernourished mothers produced offspring with increased risk of metabolic and
cardiovascular disease in adulthood270,271. The same Dutch cohort255and the Swedish Overkalix Cohort272
have also demonstrated that health outcomes of the grandchildren can be affected , suggesting that the
effects of maternal exposures persist transgenerationally.
A key component of the DOHaD is the predictive adaptive response (PAR) hypothesis273, which highlights
the concept of ‘matched environments’ between the pre and postnatal environments. If the developing
fetus can predict and adapt to the postnatal environment based on cues from the prenatal environment, it
will be better suited for the future environment. If, however, mismatch occurs, the risk of developing
adverse effects later in life increases. Epigenetic processes have been suggested as an underpinning
60
mechanism of PAR274 and several studies have supported this hypothesis275,276. For instance, a rodent study
reported that an obesogenic phenotype was prevented in offspring who were fed the same high vitamin diet
as their mothers’ (10 times the recommended level or 20 mg folic acid/kg), as male offspring born to high
folate-fed mothers and fed a normal (2 mg folic acid/kg) folate were shown to have the highest body
weight gain275. Another study, however, did not show support for the PAR hypothesis277. Female offspring
born to dams fed a high folate diet (20 mg folic acid/kg) had reduced body weight gain was associated with
smaller femoral area. Matching the high folate dam diet with a high folate diet of these female pups did not
correct for the reduced body weight gain of the high folate dam diet. Nevertheless, the mismatch in
maternal and pup diet did result in reduced femoral peak load strength and stiffness277. A possible reason
for this discrepancy could be attributed to varying critical periods of growth and development (either in
utero, postnatally, or both) of each cell, tissue type or organ278.
2.3.6 One-Carbon Nutrients and DNA Methylation
Epigenetics has emerged as a critical mechanism through which diet can influence the genome. There has
been a tremendous amount of interest in this area269,279,280 but more insight is still needed regarding the
effects of diets on epigenetic control, regulation, gene function, and ultimately human health and disease.
The influence of intrauterine one-carbon nutrients such as folate, B12, B6, and choline on DNA methylation
status and subsequent changes in phenotype and disease susceptibility later in life has garnered significant
research interest. Among these one-carbon nutrients, folate and B12 have been most extensively studied.
The primary reason for this is a biological mechanistic link that ties one-carbon nutrients with DNA
methylation, as illustrated in the folate and B12 pathway shown in Figure 2.1.3. Key elements in this link
include: SAM as the universal methyl donor in the body; MS as an enzyme that transfers a methyl group
from 5-MTHF to homocysteine to produce methionine; and B12 as a coenzyme to MS.
61
Given the involvement of folate and B12 in the transfer of methyl groups, alterations in concentrations of
these nutrients during pregnancy can potentially modulate the methyl donor ‘pool’ available for DNA
methylation reactions. Evidence from animal and human studies are summarised in Sections 2.3.7 and
2.3.8.
2.3.7 Maternal One-Carbon Nutrients on DNA Methylation and Health Outcomes in Animal Offspring
2.3.7.1 One-Carbon Nutrients on DNA Methylation on Growth and Development Outcomes
Animal studies have provided evidence of how one-carbon nutrients in utero may change DNA methylation
patterns, and how the exposure to these nutrients potentially affects offspring’s disease outcomes related to
cancer, metabolic syndrome, adiposity, insulin resistance and obesity (Table 2.3.7).
General methyl donors
One of the more well-known animal studies demonstrating the effect of maternal one-carbon nutrient
(folic acid, B12, choline and betaine) intake on offspring phenotype was conducted in the Agouti281,282.
Interestingly, CpG hypomethylation of the promoter region of Agouti gene in this model resulted in
maximal Agouti expression producing a yellow coat and more obese and prone to adult-onset obesity,
diabetes and tumorigenesis283,284. In contrast, offspring of dams fed ample one-carbon nutrients has
increased CpG methylation and showed a brown coat colour and were less obese and diabetic283,284.
Another well-cited animal model in this area is the AxinFused mouse model, where a similar methyl-rich
diet fed to dams produced a proportion of offspring with the kinked tail phenotype in the AxinFused mouse
model. This outcome was associated with higher CpG methylation in the promoter region of the AxinFused
gene284. The same group of investigators also found that a methyl donor-deficient diet administered to mice
for 60 days post-weaning resulted in loss of imprinting in IGF2, which plays an important role in growth
and development. This was considered a permanent effect on IGF2 as the loss persisted 100 days after post-
weaning even on a control diet containing methyl group donors284. Other animal studies have also shown
62
that methyl donor status in early life affects DNA methylation3,285–287 in genes related to mitochondrial
metabolism and phospholipid homeostasis, in PGC-1a (energy metabolism) 288, and in PTPN22 (immune
signalling) and PPARa (cholesterol and lipid metabolism) in rodent models289. Furthermore, methyl donor
supplementation during gestation showed differentially expressed metabolic genes in the liver and muscle,
as well as a change in IYD (related to thyroid metabolism) methylation patterns in the F2 generation of a
pig model290. The associated phenotypic effects included decreased back fat percentage, adipose tissue, and
fat thickness and increased shoulder fat percentage in the F2 generation that was fed the methylating
micronutrients290.
Folate
More than any other one-carbon nutrient, a large number of studies have investigated the effects of
maternal folic acid supplementation on changes in DNA methylation patterns and consequent phenotypic
or other health and disease outcomes. In mice heterozygous for folate binding protein gene (Folbp1+/-),
daily gavage of 5-formylTHF during periconception until 15.5 weeks gestation significantly decreased
genome-wide DNA methylation in the liver and brain of offspring291. In rats, folic acid supplementation
three-weeks prior to mating and through pregnancy and lactation decreased genome-wide DNA
methylation by 25% in the liver of pups in comparison to those from dams fed a control diet292. However,
genome-wide DNA hypomethylation was observed in placental tissue of hyperhomocysteinemic dams fed a
diet deficient in folate293. The same study found positive correlations between placental genome-wide
DNA methylation and folate concentrations in the placenta, plasma and liver293. However, not all studies
found changes in genome-wide DNA methylation in pups in response to maternal folic acid
supplementation or folate status. In a rodent study in the fetal liver maternal folate deficiency during
pregnancy did not change maternal and offspring genome-wide DNA methylation170. Nevertheless, a low-
methionine maternal diet did induce higher homocysteine concentrations and lower fetal and maternal
weight170.
63
In animal models of tumorigenesis, maternal folic acid supplementation has been shown to modify DNA
methylation and risk of cancer in the offspring. Maternal folic acid supplementation of dams with 5 mg/kg
folic acid (equivalent to ~1000 µg folic acid/day in humans) in utero during pregnancy and lactation
significantly decreased mammary genome-wide DNA methylation by 7%, and increased the risk of
mammary tumors in the offspring180. In a colorectal cancer animal model similar to the previous study by
Ly et al. 2011, contrasting findings were observed as maternal folic acid supplementation at the same level
and duration significantly increased colorectal genome-wide DNA methylation by 3%, and reduced the risk
of developing colorectal tumors in the offspring179. These findings suggest that the effect of maternal folic
acid supplementation on cancer risk may be organ specific and may be mediated in part by modifications in
genome-wide DNA methylation.
The effects of maternal diet, including folic acid supplementation, on fetal DNA methylation have also
been demonstrated at the gene-specific level. Promoter DNA methylation of the PPARg (regulator of
adipocyte differentiation) and GR (cellular proliferation and differentiation) genes was decreased by 20%
and 22.8%, respectively, while its protein expression increased in offspring from dams fed a protein
restricted diet during pregnancy294. Maternal folic acid supplementation prevented these changes294. PPARg
has recently been found to upregulate genes encoding chromatin modification enzymes such as histone
lysine methyltransferases that regulate transcription of adipogenesis295. Histone H3 at lys9 (H3K4) in
particular has been associated with resistance to obesity295. CpG methylation in the promoter of the GR
gene has been shown to be positively associated with increased adiposity and blood pressure in adulthood,
and this effect was more pronounced in children born to mothers with the most unbalanced diets based on
macronutrient intake in pregnancy296. Another study found that maternal folic acid supplementation
significantly decreased DNA methylation in the CpG sites of the promoter regions of the PPARg and ERa
(transcription factors involved in adipogenesis and glucose metabolism) genes, in exons 6 and 7 of the p53
64
gene (tumour suppressor), and in exon 15 of the APC gene (tumour suppressor) but not in p16 gene
promoter (tumour suppressor) in the liver of rat pups297. Increased maternal folic acid intake before and
during pregnancy was also found to produce loss of imprinting for IGF2 and H19 (tumour suppressor)
genes in human cord blood leukocytes, with more pronounced effects in males134.
B12
Studies examining the effects of maternal B12 status on genome-wide or gene-specific DNA methylation are
limited and are usually in combination with folic acid298. Similar to folate deficiency, B12 deficiency can
result in DNA hypomethylation as it is required for synthesis of methionine and SAM298. The effect of
maternal folic acid supplementation (4X the basal requirement) with or without B12 on placental genome-
wide DNA methylation was studied in rats299. This study found a decrease in genome-wide DNA
methylation associated with folic acid supplementation in the absence of B12, suggesting that the ratio of
folic acid and B12, rather than the actual supplemental level, may have an important role in determining
genome-wide DNA methylation patterns299. Periconceptional folate and B12 restriction was also
investigated in a sheep model, where aberrant DNA methylation patterns were reported in 4% of the 1400
CpG islands examined in the offspring in association with both maternal folate and B12. Specifically, 88% of
these CpG island sites were hypomethylated relative to the controls117. This study also found that the
offspring, most notably, males, born to mothers with folate and B12 restriction were obese and insulin
resistant, had elevated blood pressure, and exhibited altered DNA methylation at a number of sites in the
genome associated with adiposity, insulin resistance, and increased immune response117. Another study
showed that rat pups born to mothers deficient in folate and B12 during gestation and lactation were
associated with decreased birth weight, increased central fat mass, liver steatosis, and myocardium
hypertrophy300. These manifestations were linked to decreased expression of SIRT1 (intracellular protein
regulator) and PRMT1 (transfers methyl groups) and hypomethylation of PGC1-a (transcriptional regulator
of genes related to energy metabolism)300.
65
In summary, animal studies provide evidence that DNA methylation can be modulated by intrauterine
exposure to varying levels of one-carbon nutrients such as folate and B12, with resultant changes in
phenotype and other functional outcomes. However, the extrapolation of animal data to humans has
several limitations. For instance, genes investigated in mice may not be present in humans such as with Avy
and Axinfused genes283. Most animal models use inbred strains unlike human offspring that are genetically
and epigenetically diverse283. Furthermore, mouse models produce immature offspring where many of the
developmental processes that occur in humans during pregnancy are postnatal events in rats and mice, and
may as a result affect susceptibility of the animal offspring to the postnatal environment283. Despite these
limitations, animal studies contribute significant insights due commonalities of some human diseases
(diabetes, obesity, an cancer) to certain animal model systems283.
Table 2.3.7. Summary of in utero and perinatal folate and B12 nutrient supplementation on DNA methylation and related health and disease outcomes in the animal offspring Study Species One-carbon nutrient
source & levels Duration Tissue DNA
methylation Effect
Waterland et al.
2003282
Mice Avy/a
1) 5 g choline, 5 g betaine, 5 mg folic acid, 0.5 mg B12 2) 2.5 g choline, 2.5 g betaine, 2.5 mg folic acid, 0.25 mg B12
2 weeks prior to and throughout pregnancy and lactation
Tail tip Seven CpG sites in Pseudoexon 1A
CpG methylation increase (p=0.005) leading to coat color change (p=0.001)
Waterland et al.
2006284
Mice AxinFu
1) 5 g choline, 5 g betaine, 5 mg folic acid, 0.5 mg B12 2) 2.5 g choline, 2.5 g betaine, 2.5 mg folic acid, 0.25 mg B12
2 weeks prior to and throughout pregnancy and lactation
Tail tip Liver
AxinFu gene Increased CpG methylation (p=0.009) leading to less tail kinks (p<0.0001) Increased CpG methylation (p=0.05) leading to less tail kinks (p<0.0001)
Finnell et al. 2002291
Mice Folbp1+/-
25 mg/kg/d 5-formylTHF by gavage
2 weeks periconception to 15.5 wk gestation
Liver Brain
Genomic ~4-fold decrease (p<0.05)
~2-fold decrease (p<0.05)
Ly et al. 2011180
Rats 5 mg/kg folic acid vs. 2 mg/kg folic acid
3 weeks prior to mating through pregnancy and lactation
Mammary Genomic 7% decrease (P = 0.03)
66
Study Species One-carbon nutrient source & levels
Duration Tissue DNA methylation
Effect
Sie et al. 2011179
Rats 5 mg /kg folic acid vs. 2 mg/kg folic acid
3 weeks prior to mating through pregnancy and lactation
Colorectum Genomic 3% increase (P = 0.007)
Sie et al. 2013292
Rats 5 mg/kg folic acid vs. 2 mg/kg folic acid
3 weeks prior to mating through pregnancy and lactation
Liver Genomic PPARg p53 APC p16 ERα
25% decrease at weaning (p<0.001) At weaning, CpG methylation decreased in p53 exons 6 and 7, in APC in exon 15 but notp16 (p< 0.05)
Kim et al. 2009293
Rats 1) 8 mg/kg folic acid 2) 8 mg/kg folic acid & 0.3% Hcy 3) 0.3% Hcy
4 wks prior to day 20 of pregnancy
Placenta Liver
Genomic correlation with plasma and liver and placenta folate
r = 0.752 (p= 0.0003) r = 0.700 (p= 0.0012) r = 0.819 (p< 0.0001)
Lillycrop et al.
2005294
Mice 1) Control protein (180 g/kg protein plus 1 mg/kg folic acid) 2) Restricted protein (90 g/kg casein plus 1 mg/kg folic acid) 3) Restricted protein plus 5 mg/kg folic acid
Pregnancy Liver PPARg GR
PPARg decreased by 20.6% (p=0.001) GR decreased by 22.8% (p=0.01)
Maloney et al.
2007170
Rats 1) Control 2) Folate -/- 3) Folate -/- & low methionine 4) Folate -/- & low choline 5) Folate -/- & low methionine & low choline
2 weeks prior to mating to day 21 of gestation
Liver Genomic No change
Engeham et al.
2009301
Rats 1) Control: 18% casein, 1mg/kg folic acid 2) Control with folate: 18% casein, 5mg/kg folic acid 3) Low protein: 9% casein, 1mg/kg folic acid 4) Low protein with folate: 9% casein, 5mg/kg folic acid
Pregnancy Liver Genomic No change
Kulkarni et al.
2011299
Rats 8 mg/kg folic acid & B12 -/- vs. 2 mg/kg folic acid & B12 -/-
Pregnancy Placenta Genomic Decreased (P < 0.05)
Sinclair et al. 2007117
Sheep B12 & folate deficient vs. control
8 weeks prior to 6 days after conception Periconceptional
Liver 1400 CpG sites
4% of CpG sites had altered status
(P < 0.001)
Table modified from Plumptre et al. 2016 thesis33.
67
2.3.8 Maternal Folate and B12 status on DNA Methylation and Health Outcomes in Human Offspring
2.3.8.1 One-Carbon Nutrients on DNA Methylation
A limited number of human studies have investigated the effects of maternal folate and B12 concentrations
on DNA methylation patterns, but only a few looked at metabolic outcomes, in the offspring (Table
2.3.8). Focusing on studies that are based on maternal and cord blood samples, human studies have
generally reported inconsistent findings.
Folate
The majority of observational and clinical studies have investigated the relationship between maternal
folate exposures and DNA methylation of the offspring, and have shown equivocal findings. An
observational study in the Netherlands, with voluntary folic acid food fortification, reported a 4.5%
increase in DNA methylation in IGF2 DMR (differentially methylated region) derived from whole blood of
17-month old children whose mothers were supplemented with 400 µg/d folic acid compared with those
born to mothers who were not folic acid supplemented302. Independent of maternal folic acid
supplementation status, an inverse association was observed between IGF2 DMR methylation and birth
weight in the offspring302. In contrast, the recent North American Newborn Epigenetic Study (NEST)
found that DNA methylation levels at the IGF2 DMR were not significantly different between infants born
to folic acid supplemented and non-supplemented mothers303. However, the same investigators found
significantly lower H19 DMR methylation in infants born to mothers supplemented with folic acid
compared with those born to mothers not supplemented with folic acid303. In the Generation R study in the
Netherlands, a positive association between IGF2 DNA methylation and maternal folic acid
supplementation after the first trimester (>12 weeks gestation), but not with preconceptional or first
trimester (<12 weeks gestation) folic acid supplementation, has been reported304. However, this study
found no association between maternal folic acid supplementation and DNA methylation in 11 regions of
seven genes (NR3C1, DRD4, 5-HTT, KCNQ1OT1, MTHFR, IGF2 DMR, and H19)304. With similar
68
fortification practice as the Netherlands, a study in France did not find a correlation between the
combination of folic acid supplementation with B12 and B6 intake before or during pregnancy and DNA
methylation of the PLAGL1 DMR305. However, an American mother-child cohort found maternal folic acid
supplementation to be positively associated with PLAGL1 DNA methylation in cord leukocytes306. In a study
based in the United Kingdom, with similar mandatory folic acid food fortification as the US and Canada,
they reported an inverse association between long interspersed nucleotide element-1 (LINE-1) DNA
methylation and PEG3 methylation with maternal folic acid supplementation, only after the first
trimester307. Overall, the varying outcomes of these studies may be due to the timing of prenatal folic acid
supplementation during the gestational periods or from the folate status of the population that vary
depending on the country’s folic acid fortification practice.
B12
Some population-based studies have suggested that maternal B12 intake may have a stronger influence on
DNA methylation of the offspring than folate13,238,300, and are associated with infant growth and
development. A cross-sectional study reported that methylation of the P2 promoter of the imprinted IGF2
gene was inversely associated with maternal serum B12 but not with maternal or cord blood RBC folate
concentrations308. Increased maternal B12 during pregnancy was associated with decreased genome-wide
DNA methylation in newborns while increased B12 in newborns was associated with reduced DNA
methylation of IGFBP3 (involved in intrauterine growth) gene. Promoter regions of the SREBF1 and LDLR
(two key regulators of cholesterol biosynthesis) genes were hypomethylated in mothers with B12
deficiency, and as a result, the expression of downstream effector genes, including IDI1, SQLE, SC4MOL
and INSIF1, STarD4, and cholesterol biosynthesis were significantly upregulated235. Two US cohort studies
found that the association between BMI and DNA methylated HIF3A consistently increased across tertiles
of total B12, folate, and B2 intake, suggesting a potential causal relation between DNA methylation and
adiposity through interactions with B2, B12 and folate total or supplemental intakes309.
69
2.3.8.2 DNA Methylation on Health Outcomes
Several studies have investigated the effects of DNA methylation on offspring health outcomes with a
specific focus on metabolic diseases (Table 2.3.8).
Three recent reviews investigating the epigenetic effects on obesity310–312 and type 2 diabetes311,312 have
reported inconsistent associations between genome-wide DNA methylation and obesity and type 2
diabetes, although multiple obesity-associated differentially methylated sites in peripheral and cord blood
cells were identified. Gene-specific DNA methylation studies have identified DNA alterations in various
genes that are implicated in obesity, appetite control and metabolism, insulin signalling, immunity
inflammation, growth and circadian regulation295,313,314. Studies have also examined measures related to
metabolic syndrome and found candidate genes315–317 primarily associated with cortisol, adiposity and
inflammation315that are epigenetically modulated.
Genes related to BMI and body fat distribution have also been shown to be associated with changes in DNA
methylation318–323. For instance, DNA hypermethylation of the HIF3A gene (related to neonatal and infant
anthropometry) in blood and adipose tissue has been associated with greater infant weight and adiposity322.
In another study, 23 birth weight-associated CpG sites in 14 genes from cord blood were associated with
developmental phenotypic characteristics in childhood324. However, beyond early childhood, the study
found a lack of persistence in DNA methylation differences324. DNA methylation was also shown to affect
genes related to the downstream effectors of PPARy-induced genes that determine mesenchymal stem cell
differentiation into osteocytes and adipocytes, which may predispose the infant to fat accumulation325.
Similarly, studies have shown other measures such as waist-hip ratio326, blood lipids296,326,327, adiposity
distribution at 5 years of age328 and at young adulthood329 associated with DNA methylation changes.
70
While some of the aforementioned studies reported significant changes in DNA methylation patterns as
well as suggested possible functional outcomes in relation to birth weight, tumour development, obesity,
metabolic syndrome, and other chronic diseases; the potential link between the observed DNA
methylation changes and the aforementioned health outcomes have not been unequivocally established.
Furthermore, it is difficult to make clear conclusions regarding the associations between one-carbon
nutrient-mediated epigenetic changes in utero and the risk of chronic disease later in life due to various
factors that come into play. These factors include differences in study design, research models, techniques,
and populations’ one-carbon nutrient status and/or supplementation habits.
Table 2.3.8. Summary of in utero and perinatal folate and B12 nutrient status on DNA methylation and potentially related health outcomes in human offspring
Study One-carbon
nutrient source & levels
Duration Tissue DNA methylation
Effect on DNA methylation
Outcomes on Offspring
Fryer et al. 2009330
Folic acid supplement, 400 µg/day
Pregnancy
Cord blood lymphocytes
LINE-1
r = 0.364 (p = 0.08)
Birth weight
Cord serum folate, 15.8 µmol/l
r = 0.209 (p > 0.05)
Cord Hcy, 10.8 µmol/l
r = –0.688 (p=0.001)
Steegers-Theunissen et al.
2009302
Folic acid supplement, 400 µg/day
Periconceptional Whole blood at 17 months old
IGF2 4.5% higher, (p=0.014)
Inverse association between IGF2 and birth weight (1.7% methylation per SD birth weight; p=0.034)
Fryer et al. 2011331
Folic acid supplement, 400 µg/day
Pregnancy Cord blood lymphocytes
27,578 CpG loci associated with 14,496 genes
No association with supplementation CpG methylation correlated with plasma Hcy (p=0.04), and LINE-1 methylation (p=0.03)
CpG methylation correlated with birth weight percentile (P=0.02)
Ba et al. 2011332 Cord serum folate and B12, 7.29 ng/ml Maternal serum folate and B12, 2.29 ng/ml
N/A (Observational design)
Cord blood MNCs
IGF2 promoter2 IGF2 promoter3
No association with folate; P3 methylation associated with maternal serum B12 (mean change=-0.22, p=0.0014)
Not determined
Hoyo et al. 2011303
Folic acid users vs. non-users before or during pregnancy
N/A (Observational design)
Cord blood leukocytes
IGF2 DMR H19 DMR
No association 2.8% decrease (p = 0.03) before pregnancy 4.9% decrease (p = 0.04) during pregnancy
Chronic diseases related to loss of IGF2 imprinting* (high methylation, low birth weight)
71
Study One-carbon
nutrient source & levels
Duration Tissue DNA methylation
Effect on DNA methylation
Outcomes on Offspring
Boeke et al. 2012333
Dietary intakes of folic acid, B12, choline, betaine
Pregnancy Cord blood leukocytes
LINE-1 No association Not determined
McKay et al. 2012334
Maternal RBC folate, 379 ng/ml Maternal B12, 283 pg/ml
N/A (Observational design)
Cord blood LUMA Maternal B12 inversely associated with cord genome-wide DNA methylation (P=0.04).
Not determined
Haggarty et al. 2013307
Folic acid supplement users
Before and after 12 weeks gestation
Cord blood LINE-1 IGF2 PEG3 SNRPN
Supplementation >12 wks gestation associated with -0.3% (P=0.03) +0.7% (P=0.04) -0.5% (P=0.02) No association
Genes related to fetal growth, imprinting syndromes, obesity, metabolic syndrome, tumour development, and neurogenetic disorders
van Mil et al. 2014304
Folic acid supplement use Plasma folate, 19.1±9.0 nmol/L Plasma Hcy, median (90% range) 7.0 (4.9-10.9) µmol/L
N/A (Observational design)
Cord blood leukocytes
11 regions of seven genes NR3C1, DRD4, 5-HTT, IGF2 DMR, H19, KCNQ1OT1, MTHFR
No association with maternal homocysteine and folate concentrations, folic acid supplement use, and the MTHFR genotype
Genes for fetal growth and neurodevelopment candidate genes
Azzi et al. 2014305 Supplementation with folic acid alone, or the combo of folic acid, B6 & B12
Before or during pregnancy
Cord blood leukocytes
PLAGL1 No association with PLAG1 or C-peptide; positively associated with estimated birth weight at 32 wk gestation (r=0.15 P=0.01), weight at birth (r=0.09 P=0.15), and weight at 1 yrs (r=0.14 P=0.03).
Hoyo et al. 2014306
Quartiles of RBC folate concentrations
N/A (Observational design)
Cord blood leukocytes
IGF2 H19 DLK1/MEG3 PEG1/MEST PEG3 PEG10/SGCE PLAGL1 NNAT
Q2 < Q1 (P= 0.004) No association Q4 > Q1 (P= 0.001) No association Q2 > Q1 (P= 0.03) No association Q3 < Q1 (P= 0.01) No association
Birth weight (higher folate higher birth weight)
Dominguez-Salas et al. 2014335
Plasma folate, B12, B6, ACTB12 (holotranscobalamin), choline, betaine, methionine, homocysteine, SAM, SAH, DMG
N/A (Observational design)
Maternal blood near conception (one month prior to pregnancy)
BOLA3 LOC654433 EXD3 ZFYVE28 RBM46 ZNF67
Mean methylation higher in 2- to 8-month infants conceived in the rainy (nutrient-rich) season
Not determined
Adaikalakoteswari et al. 2015235
Supplementation of folic acid and B12 dietary intake
N/A (Observational design)
Cord blood Promoter regions of SREBF1 and LDLR IDI1, SQLE, SC4MOL and INSIF1, STarD4
Hypomethylated and upregulated in response to B12
insufficiency
Genes associated with cholesterol biosynthesis and transport regulators
Table modified from Plumptre et al. 2016 thesis33.
72
2.4 The PREFORM Study
A recent prospective study, the PREnatal FOlic acid exposuRe on DNA Methylation in the newborn infant
(PREFORM) Study, investigated the effect of maternal intakes and blood levels of one-carbon nutrients
including folate/folic acid and B12 on genome-wide DNA methylation in cord blood mononuclear cells
(MNCs). The PREFORM study is the parent study from which this present study was conducted. Briefly,
the PREFORM study aimed at determining the effects of maternal intakes and blood levels of one-carbon
nutrients on genome-wide DNA methylation (5mC) and hydroxymethylation (5hmC) markers in cord
blood MNCs. Assessment of baseline maternal and infant characteristics and blood concentrations occurred
during early pregnancy (12-16 wks of gestation) and preconception (previous 3 months). A 110-food item
modified and validated Block FFQ (NutritionQuest, Berkeley, CA)336 and Demographic and Health
Questionnaire (DHQ) were used to assesses baseline demographics and consisted of sections of the
Canadian Community Health Survey (CCHS) cycle 2.2195. Infant characteristics such as infant sex, birth
weight, length, head circumference, gestational age, size for gestational age, and 1 and 5-minute Apgar
scores were obtained from medical records. RBC folate was assessed using Elecsys Folate Assay (Roche
Diagnostics, Mannheim, Germany) and B12 by Access competitive-binding immunoenzymatic assay
(Beckman Coulter, Brea, CA). See Appendix A for analysis of blood biomarkers in maternal and
umbilical cord blood.
A summary of RBC folate and serum B12 concentrations at recruitment and delivery and in cord blood are
shown in Table 2.4a. As shown, maternal and cord blood concentrations of RBC folate are high, while
serum B12 concentrations are shown to be adequate.
Table 2.4a. Maternal and cord blood concentrations of folate and B12
Nutrient concentration At recruitment (12-16 wk gestation)
At delivery (37-42 wk gestation)
Cord blood
RBC folate (nmol/L) 2416 (2361, 2472) 2789 (2717, 2864) 2681 (2607, 2758) Serum B12 (pmol/L) 218 (208, 227) 168 (161, 175) 320 (299, 342)
Reprinted with permission from Plumptre et al. 2015142
73
The main findings of the PREFORM study have been published previously in Plumptre et al. 2015 and
Visentin et al 2016143,204. This study found that maternal serum B12 concentrations in early pregnancy were
positively correlated with 5mC content in cord blood MNCs (r=0.25, p=0.002), while RBC folate
concentrations in early pregnancy were positively correlated with 5-hmC (r=0.16, p=0.04). Additionally,
plasma betaine concentrations (r=-0.3, p=0.01) in early pregnancy were positively, whereas cord plasma
UMFA concentrations (r=-0.23, p=0.004) at early pregnancy were negatively, correlated with 5-hmC
content. In a sub-analysis, mothers with early pregnancy RBC folate concentrations in the highest folate
quartile (>2860 nmol/L) and serum B12 (<167 pmol/L) concentrations in the lowest quartile had
significantly lower 5mC content in cord blood MNC and higher birth weight in comparison with those
with lower RBC folate (quartiles 1-3) and higher serum B12 (quartiles 2-4) concentrations (Table 2.4b).
Table 2.4b. Genome-wide DNA methylation and hydroxymethylation and infant birth weight High Folate/Low
B12 (n=13) Moderate Folate/Marginal to
Adequate B12 (n=266) p-value*
Infant 5mC 5.16 � 0.28 % 5.50 � 0.41 % 0.0069 Infant 5hmC 0.025 � 0.015% 0.032 � 0.016% 0.208 Infant Birth Weight 3697 � 455 g 3393 � 451 g 0.001
*p-value represents statistical significance of differences in cord MNCs 5mC, 5hmC content and infant birth weight between groups of mothers using a t-test. High RBC folate >2860 nmol/L (quartile 4), low B12 <167 pmol/L (quartile 1). Quartiles 1-3 and 2-4 belong to the moderate folate/marginal to adequate B12 group.
In summary, the literature of my thesis highlights the relationship between nutrient exposure of folate and
B12 and infant health and disease through DNA methylation. Few studies have investigated this tripartite
relationship, specifically in combination with folate and B12, in humans. Also, few studies have investigated
this in a gene-specific manner in cord blood MNCs. As the PREFORM study examined the effects of
genome-wide DNA methylation and hydroxymethylation on one-carbon nutrients, this study takes a step
further and address the gaps in the literature by investigating folate and B12 on gene-specific DNA
methylation and its biological and clinical functional relevance to infant health and disease.
74
3- Rationale, Research Questions, Hypothesis and Objectives
3.1 Rationale
Folate and folic acid intakes and blood concentrations are higher presently in Canada than they have been in
the past 20 years65,133. This is mainly attributed to both the 1998 mandatory folic acid fortification and
widespread use of folic acid supplements193. In addition, WCBA are advised to take folic acid supplement
before and during pregnancy. The Canadian and US nationwide data indicate that the majority of women
take prenatal supplements mostly containing 1000 µg folic acid337. Similarly, B12 intakes are shown to be
adequate in pregnant women through intake of prenatal supplements containing B12. Nevertheless, studies
including our PREFORM study135,204, have found that despite adequate intakes, marginal to inadequate B12
concentrations were seen in pregnant women203,207. Exposure to high folate and low B12 has been associated
with adverse maternal and offspring health outcomes. Particularly, low B12 amidst a high folate
environment have been related to an increased risk for small-for-gestational-age babies12 and childhood
adiposity and insulin resistance at age six13. Mothers with the highest folate and lower B12 blood
concentration group reported to have children who were the most insulin resistant13. Despite these
findings, current evidence is insufficient to establish strong associations between maternal folate and B12
status and health outcomes. Folate and B12 are integral factors in the one-carbon metabolic pathway that
have the potential to affect DNA methylation through modulation of methionine synthase activity and the
concentration of S-adenosyl methionine, the universal methyl group donor for DNA methylation and over
100 other biological methylation reactions45. DNA methylation is an epigenetic process that affects gene
expression and stability, and is therefore important in human health and disease susceptibility252. The
Developmental Origins of Health and Disease (DOHaD) hypothesis suggests that intrauterine
environmental exposures can influence chronic disease susceptibility in the offspring later in life through
hormonal, metabolic or epigenetic changes268. DNA methylation is programmed in embryogenesis in order
to establish a unique DNA methylation pattern in the fetal genome6. Animal studies have shown that
75
maternal diet, including one-carbon nutrients, can alter DNA methylation in the offspring and affect
permanent phenotypic changes and disease susceptibility related to birth weight, adiposity and metabolic
disorder32,117,180,282,284. Human studies have also shown that maternal diet, including one-carbon nutrients,
can alter DNA methylation patterns in the offspring and affect infant health in terms of birth weight,
metabolism, and developmental processes235,304,306,307,330. However, the role of genome-wide and gene-
specific DNA methylation in linking intrauterine environmental dietary exposure and infant health and
disease has not been clearly elucidated.
Given the aforementioned considerations and access to DNA from the PREFORM study, the
overarching research question is: Can maternal B12 in combination with a high folate environment have an
epigenetic effect in the offspring, and influence infant health and disease outcomes?
76
3.2 Hypothesis
I hypothesize that:
1) Low maternal B12 will decrease gene-specific DNA methylation patterns in cord blood MNCs in a
high folate environment.
2) CpG methylation alterations associated with maternal B12 and folate status will be observed
primarily in genes related to intrauterine growth and development.
3) The observed CpG methylation alterations involved in intrauterine growth and development will
affect the expression of these genes.
3.3 Objectives
1) To determine gene-specific DNA methylation alterations in newborn infants born to mothers with
low B12 status in comparison to those born to mothers with higher B12 status in a high folate
environment.
2) To determine functional, biological, and clinical ramifications of altered gene-specific DNA
methylation patterns associated with the maternal low B12 and high folate status.
77
4- Effects of Maternal Vitamin B12 Status in Combination with High Folate Status on Gene-Specific DNA Methylation in Cord Blood Mononuclear Cells 4.1 Subjects and Methods
4.1.1 Study Design and Sample Selection for the Present Study
Using the PREFORM cohort, mothers were separated to cases and controls based on serum B12 and RBC
folate blood concentrations and matched based on a specific criteria shown in Table 4.1.1. The matched
cord blood DNA samples were analyzed for gene-specific DNA methylation using Illumina Infinium
MethylationEPIC array. CpG sites and their associated genes were identified with bioinformatics analysis
for methylation levels in cases and controls. Ingenuity Pathway Analysis was used for functional analysis to
identify biological, disease and physiological functions that are most significant to genes that are
differentially methylated. mRNA expression for candidate genes that are shown to be regulated by DNA
methylation was confirmed by Droplet Digital PCR (ddPCR) (Figure 4.1.1a). Cord blood MNCs RNA
extraction, quantity and quality were determined using NanoDrop-1000 (Thermo Scientific)
Spectrophotometer and Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, USA).
Figure 4.1.1a. Study Design. Baseline characteristics, dietary and blood sample data at early pregnancy was obtained from the PREFORM study. Data was used for sample selection, DNA methylation, functional and gene expression analyses. Adapted with permission from Plumptre et al. 2015143.
78
A total of 44 women were selected from the original PREFORM cohort using the propensity score. From
the 364 mothers, 15% (n=55) were lost to follow up mainly due to their transfer to another centre143 to
give 309 mother-child pairs (Figure 4.1.1b). They were matched based on RBC folate and serum B12
concentration cutoffs as well as a specific matching criterion. Cases were selected based on ‘inadequate’
serum B12 blood concentrations that are <148 pmol/L338 and controls based on ‘adequate’ B12 >148
pmol/L. This cutoff is used as a diagnostic cutpoint for B12 deficiency based on the 1999-2004 NHANES338
and is currently used in the CHMS339. RBC folate concentrations were >1800 nmol/L for both cases and
controls. This cutoff is based on1820 nmol/L post-fortification RBC folate concentration of the 90th
percentile of the 2005—2010 NHANES65. Using this cutoff allows for more consistent measurement of a
high folate status in WCBA, and subsequently, better evaluation of their health outcomes and their
potential offspring65. Furthermore, in this present study, 1800 nmol/L cutoff was used over a lower cutoff
because of the generally very high RBC folate concentration of mothers in the PREFORM study. A higher
cutoff was also not chosen in order to be more generalizable to cutoffs found in the literature investigating
maternal RBC folate status and offspring health and disease outcomes. Propensity score matching using SAS
9.4 (SAS Institute Inc., Cary, NC) was used to control for potential confounding variables for DNA
methylation between the cases and controls and determine the final number of matches (Table 4.1.1).
Missing DNA and RNA samples were removed from the matches, and together with a final propensity
score model matched by case to control ratio of 1:1, gave a total of 44 matched samples.
79
Figure 4.1.1b. Final sample selection and flowchart of the PREFORM participants. Cases and controls were matched based on specific criteria from the original PREFORM cohort of 309 mother-child pairs. Adapted with permission from Plumptre et al. 201633.
Seven variables were chosen a priori and controlled for by the propensity score as they have been shown in
the literature to affect DNA methylation. These variables are maternal plasma choline and B6
concentrations at early pregnancy143, maternal age340, maternal BMI at early pregnancy341, mode of
delivery342 and infant sex343 (Table 4.1.2). The Infant gestational age was removed from the matching
criteria as all preterm births <37 weeks were removed from the cohort. The remaining two variables of
maternal ethnicity344 and smoking345 during pregnancy were not controlled for as it did not satisfy the
Standardized Differences Calculation for continuous variables test and it restricted the propensity score
matches.
Delivered at SMH
n = 309
Case - control matches n = 44
Loss to Follow up
n = 55 (15%)
Enrolled n = 364
Consent to future contact 87%
The PREFORM: Case-Control Study
The PREFORM Study
Propensity Score matching Matching criteria
80
Table 4.1.1. Matching criteria for cases and controls
Variables Maternal Age Maternal BMI
Maternal plasma B6 Maternal plasma Choline
Mode of Delivery Infant Sex
Infant Gestational Age Maternal Ethnicity Pregnancy Smoking
The way the propensity score matches cases to the controls is through an algorithm that uses a ‘caliper’
value that is +/- 0.2*standard deviation346. The propensity score also provides a c-statistic as a measure of
how well the criteria, expressed in a regression equation, predicts case or control. Values range from 0.5
to 1.0, where a value close to 1 indicates that the cases and controls are very different from one another
and thus any differences in their outcomes may be due to differences in other variables instead of folate and
B12. A c-Statistic of 0.7 was produced by the final propensity score model shown below. This c-statistic
indicates that cases and controls are quite different; nevertheless, there is a fair amount of overlap in the
propensity scores between them. The final propensity score model used to generate the matched case with
controls and the total number of matches is shown in Appendix B.
In order to produce this final propensity score model with a c-statistic of 0.7, all nine variables were
assessed for redundancy to ensure that none of them was a restatement of the other. This was assessed
through understanding of each variable biologically and more formally with statistical tests. Chi-Squared
tests for dichotomous and T-tests for normally distributed continuous variables found significance with
pregnancy smoking (p = 0.0675), BMI (p=0.0507), and B6 (p=0.0002). Other variables were non-
significant with ethnicity (p=0.2233), infant sex (p=0.4187), mode of delivery (p=0.2432), maternal age
(p=0.7042), and choline (p=0.9136). Hence, these three variables were the focus of the propensity score
regression model equation to be controlled. Additionally, a Standardized Difference calculation was
performed in order to verify whether the propensity score matched accurately.
81
A three-interaction term was created in the regression model equation in order to control for pregnancy
smoking; however, the propensity score was unable to effectively control for this variable as the
Standardized Difference was > 0.1. Similarly, maternal ethnicity and BMI did not match accurately. BMI
was thus included in the ‘mvars’ term of the propensity score with a ‘dmaxk’ of 3. ‘Mvars’ creates BMI as
a covariate to the propensity score where the ‘dmaxk’ uses an exact matching method with a range of 3 to
pair cases with control. As BMI is depicted with ranges for each lean, overweight, and obese categories,
matching a ‘dmaxk’ by 3 is clinically acceptable. The two-way interaction terms were not included in the
regression model as the purpose was solely for matching and not for model creation.
4.1.2 Genome-Wide and Gene-Specific DNA Methylation and Functional Analyses
4.1.2.1 Illumina Infinium MethylationEPIC Array
The Illumina Infinium MethylationEPIC array was used for gene-specific methylation analysis,
interrogating 850,000 CpG methylation sites quantitatively across the genome at single-nucleotide
resolution for >99% of RefSeq genes347. DNA of each sample were processed at the SickKids TCAG Core
Facility (Toronto, ON). Briefly, 1 µg of genomic DNA was bisulfite-converted using the EZ-96 DNA
Methylation Kit (Zymo Research) according to manufacturer’s instructions. Unmethylated cytosines are
converted to uracil in the presence of bisulfite, while methylated cytosines are unaffected by deamination
of bisulfite and remain as cytosine. The bisulfite conversion included a thermoycling program with a short
denaturation step (16 cycles of 95°C for 30 seconds followed by 50°C for 1 hour). The amount of
bisulfite-converted DNA and completeness of bisulfite conversion were assessed using a panel of
MethylLight-based quality control (QC) reactions as previously described249. All samples passed QC tests
and were used for the Infinium HD Methylation Assay pipeline for whole-genome amplification (WGA)
and endpoint enzymatical fragmentation. The bisulfite-converted WGA-DNA samples were purified and
applied to the BeadChips and were incubated in order to hybridize. During hybridization, the WGS-DNA
molecules annealed to locus-specific DNA oligomers linked to individual bead types. The two bead types
82
corresponded to a methylated (C) or unmethylated (T) state of each CpG locus. This allele-specific primer
annealing was followed by single-base extension of detectable labels of DNP- and Biotin-labeled ddNTPs.
Both bead types for the same CpG locus incorporated the same type of labeled nucleotide, determined by
the base preceding the interrogated “C” in the CpG locus, and therefore would be detected in the same
color channel (Figure 4.1.2.1). After extension, the BeadChip was scanned using a HiScan System and the
intensities of the unmethylated and methylated bead types measured from the light emitted from the
fluorophores348.
Figure 4.1.2.1. The HD Infinium assay for methylation. Methylations status at each CpG loci is determined by M or U bead-bound probes that detect presence of C or T of the bisulfite-converted DNA. Methylated C is protected from conversion (left), while unmethylated C is converted to T (right). By hybridization, methylated and unmethylated bead-bound probes are used to identify CpG loci methylation. This is followed by single-base extension with a labelled nucleotide347. (https://www.illumina.com/ content/dam/illumina-marketing/documents/products/datasheets/humanmethylationepic-data-sheet-1070-2015-008.pdf)
4.1.2.2 Bioinformatics
A measure of the level of DNA methylation at each CpG site was scored as beta (b)-values using
IlluminaGenomeStudio 2011.1 software (Illumina Inc., San Diego, CA). DNA methylation b-values
represent the ratio of the intensity of the methylated bead type to the combined (methylated and
unmethylated) locus intensity ranging from 0 to 1. Values close to 0 indicate low levels of methylation,
while values close to 1 indicate high levels of DNA methylation267. Raw scanned data were normalized and
83
filtered based on detection p-values, non-CpG sites, cross-reactive probes, and sex chromosomes giving
804, 449 CpG sites. The detection p-values for each b-value measure the difference of signal intensities at
the interrogated CpG site and is compared to the standard deviation of 600 negative control probes to
establish detection limits for methylation probes. All data points with a detection p-value >0.05 were
considered not statistically significantly different from background measurements, and therefore filtered
out to ensure inclusion of only high-confidence probes. CpG sites that are differentially methylated were
determined using a Mann-Whitney non-parametric test due to the bimodal distribution of having either
hyper- and hypomethylated patterns349. Total CpG site data were filtered to 921 CpG sites (521 CpG-site-
associated genes) by methylation difference score (DiffScore) ³13 and £-13 and delta b-values ³0.1 and £-
0.1 based on intergroup differences in methylation levels between cases and controls. Average b-values for
each case and control group is a measure of the ratio of signal from the methylated probe to the sum of
both methylated and unmethylated signals. A Student’s T-test was used to obtain 98 CpG sites (69 CpG-
site-associated genes) that differed most between cases and controls, where statistical significance was
assessed with p-values and corrected for multiple comparison testing using a Benjamini-Hochberg
correction giving q-values350. A q-value is an ‘adjusted p-value’ indicating the level of false positives for a
given p-value and is thus used to control for false positives. A q-value greater than 0.05 implies that more
than 5% of significant tests are false positives. Fold change was used to determine methylation status of
each CpG site. Fold change represents the average difference to the power of two between cases and
controls that are part of the T-test. Statistical analysis was carried out using the SAS 9.4 software. The
Illumina Infinium MethylationEPIC DNA methylation b-values were represented graphically using heat
maps and boxplots generated by the Qlucore Omics Explorer (Qlucore, New York, NY) and
GenomeStudio 2011.1 softwares.
84
4.1.2.3 Gene Selection for Gene Expression Analysis
Based on bioinformatic and functional analyses, 7 hypomethyated and 7 hypermethylated genes in the cases
were selected for gene expression analysis. The genes were selected based on the most significant
DiffScore, greatest magnitude delta b-values, and most relevant biological functions to growth and
development.
4.1.3 Functional Analysis
4.1.3.1 Ingenuity Pathway Analysis
The functional analysis was performed using Ingenuity Pathway Analysis (IPA, Ingenuity® Systems,
Redwood City, CA; http://www.ingenuity.com) to identify diseases, networks, biological functions and
genes that were most differentially methylated between cases and controls. The Ingenuity Knowledge Base
contains findings and annotations from the literature as well as other sources including EntrezGene,
RefSeq, OMIM, GWAS database, Gene Ontology, Tissue Expression Body Atlas, NCI-60 Cell Line
Expression Atlas, KEGG Metabolic pathway information. The right-tailed Fisher’s exact test was used to
calculate a p-value determining the likelihood that each biological function, network and/or disease
assigned to the focus data set is due to chance alone351. For network analysis, IPA computes a score for each
network according to the fit of that network to the user-defined data set of focus genes based on a p-value
that is displayed as the negative log of that value. The score indicates the likelihood of the focus genes in a
network being found together due to random chance, where a score of 2 indicates that there is a 1 in 100
chance that the focus genes in a network are due to chance. Hence, scores of 2 or higher have at least a
99% confidence of not being due to random chance alone.
85
4.1.4 Gene Expression Analysis
4.1.4.1 RNA Extraction
RNAs from cord blood MNCs were extracted using RNeasy® Mini Kit (Qiagen, Cat#: 74104) according
to the manufacturer’s protocol. The protocol involved loosening MNCs by passing lysate through a 20-
gauge needle in a phenol/guanidine-based lysis buffer and subsequently with ethanol. RNA was recovered
using a RNA-binding-specific silica membrane within a spin-column. On-column DNA digestion using the
RNase-free DNA set (Qiagen) was also performed to ascertain prevention of DNA contamination within
the samples. RNA concentration was measured using a NanoDrop-1000 Spectrophotometer through
measurement of pertinent absorbance values and calculation of absorbance ratios. All samples had a ratio of
A260:A280 nm between 2.0 and 2.09, indicating pure RNA. The purified RNA was stored at -20°C for
further analysis.
RNA concentration and integrity was assessed using the Agilent RNA 6000 Nano Kit and the 2100
Bioanalyzer (Agilent Technologies) instrument, a microfluidics-based platform, was used for quantification
and quality of RNA. The RNA Integrity Number (RIN) score was generated on the Agilent software. For
gene expression analysis, 26 cord lymphocytes samples were available and only 8 samples had more usable
RINs. The RNA quality for the 4 cases matched to its 4 controls had variable RINs, ranging from 3.5 to
9.7, and was attributed to the collection and storage variability upon collection of cord blood MNCs at
birth.
4.1.4.2 Droplet Digital Polymerase Chain Reaction
cDNA was synthesized using the QuantiTect® Reverse Transcription Kit (Qiagen) according to the
manufacturer’s protocol. Droplet DigitalTM Polymerase Chain Reaction (ddPCR) was performed using the
QX200TM Droplet DigitalTM PCR System (Bio-Rad) according to the manufacturer’s instructions. Briefly,
ddPCR is a technique that partitions a sample into 20,000 smaller reaction vessels or ‘droplets’ and
86
measures fluorescence of the amplified target gene in each droplet. Fluorescence measurement thus gives a
positive and negative population of binary readouts, where ones (positive) indicates gene presence and
zeroes (negative) indicates gene absence. Since the presence of a gene in a given droplet is a random event,
the data generated fits a Poisson distribution. This removes the need for standard curves and allows for a
robust estimation of data dispersion from technical replicates in absolute quantities352. Compared to RT-
qPCR, ddPCR allows for increased sensitivity and accuracy in detecting gene expression when the gene
concentration is low and when the gene expression is previously known to be low in the target cell- or
tissue-species. Furthermore, absolute quantification removes the necessity for a housekeeping gene.
ddPCR reactions were performed in a 10 µl volume with EvaGreen with ddPCR supermix (Bio-Rad), 2 µl
of cDNA template (1 ng/µl), and 2 µl of candidate gene primers diluted 4X to 2.5 µM for FRMD4B,
ARGHEF12, and 10X to 1.25 µM for HLAE, PRKCH, SLC25A24 and CYFIP1. Primers and their forward and
reverse sequences are shown in (Table 4.1.4.2). cDNA and primer concentrations and volumes were
optimized through repeated adjustments until ‘rain’ or noise in the ddPCR plots were minimized and all
samples had approximately 20,000 droplets. Negative controls that do not contain sample cDNA should
have small fluorescence ranges for the negative population and positive population with low or no droplets.
20µl of the ddPCR reaction was loaded into the droplet generator to partition the sample into droplets.
The cycling conditions included: 2 mins at 50°C and 10 mins at 95°C which was followed by 40 cycles
involving 15 second denaturation at 95°C and 1 min primer annealing and fragment elongation at 62°C.
Three genes (DOCK10, SCD and APBB2) were run in slightly different conditions with 61°C annealing for 1
minute for optimized runs. QuantaSoft software (Bio-Rad, Hercules, CA) was used for analysis of
fluorescence based on positive and negative readings in each sample. The fraction of positive droplets
determines concentration of gene in copies/µl353. The mean (± stdev) concentration of gene copies/µl was
obtained from the 8 cases and 8 control samples, and their difference was calculated for each gene.
87
Table 4.1.4.2. Genes-of-Interest and Housekeeping gene primer sequences for ddPCR
Gene Symbol Sequence
HLA-E Forward: 5’-AGC TGT CTC AGG CTT TAC AAG-3’ Reverse: 5’-CTG TGA TAT GGA GGA AGA AGA GC -3’
LIF Forward: 5’-TCC AGT GCA GAA CCA ACA G-3’ Reverse: 5’-GAG CAT GAA CCT CTG AAA ACT G-3’
FRMD4B Forward: 5’- GTT TCC ATC TGT ACC ATC CAG T-3’ Reverse: 5’- TCC AGC ACA TTC TTC TTA CAC C-3’
MYT1L Forward: 5’- GCT CTG TGC TAT CCT GAT ACC -3’ Reverse: 5’-GCT GTG ATG GTT CTG GAC AT -3’
CYFIP1 Forward: 5’-CAT GCA GAT ACT GAG GCT GT -3’ Reverse: 5’-CTA TGA CTT CCT GCC CAA CTA C -3’
GRIN1A Forward: 5’- GCA GAA ACA ATG AGC AGC ATC-3’ Reverse: 5’- CAA GAA GTA ATG GCA CCG TCT-3’
DOCK10 Forward: 5’- GCT CTG CAT CTT CAG GTA CTG-3’ Reverse: 5’-GAT CTC TTG TTC TTC CCC AGT G -3’
PPAP2B Forward: 5’-GCT CTC ATC ATT GCA GTA AAA CC -3’ Reverse: 5’-TCG ACC TCT TCT GCC TCT T -3’
DRD4 Forward: 5’- GAA CCT GTC CAC GCT GAT G-3’ Reverse: 5’- TGC TGC CGC TCT TCG TC-3’
SCD Forward: 5’-GGA ATT ATG AGG ATC AGC ATG TG -3’ Reverse: 5’- CTC TTT CTG ATC ATT GCC AAC AC-3’
ARHGEF12 Forward: 5’-ACC AGA GTT CCA TTC ACC TTG -3’ Reverse: 5’- ACA ATC CAG TCT TCG TAC AGT C-3’
SLC24A24 Forward: 5’-CAG TGC TTG TTC GAG AGA CA -3’ Reverse: 5’- GCT TAA CTA TTC CAG ATG AAT TCA CG-3’
PRKCH Forward:5’-GTG CTT CAT CAA AAC GAC GAG -3’ Reverse: 5’-CAG TTG TTC TGC TGC TTT CAG -3’
APBB2 Forward: 5’- CTG ATG CCA CTG TGA CTG TC-3’ Reverse: 5’- CAT GAT GAA GGC AAA TGT GTG G-3’
88
4.1.5 Statistic Analysis
For sample selection of maternal dietary and supplemental intakes, and blood concentrations, a paired T-
test (for continuous variables) and a McNemar’s test (for categorical characteristics) was used for
comparison between case and control means. A Wilcoxon Signed-Rank Sum test was used for skewed
distributions. For genome-wide DNA methylation CpG and non-CpG island analyses, statistical
significance was determined using a paired T-test for comparison of average b-values for case and control.
The results were considered statistically significant if the p-values were < 0.05. A Wilcoxon Signed-Rank
Sum test was also used for total and non-CpG island methylation due to its skewed distribution. Analyses
and graphical representations were done using GenomeStudio 2011.1 and SAS 9.4. For gene-specific
methylation, detailed information on data filtering, normalization and statistical analysis is presented in
Section 4.1.2.3. Qlucore Omics Explorer was used to calculate the largest difference between the cases
and controls, and a paired T-test was used to compare difference between the groups. For gene expression
analysis for ddPCR, results were analyzed using mean copy number difference using a paired T-test.
89
4.2 Results
4.2.1 Subject Characteristics
4.2.1.1 Matched Cases and Controls
From the original PREFORM cohort of 309 women, 22 cases were matched to 22 controls based on RBC
folate and serum B12 concentration cutoffs as well as the specific matching criteria. Cases were selected
based on ‘inadequate’ B12 blood concentrations that are <148 pmol/L338 and controls based on ‘adequate’
B12 >148 pmol/L. RBC folate concentrations were >1800 nmol/L for both cases and controls. Mean
dietary and supplemental intakes and blood concentrations of folate and B12 in early pregnancy for cases and
controls are shown in Table 4.2.1.2 b and c.
4.2.1.2 Maternal and Infant Characteristics
Baseline maternal characteristics for cases and controls are listed in Table 4.2.1.2a. Maternal age, BMI,
and plasma B6 and choline concentrations were matched based on propensity score. On average, mothers
were 32 years of age and had a BMI of 27 at recruitment. The ethnicities of mothers were Caucasian
(59%), Asians (16%) included Chinese, Filipino, Japanese, Korean and Southeast Asian, Latin Americans
(9%), and other ethnicities (16%) consisted of Black and Aboriginal. Twenty-three % of women reported
completing a college/vocational diploma and 59% completed a university degree or more and are thus
generally well educated. Few reported household income below the poverty line (9%). Even fewer were
single mothers (7%). Seventy percent of mothers experienced their first pregnancy (gravidity = 1) and
36% were nulliparous (parity = 0). Mothers who smoked during pregnancy were minimal (7%). No
significant differences were found between cases and controls using a paired T-test and a McNemar’s test.
90
Table 4.2.1.2a. Baseline maternal characteristics Descriptive characteristics Cases (n=22) Controls (n=22)
Age (y) [mean (95% CI)] • 31 (29, 34) 33 (31, 34) Gestational Age at recruitment (wks) BMI at recruitment (kg/m2) [mean (95% CI)] •
13.3 (12.5, 14.0) 26.7 (24.3, 29.2)
13.5 (12.6, 14.4) 26.7 (24.1, 29.2)
Ethnicity (n) Caucasian 14 12 Asian 2 5 Latin American 2 2 Other 4 3
Born in Canada (n) 8 10 Education Level (highest degree/diploma attained) (n)
High School Diploma College/Vocational Diploma University degree or higher
6 6
10
2 4
16 Household income below poverty line (n) 2 2 Single parent family 2 1 Gravidity
>1
14
17 Parity (%)
0 1 >1
8 7
14
8
13 1
Smoker 2 1 Marked (•) characteristics belong to the matching criteria used in the propensity score analysis to matched cases with controls. p>0.05 based on a paired T-test and a McNemar’s test
Maternal dietary intakes and blood concentrations in early pregnancy are listed in Table 4.2.1.2b and c.
Relative dietary intakes (naturally occurring and folic acid as a fortificant) without supplementation of one-
carbon nutrients, except for B12, were shown to be higher in the cases compared to the controls. Statistical
significance was found using a paired T-test for folate with a mean difference of 218 µg DFE/d (p=0.004),
B6 with 0.7 mg/d (p=0.01), and choline with 113 mg/d (p=0.006). B12 was not significantly different
between cases and controls (p > 0.05). Focusing mainly on folate and B12 relative intakes, prevalence of
inadequacy from diet alone during pregnancy was 18 of 44 (41%) and 3 of 44 (7%), respectively.
However, adequate folate and B12 intakes were observed in all mothers owing to the use of prenatal
supplement during pregnancy. Supplemental intakes (mean (95%CI)) of folic acid were 815 µg/day (578,
1052) for cases and 1423 µg/day (788, 2058) for controls. For B12, they were 11.9 (8.9, 32.6) µg/day for
91
cases and 5.0 µg/day (2.9, 7.1) for controls. For B6, supplemental intake was also adequate with 1.8 mg/d
(0.9, 2.8) for cases and 3.9 mg/d (2.1, 5.8) for controls. There was no recorded information on choline
supplemental intake as it was not included in the supplements. Based on paired T-tests for folic acid, B12
and B6, there were no significant differences observed between case and control supplemental intakes
(p>0.05).
Table 4.2.1.2b. Maternal dietary intakes in early pregnancy
Dietary Intake Cases
[Mean (95% CI)] Controls
[Mean (95% CI)] EAR for
Pregnancy Folate (µg DFE/d)* 674 (564, 785) 456 (399, 512) 520
B12 (µg/d) 5.6 (4.3, 7.0) 4.4 (3.6, 5.1) 2.2 B6 (mg/d)* 2.2 (1.8, 2.7) 1.5 (1.3, 1.7) 1.6
Choline (mg/d)* 353 (286, 420) 250 (216, 264) 450 (AI) DFE, dietary folate equivalent; EAR, estimated average requirement; AI, adequate intake. * p<0.05 based on a paired T-test
In early pregnancy, overall use of prenatal B-vitamin supplements containing folic acid, B12, B6, other
vitamins and minerals for cases and controls were 71% and 91%. Use of B vitamin-containing supplement
of folic acid, B12, B6 containing supplement designed for non-pregnant adults at least once each week for
cases and controls was 81% and 100%, multivitamins 5% and 0%, and folic acid supplementation alone
15% and 9%, respectively. The median supplemental intakes reported during early pregnancy was 1000
µg/d of folic acid and 2.6 mg/d of B12 for cases and controls. Women consuming a combination of ≥2 of
the B vitamin–containing supplements were >77% for both cases and controls.
Adequacy of nutrient status, considering dietary and supplemental intakes, are assessed using blood
concentrations. Mean (95%CI) RBC folate and serum B12 concentrations in early pregnancy were 2682
nmol/L (2506, 2858) and 120 pmol/L (112, 128) for cases and 2512 nmol/L (2324, 2700) and 253
pmol/L (220, 286) for controls, respectively. Folate, B6 and choline are above while only B12 control are
below the inadequacy cutoffs. Serum B12 concentration was significantly different between cases and
controls with a mean difference of -133 pmol/L (p<0.0001).
92
Table 4.2.1.2c. Maternal blood concentrations in early pregnancy
Blood Concentration Cases
[Mean (95% CI)] Controls
[Mean (95% CI)]
Inadequacy Cutoff
RBC folate (nmol/L) 2682 (2506, 2858) 2512 (2324, 2700) <305 Serum B12 (pmol/L)* 120 (112, 128) 253 (220, 286) <148 Plasma B6 (nmol/L) • 81 (67, 95) 81(58, 103) <20
Plasma choline (nmol/mL) • 7.2 (6.7, 7.9) 7.1 (6.4, 7.8) n/d
Marked (•) characteristics belong to the matching criteria used in the propensity score analysis to matched cases with controls. n/d (not determined); there is currently no set cutoff for plasma choline204 p<0.0001 based on a paired T-test
Newborn characteristics are listed in Table 4.2.1.2d. Overall, all newborns are born full term (>37
weeks gestation) with normal Apgar score. Infants in the case group are shown to have lower Apgar five
(p=0.03) scores based on the Wilcoxon Signed-Rank Sum test, and lower Apgar one (p=0.05) based on a
paired T-test. There were no significant differences in birth weight, head circumference and birth length
between cases and controls. Birth weight percentile was found to be borderline significant based on the
Wilcoxon Signed-Rank Sum test with a p-value of 0.055. None were diagnosed with gestational diabetes
and pregnancy induced hypertension, but one woman had an NTD-affected pregnancy.
Table 4.2.1.2d. Newborn characteristics
Descriptive characteristics Case (n=22) Controls (n=22) Gestational age at birth (wks) • 39.8 (39.3, 40.3) 39.9 (39.5, 40.3) Caesarian delivery, n• 3 3 Male, n• 15 14 Birth weight (g) 3634 (3429, 3840) 3465 (3304, 3627) Birth weight percentile (%)*t 80 (70, 89) 69 (58, 80) Birth length (cm) 52.9 (51.7, 54.2) 52.4 (51.3, 53.6) Head circumference (cm) 34.7 (33.9, 35.4) 34.2 (33.7, 34.8) 1 min. Apgar score*t 7.9 (6.9, 8.9) 8.9 (8.8, 9.0) 5 min Apgar score* 8.6 (7.9, 9.3) 9.1 (8.9, 9.3)
Marked (•) characteristics belong to the matching criteria used in the propensity score analysis to matched cases with controls. Values expressed at mean (95% CI) *p ≤ 0.05 based on a Wilcoxon Signed-Rank Sum test for 5 min Apgar score *tp ≤ 0.05 paired T-test for 1 min Apgar score and birth weight percentile.
93
4.2.2 Genome-wide DNA Methylation
Since the Illumina Infinium MethylationEPIC array contains additional probes that are in non-CpG islands,
which are more hypermethylated250 (see Appendix C), a separate genome-wide methylation was
performed for CpG islands and non-CpG islands. Total CpG site refers to the entire CpG site probes that
are for both CpG islands, including sites at the north and south shores and shelves, and non-CpG islands.
There was a statistically significant difference in genome-wide DNA methylation between cases and
controls for total CpG sites, and when separately analysing CpG island and non-CpG island methylation. A
Wilcoxon Signed-Rank Sum test was used for total and non-CpG island methylation due to its skewed
distribution. We found a mean difference of –0.0059 ±0.0147 (p= 0.013) for the total 804,449 CpG sites
and –0.0078 ±0.0147 (p=0.015) for non-CpG islands only. Significance was also found for CpG islands
using a Student’s paired T-test with a mean difference of –0.0036 ±0.0069 (p= 0.022) (Table 4.2.2.1).
Table 4.2.2.1. Genome-wide methylation mean ß-value difference between cases and control for the total, CpG island and non-CpG island based on a paired T-test and Wilcoxon Signed-Rank Sum test Mean ß-value Difference (n=22) p-value Total (804,449) -0.0059±0.0147 0.013* CpG Island (351,400) -0.0036±0.0069 0.022*t Non-CpG Island (453,049) -0.0078±0.0147 0.015*
*p-value based on a Wilcoxon Signed-Rank Sum test due to skewed distribution. *t p- value based on the paired-T test.
4.2.3 Gene-Specific Methylation
4.2.3.1 Genes differentially methylated between the cases and controls in cord blood MNCs
Qlucore Omics Explorer was used to calculate the largest difference between the cases and controls, and a
paired T-test was used to compare difference between the groups. 98 CpG sites and 66 CpG-associated
genes (Table 4.2.3.1a) were identified that were differentially methylated with a p-value of 0.007. A
Benjamini-Hochberg correction for multiple testing was applied to each CpG site (804,449 sites) resulting
in a q-value of 0.999.
94
Table 4.2.3.1a. Summary of the number of genes differentially methylated in cord blood MNCs relative to the control based on a T-test
Hypermethylated Hypomethylated Total T-test (p=0.007 q= 0.999)
21 (0.08%)
45 (0.17%)
66 (0.25%)
The numbers in brackets indicate the percentage of genes differentially methylated relative to total genes (25,159) targeted in the Illumina Infinium MethylationEPIC Array.
DNA methylation patterns of ß-values for CpG sites and their associated genes ‘CpG-site-associated-genes’
for the cases and controls are illustrated using an unsupervised hierarchical clustering approach, using the
Euclidean distance method for clustering, and a heat map (Figure 4.2.3.1a). The heat map revealed
distinctively different DNA methylation profiles between the cases (dark green) and controls (light green).
Cases are shown to be hypomethylated (blue squares) in 75% of the CpG sites. Infant anthropometric
measures were compared using a cutoff based on the mean of the 44 subjects, where 1 indicates below the
mean and 2 indicates above the mean. Infant sex, infant gestational age and chip are distributed randomly
between cases and controls indicating no sex, gestational age and batch effect between groups. Head
circumference (‘HC’), birth weight (‘BW’), and birth weight percentile (‘BW%ile’) are shown to be
greater in the cases than in the controls (Figure 4.2.3.1a). Although, borderline significance was only
observed for birth weight percentile based on a Wilcoxon Rank-Sum test. Birth length (‘BL’) did not show
patterns in distribution in either the case or control groups.
95
Figure 4.2.3.1a. Unsupervised hierarchical clustering and heat map of DNA methylation ß-value in the case and control groups. Each row is a CpG site and each column is either a case or a control subject. Groups in dark green are the cases and light green are the controls. Blue squares within the heat map indicate hypomethylation and yellow squares indicate hypermethylation. For infant sex, males are light blue and females are burgundy. For head circumference, pale green indicates below the cutoff, orange is above, and white indicates missing. For birth weight, magenta indicates below the cutoff and light purple for above. For birth weight percentile, sky blue indicated below the cutoff and dark orange for above. Subject identification is shown in the bottom axis and CpG-site-associated-genes are shown in vertical axis.
Subject ID
Genes
96
In addition to the T-test, a Mann-Whitney U-test (MWU) was used to filter the most significant and largest
magnitude CpG sites based on diffscore and Ƨ values, respectively. 912 CpG sites were identified with
552 sites hypomethylated and 360 hypermethylated. 381 genes were associated with CpG sites with 219
genes hypomethylated and 162 genes hypermethylated (Table 4.2.3.1b).
Table 4.2.3.1b. Summary of the number of genes differentially methylated in cord blood MNCs relative to the control based on a Mann-Whitney U-test
Hypermethylated Hypomethylated Total Mann-Whitney U-test (DiffScore = >|13| Ƨ = >|0.1|)
162 (0.6%)
219 (0.9%)
381 (0.15%)
A principal component analysis (PCA) plot was used to identify the largest variation in the data. The first
principal component (PCA 1) explains most of the variation, the second (PCA 2) or any other subsequent
principal components accounts for any remaining variation. Our PCA analysis demonstrated that 23% of
the variation in the data is explained by the difference in B12 status of cases and controls (Figure 4.2.3.1c).
PCA 1 (x-axis) represented the largest variation in the data where the cases (dark green) and controls (light
green) are dispersed in two clusters. PCA 2 (y-axis) represented 6% of the variation in the data. These can
be due to other environmental factors such as maternal or infant characteristics.
97
Figure 4.2.3.1c. Principal component analysis plot of DNA methylation. PCA plot indicates the largest variation clusters in the data of ß-values for 804,449 CpG sites. Points represent all 44 subjects where blue dots are the cases and yellow dots are the controls. PCA 1 is the x-axis and PCA 2 is the y-axis.
4.2.3.2 Functional analysis on differentially methylated genes of the cases and controls in cord blood MNCs
Functional analysis on the 66 (T-test-generated), 381 (total MWU test-generated), 219 (hypomethylated
MWU-test-generated), and 162 (hypermethylated MWU-test-generated) genes were performed using IPA
to identify disease, molecular, cellular and physiological processes most relevant to the differentially
methylated genes. Some genes were assigned to more than one category. These four datasets were
generated, as mentioned in Table 4.2.3.1a and b, using the T-test and the MWU test. The differences
between the three MWU-test-generated datasets is in that functional analysis was performed in the total
number of genes (381), and separately as hypomethylated (219) and hypermethylated (162) genes in order
to better elucidate functional processes specific to gene methylation. The top five molecular, cellular,
physiological functions and disease processes for the 66 genes are presented based on significance. Cell and
molecular functions were associated with genes in metabolism, cell cycle, molecular transport and small
2 (6%)
1 (23%)
98
molecule biochemistry. Genes related to behaviour, endocrine and nervous system, and embryonic and
hair and skin development were physiological system development and functions. Diseases and disorders
were associated with genes in cancer, neurological, cardiovascular, hematological diseases and organismal
abnormalities (Table 4.2.3.2a).
Table 4.2.3.2a. The top molecular, cellular, physiological functions and disease processes for the 66 hyper- and hypomethylated genes
Category p-value range No. of Genes Molecular and Cellular Functions
Carbohydrate metabolism 2.22-2 – 1.08-3 3 Lipid metabolism 3.47-2 – 1.08-3 7 Molecular transport 2.53-2 – 1.08-3 11 Small molecule biochemistry 3.47-2 – 1.08-3 11 Cell Cycle 2.84-2 – 3.20-3 4
Physiological System Development and Function Behaviour 2.84-2 – 3.20-3 3 Embryonic 4.39-2 – 3.20-3 5 Endocrine 3.20 -2 – 3.20-3 1 Hair and Skin 2.84-2 – 3.20-3 2 Nervous system 4.39-2 – 3.20-3 3
Diseases and Disorders Cancer 4.60-2 – 3.54-4 59 Neurological 4.39-2 – 3.54-4 24 Organismal Injury and Abnormalities 4.60-2 – 3.54-4 60 Cardiovascular 4.39-2 – 1.24-3 8 Hematological 4.70-2 – 1.24-3 19
Five networks were identified and ranked by a network score in the p-value calculation of the IPA, which
ranged from 2 to 63. The highest-scoring network of 63 revealed a significant link with cell death and
survival, cellular development, cellular growth, and proliferation. The remaining three networks had a
score of 2 and were linked to cancer, hematological, immunological, cardiovascular disease, cellular
assembly and organization, DNA replication, recombination and repair, and cell signalling. The network
score indicates the likelihood of the Focus Genes in a network being found together due to random chance.
The Focus Genes indicate the uploaded genes of interest for which information is available in the Ingenuity
Knowledge Base.
99
The top five molecular, cellular, physiological functions and disease processes for the 219 hypomethylated
and 162 hypermethylated genes are presented based on significance. Common cell and molecular functions
are associated with genes in cell cycle. Physiological system development and functions are associated with
genes in connective tissue development. Diseases and disorders are associated with genes in cancer,
organismal abnormalities, gastrointestinal disease and endocrine system disorders (Table 4.3.3.2b).
Table 4.2.3.2b. The top molecular, cellular, physiological functions and disease processes for the 219 hypomethylated and 162 hypermethylated genes
Hypomethylated Hypermethylated
Category p-value range No. of Genes Category p-value range
No. of Genes
Molecular and Cellular Functions
Cellular movement
4.76-2 – 2.75-5 32
Cellular growth and
proliferation
4.85-2 – 2.95-4 24
Lipid metabolism 4.76-2 – 2.75-4 11 Cell cycle 4.85-2 – 1.70-3 10 Molecular transport
4.76-2 – 2.75-4 9 Cell death and survival
4.23-2 – 2.15-3 18
Small molecule biochemistry
4.76-2 – 2.75-4 21 Gene expression
7.07-3 – 4.47-3 6
Cell Cycle 4.76-2 – 4.42-4 17 Cellular development
4.85-2 – 4.70-3 23
Physiological System Development and Function Tissue 4.06-2 – 2.47-3 17 Hematological 4.52-2 – 2.95-4 9 Connective tissue 3.83-2 – 4.86-3 6 Lymphoid 2.38-2 – 4.70-3 8 Skeletal and Muscular
3.83-2 – 4.86-3 4 Reproductive 3.49 -2 – 5.58-3 4
Cardiovascular 4.76-2 – 5.05-3 16 Immune cell trafficking
4.52-2 – 7.05-3 7
Organismal 4.76-2 – 5.05-3 19 Connective tissue
4.85-2 – 7.07-3 3
Diseases and Disorders Cancer 4.72-2 – 1.60-6 176 Cancer 4.85-2 – 6.56-5 122 Organismal Injury and Abnormalities
4.76-2 – 1.60-6 180
Organismal Injury and
Abnormalities
4.85-2 – 6.56-5 126
Gastrointestinal 4.66-2 – 1.92-6 156 Endocrine 4.85-2 – 9.26-5 66 Hepatic 4.66-2 – 1.92-6 94 Gastrointestinal 4.85-2 – 9.26-5 116 Endocrine 4.72-2 – 1.10-4 74 Metabolic 4.85-2 – 9.26-5 28
100
Five networks were identified and were ranked by a network score in the p-value calculation of the IPA,
which ranged from 2 to 67 for hypomethylated genes and 60 for hypermethylated genes. The highest-
scoring networks for hypomethylated genes of 67 and 65 are linked with cell death and survival, cellular
development, growth and proliferation, and cellular movement. The remaining three networks have a
score of 17 and 2 and show similar related links in addition to organismal injury and abnormalities, free
radical scavenging, and molecular transport. For hypermethylated genes, network scores of 60 and 56 are
linked with organismal, cellular growth and proliferation, gene expression, cell-to-cell signaling and
interaction, and cardiovascular system development and function. Lower-ranking network scores of 2 are
linked to cellular assembly and organisation, DNA replication, recombination, repair, cell cycle, cancer,
cell death and survival and morphology. Summary table of IPA functions for common genes for all three
gene groups, including 371 total MWU-test functional analysis, are shown in Appendix D.
Common genes between generated by the T-test and MWU-test are illustrated using a Venn diagram
(Figure 4.2.3.2). 27 genes overlap where 15 are hypomethylated and 12 are hypermethylated in the case
group.
Figure 4.2.3.2. Number of genes generated by T-test and MWU-test that are commonly differentially methylated in cord blood MNCs.
381 27 66 MWU-test genes T-test genes
101
4.2.4 Gene Expression
4.2.4.1 Candidate genes selected for gene expression analysis
A total of 14 genes were selected for gene expression analysis, of which 7 are hypo- and 7 are
hypermethylated relative to the controls (Table 4.2.4.1). Genes were selected from the 27 overlap genes
common to both the T-test and MWU-test generated genes. The most hypo- and hypermethylated genes
based on fold change were selected. Genes were also chosen depending on functional relevance to growth,
development and metabolism based on IPA analyses. Gene expression in cord MNCs was estimated from
lymphocytes, monocytes and B cells in peripheral blood using BioGPS database (http://biogps.org). This
was assessed for all candidate genes to determine whether the genes were expressed in MNCs and in what
level, so that we can estimate what gene expression level we can expect for ddPCR.
102
Table 4.2.4.1. Hypo- and hypermethylated genes selected for gene expression analysis
Gene names and biological functions were obtained from PANTHER database354
Genes
Gene Names Top Biological Function(s)
Hypomethylated PPAP2B
Lipid phosphate phosphohydrolase 3
biosynthetic process; nitrogen and phospholipid compound metabolic processes; regulation of biological process; response to stimulus; signal transduction
DRD4
D(4) dopamine receptor heterotrimeric G-protein signalling pathway; nicotine pharmacodynamics pathway; dopamine receptor mediated signalling pathway
SCD
Acyl-CoA desaturase gene expression regulation; lipogenesis; regulation of mitochondrial fatty acid oxidation; energy homeostasis; biosynthesis of membrane phospholipids; cholesterol esters and triglycerides
ARHGEF12 Rho guanine nucleotide exchange factor 12
protein binding; small GTPase regulator activity; guanyl-nucleotide exchange factor activity
SLC25A24
Calcium-binding mitochondrial carrier protein
biosynthetic and cellular process; translation; amino acid transporter
PRKCH
Protein kinase C eta type intracellular signal transduction; phosphate-containing compound metabolic process; regulation of biological process; response to stimulus; pathways related to cardiovascular, nervous systems and growth
APBB2
Amyloid beta A4 precursors protein-binding family B member 2
biosynthetic and cellular process; nitrogen compound metabolic process; regulation of nucleobase-containing compound metabolic process; DNA-dependent transcription
Hypermethylated HLA-E
HLA class I histocompatibility antigen, alpha chain E
antigen processing and presentation; major histocompatibility complex antigen
LIF Leukemia inhibitory factor antigen processing and presentation
FRMD4B
FERM domain-containing protein
cellular component morphogenesis and process
MYT1L
Myelin transcription factor 1-like protein
regulation of transcription from RNA polymerase II promoter
CYFIP1 Cytoplasmic FMR1-interacting protein 1
signal transduction; Huntington disease pathway (tumour protein p53)
GRIN2A
Glutamate receptor ionotropic, mitochondrial
signal transduction; Huntington disease pathways (ionotropic and metabotropic glutamate receptor pathways)
DOCK10
Dedicator of cytokinesis protein 10
cellular component movement; intracellular protein transport and transduction; phagocytosis
103
4.2.4.2 Gene expression analysis based on ddPCR
All genes showing expression are based on a mean copy number difference that is relative to the control.
Nine out of 14 genes showed an increase in gene expression regardless of whether the DNA methylation
status was hypo- or hypermethylated. Three hypomethylated genes (PPAP2B, DRD4 and SCD) and two
hypermethylated genes (GRIN2A and DOCK10) showed decreased gene expression. ARHGEF12 is the only
gene shown to have a statistically significant increase in gene expression with a mean copy difference of
169.5 ± 54.1 (p=0.008) based on a paired T-test (Table 4.2.4.2). Bar graphs are shown in Figures
4.2.4.2a,b,c as a visual representation of the changes in gene expression per gene and their variation.
ARHGEF12 was the second most highly expressed gene in cord blood MNCs, after HLA-E. Generally, our
selected genes are not highly expressed in cord blood MNCs. Nevertheless, in comparison with each other,
7 genes show a greater level of expression while the other half show a lower level of gene expression
regardless of DNA methylation status.
Table 4.2.4.2. ddPCR gene expression analysis for candidate genes Genes Gene expression (n=4) Mean copy difference (± stdev) Significance
Hypomethylated PPAP2B ¯ -3.57 ± 8.75 p=0.474 DRD4 ¯ -0.07 ± 1.11 p=0.914 SCD ¯ -2.05 ± 6.62 p=0.580 ARHGEF12* 169.5 ± 54.1 p=0.008 SLC25A24 28.0 ± 40.6 p=0.262 PRKCH 103.5 ± 67.0 p=0.054 APBB2 0.18 ± 0.95 p=0.738
Hypermethylated HLA-E 755.5 ± 553.1 p=0.072 LIF 0.38 ± 0.43 p=0.433 FRMD4B 2.9 ± 13.3 p=0.693 MYT1L 0.27 ± 0.19 p=0.063 CYFIP1 17.1 ± 15.9 p=0.121 GRIN2A ¯ -0.34 ± 0.50 p=0.275 DOCK10 ¯ -3.88 ± 43.8 p=0.8709
* indicates significance (p<0.05)
104
Figure 4.2.4.2a. Gene expression on candidate gene mean copy number. Absolute measurement of mRNA expression shown as mean copy number (± standard deviation) between cases and controls. Bars represent average copy number with standard deviation. Asterisk (*) represents significant differences between cases and controls (n=4). Significance was determined if p<0.05.
ARGEF12, PRKCH and SLC25A24 for hypomethylated and HLA-E then DOCK10, CYFIP1 and FRMD4B for
hypermethylated genes show a greater level of gene expression in both cases and controls (Figure
4.2.4.2b).
-500
0
500
1000
1500
2000
2500
Mean Copy Number of Hypo- and Hypermethylated Genes
control case
-100
0
100
200
300
400
500
600
700
800
ARGHEF12 PRKCH SLC25A24
High Expressing Hypomethylated Genes
control case
*
*
105
Figure 4.2.4.2b. Gene expression of high expressing hypo- and hypermethylated genes. Absolute measurement of mRNA expression of cases and controls. Bars represent average copy number with standard deviation. Asterisks (*) represent significant differences between cases and controls (n=4). Significance was determined if p<0.05.
0
10
20
30
40
50
60
70
DOCK10 CYFIP1 FRMD4B
High Expressing Hypermethylated Genes
control case
0
500
1000
1500
2000
2500
HLAE
High Expressing of HLA-E Hypermethylated Gene
control case
106
PPAP2B, DRD4, SCD, and APBB2 for hypomethylated and LIF, MYT1L and GRIN2A for hypermethylated
genes show a lower level of gene expression in cases and controls (Figure 4.2.4.2c).
Figure 4.2.4.2c. Gene expression of low expressing hypo- and hypermethylated genes. Absolute measurement of mRNA expression of cases and controls. Bars represent average copy number with standard deviation.
-2
0
2
4
6
8
10
12
14
16
PPAP2B DRD4 SCD APBB2
Low Expressing Hypomethylated Genes
control case
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
LIF MYT1L GRIN2A
Low Expressing Hypermethylated Genes
control case
107
4.3 Discussion
In the present study, we explored 1) genome-wide and 2) gene-specific DNA methylation in cord blood
MNCs comparing infants born to mothers with low B12 status with those of infants born to mothers with
high B12 status in a high folate environment. We further investigated 3) functional, biological and clinical
ramifications associated with the observed DNA methylation changes. Our exploratory findings are
discussed below.
4.3.1 Lower genome-wide DNA methylation in infants born to mothers with low B12 status
Infants born to mothers with low B12 status in a high folate environment during early pregnancy were
found to have significantly lower genome-wide DNA methylation (total CpG, CpG islands and non-CpG
islands sites) in cord blood MNCs compared to infants born to mothers with high B12 status (Table
4.3.2.1). The observed lower genome-wide DNA methylation in cord blood MNCs in infants born to
mothers with low B12 status is likely related to the methyl trap. As B12 functions as a coenzyme to MS,
inadequate B12 levels would reduce MS activity, which catalyses the transfer of a methyl group from 5-
MTHF to remethylate homocysteine to methionine, resulting in folate that is ‘trapped’ in the form of 5-
MTHF23. This results in lower levels of SAM, the universal methyl group donor to DNA, proteins, lipids
and metabolites for most biological reactions including DNA methylation, and thus lower DNA
methylation52(Figure 2.1.3). Animal studies have more consistently shown a reduction of DNA
methylation with B12 deficiency117,355–357.
With high RBC folate concentrations, hypomethylation may also be observed from two potential
mechanisms. Firstly, the observed high RBC folate status attributable to prenatal folic acid supplementation
can lead to excess DHF, which inhibits MTHFR, and thereby reduces 5-MTHF. Since 5-MTHF donates its
methyl to SAM for DNA methylation, a reduction would subsequently lead to DNA hypomethylation358.
Secondly, preferential shuttling from the DNA methylation pathway to the DNA synthesis pathway
108
through upregulation of TS292 may occur with high folate concentrations. Folic acid has been shown to
upregulate DHFR in certain situations359, and therefore also upregulate TS activity as transcription of both
DHFR and TS genes are coregulated360. As a result, an increase in thymidylate may be observed at the
expense of DNA methylation361 (Figure 2.1.3). Hypomethylation has been associated with folic acid
supplementation in animal studies291–293, albeit inconsistently, and can be attributed to the varying folate
concentrations, duration of study and methods of assessment utilized in animal studies.
In contrast to our findings, a human study looking at the effects of RBC folate and serum B12 found that
increased maternal B12 during pregnancy was associated with decreased genome-wide DNA methylation in
newborns334. Reasons for this may be due to the differences in serum B12 and RBC folate concentrations
between our studies. The McKay et al. 2012334 study correlated serum B12 with DNA methylation based on
a higher mean B12 concentration of 209 (226, 389) pmol/L compared to our study’s low B12 group of 120
(112, 128) pmol/L, as well as the <148 pmol/L deficiency cutoff for serum B12. It may be that the B12
concentration they reported is too high to have decreased MS and SAM activity, and thereby also decrease
methylation. Their study also reported RBC folate concentration to be 859 (675, 1160) nmol/L, which is
approximately three times lower than our study’s folate concentration for the low B12 group of 2682
(2506, 2858) nmol/L, and is also below the 1000 nmol/L cutoff for maximal protection against NTDs.
This concentration may not have been high enough to have increased DHF and reduce MTHFR activity.
Similarly, preferential shuttling to the DNA synthesis pathway may also not have occurred.
One of the major findings of the PREFORM study showed similar findings with ours. Their study found
that lower genome-wide DNA methylation was observed in cord blood MNCs in infants born to mothers
with high RBC folate (≥2860 nmol/L) and low serum B12 (≤167 pmol/L) concentrations than infants born
to mothers with lower RBC and higher serum B12 concentrations33. We observed very high RBC folate
concentrations and lower serum B12 concentrations in the PREFORM study compared to McKay et al.
109
2012334. The PREFORM study also had similar B12 and folate concentrations to the present study. Hence, a
higher folate and lower B12 concentration may be needed to see a decrease in genome-wide DNA
methylation. In addition to the actual supplemental levels of B12 and folate, the ratio of B12 to folate may be
also important in modulating genome-wide DNA methylation. In a rat model study, where maternal folic
acid supplementation (4X the basal requirement) was studied with or without B12 on placental genome-
wide DNA methylation, a decrease in genome-wide DNA methylation in the absence of B12 was
observed299. This suggests that the ratio of folic acid and B12, in addition to the actual supplemental levels,
may have an important role in determining genome-wide DNA methylation patterns. In our study and in
the PREFORM study, our B12 to folate ratio and our genome-wide DNA methylation findings resemble the
findings in the rat study. Although our sample population may not be fully representative of the Canadian
population, our findings provide relevant insight to our Canadian WCBA population who have been shown
to have high concentrations of RBC folate65,133,143, and a proportion of them with suboptimal B12 status
despite adequate intakes69,205,362,363. Hence, both the supplemental levels and the ratio of B12 to folate
highlight the importance of the current recommendations for B12 and folate, and the potential need to
change these recommendations in order to provide adequate B12 throughout each trimester and perhaps a
more conservative dose recommendation for prenatal folic acid supplementation.
4.3.2 Differentially regulated genes are more hypomethylated in mothers with low B12 status
On a gene-specific DNA methylation scale, we observed findings similar to the genome-wide DNA
methylation data. We interrogated DNA methylation status of 804,449 individual CpG sites of 25,158
genes using the Illumina Infinium MethylationEPIC array. Cord blood MNC DNA from infants born to
mothers with low B12 had more hypomethylated genes (p=0.007, q=0.999) than hypermethylated genes,
45 and 21, respectively. Significance based on the T-test p-value was observed without bias of batch
number, infant sex and infant gestational age (Figure 4.2.3.1a). However, multiple comparison testing
did not reach statistical significance (Table 4.2.3.1a). Similar findings, in that there were a greater number
110
of hypo- than hypermethylated genes, were observed with genes obtained from the non-parametric MWU-
test (219 hypo- and 162 hypermethylated) (Table 4.2.3.1b) and overlapping genes (15 hypo- and 12
hypermethylated) (Figure 4.2.3.2). To address the extent of how DNA methylation alterations between
the cases and controls differed, we performed PCA according to the DNA methylation status of the
interrogated CpG sites and its associated genes. A two-dimensional PCA plot detected two distinct
variations of 23% and 6%. PCA 1 with 23% of the variation is explained by the differences in maternal B12
status of the cases and controls in the presence of high folate status (Figure 4.2.3.1c). A change in
methylation ≥5% has been considered biologically significant364,365.
4.3.3 Ramifications of B12 and folate induced changes in DNA methylation on functional and clinical
significance of hypo- and hypermethylated genes
We also performed functional analyses using IPA in order to elucidate the functional ramifications of the
hypo- and hypermethylated genes that differed most between the cases and controls. Functional analyses
revealed several molecular and cellular functions, including macromolecule metabolism and cell cycle,
were affected by changes in B12 and folate. This finding in particular was observed for genes that were
hypomethylated. Hypermethylated genes were associated with cellular growth, proliferation, death and
survival, development, and gene expression (Table 4.3.4). Additionally, functional analyses provided
insight into associated physiological system development and functions as well as diseases and disorders for
the hypo- and hypermethylated genes that were influenced by maternal B12 status in a high folate
environment. These genes are shown to be involved in organ and organismal development, in diseases
related to cancer and organismal injury and abnormalities, and metabolic disorders (Table 4.3.4).
Although exploratory and preliminary, our functional analyses provided insight into the molecular,
cellular, and physiological processes potentially involved in metabolic diseases such as obesity, type 2
diabetes, and metabolic syndrome.
111
From our clinical anthropometric measures, we found a trend towards a higher birth weight, birth weight
percentile, and larger head circumference in infants born to mothers with low B12 status in a high folate
environment compared with those born to mothers with high B12 status based on the 66 differentially
methylated genes generated by the T-test (Figure 4.3.3.1a). While the small sample size and subsequent
lack of power limited our study, birth weight percentile showed borderline significance (p=0.055) based
on a Wilcoxon Rank-Sum test (Table 4.2.1.2d). However, this finding may lose its significance using the
z-score method for calculating birth weight percentile. Nevertheless, our observed trend towards a larger-
sized infant from mothers with low B12 status amidst a high folate environment supports the finding of the
PREFORM study where infants born to mothers with high RBC folate and low serum B12 concentrations
had a higher birth weight (3697 ± 455 g) in comparison to infants born to mothers with lower RBC and
higher serum B12 concentrations (3393 ± 451 g) (p=0.001)33. Furthermore, despite the lack of statistical
significance observed for an increase in infant birth weight, it is possible that changes in body weight may
be observed after birth. Animal studies have shown that although no initial change in birth weight was
observed in the offspring, dams fed a high multivitamin supplement (including folic acid) had male171 and
female pups that had a higher post-weaning body weight than controls366,367.
Studies solely investigating B12, associated a low B12 status with a lower birth weight and a higher risk for
SGA226,368. Similarly, low B12 intakes in the presence of high total folate intake was associated with low
birth weight and higher risk for SGA12. Smaller birth length, head and chest circumference were also
observed with a high plasma folate to B12 ratio362. Although these studies may differ from our findings, our
studies varied due to the differences in nutrition status of the populations of the studies, infant mean birth
weight, and measurement of B12 and folate concentrations. The studies focusing on B12 alone, and in
combination with folate, were based on mothers who were predominantly vegetarian, and were reported
to have inadequate intakes of dietary B12. Additionally, plant-based diets are known to be low in other
nutrients such as vitamin D and zinc, which have been associated with low birth weight369,370. The
112
inadequacy of serum B12 in our study was not due to insufficient intakes, as mothers had dietary intakes
above the EAR and most supplemented with a form of B12-containing vitamin, but can rather be attributed
to the demands of the infant during pregnancy14,206. This dissimilarity in diet may also contribute to the
differences seen between our study’s infant anthropometry. The birth weights between our study’s may
also not be comparable as our infant birth weight was larger by approximately 1000 g. In addition, these
studies, except for Muthayya et al. 2006226, measured either plasma folate and B12362 or only intakes12,368.
Finally, while our study focused on early pregnancy B12 and folate concentrations, other studies reported
measurements at different periods during the pregnancy.
Although it has been shown that B12 and birth weight are positively associated, folate has been shown to be
even more consistently associated with birth weight115,116,165,169, and negatively with risk for SGA167 and
IUGR168. Since our study focused on a combination with high folate status, it is possible that folate drives
changes in birth weight. Moreover, similar to the effect on DNA methylation as previously described, the
ratio of B12 to folate may be important in affecting infant birth weight. An observational study found that
the ratio of folate to B12 was associated with neonatal anthropometry, whereas maternal plasma folate and
B12 concentrations were not362.
113
4.3.4 Upregulation of B12 and folate induced hyper- and hypomethylated genes
Based on functional relevance and the greatest change in DNA methylation and p-value significance of
genes generated by the T-test, MWU-test and its overlap, 14 genes were selected for gene expression
analysis. Functional relevance of the selected genes was primarily associated with lipid metabolism,
molecular transport and small molecule biochemistry. This association was more evident in the
hypomethylated genes.
ddPCR was utilized to provide absolute measurement of gene expression data without the need for a
housekeeping gene and could better detect changes in low expression levels. Although we did not find
statistical significance between the mean copy differences except for ARHGEF12 (169.5 ± 54.1 (p=0.008)),
this is likely due to our small sample size, and can also be attributed to the fact that gene expression levels
vary across different tissues. Even if DNA methylation changes occur across all tissues, gene expression
changes may occur in specific tissues. We only examined gene expression in cord blood MNCs in the
present study. Nevertheless, an increase in gene expression was observed in 9 out of 14 genes (Table
4.2.4.2). This can be expected as hypomethylated CpG sites often allow for transcription to occur due to
its effect on transcription machinery260–262. A study investigating B12 deficiency reported upregulation of
hypomethylated genes of cholesterol biosynthesis regulators in women235. Another study based on a pig
model found that methylating micronutrient supplementation showed differentially expressed metabolic
genes after a change in methylation. This study also reported phenotypic changes associated with decreased
back fat percentage, adipose tissue, fat thickness and increased shoulder percentage in the F2 generation290.
ARHGEF12 particularly has been associated with myeloid leukemia371, and shown to be expressed in
prostate cancer cell lines372 and squamous cell carcinoma (HSC-3) cells373. Interestingly, leukemia-
associated ARHGEF12 has been shown to increase cell growth when it is in complex with CD44, a cell-
surface glycoprotein, and EGFR, an epidermal growth factor, in HSC-3 cells373. Upregulation of ARHGEF12
may thus be potentially involved with infant growth and development, and with the observed increase in
114
infant anthropometric measures. Another gene that is upregulated is FRMD4B, which has been shown to
play a critical role in insulin receptor signaling374. Similarly, SCD was shown to be involved in diet-induced
hepatic insulin resistance, where downregulation of SCD decreased glucose production by 75%, and
decrease both gluconeogenesis and glycogenolysis via reduced activity of glucose-6-phosphotase and
phosphoenolpyruvate carboxykinase375. SCD has also been inversely associated with hypertriglyceridemia in
mice and humans through impaired triglyceride and cholesterol biosynthesis376. The relevance of these
genes to clinical measures related to metabolic disease highlights the need of these B12 and folate induced
DNA methylated genes to potentially influence infant health and disease outcomes at birth and later in life.
Although gene expression analysis was limited by RNA quality, expression in MNCs and sample size, our
findings provide potentially interesting information regarding the influence of B12 and folate induced
changes in DNA methylation on gene expression regulation. Future studies are needed to confirm the
aforementioned associations and to further explore the effect of B12 and folate on DNA methylation and
gene expression.
Table 4.3.4. Summary table of gene expression and functional analyses of hypo- and hypermethylated genes
Genes Methylation status Gene expression
Top Molecular/Cellular Function
PPAP2B ¯ ¯ Carbohydrate metabolism Lipid metabolism
Molecular Transport Small Molecule Biochemistry
Behaviour Cell Cycle
Cellular movement
DRD4 ¯ ¯ SCD ¯ ¯ ARHGEF12 ¯ SLC25A24 ¯ PRKCH ¯ APBB2 ¯ HLA-E
Cell growth and proliferation Cell death and survival
Cell cycle Cell development Lipid metabolism
Molecular Transport Small Molecule Biochemistry
Behaviour Gene Expression
LIF FRMD4B MYT1L CYFIP1 GRIN2A ¯ DOCK10 ¯
115
4.4 Conclusion
The PREFORM study is a large observational study in a demographically diverse group of pregnant
Canadian women and their infants. We explored whether or not changes in DNA methylation were
influenced by maternal low serum B12 and high RBC folate concentrations at early pregnancy. We also
aimed to understand whether or not these B12 and folate induced changes in DNA methylation have the
potential to affect infant health and disease outcomes at birth and in adulthood. We thus examined
functional relevance of differentially methylated genes, changes in gene expression, as well as in
measurement of infant anthropometry (Figure 5.1).
We found that both genome-wide and gene-specific DNA methylation is reduced in cord blood MNCs of
infants born to mothers with a low B12 amidst a high folate environment. Genome-wide methylation
analysis showed significance for mean ß-value difference of the total (-0.0059±0.0147, p= 0.013), CpG
island only (-0.0036±0.0069, p=0.022), and non-CpG island only (-0.0078±0.0147, p=0.015) analyses.
A total of 66 genes were found to be differentially methylated (p=0.007, q= 0.999), 45 of which were
hypomethylated and 21 relative to the control. A 23% difference in DNA methylation was observed
between the cases and the control groups based on the PCA. Of the 14 selected genes, 9 showed
upregulation and 5 showed downregulation regardless of DNA methylation status. ARHGEF12 was shown
to be significantly upregulated with a mean copy difference of 169.5 ± 54.1 (p=0.008) relative to the
control. Functional analyses reveal that these genes are associated with metabolism and development.
Infants born to mothers with a low B12 status are shown to have a higher birth weight and birth weight
percentile and a larger head circumference. Taken together, our exploratory findings support the notion
that environmental nutrient exposure can be affected by epigenetic changes to potentially influence infant
health and disease outcomes later in life, as described in the DOHaD hypothesis (Figure 5.1).
116
Future recommendations for B12 and folate should be reconsidered in an attempt to lower RBC folate and
raise serum B12 concentrations in pregnant women. Future studies are warranted to confirm our findings
with a larger sample size, and elucidate clinical ramifications on chronic disease risk-related factors such as
adiposity and insulin resistance on children at 5 years of age and older.
117
5- Overall Conclusions and Future Directions
5.1 Summary and General Discussion
The primary objective of this thesis was to determine whether or not maternal low B12 and high folate
status affects DNA methylation and gene expression in newborn infants’ cord blood MNCs. Based on the
methyl trap theory, it was hypothesized that low maternal serum B12 concentrations in combination with
high folate status reduced genome-wide DNA methylation as well as gene-specific DNA methylation in
genes that are related to fetal and infant growth and development (Figure 5.1).
Figure 5.1. Summary of thesis conclusion. Similar to the DOHaD model, a tripartite relationship of maternal nutrient exposure, infant DNA methylation and possible future effects on chronic disease risk was observed. We indeed demonstrated that infants born to mothers with low B12 status amidst the presence of a high
folate status had lower genome-wide DNA methylation and more hypomethylated genes compared with
those born to mothers with higher B12 status. Their molecular, physiological and functional relevance were
shown to be associated with metabolism, organ and organismal development. We also found that these B12
and folate induced hypomethylation was associated with upregulation of genes related to metabolism and
Maternal B12
and Folate
Infant cord MNC genome-wide and gene-specific DNA methylation
Gene expression Infant anthropometry
Functional relevance
118
development in the newborn infant, which may lead to more pronounced phenotypic effects in early
childhood. Our study provides novel information regarding the relationships between maternal B12 and
folate status at a critical period of development and DNA methylation. With currently high levels of folic
acid in the Canadian WCBA population, and a proportion of pregnant women with suboptimal B12 status
despite adequate intakes, our findings are important in bringing to light possible changes in
recommendations for supplementation during preconception and throughout each trimester. Although it is
important to have recommendations that encourage preconceptional folic acid supplementation to prevent
NTDs, previous studies including ours have shown high maternal RBC folate concentrations to have
exceeded presently designated ‘high’ cutoffs. Their effects on infant health and disease have been shown to
be adverse in relation to insulin resistance, adiposity, and risk for developing metabolic disorders;
although, current evidence is inconclusive. B12 dietary and supplemental intakes, on the other hand, appear
to be at or exceeding the EAR of 2.6 µg/d for pregnant women. However, B12 status has been shown to be
suboptimal in some women; hence, modified recommendations for folate and B12 should be considered.
119
5.2 Strengths and Limitations
To date, the PREFORM study is the largest Canadian observational study assessing maternal one-carbon
nutrients including B12 and folate concentrations on DNA methylation in their newborn infants143. While
the PREFORM original cohort is highly reflective of the multiethnic population of Toronto, our findings,
using a smaller subsample of matched cases and controls, may not represent all Canadians135. Nonetheless,
our data significantly adds to the growing body of evidence that suggests that maternal nutrition may have
epigenetic effects in the offspring. Also, we are one of the few studies that measured the effects of B12 and
folate status on genome-wide and gene-specific DNA methylation in cord blood MNCs using the Illumina
Infinium MethylationEPIC array. Another strength of our study is our exploratory analysis on elucidating
biological ramifications of nutrient-induced epigenetic changes with functional and gene expression
analyses using Ingenuity Pathway Analysis, ddPCR, and infant anthropometry. Our study is one of the few
studies which investigated a tripartite relationship between nutrient status, DNA methylation and gene
expression in relation to infant growth and development. The method of gene expression was also a
strength of this study as ddPCR allowed for more accurate and reliable measurement of gene expression
through absolute (without use of a housekeeping gene) instead of comparative analysis, as seen with RT-
qPCR. In addition, we have amassed a large dataset of gene-specific DNA methylation patterns from cord
blood MNCs that differ based on maternal B12 status in a high folate environment, which can be used for
future analyses.
There are a number of limitations to this study. The cases and controls were matched using propensity
score analysis based on seven confounding variables that affect DNA methylation; however, the sample size
was too small to correct other potential confounders in the regression. Hence, maternal ethnicity and
smoking during pregnancy were not controlled for, among other residual confounders. For smoking
during pregnancy, however, a small number of women smoked during pregnancy (5 out of 44) and were
relatively balanced between cases and controls. Another limitation pertains to the cell type for DNA
120
methylation analysis. Immune cells such as lymphocytes were not separated from other MNC’s when
assessing DNA methylation. This made it difficult to verify expression of candidate genes by cell type for
gene expression analysis. Furthermore, since DNA methylation are cell and tissue-specific, future studies
should ensure that DNA methylation analysis is performed in a specific cell type and avoid contamination
with other cell types. Gene expression in cord MNCs were generally quite low, making detection difficult.
Nevertheless, our ddPCR method allowed for more accurate measurement of low gene expression as well
as low cDNA concentrations in comparison to other methods such as RT-qPCR. Hypo- and
hypermethylated genes that were generated by a T-test did not show statistical significance after multiple
comparison testing using the Benjamini-Hochberg correction; however, we mitigated this limitation by
testing for changes in gene expression through ddPCR. Another limitation may be with the calculation of
birth weight percentile. Despite having found borderline statistical significance, the calculation was
performed manually using a growth chart and may confer a false positive finding. Using z-scores from the
World Health Organization’s database may allow for more precise birthweight percentile measurement.
This calculation is currently underway. Our sample size for gene-specific DNA methylation analysis was
small primarily due to matching cases with controls, the availability of DNA sample and cost. The sample
size for gene expression analysis was even smaller due to poor RNA quality obtained from cord MNCs.
Finally, a larger sample size would have been needed to elucidate a clearer picture on the associations with
B12 and folate on DNA methylation and on gene expression. Nevertheless, our exploratory approach in this
study may be used as a foundation for future studies in order to confirm our findings.
121
5.3 Future Directions
Our data provides a framework for future large-scale studies aimed at elucidating the relationship between
maternal B12 and folate status and DNA methylation in human offspring and their functional, biological and
clinical ramifications and long-term health consequences. Some potential future studies include: 1)
collection of a larger sample of Canadian pregnant women, with higher quality infant RNA from cord
blood lymphocytes, and confirm DNA methylation status with bisulphite pyrosequencing of the fourteen
genes identified as most relevant to infant growth and development, and gene expression with ddPCR; 2)
design and implement follow up studies of the infants at early childhood to investigate the effect of B12 and
folate on DNA methylation, anthropometric and metabolic outcomes, and health and disease outcomes
related to growth and development such as with insulin resistance and adiposity; and 3) design a similar
study in an animal model to investigate more causal relationships between B12 and folate status, DNA
methylation and gene expression.
122
5.4 Final Conclusions
To conclude, this research project provides novel information regarding the effect of B12 and folate status
on DNA methylation and genes associated with infant growth and development. We measured genome-
wide and gene-specific methylation in cord blood MNCs, as well as gene expression on selected genes with
functional significance to metabolism and development. Overall, our findings support the importance of
nutrient induced epigenetic changes that can affect infant health at birth, early childhood, and can
potentially modulate chronic disease risk in adulthood (Figure 5.1).
123
References
1. Ly, A., Hoyt, L., Crowell, J. & Kim, Y.-I. Folate and DNA Methylation. Antioxid. Redox Signal. 17, 302–326 (2012).
2. Depeint, F., Bruce, W. R., Shangari, N., Mehta, R. & O’Brien, P. J. Mitochondrial function and toxicity: Role of B vitamins on the one-carbon transfer pathways. Chem. Biol. Interact. 163, 113–132 (2006).
3. Lee, H. S. Impact of maternal diet on the epigenome during in utero life and the developmental programming of diseases in childhood and adulthood. Nutrients 7, 9492–9507 (2015).
4. Smith, D. et al. Is folic acid good for everyone? Am. J. Clin. Nutr. 87, 517–533 (2008). 5. Warzyszynska, J. E. & Kim, Y.-I. J. Folate in Human Health and Disease. eLS 1–14 (2014).
doi:10.1002/9780470015902.a0002268.pub2 6. W, R., W, D. & J, W. Epigenetic reprogramming in mammalian development. Science (80-. ). 1,
19–25 (2007). 7. Robertson, K. D. DNA methylation and human disease. Nat. Rev. Genet. 6, 597–610 (2005). 8. Waterland, R. A. & Jirtle, R. L. Early Nutrition , Epigenetic Changes at Transposons and
Imprinted Genes , and Enhanced Susceptibility to Adult Chronic Diseases. 63–68 (2004). doi:10.1016/j.nut.2003.09.011
9. Sharland, E., Montgomery, B. & Granell, R. Folic acid in pregnancy: Is there a link with childhood asthma or wheeze? Aust. Fam. Physician 40, 421–424 (2011).
10. Pfeiffer, C. M. et al. Trends in blood folate and vitamin B-12 concentrations in the United States, 1988-2004. Am. J. Clin. Nutr. 86, 718–727 (2007).
11. Yetley, E. & Pfeiffer, C. Biomarkers of vitamin B-12 status in NHANES: a roundtable summary. Am. J. Clin. Nutr. 94, 313–321 (2011).
12. Dwarkanath, P. et al. High folate and low vitamin B-12 intakes during pregnancy are associated with small-for-gestational age infants in South Indian women: a prospective observational cohort study`. Am. J. Clin. Nutr. 98, 1450–1458 (2013).
13. Yajnik, C. S. et al. Vitamin B12 and folate concentrations during pregnancy and insulin resistance in the offspring: the Pune Maternal Nutrition Study. Diabetologia 51, 29–38 (2008).
14. Medicine, I. of. in (The National Academies Press, 1998). 15. Kim, T. H., Yang, J., Darling, P. B., Connor, D. L. O. & Al, K. I. M. E. T. Human Nutrition and
Metabolism A Large Pool of Available Folate Exists in the Large Intestine of Human Infants and Piglets 1. 1389–1394 (2004).
16. Rong, N., Selhub, J., Goldin, B. R. & Rosenberg, I. H. Bacterially synthesized folate in rat large intestine is incorporated into host tissue folyl polyglutamates. J. Nutr. 121, 1955—1959 (1991).
17. Aufreiter S, Gregory JF, Pfeiffer CM, Fazili Z, Kim YI, Marcon N, Kamalaporn P, Pencharz PB, O. D. Folate is absorbed across the colon of adults: evidence from cecal infusion of C-13-labeled 6S -5-formyltetrahydrofolic acid. Am J Clin Nutr 9, 116–23 (2009).
18. Lynn B. Bailey. Folate in health and disease. (Taylor & Francis, 2010). at <http://search.library.utoronto.ca/details?7018847>
19. Visentin, M., Diop-Bove, N., Zhao, R. & Goldman, D. The intestinal absorption of folates. Am. J. Clin. Nutr. 32, 846–855 (2014).
20. Aufreiter, S. et al. Folate is absorbed across the colon of adults : evidence from cecal infusion of 13 C-labeled [ 6S ] -5-formyltetrahydrofolic acid 1 – 3. 116–123 (2009). doi:10.3945/ajcn.2008.27345.INTRODUCTION
21. Lakoff, A. et al. Folate is absorbed across the human colon: Evidence by using enteric-coated caplets containing13C-labeled [6S]-5-formyltetrahydrofolate1-4. Am. J. Clin. Nutr. 100, 1278–1286 (2014).
22. Wright, A. J. a, Dainty, J. R. & Finglas, P. M. Folic acid metabolism in human subjects revisited:
124
potential implications for proposed mandatory folic acid fortification in the UK. Br. J. Nutr. 98, 667–75 (2007).
23. Obeid, R., Kasoha, M., Kirsch, S. H., Munz, W. & Herrmann, W. Concentrations of unmetabolized folic acid and primary folate forms in pregnant women at delivery and in umbilical cord blood. Am. J. Clin. Nutr. 92, 1416–1422 (2010).
24. Boilson, A. et al. Unmetabolized folic acid prevalence is widespread in the older Irish population despite the lack of a mandatory fortification program. Am. J. Clin. Nutr. 96, 613–621 (2012).
25. Bailey, R. L. et al. Serum unmetabolized folic acid in a nationally representative sample of adults ≥60 years in the United States, 2001–2002. Food Nutr. Res. 56, 10–12 (2012).
26. Kalmbach, R. & Paul, L. Determination of unmetabolized folic acid in human plasma using affinity HPLC. Am. J. Clin. 5–9 (2011). doi:10.3945/ajcn.111.013433.
27. Shane, B. Folate status assessment history: Implications for measurement of biomarkers in NHANES. Am. J. Clin. Nutr. 94, (2011).
28. Kim, Y.-I. I. Folate and colorectal cancer: An evidence-based critical review. Mol. Nutr. Food Res. 51, 267–292 (2007).
29. Shane, B. in Folate in Health and Disease 11, 1–24 (CRC Press, 2009). 30. Zhao, R., Matherly, L. H. & Goldman, I. D. Membrane transporters and folate homeostasis:
intestinal absorption and transport into systemic compartments and tissues. Expert Rev. Mol. Med. 11, (2009).
31. Stover, P. J. Physiology of folate and vitamin B12 in health and disease. Nutr. Rev. 62, S3–12; discussion S13 (2004).
32. Sie, K. K. Y. et al. Effect of maternal and postweaning folic acid supplementation on global and gene-specific DNA methylation in the liver of the rat offspring. Cancer Res. 71, 988–997 (2013).
33. Plumptre, L. Maternal one-carbon nutrient status and effects on DNA methylation and hydroxymethylation in newborn infants. (University of Toronto, 2016).
34. Lucock, M. Folic acid: nutritional biochemistry, molecular biology, and role in disease processes. Mol. Genet. Metab. 71, 121–138 (2000).
35. Blom, H. J., Shaw, G. M., den Heijer, M. & Finnell, R. H. Neural tube defects and folate: case far from closed. Nat Rev Neurosci 7, 724–731 (2006).
36. Tibbetts, A. S. & Appling, D. R. Compartmentalization of Mammalian folate-mediated one-carbon metabolism. Annu. Rev. Nutr. 30, 57–81 (2010).
37. Ashokkumar, B., Mohammed, Z. M., Vaziri, N. D. & Said, H. M. Effect of folate oversupplementation on folate uptake by human intestinal and renal epithelial cells. Am. J. Clin. Nutr. 86, 159–166 (2007).
38. Bailey, S. W. & Ayling, J. E. The extremely slow and variable activity of dihydrofolate reductase in human liver and its implications for high folic acid intake. Proc. Natl. Acad. Sci. U. S. A. 106, 15424–9 (2009).
39. Kelly, P., McPartlin, J., Goggins, M., Weir, D. G. & Scott, J. M. Unmetabolized folic acid in serum: acute studies in subjects consuming fort...: EBSCOhost. (1997). at <http://web.a.ebscohost.com.buproxy.bastyr.edu:2048/ehost/pdfviewer/pdfviewer?sid=66d24fe9-a997-413c-a94c-46ccd26636c1@sessionmgr4003&vid=1&hid=4101>
40. Ueland, P. M., Hustad, S., Schneede, J., Refsum, H. & Vollset, S. E. Biological and clinical implications of the MTHFR C677T polymorphism. Trends Pharmacol. Sci. 22, 195–201 (2001).
41. Chen, L. H. et al. Human methionine synthase reductase. J. Biol. Chem. 276, 35558–63 (2001). 42. Leclerc, D. et al. Cloning and Mapping of a cDNA for Methionine Synthase Reductase, a
Flavoprotein Defective in Patients with Homocystinuria. Proc. Natl. Acad. Sci. U. S. A. 95, 3059–3064 (1998).
43. Kim, Y.-I. Nutritional Epigenetics: Impact of Folate Deficiency on DNA Methylation and Colon Cancer Susceptibility. J. Nutr. 135, 2694–2697 (2005).
125
44. Kim, Y. Folate and DNA Methylation : A Mechanistic Link between Folate Deficiency and Colorectal Cancer ? Folate and DNA Methylation : A Mechanistic Link between Folate Deficiency and Colorectal Cancer ? 13, 511–519 (2004).
45. Smith, D., Kim, Y. & Refsum, H. Is folic acid good for everyone? Am. J. Clin. Nutr. 87, 517–533 (2008).
46. Bailey, L. B. et al. Biomarkers of Nutrition for Development— Folate Review 1–5. J. Nutr. 1–45 (2015). doi:10.3945/jn.114.206599.1S
47. Prinz-Langenohl, R., Fohr, I. & Pietrzik, K. Beneficial role for folate in the prevention of colorectal and breast cancer. Eur. J. Nutr. 40, 98–105 (2001).
48. Otten, J. J., Hellwig, J. P. & Linda, D. Dietary DRI Reference Intakes. (2006). 49. Branum, A. M., Bailey, R. & Singer, B. J. Dietary Supplement Use and Folate Status during
Pregnancy in the United States. J. Nutr. Nutr. Epidemiol. (2013). doi:10.3945/jn.112.169987.the 50. Bailey, L. B. & Gregory, J. F. Recent Advances in Nutritional Science Folate Metabolism and
Requirements 1,2. 51. Molloy AM, Mills JL, McPartlin J, Kirke PN, Scott JM, D. S. Maternal and fetal plasma
homocysteine concentrations at birth: The influence of folate, vitamin B12, and the 5, 10-methylenetetrahydrofolate reductase 677C→T variant. Am J Obs. Gynecol 186, 499–503 (2002).
52. Molloy, A. M. Folate-vitamin B12 interrelationships; links to disease risk. Folate Heal. Dis. 2nd ed. Boca Rat. CRC Press. Taylor Fr. Gr. 381–408 (2009).
53. Selhub, J. & Rosenberg, I. H. Excessive folic acid intake and relation to adverse health outcome. Biochimie 126, 71–78 (2016).
54. Sawaengsri, H. et al. High folic acid intake reduces natural killer cell cytotoxicity in aged mice. J. Nutr. Biochem. 30, 102–107 (2016).
55. Paniz, C. et al. A Daily Dose of 5 mg Folic Acid for 90 Days Is Associated with Increased Serum Unmetabolized Folic Acid and Reduced Natural Killer Cell Cytotoxicity in Healthy Brazilian Adults. 1677–1685 (2017). doi:10.3945/jn.117.247445
56. Yetley, E. A. & Johnson, C. L. Folate and vitamin B-12 biomarkers in NHANES: History of their measurement and use. Am. J. Clin. Nutr. 94, 322S–331S (2011).
57. Pfeiffer, C. M. et al. Unmetabolized folic acid is detected in nearly all serum samples from US children, adolescents, and adults. J. Nutr. 145, 520–31 (2015).
58. Yetley, E. A., Coates, P. M. & Johnson, C. L. Overview of a roundtable on Nhanes monitoring oif biomarkers of folate and vitamin B12 status.pdf. 94, 297–302 (2011).
59. MR, S. et al. Persistent circulating unmetabolised folic acid in a setting of liberal voluntary folic acid fortification. Implications for further mandatory fortification? BMC Public Health 9, 295 (2009).
60. Kalmbach, R. D. et al. Circulating folic acid in plasma: relation to folic acid fortification. Am J Clin Nutr 88, 763–768 (2008).
61. Sweeney, M. M. R. et al. Persistent circulating unmetabolised folic acid in a setting of liberal voluntary folic acid fortification. Implications for further mandatory fortification? BMC Public Health 9, 295 (2009).
62. Sweeney, M. R. et al. Evidence of unmetabolised folic acid in cord blood of newborn and serum of 4-day-old infants. Br. J. Nutr. 727–730 (2005).
63. Blackmore, S., Pfeiffer, C. M., Lee, A., Fazili, Z. & Hamilton, M. S. Isotope dilution--LC-MS/MS reference method assessment of serum folate assay accuracy and proficiency testing Consensus Mean. Am. Assoc. Clin. Chem. 986 (2011).
64. Colapinto, C. K. et al. The direction of the difference between Canadian and American erythrocyte folate concentrations is dependent on the assay method employed: a comparison of the Canadian Health Measures Survey and National Health and Nutrition Examination Survey. Br. J. Nutr. 112, 1873–1881 (2014).
126
65. Colapinto, C. K., O’Connor, D. L., Dubois, L. & Tremblay, M. S. Prevalence and correlates of high red blood cell folate concentrations in the Canadian population using 3 proposed cut-offs. Appl. Physiol. Nutr. Metab. = Physiol. Appl. Nutr. metabolisme 40, 1025–30 (2015).
66. Daly, L. E. Folate Levels and Neural Tube Defects. Jama 274, 1698 (1995). 67. Crider, K. S. et al. Population red blood cell folate concentrations for prevention of neural tube
defects: Bayesian model. Bmj 349, g4554 (2014). 68. Colapinto, C. K., O’Connor, D. L., Sampson, M., Williams, B. & Tremblay, M. S. Systematic
review of adverse health outcomes associated with high serum or red blood cell folate concentrations. J. Public Health (Oxf). 38, e84-97 (2016).
69. Macfarlane, A. J., Greene-finestone, L. S. & Shi, Y. Vitamin B-12 and homocysteine status in a folate-replete population : results from the Canadian Health Measures Survey 1 – 3. Am. J. Clin. Nutr. 4, 1079–87 (2011).
70. Pfeiffer, C. M. et al. Estimation of trends in serum and RBC folate in the U.S. population from pre- to postfortification using assay-adjusted data from the NHANES 1988-2010. J. Nutr. 142, 886–93 (2012).
71. Food and Nutrition Board , Institute of Medicine, N. A. Dietary Reference Intakes ( DRIs ): Estimated Average Requirements. 1–4 (2011).
72. Clarke, R., Halsey, J., Bennett, D. & Lewington, S. Homocysteine and vascular disease: Review of published results of the homocysteine-lowering trials. J. Inherit. Metab. Dis. 34, 83–91 (2011).
73. Homocysteine, S. C. Homocysteine and Risk of Ischemic Heart Disease and Stroke: A Meta-analysis. JAMA J. Am. Med. Assoc. 288, 2015–2022 (2002).
74. Albert, C. M. et al. Effect of folic acid and B vitamins on risk of cardiovascular events and total mortality among women at high risk for cardiovascular disease: A randomized trial. 299, 2027–2036 (2008).
75. Miller, E. R. et al. Meta-analysis of folic acid supplementation trials on risk of cardiovascular disease and risk interaction with baseline homocysteine levels. Am. J. Cardiol. 106, 517–527 (2010).
76. Yang, H.-T., Lee, M., Hong, K.-S., Ovbiagele, B. & Saver, J. L. Efficacy of folic acid supplementation in cardiovascular disease prevention: an updated meta-analysis of randomized controlled trials. Eur. J. Intern. Med. 23, 745–54 (2012).
77. Zhang, C., Wang, Z.-Y., Qin, Y.-Y., Yu, F.-F. & Zhou, Y.-H. Association between B vitamins supplementation and risk of cardiovascular outcomes: a cumulative meta-analysis of randomized controlled trials. PLoS One 9, e107060 (2014).
78. Clarke, R. et al. Effects of Lowering Homocysteine Levels With B Vitamins on Cardiovascular Disease, Cancer, and Cause-Specific Mortality<subtitle>Meta-analysis of 8 Randomized Trials Involving 37 485 Individuals</subtitle><alt-title>Effects of Lowering H. Arch. Intern. Med. 170, 1622 (2010).
79. Zhou, Y.-H. et al. Effect of folic acid supplementation on cardiovascular outcomes: a systematic review and meta-analysis. PLoS One 6, e25142 (2011).
80. Huo, Y. et al. Efficacy of Folic Acid Supplementation in Stroke Prevention: New Insight From a Meta-Analysis. Heart 98, E149.3-E150 (2012).
81. Zeng, R., Xu, C.-H., Xu, Y.-N., Wang, Y.-L. & Wang, M. The effect of folate fortification on folic acid-based homocysteine-lowering intervention and stroke risk: a meta-analysis. Public Health Nutr. 18, 1514–21 (2015).
82. Kim, Y.-I. Folic Acid Supplementation and Cancer Risk: Point. Cancer Epidemiol. Biomarkers Prev. 17, 2220–2225 (2008).
83. Kim, Y. Review Folate and carcinogenesis : Evidence , mechanisms , and implications. Rev. Lit. Arts Am. 2863, 66–88 (1999).
84. Kim, Y. International Research Conference on Food , Nutrition , and Cancer Role of Folate in Colon Cancer Development and Progression 1. 3731–3739 (2003).
127
85. Kim, Y. I. Will mandatory folic acid fortification prevent or promote cancer? Am. J. Clin. Nutr. 80, 1123–1128 (2004).
86. Kennedy, D. A. et al. Folate intake and the risk of colorectal cancer: A systematic review and meta-analysis. Cancer Epidemiol. 35, 2–10 (2011).
87. Cole, B. F. et al. Folic Acid for the Prevention of Colorectal Adenomas. JAMA 297, 2351 (2007). 88. Jaszewski, R. et al. Folic acid supplementation inhibits recurrence of colorectal adenomas: A
randomized chemoprevention trial. World J. Gastroenterol. 14, 4492–4498 (2008). 89. Logan, R. F. A., Grainge, M. J., Shepherd, V. C., Armitage, N. C. & Muir, K. R. Aspirin and
Folic Acid for the Prevention of Recurrent Colorectal Adenomas. Gastroenterology 134, 29–38 (2008).
90. Wu, K. et al. A randomized trial on folic acid supplementation and risk of recurrent. Am. J. Clin. Nutr. 1623–1631 (2009). doi:10.3945/ajcn.2009.28319.INTRODUCTION
91. Reynolds, E. H. The neurology of folic acid deficiency. Handbook of Clinical Neurology 120, (Elsevier B.V., 2014).
92. Reynolds, E. Vitamin B12 Deficiency and the Nervous System. Neurologist 12, 169–170 (2006). 93. Czeizel, A. E. & Dudas, I. Prevention of the first occurrence of neural-tube defects by
periconceptional vitamin supplementa- tion. N. Engl. J. Med. 132, 2393S–2400S (1992). 94. Bailey, L. B. & Berry, R. J. Folic acid supplementation and the occurrence of congenital heart
defects, orofacial clefts, multiple births, and miscarriage. Am. J. Clin. Nutr. 81, 1213S–1217S (2005).
95. Hughes, C. F., Ward, M., Hoey, L. & McNulty, H. Vitamin B12 and ageing: current issues and interaction with folate. Ann. Clin. Biochem. 50, 315–29 (2013).
96. McCracken, C., Hudson, P., Ellis, R. & McCaddon, A. Methylmalonic acid and cognitive function in the medical research council cognitive function and ageing study. 84, 1406–1411 (2006).
97. Clarke, R. et al. Low vitamin B-12 status and risk of cognitive decline in older adults. 86, 1384–1391 (2007).
98. Clarke, R., Smith, A. D., Jobst, K. A. & Refsum, H. Folate, Vitamin B 12 , and Serum Total Homocysteine Levels in Confirmed Alzheimer Disease. (1998).
99. Selhub, J., Morris, M. S., Jacques, P. F. & Rosenberg, I. H. Folate – vitamin B-12 interaction in relation to cognitive impairment , anemia , and biochemical indicators of vitamin B-12 deficiency. Am. J. Clin. Nutr. 89, 702–707 (2009).
100. Paspatis, G. & Karamanolis, D. Folate Supplementation and Adenomatous Colonic Polyps. Dis Colon Rectum 11, 1340–1341 (1994).
101. Donnenfeld, M. et al. Prospective association between dietary folate intake and skin cancer mentation en Vitamines et Min e raux risk : results from the Suppl e Antioxydants cohort. Am. J. Clin. Nutr. 102, 471–8 (2015).
102. Ebbing, M. Cancer Incidence and Mortality After Treatment With Folic Acid and Vitamin B 12. 302, 2119–2126 (2010).
103. Vollset, S. E. et al. Effects of folic acid supplementation on overall and site-specifi c cancer incidence during the randomised trials : meta-analyses of data on 50 000 individuals. Lancet 381, 1029–1036 (2013).
104. Mason, J. B. et al. A temporal association between folic acid fortification and an increase in colorectal cancer rates may be illuminating important biological principles: A hypothesis. Cancer Epidemiol. Biomarkers Prev. 16, 1325–1329 (2007).
105. Hirsch, S. et al. Colon cancer in Chile before and after the start of the flour fortification program with folic acid. Eur J Gastroenterol Hepatol Wolters Kluwer Heal. Eur. J. Gastroenterol. Hepatol. 0, 0–0 (2008).
106. Stevens, V. L. et al. High levels of folate from supplements and fortification are not associated with increased risk of colorectal cancer. Gastroenterology 141, 98–105.e1 (2011).
128
107. Gibson, T. M. et al. Pre- and postfortification intake of folate and risk of colorectal cancer in a large prospective cohort study in the United States 1 – 3. Am. J. Clin. Nutr. 94, 1053–1062 (2011).
108. Dervieux, T. et al. Pharmacogenetic and metabolite measurements are associated with clinical status in patients with rheumatoid arthritis treated with methotrexate: results of a multicentred cross sectional observational study. Ann. Rheum. Dis. 64, 1180–5 (2005).
109. Khanna, D. et al. Reduction of the efficacy of methotrexate by the use of folic acid: Post hoc analysis from two randomized controlled studies. Arthritis Rheum. 52, 3030–3038 (2005).
110. Zhao, R. & Goldman, I. D. Resistance to antifolates. Oncogene 22, 7431–7457 (2003). 111. Robien, K. Folate during antifolate chemotherapy: what we know... and do not know. Nutr. Clin.
Pract. 20, 411–22 (2005). 112. Chattopadhyay, S. et al. Commentary: a case for minimizing folate supplementation in clinical
regimens with pemetrexed based on the marked sensitivity of the drug to folate availability. Oncologist 12, 808–15 (2007).
113. Travers, P., Walport, M. & Capra, J. D. Immunobiology: the immune system in health and disease. Curr. Biol. Ltd (1999).
114. Troen, A. M. et al. Nutritional Immunology Unmetabolized Folic Acid in Plasma Is Associated with Reduced Natural Killer Cell Cytotoxicity among Postmenopausal Women 1. 189–194 (2006).
115. Keating, E. et al. Excess perigestational folic acid exposure induces metabolic dysfunction in post-natal life. 245–259 (2015). doi:10.1530/JOE-14-0448
116. Huot, P. S. P. et al. Maternal and postweaning folic acid supplementation interact to influence body weight , insulin resistance , and food intake regulatory gene expression in rat offspring in a sex-specific manner. 420, 411–420 (2016).
117. Sinclair, K. D. et al. DNA methylation, insulin resistance, and blood pressure in offspring determined by maternal periconceptional B vitamin and methionine status. Proc. Natl. Acad. Sci. U. S. A. 104, 19351–6 (2007).
118. Krishnaveni, G. V & Veena, S. R. Association between maternal folate concentrations during pregnancy and insulin resistance in Indian children. 110–121 (2014). doi:10.1007/s00125-013-3086-7
119. MRC VITAMIN STUDY RESEARCH GROUP. Prevention of neural tube defects: Results of the Medical Research Council Vitamin Study. Lancet 338, 131–137 (1991).
120. Smithells, R. W., Sheppard, S. & Schorah, C. J. Vitamin deficiencies and neural tube defects. Arch. Dis. Child. 51, 944–50 (1976).
121. BERRY, R. J. et al. PREVENTION OF NEURAL-TUBE DEFECTS WITH FOLIC ACID IN CHINA R. N. Engl. J. Med. 341, 69–75 (1999).
122. De Wals, P. et al. Spina bifida before and after folic acid fortification in Canada. Birth Defects Res. Part A - Clin. Mol. Teratol. 82, 622–626 (2008).
123. Refsum, H. Folate, vitamin B12 and homocysteine in relation to birth defects and pregnancy outcome. Br. J. Nutr. 85 Suppl 2, S109–S113 (2001).
124. Mills, J. L. & Signore, C. Neural tube defect rates before and after food fortification with folic acid. Birth Defects Res. Part A - Clin. Mol. Teratol. 70, 844–845 (2004).
125. Mathews, T. J., Honein, M. & Erickson, J. D. Spina Bifida and Anencephaly Prevalence. MMWR. Recomm. Rep. 51, 9–11 (2002).
126. Eoin P, Q. & Jesse F, G. The impact of food fortification on folic acid intake in Canada. Can J Public Heal. 94, (2003).
127. De Wals, P. et al. Reduction in neural-tube defects after folic acid fortification in Canada. N. Engl. J. Med. 357, 135–42 (2007).
128. Shakur, Y. a., Garriguet, D., Corey, P. & O’Connor, D. L. Folic acid fortification above mandated levels results in a low prevalence of folate inadequacy among Canadians. Am. J. Clin. Nutr. 92, 818–825 (2010).
129
129. Choumenkovitch, S. F. et al. Folic acid intake from fortification in United States exceeds predictions. J. Nutr. 132, 2792–2798 (2002).
130. Pfeiffer, C. et al. The CDC’s Second National Report on Biochemical Indicators of Diet and Nutrition in the U . S . Population. J Nutr 143, 1–484 (2013).
131. Colapinto, C. K., O’Connor, D. L., Dubois, L. & Tremblay, M. S. Folic acid supplement use is the most significant predictor of folate concentrations in Canadian women of childbearing age. Appl. Physiol. Nutr. Metab. 284–292 (2011).
132. Dietrich, M., Brown, C. J. P. & Block, G. The effect of folate fortification of cereal-grain products on blood folate status, dietary folate intake, and dietary folate sources among adult non-supplement users in the United States. J. Am. Coll. Nutr. 24, 266–74 (2005).
133. Colapinto, C. K., Connor, D. L. O. & Tremblay, M. S. Folate status of the population in the Canadian Health. 183, (2011).
134. Hoyo, C. et al. Folic acid supplementation before and during pregnancy in the Newborn Epigenetics STudy (NEST). BMC Public Health 11, 46 (2011).
135. Masih, S. P. et al. Pregnant Canadian Women Achieve Recommended Intakes of One-Carbon Nutrients through Prenatal Supplementation but the Supplement Composition , Including Choline , Requires Reconsideration 1 – 3. J. Nutr. 145, 1824–1834 (2015).
136. Sherwood, K. L., Houghton, L. A., Tarasuk, V. & O’Connor, D. L. One-third of pregnant and lactating women may not be meeting their folate requirements from diet alone based on mandated levels of folic acid fortification. J. Nutr. 136, 2820–6 (2006).
137. Fayyaz, F. et al. Folate, vitamin B12, and vitamin B6 status of a group of high socioeconomic status women in the Alberta Pregnancy Outcomes and Nutrition (APrON) cohort. Appl. Physiol. Nutr. Metab. 39, 1402–8 (2014).
138. Yeung, L., Yang, Q. & Berry, R. J. Contributions of Total Daily Intake of Folic Acid. JAMA 300, 2486–2487 (2008).
139. Guerra-Shinohara, E. M. et al. Relationship between total homocysteine and folate levels in pregnant women and their newborn babies according to maternal serum levels of vitamin B12. BJOG 109, 784–91 (2002).
140. Obeid, R., Munz, W., Jäger, M., Schmidt, W. & Herrmann, W. Biochemical indexes of the B vitamins in cord serum are predicted by maternal B vitamin status. Am. J. Clin. Nutr. 82, 133–139 (2005).
141. Takimoto, H. et al. Relationship between dietary folate intakes, maternal plasma total homocysteine and B-vitamins during pregnancy and fetal growth in Japan. Eur. J. Nutr. 46, 300–306 (2007).
142. Hure, A. J., Collins, C. E. & Smith, R. A longitudinal study of maternal folate and vitamin B12 status in pregnancy and postpartum, with the same infant markers at 6 months of age. Matern. Child Health J. 16, 792–801 (2012).
143. Plumptre, L. et al. High concentrations of folate and unmetabolized folic acid in a cohort of pregnant Canadian women and umbilical cord blood 1 – 2. Am. J. Clin. Nutr. 1–10 (2015). doi:10.3945/ajcn.115.110783.
144. Chalmers, B., Dzakpasu, S., Heaman, M. & Kaczorowski, J. The Canadian Maternity Experiences Survey: An Overview of Findings. J. Obstet. Gynaecol. Canada 31, 303 (2008).
145. Pentieva, K. et al. Evidence from a Randomized Trial That Exposure to Supplemental Folic Acid at Recommended Levels during Pregnancy Does Not Lead to Increased Unmetabolized Folic Acid Concentrations in Maternal or Cord Blood. J. Nutr. 146, 494–500
146. Barua, S., Kuizon, S. & Junaid, M. a. Folic acid supplementation in pregnancy and implications in health and disease. J. Biomed. Sci. 21, 77 (2014).
147. Bodnar, L. M., Tang, G., Ness, R. B., Harger, G. & Roberts, J. M. Periconceptional multivitamin use reduces the risk of preeclampsia. Am. J. Epidemiol. 164, 470–477 (2006).
130
148. Wen, S. W. et al. Folic acid supplementation in early second trimester and the risk of preeclampsia. Am. J. Obstet. Gynecol. 198, 45.e1-7 (2008).
149. Timmermans, S. et al. Folic acid is positively associated with uteroplacental vascular resistance: The Generation R Study. Nutr. Metab. Cardiovasc. Dis. 21, 54–61 (2011).
150. Li, Z. et al. Folic acid supplementation during early pregnancy and the risk of gestational hypertension and preeclampsia. Hypertension 61, 873–879 (2013).
151. Wang, Y. et al. Folic acid supplementation and dietary folate intake, and risk of preeclampsia. Eur J Clin Nutr 33, 395–401 (2015).
152. Zheng, J.-S. et al. Pre-conceptional intake of folic acid supplements is inversely associated with risk of preterm birth and small-for-gestational-age birth: a prospective cohort study. Br. J. Nutr. 115, 509–516 (2016).
153. Wen, S. W. et al. Folic acid supplementation in pregnancy and the risk of pre-eclampsia-A cohort study. PLoS One 11, 1–11 (2016).
154. Martinussen, M. P., Bracken, M. B., Triche, E. W., Jacobsen, G. W. & Risnes, K. R. Folic acid supplementation in early pregnancy and the risk of preeclampsia, small for gestational age offspring and preterm delivery. Eur. J. Obstet. Gynecol. Reprod. Biol. 195, 94–99 (2015).
155. Czeizel, A. E. Reduction of Urinary Tract and Cardiovascular Defects by Periconceptional Multivitamin Supplementation. 183, 179–183 (1996).
156. Shaw, G. M., Malley, C. D. O., Wasserman, C. R., Marie, M. & Lamer, E. J. American Journal of Medical Genetics 59536-545 (1995). 545, (1995).
157. Botto, L. D., Khoury, M. J., Mulinare, J. & Erickson, J. D. Periconceptional multivitamin use and the occurrence of conotruncal heart defects: results from a population-based, case-control study. Pediatrics 98, 911–917 (1996).
158. Van Beynum, I. M. et al. Protective effect of periconceptional folic acid supplements on the risk of congenital heart defects: A registry-based case-control study in the northern Netherlands. Eur. Heart J. 31, 464–471 (2010).
159. Li, X. et al. The association between periconceptional folic acid supplementation and congenital heart defects: A case-control study in china. Prev. Med. (Baltim). 56, 385–389 (2013).
160. Czeizel, A. E., Vereczkey, A. & Szabó, I. Folic acid in pregnant women associated with reduced prevalence of severe congenital heart defects in their children: A national population-based case-control study. Eur. J. Obstet. Gynecol. Reprod. Biol. 193, 34–39 (2015).
161. De-regil, L. M., Fernández-gaxiola, A. C., Dowswell, T. & Peña-, J. P. Effects and safety of periconceptional fotale supplementation for preventing birth defects. Cochrame Database Syst Rev. 2, 1–135 (2015).
162. Ba, H. & Za, B. Multiple-micronutrient supplementation for women during pregnancy ( Review ) SUMMARY OF FINDINGS FOR THE MAIN COMPARISON. (2016). doi:10.1002/14651858.CD004905.pub4.www.cochranelibrary.com
163. Sengpiel, V. et al. Folic acid supplementation, dietary folate intake during pregnancy and risk for spontaneous preterm delivery: a prospective observational cohort study. BMC Pregnancy Childbirth 13, 160 (2013).
164. Chen, L. W. et al. Maternal folate status, but not that of vitamins B-12 or B-6, is associated with gestational age and preterm birth risk in a multiethnic Asian population. J Nutr 145, 113–120 (2015).
165. Papadopoulou, E. et al. The effect of high doses of folic acid and iron supplementation in early-to-mid pregnancy on prematurity and fetal growth retardation: The mother-child cohort study in Crete, Greece (Rhea study). Eur. J. Nutr. 52, 327–336 (2013).
166. Kalhan, S. C. One carbon metabolism in pregnancy: Impact on maternal, fetal and neonatal health. Mol. Cell. Endocrinol. 435, 48–60 (2016).
167. Bergen, N. E. et al. Homocysteine and folate concentrations in early pregnancy and the risk of
131
adverse pregnancy outcomes: The generation R study. BJOG An Int. J. Obstet. Gynaecol. 119, 739–751 (2012).
168. Lindblad, B. et al. Folate, vitamin B12, and homocysteine levels in South Asian women with growth-retarded fetuses. Acta Obstet. Gynecol. Scand. 84, 1055–1061 (2005).
169. van Uitert, E. M. & Steegers-Theunissen, R. P. M. Influence of maternal folate status on human fetal growth parameters. Mol. Nutr. Food Res. 57, 582–595 (2013).
170. Maloney, C. a, Hay, S. M. & Rees, W. D. Folate deficiency during pregnancy impacts on methyl metabolism without affecting global DNA methylation in the rat fetus. Br. J. Nutr. 97, 1090–1098 (2007).
171. Giudicelli, F., Brabant, A., Grit, I., Parnet, P. & Amarger, V. Excess of Methyl Donor in the Perinatal Period Reduces Postnatal Leptin Secretion in Rat and Interacts with the Effect of Protein Content in Diet. 8, 1–10 (2013).
172. Tsang, V. et al. The Epigenetic Effects of a High Prenatal Folate Intake in Male Mouse Fetuses Exposed In Utero to Arsenic. Toxicol Appl Pharmacol 264, 439–450 (2012).
173. Greenop, K. R. et al. Maternal dietary intake of folate and vitamins B6 and B12 during pregnancy and risk of childhood brain tumors. Nutr. Cancer 66, 800–809 (2014).
174. Metayer, C. et al. Maternal supplementation with folic acid and other vitamins and risk of leukemia in offspring: a childhood leukemia international consortium study. Epidemiology 25, 811–22 (2014).
175. Ingrid Goh, Y., Bollano, E., Einarson, T. R. & Koren, G. Prenatal Multivitamin Supplementation and Rates of Congenital Anomalies: A Meta-Analysis. J. Obstet. Gynaecol. Canada 28, 680–689 (2006).
176. Schüz, J., Weihkopf, T. & Kaatsch, P. Medication use during pregnancy and the risk of childhood cancer in the offspring. Eur. J. Pediatr. 166, 433–441 (2007).
177. Orjuela, M. A. et al. Risk of retinoblastoma is associated with a maternal polymorphism in dihydrofolatereductase (DHFR) and prenatal folic acid intake. Cancer 118, 5912–5919 (2012).
178. Milne, E. et al. Folate Pathway Gene Polymorphisms, Maternal Folic Acid Use, and Risk of Childhood Acute Lymphoblastic Leukemia. Cancer Epidemiol. Biomarkers Prev. 24, 48–56 (2015).
179. Sie, K. K. et al. Effect of maternal and postweaning folic acid supplementation on colorectal cancer risk in the offspring. Gut 60, 1687–1694 (2011).
180. Ly, A. et al. Effect of maternal and postweaning folic acid supplementation on mammary tumor risk in the offspring. Cancer Res. 71, 988–997 (2011).
181. Chatzi, L. et al. Effect of high doses of folic acid supplementation in early pregnancy on child neurodevelopment at 18 months of age: the mother–child cohort ‘Rhea’ study in Crete, Greece. Public Health Nutr. 15, 1728–1736 (2012).
182. Roth, C., Magnus, P., Schjølberg, S., Reichborn-kjennerud, T. & Susser, E. Folic Acid Supplements in Pregnancy and Severe Language Delay in Children Christine. JAMA J. Am. Med. Assoc. 306, 1566–1573 (2013).
183. Suren, P. et al. Association between maternal use of folic acid supplements and risk of autism spectrum disorders in children. J. Am. Med. Assoc. 309, 570–577 (2013).
184. Braun, J. M. et al. Are autistic-behaviors in children related to prenatal vitamin use and maternal whole blood folate concentrations? Joseph. J. Autism Dev Disord. 44, 2602–2607 (2014).
185. Steenweg-de Graaff, J., Ghassabian, A., Jaddoe, V. W. V, Tiemeier, H. & Roza, S. J. Folate concentrations during pregnancy and autistic traits in the offspring. the Generation R Study. Eur. J. Public Health 25, 431–433 (2015).
186. Crider, K. S. et al. Prenatal folic acid and risk of asthma in children : a systematic review. Clin. Nutr. 2, 1272–1281 (2013).
187. Brown, S. B., Reeves, K. W. & Bertone-Johnson, E. R. Maternal folate exposure in pregnancy and childhood asthma and allergy: A systematic review. Nutr. Rev. 72, 55–64 (2014).
188. Wang, T. et al. Is Folate Status a Risk Factor for Asthma or Other Allergic Diseases? Allergy. Asthma
132
Immunol. Res. 7, 538 (2015). 189. Whitrow, M. J., Moore, V. M., Rumbold, A. R. & Davies, M. J. Effect of supplemental folic acid
in pregnancy on childhood asthma: A prospective birth cohort study. Am. J. Epidemiol. 170, 1486–1493 (2009).
190. Watanabe, F. Vitamin B12 Sources and Bioavailability. Exp. Biol. Med. 232, (2007). 191. Masih, S. Maternal Intakes and Sources of Folate and other One-carbon Nutrients in the Post-
fortification Era. (University of Toronto, 2013). at <https://tspace.library.utoronto.ca/handle/1807/43178>
192. Yang, Q. et al. Folic acid source, usual intake, and folate and vitamin B-12 status in US adults: National Health and Nutrition Examination Survey. Am. J. Clin. Nutr. 91, 64–72 (2010).
193. Shakur, Y. a., Tarasuk, V., Corey, P. & O’Connor, D. L. A comparison of micronutrient inadequacy and risk of high micronutrient intakes among vitamin and mineral supplement users and nonusers in Canada. J. Nutr. 142, 534–540 (2012).
194. Lindenbaum, J., Rosenberg, I. H., Wilson, P. W., Stabler, S. P. & Allen, R. H. Prevalence of cobalamin deficiency in the Framingham elderly population. Am. J. Clin. Nutr. 60, 2–11 (1994).
195. Carmel, R., Green, R., Rosenblatt, D. S. & Watkins, D. Update on cobalamin, folate, and homocysteine. Hematology Am. Soc. Hematol. Educ. Program 62–81 (2003). doi:10.1182/asheducation-2003.1.62
196. Molloy, A. M. et al. Maternal vitamin B12 status and risk of neural tube defects in a population with high neural tube defect prevalence and no folic acid fortification. Pediatrics 123, 917–923 (2009).
197. Molloy, a M., Kirke, P., Hillary, I., Weir, D. G. & Scott, J. M. Maternal serum folate and vitamin B12 concentrations in pregnancies associated with neural tube defects. Arch. Dis. Child. 60, 660–665 (1985).
198. Selhub, J., Jacques, P. F., Dallal, G., Choumenkovitch, S. & Rogers, G. The use of blood concentrations of vitamins and their respective functional indicators to define folate and vitamin B12 status. Food Nutr. Bull. 29, 67–73 (2008).
199. Carmel, R. Biomarkers of cobalamin (vitamin B-12) status in the epidemiologic\rsetting: a critical overview of context, applications, and performance\rcharacteristics of cobalamin, methylmalonic acid, and\r\nholotranscobalamin II. (2011). doi:10.3945/ajcn.111.013441.
200. Ryan-Harshmann, M. & Aldoori, W. Vitamin B12 and Health. Can. Fam. Physician 54, 536–41 (2008).
201. Wyckoff, K. F. & Ganji, V. Proportion of individuals with low serum vitamin B-12 concentrations without macrocytosis is higher in the post folic acid fortification period than in the pre folic acid fortification period. Am. J. Clin. Nutr. 86, 1187–92 (2007).
202. Gadgil, M. et al. Imbalance of folic acid and vitamin B12 is associated with birth outcome: an Indian pregnant women study. Eur. J. Clin. Nutr. 68, 726–729 (2014).
203. Cikot, R. J. L. M. et al. Longitudinal vitamin and homocysteine levels in normal pregnancy. Br. J. Nutr. 85, 49 (2007).
204. Visentin, C. E. et al. Low Serum Vitamin B-12 Concentrations Are Prevalent in a Cohort of Pregnant Canadian Women. J. Nutr. 146, 1035–42 (2016).
205. Wu, B. T. F., Innis, S. M., Mulder, K. A., Dyer, R. A. & King, D. J. Low plasma vitamin B-12 is associated with a lower pregnancy-associated rise in plasma free choline in Canadian pregnant women and lower postnatal growth rates in their male infants. Am. J. Clin. Nutr. 98, 1209–1217 (2013).
206. Obeid, R., Morkback, A. L., Windried, M., Ebba, N. & Herrmann, W. The Cobalamin-Binding Proteins Transcobalamin and Haptocorrin in Maternal and Cord Blood Sera at Birth. Clin. Chem. 52, 263–9 (2006).
207. Baker, H., DeAngelis, B., Holland, B., Gittens-Williams, L. & Barrett, T. Vitamin profile of 563 gravidas during trimesters of pregnancy. J. Am. Coll. Nutr. 21, 33–7 (2002).
133
208. Yetley, E. A. et al. Biomarkers of folate status in NHANES: A roundtable summary1-6. Am. J. Clin. Nutr. 94, (2011).
209. Herran, O. F., Ward, J. B. & Villamor, E. Vitamin B12 serostatus in Colombian children and adult women: results from a nationally representative survey. Public Health Nutr. 18, 836–843 (2015).
210. Graber, S. E., Scheffel, U., Hodknson, B. & Mcinmyre, P. A. Placental Transport of Vitamin B12 the Pregnant Rat. 50, 1000–1004 (1971).
211. Imbard, A., Benoist, J. & Blom, H. J. Neural Tube Defects , Folic Acid and Methylation. 4352–4389 (2013). doi:10.3390/ijerph10094352
212. Afman, L. a, Van Der Put, N. M., Thomas, C. M., Trijbels, J. M. & Blom, H. J. Reduced vitamin B12 binding by transcobalamin II increases the risk of neural tube defects. QJM 94, 159–66 (2001).
213. Kirke, P. N. et al. Maternal plasma folate and vitamin B12 are independent risk factors for neural tube defects. Q. J. Med. 86, 703–8 (1993).
214. van der Put, N. et al. Altered folate and vitamin B metabolism in families with spina bifida. 505–510 (1997).
215. Suarez, L., Hendricks, K., Felkner, M. & Gunter, E. Maternal serum homocysteine and risk for neural tube defects in a Texas-Mexico border population. Birth Defects Res. Part A - Clin. Mol. Teratol. 85, 574–581 (2003).
216. Groenen, P. M. W. et al. Marginal maternal vitamin B12 status increases the risk of offspring with spina bifida. Am. J. Obstet. Gynecol. 191, 11–17 (2004).
217. Ray, J. G. et al. Vitamin B12 and the risk of neural tube defects in a folic-acid-fortified population. Epidemiology 18, 362–366 (2007).
218. Acilmis, Y. G. et al. Homocysteine, folic acid and vitamin B12 levels in maternal and umbilical cord plasma and homocysteine levels in placenta in pregnant women with pre-eclampsia. J. Obstet. Gynaecol. Res. 37, 45–50 (2011).
219. Makedos, G. et al. Homocysteine, folic acid and B12 serum levels in pregnancy complicated with preeclampsia. Arch. Gynecol. Obstet. 275, 121–124 (2007).
220. Patrick, T. E., Powers, R. W., Daftary, A. R., Ness, R. B. & Roberts, J. M. Homocysteine and folic acid are inversely related in black women with preeclampsia. Hypertension 43, 1279–1282 (2004).
221. Ronnenberg, A. G. et al. Preconception homocysteine and B vitamin status and birth outcomes in Chinese women. Am. J. Clin. Nutr. 76, 1385–1391 (2002).
222. Bondevik, G. T. et al. Homocysteine and methylmalonic acid levels in pregnant Nepali women. Should cobalamin supplementation be considered? Eur. J. Clin. Nutr. 55, 856–64 (2001).
223. Dhobale, M. et al. Reduced folate, increased vitamin B12 and homocysteine concentrations in women delivering preterm. Ann. Nutr. Metab. 61, 7–14 (2012).
224. Furness, D. et al. Folate, Vitamin B12, Vitamin B6 and homocysteine: Impact on pregnancy outcome. Matern. Child Nutr. 9, 155–166 (2013).
225. Muthayya, S. et al. Low maternal vitamin B12 status is associated with intrauterine growth retardation in urban South Indians. Eur. J. Clin. Nutr. 60, 791–801 (2006).
226. Muthayya, S. et al. The relationship of neonatal serum vitamin b12 status with birth weight. Asia Pac J Clin Nutr 15, 538–43 (2006).
227. Sukumar, N. et al. Vitamin B12 Status among Pregnant Women in the UK and Its Association with Obesity and Gestational Diabetes. 6–15 (2016). doi:10.3390/nu8120768
228. Frelut, M.-L., Nicolas, J.-P., Guilland, J.-C. & de Courcy, G. P. Methylenetetrahydrofolate reductase 677 C->T polymorphism: a link between birth weight and insulin resistance in obese adolescents. Int. J. Pediatr. Obes. 6, e312–e317 (2011).
229. Yajnik, C. S. et al. Maternal homocysteine in pregnancy and offspring birthweight: Epidemiological associations and Mendelian randomization analysis. Int. J. Epidemiol. 43, 1487–1497 (2014).
230. Bhate, V. et al. Vitamin B12 status of pregnant Indian women and cognitive function in their 9-year-
134
old children. Food Nutr. Bull. 29, 249–254 (2008). 231. Bhate, V. K. et al. Vitamin B12 and folate during pregnancy and offspring motor, mental and social
development at 2 years of age. J. Dev. Orig. Health Dis. 3, 123–130 (2012). 232. Lövblad, K. O. et al. Retardation of myelination due to dietary vitamin B12 deficiency: cranial MRI
findings. Pediatr. Radiol. 27, 155–158 (1997). 233. Black, M. M. Effects of vitamin B12 and folate deficiency on brain development in children. Food
Nutr. Bull. 29, S126--31 (2008). 234. Doyle, J. J., Langevin, A. M. & Zipursky, A. Nutritional vitamin B12 deficiency in infancy: three
case reports and a review of the literature. Pediatr. Hematol. Oncol. 6, 161–172 (1989). 235. Adaikalakoteswari, A. et al. Low maternal vitamin B12 status is associated with lower cord blood
HDL cholesterol in white caucasians living in the UK. Nutrients 7, 2401–2414 (2015). 236. Horan, M. K., McGowan, C. A., Gibney, E. R., Donnelly, J. M. & Fionnuala M McAuliffe. The
association between maternal dietary micronutrient intake and neonatal anthropometry - secondary analysis from the ROLO study. Nutr. J. 14, 105 (2015).
237. Horan, M. K. et al. Maternal nutrition and glycaemic index during pregnancy impacts on offspring adiposity at 6 months of age???analysis from the ROLO randomised controlled trial. Nutrients 8, 13–15 (2016).
238. Stewart, C. P. et al. Low maternal vitamin B-12 status is associated with offspring insulin resistance regardless of antenatal micronutrient supplementation in rural Nepal. J. Nutr. 141, 1912–1917 (2011).
239. Aleliunas, R. E. et al. Folic Acid Supplementation of Female Mice , with or without Vitamin B-12 , before and during Pregnancy and Lactation Programs Adiposity and Vascular Health in Adult Male Offspring 1 – 3. 1–9 (2016). doi:10.3945/jn.115.227629.1
240. Danenberg, P. V & Danenberg, K. D. Effect of 5, 10-methylenetetrahydrofolate on the dissociation of 5-fluoro-2’-deoxyuridylate from thymidylate synthetase: evidence for an ordered mechanism. Biochemistry 17, 4018–4024 (1978).
241. Jiang, C. & Pugh, B. F. Nucleosome positioning and gene regulation: advances through genomics. 10, 161 (2009).
242. Saito, Y. & Jones, P. M. Epigenetic Activation of Tumor Suppressor MicroRNAs in Human Cancer Cells. Cell Cycle 5, 2220–2222 (2006).
243. Fabbri, M. et al. MicroRNA-29 family reverts aberrant methylation in lung cancer by targeting DNA methyltransferases 3A and 3B. 104, 15805–15810 (2007).
244. Friedman, J. M. et al. The Putative Tumor Suppressor microRNA-101 Modulates the Cancer Epigenome by Repressing the Polycomb Group Protein EZH2. 2623–2630 (2009). doi:10.1158/0008-5472.CAN-08-3114
245. Reik, W. Stability and flexibility of epigenetic gene regulation in mammalian development. Nature 447, 425–432 (2007).
246. Jang, H. S., Shin, W. J., Lee, J. E. & Do, J. T. CpG and Non-CpG Methylation in Epigenetic Gene Regulation and Brain Function. Genes (Basel). 8, 1–20 (2017).
247. Esteller, M. Cancer epigenomics: DNA methylomes and histone-modification maps. Nat. Rev. Genet. 8, 286–298 (2007).
248. Herman, J. G. & Baylin, S. B. Gene Silencing in Cancer in Association with Promoter Hypermethylation. N. Engl. J. Med. 349, 2042–2054 (2003).
249. Laird, P. W. Principles and challenges of genomewide DNA methylation analysis. Nat. Rev. Genet. 11, 191–203 (2010).
250. Wu, H. & Zhang, Y. Reversing DNA methylation: Mechanisms, genomics, and biological functions. Cell 156, 45–68 (2014).
251. Robertson, K. D. & Wolffe, A. P. DNA methylation in health and disease. Nat. Rev. Genet. 1, 11–19 (2000).
135
252. Jones, P. a & Baylin, S. B. The fundamental role of epigenetic events in cancer. Nat. Rev. Genet. 3, 415–28 (2002).
253. Michel, S. et al. Farm exposure and time trends in early childhood may influence DNA methylation in genes related to asthma and allergy. Allergy Eur. J. Allergy Clin. Immunol. 68, 355–364 (2013).
254. Duthie, S. J. Epigenetic modifications and human pathologies: cancer and CVD. Proc Nutr Soc 70, 47–56 (2011).
255. Veenendaal, M. V. E. et al. Transgenerational effects of prenatal exposure to the 1944-45 Dutch famine. BJOG An Int. J. Obstet. Gynaecol. 120, 548–553 (2013).
256. Moen, E. L. et al. New themes in the biological functions of 5-methylcytosine and 5-hydroxymethylcytosine. Immunol. Rev. 263, 36–49 (2015).
257. James, S. J., Melnyk, S., Pogribna, M., Pogribny, I. P. & Caudill, M. A. Elevation in S-Adenosylhomocysteine and DNA Hypomethylation: Potential Epigenetic Mechanism for Homocysteine-Related Pathology. 2361–2366 (2002).
258. Zhu, J.-K. Active DNA demethylation mediated by DNA glycosylases. Annu. Rev. Genet. 43, 143–66 (2009).
259. Crider, K. S., Yang, T. P., Berry, R. J. & Bailey, L. B. Folate and DNA Methylation : A Review of Molecular Mechanisms and the Evidence for Folate ’ s Role. Am. Soc. Nutr. 3, 21–38 (2012).
260. Bird, A. DNA methylation patterns and epigenetic memory DNA methylation patterns and epigenetic memory. Genes Dev. 6–21 (2002). doi:10.1101/gad.947102
261. Bird, A. P. & Wolffe, A. P. Methylation-Induced Repression— Belts, Braces, and Chromatin. Cell 99, 451–454 (1999).
262. Baylin, S. DNA methylation and gene silencing in cancer. Nat. Clin. Pract. Oncol. 2, 21231 (2005). 263. Seisenberger, S. et al. Reprogramming DNA methylation in the mammalian life cycle : building and
breaking epigenetic barriers. 95, 1–11 (2012). 264. Li E & R, J. in DNA Alterations in Cancer: Genetic and Epigenetic Changes (ed. M, E.) 351–65 (Natick,
MA: Eaton Publishing, 2000). 265. Lister, R. et al. Human DNA methylomes at base resolution show widespread epigenomic
differences. Nature 462, 315–22 (2009). 266. Laird, P. W. THE POWER AND THE PROMISE OF DNA METHYLATION MARKERS. Nat.
Rev. Caner 3, 253–266 (2003). 267. Bibikova, M. et al. Genome-wide DNA methylation profiling using Infinium ® assay. Epigenomics 1,
177–200 (2009). 268. Wadhwa, P. D., Buss, C., Entringer, S. & Swanson, J. M. Developmental Origins of Health and
Disease: Brief History of the Approach and Current Focus on Epigenetic Mechanisms. Natl. Inst. Heal. 27, 358–368 (2010).
269. Gillman, M. W. Developmental Origins of Health and Disease. N Engl J Med 353, 1848–1850 (2005).
270. De Rooij, S. R. et al. Impaired insulin secretion after prenatal exposure to the Dutch famine. Diabetes Care 29, 1897–1901 (2006).
271. Painter, R. C. et al. Early onset of coronary artery disease after prenatal exposure to the. 322–327 (2006).
272. Bygren, L. O. et al. Change in paternal grandmothers’ early food supply influenced cardiovascular mortality of the female grandchildren. BMC Genet. 15, 12 (2014).
273. Gluckman, P. D., Hanson, M. A. & Spencer, H. G. Predictive adaptive responses and human evolution. Trends Ecol. Evol. 20, 527–533 (2005).
274. Bateson, P., Gluckman, P. & Hanson, M. The biology of developmental plasticity and the Predictive Adaptive Response hypothesis. J. Physiol. 592, 2357–68 (2014).
275. Cho, C. E. et al. High folate gestational and post-weaning diets alter hypothalamic feeding pathways by DNA methylation in Wistar rat offspring. Epigenetics 8, 710–719 (2013).
136
276. Jones, a P. & Friedman, M. I. Obesity and adipocyte abnormalities in offspring of rats undernourished during pregnancy. Science 215, 1518–9 (1982).
277. Huot, P. S. P. et al. High folic acid intake during pregnancy lowers body weight and reduces femoral area and strength in female rat offspring. J. Osteoporos. 2013, (2013).
278. Selevan, S. G., Kimmel, C. A. & Mendola, P. Identifying critical windows of exposure for children’s health. Environ. Health Perspect. 108, 451–455 (2000).
279. Cooper, C., Harvey, N., Cole, Z., Hanson, M. & Dennison, E. in Early Nutrition Programming and Health Outcomes in Later Life 31–39 (Springer, 2009).
280. Hanson, M. a & Gluckman, P. D. Early developmental conditioning of later health and disease: physiology or pathophysiology? Physiol. Rev. 94, 1027–76 (2014).
281. Wolff GL, Kodell RL, Moore SR, C. C. Maternal epigenetics and methyl supplements affect agouti gene expression in Avy/a mice. Faseb J 12, 949–57 (1998).
282. Waterland, R. A. & Jirtle, R. L. Transposable elements: targets for early nutritional effects on epigenetic gene regulation. Mol. Cell. Biol. 23, 5293–300 (2003).
283. Rosenfeld, C. S. Animal models to study environmental epigenetics. Biol. Reprod. 82, 473–88 (2010).
284. Waterland, R. A. et al. Maternal methyl supplements increase offspring DNA methylation at Axin Fused. Genesis 410–6 (2006).
285. Gicquel, C. & Bouc, Y. Le. Epigenetic regulation and fetal programming. 22, 1–16 (2008). 286. Maclennan, N. K. et al. Uteroplacental insufficiency alters DNA methylation , one-carbon
metabolism , and histone acetylation in IUGR rats. 1, 43–50 (2004). 287. Vanhees, K., Vonhögen, I. G. C., Schooten, F. J. Van & Godschalk, R. W. L. You are what you
eat , and so are your children : the impact of micronutrients on the epigenetic programming of offspring. (2013). doi:10.1007/s00018-013-1427-9
288. Guéant, J., Elakoum, R. & Ziegler, O. Nutritional models of foetal programming and nutrigenomic and epigenomic dysregulations of fatty acid metabolism in the liver and heart. 833–850 (2014). doi:10.1007/s00424-013-1339-4
289. Schaible, T. D., Harris, R. A., Dowd, S. E., Smith, C. W. & Kellermayer, R. Maternal methyl-donor supplementation induces prolonged murine offspring colitis susceptibility in association with mucosal epigenetic and microbiomic changes. 20, 1687–1696 (2011).
290. Braunschweig, M., Jagannathan, V., Gutzwiller, A. & Bee, G. Investigations on Transgenerational Epigenetic Response Down the Male Line in F2 Pigs. 7, (2012).
291. Finnell, R. H. et al. DNA methylation in Folbp1 knockout mice supplemented with folic acid during gestation. J. Nutr. 132, 2457S–2461S (2002).
292. Sie, K. K. Y. et al. Effect of maternal and postweaning folic acid supplementation on global and gene-specific DNA methylation in the liver of the rat offspring. Mol. Nutr. Food Res. 57, 677–685 (2013).
293. Kim, J.-M., Hong, K., Lee, J. H., Lee, S. & Chang, N. Effect of folate deficiency on placental DNA methylation in hyperhomocysteinemic rats. J. Nutr. Biochem. 20, 172–176 (2009).
294. Lillycrop, K. A., Phillips, E. S., Jackson, A. A., Hanson, M. A. & Burdge, G. C. Dietary protein restriction of pregnant rats induces and folic acid supplementation prevents epigenetic modification of hepatic gene expression in the offspring. J. Nutr. 135, 1382–1386 (2005).
295. Okamura, M., Inagaki, T., Tanaka, T. & Sakai, J. Role of histone methylation and demethylation in adipogenesis and obesity. Organogenesis 6, 24–32 (2010).
296. Drake, A. J. et al. An unbalanced maternal diet in pregnancy associates with offspring epigenetic changes in genes controlling glucocorticoid action and foetal growth. Clin. Endocrinol. (Oxf). 77, 808–815 (2012).
297. Janesick, A. & Blumberg, B. Obesogens, Stem Cells and the Developmental Programming of Obesity. Int J Androl 35, 437–448 (2013).
137
298. Vickers, M. H. Early life nutrition, epigenetics and programming of later life disease. Nutrients 6, 2165–2178 (2014).
299. Kulkarni, A. et al. Effects of altered maternal folic acid, vitamin B12 and docosahexaenoic acid on placental global DNA methylation patterns in Wistar rats. PLoS One 6, e17706 (2011).
300. Gueant, J. L., Namour, F., Gueant-Rodriguez, R. M. & Daval, J. L. Folate and fetal programming: A play in epigenomics? Trends Endocrinol. Metab. 24, 279–289 (2013).
301. Engeham, S. F., Haase, A. & Langley-Evans, S. C. Supplementation of a maternal low-protein diet in rat pregnancy with folic acid ameliorates programming effects upon feeding behaviour in the absence of disturbances to the methionine-homocysteine cycle. Br. J. Nutr. 103, 996–1007 (2010).
302. Steegers-Theunissen, R. P. et al. Periconceptional maternal folic acid use of 400 μg per day is related to increased methylation of the IGF2 gene in the very young child. PLoS One 4, 1–5 (2009).
303. Hoyo, C. et al. Methylation variation at IGF2 differentially methylated regions and maternal folic acid use before and during pregnancy. Epigenetics 6, 928–936 (2011).
304. van Mil, N. H. et al. Determinants of maternal pregnancy one-carbon metabolism and newborn human DNA methylation profiles. Reproduction 148, 581–92 (2014).
305. Azzi, S. et al. Degree of methylation of ZAC1 (PLAGL1) is associated with prenatal and post-natal growth in healthy infants of the EDEN mother child cohort. Epigenetics 9, 338–345 (2013).
306. Hoyo, C. et al. Erythrocyte folate concentrations, CpG methylation at genomically imprinted domains, and birth weight in a multiethnic newborn cohort. Epigenetics 9, 1120–1130 (2014).
307. Haggarty, P. et al. Folate in pregnancy and imprinted gene and repeat element methylation in the offspring. Am. J. Clin. Nutr. 97, 94–99 (2013).
308. Ba, Y. et al. Relationship of folate, vitamin B12 and methylation of insulin-like growth factor-II in maternal and cord blood. Eur. J. Clin. Nutr. 65, 480–5 (2011).
309. Huang, T. et al. DNA methylation variants at HIF3A locus, B-vitamin intake, and long-term weight change: Gene-diet interactions in two U.S. Cohorts. Diabetes 64, 3146–3154 (2015).
310. Lopomo, A., Burgio, E. & Migliore, L. Epigenetics of Obesity. Genet. Monogenic Syndr. Obes. 140, 151–184 (2016).
311. van Dijk, S. J., Molloy, P. L., Varinli, H., Morrison, J. L. & Muhlhausler, B. S. Epigenetics and human obesity. Int. J. Obes. 39, 85–97 (2015).
312. Burgio, E., Lopomo, A. & Migliore, L. Obesity and diabetes: from genetics to epigenetics. Mol. Biol. Rep. 42, 799–818 (2015).
313. Pinnick, K. E. & Karpe, F. Symposium 1 : Nutrition and epigenetics DNA methylation of genes in adipose tissue Proceedings of the Nutrition Society. 57–63 (2011). doi:10.1017/S0029665110004015
314. van Dijk, S. J., Tellam, R. L., Morrison, J. L., Muhlhausler, B. S. & Molloy, P. L. Recent developments on the role of epigenetics in obesity and metabolic disease. Clin. Epigenetics 7, 66 (2015).
315. Lemche, E., Chaban, O. S. & Lemche, A. V. Neuroendorine and Epigentic Mechanisms Subserving Autonomic imbalance and HPA dysfunction in the metabolic syndrome. Front. Neurosci. 10, 1–27 (2016).
316. Whincup, P. H. et al. Birth Weight and Risk of Type 2 Diabetes A Systematic Review. 300, (2017). 317. Obermann-borst, S. A. et al. Duration of breastfeeding and gender are associated with methylation
of the LEPTIN gene in very young children. 1–6 (2013). doi:10.1038/pr.2013.95 318. Hermsdorff, H. H. et al. Cytokine TNF-alpha promoter methylation in peripheral white blood
cells : Relationship with circulating TNF a , truncal fat and n-6 PUFA intake in young women. Cytokine (2013). doi:10.1016/j.cyto.2013.05.028
319. Stepanow, S. et al. Methylation of Human MCHR1. 6, (2011). 320. Barres, R. et al. Report Weight Loss after Gastric Bypass Surgery in Human Obesity Remodels
138
Promoter Methylation. (2013). doi:10.1016/j.celrep.2013.03.018 321. Agha, G. et al. Adiposity is associated with DNA methylation profile in adipose tissue. Int. J.
Epidemiol. 44, 1277–1287 (2015). 322. Pan, H. et al. HIF3A association with adiposity: the story begins before birth. Epigenomics 7, 937–
950 (2015). 323. Liu, X. et al. Maternal Preconception Body Mass Index and Offspring Cord Blood DNA
Methylation: Exploration of Early Life Origins of Disease Xin. 55, 223–230 (2015). 324. Simpkin, A. J. et al. Longitudinal analysis of DNA methylation associated with birth weight and
gestational age. Hum. Mol. Genet. 24, 3752–3763 (2015). 325. Mol, H. & Clin, B. Adipose tissue and bone : role of PPAR γ in adipogenesis and osteogenesis. 15,
105–113 (2013). 326. Milagro, F. I., Mansego, M. L., De Miguel, C. & Martinez, J. A. Dietary factors, epigenetic
modifications and obesity outcomes: Progresses and perspectives. Mol. Aspects Med. 34, 782–812 (2013).
327. Zhao, J., Goldberg, J. & Vaccarino, V. Promoter methylation of serotonin transporter gene is associated with obesity measures : a monozygotic twin study. Int. J. Obes. 37, 140–145 (2012).
328. Kuehnen, P. et al. An Alu Element – Associated Hypermethylation Variant of the POMC Gene Is Associated with Childhood Obesity. 8, 1–12 (2012).
329. Huang, R. et al. DNA methylation of the IGF2 / H19 imprinting control region and adiposity distribution in young adults. 1–11 (2012).
330. Fryer, A. a. et al. LINE-1 DNA methylation is inversely correlated with cord plasma homocysteine in man: A preliminary study. Epigenetics 4, 394–398 (2009).
331. Fryer, A. A. et al. Quantitative, high-resolution epigenetic profling of CpG loci identifies associations with cord blood plasma homocysteine and birth weight in humans. Epigenetics 6, 86–94 (2011).
332. Ba, Y. et al. Relationship of folate, vitamin B12 and methylation of insulin-like growth factor-II in maternal and cord blood. Eur. J. Clin. Nutr. 65, 480–5 (2011).
333. Boeke, C. E. et al. Gestational intake of methyl donors and global LINE-1 DNA methylation in maternal and cord blood: Prospective results from a folate-replete population. Epigenetics 7, 253–260 (2012).
334. McKay, J. A. et al. Genetic and non-genetic influences during pregnancy on infant global and site specific DNA methylation: role for folate gene variants and vitamin B12. PLoS One 7, e33290 (2012).
335. Dominguez-Salas, P. et al. Maternal nutrition at conception modulates DNA methylation of human metastable epialleles. Nat. Commun. 5, 3746 (2014).
336. Saldana, T. M., Siega-Riz, A. M. & Adair, L. S. Effect of macronutrient intake on the development of glucose intolerance during pregnancy. Am. J. Clin. Nutr. 79, 479–486 (2004).
337. Wilson, R. D. Pre-conceptional Vitamin/Folic Acid Supplementation 2007: The Use of Folic Acid in Combination With a Multivitamin Supplement for the Prevention of Neural Tube Defects and Other Congenital Anomalies. J. Obstet. Gynaecol. Canada 29, 1003–1013 (2007).
338. Carmel, R. Subtle cobalamin deficiency. Ann. Intern. Med. 124, 338–340 (1996). 339. Statistics Canada. Vitamin B12 status of Canadians, 2009 to 2011. (2013). at
<https://www.statcan.gc.ca/pub/82-625-x/2012001/article/11731-eng.htm#n1> 340. Markunas, C. A. et al. Maternal Age at Delivery Is Associated with an Epigenetic Signature in Both
Newborns and Adults. 1–11 (2016). doi:10.1371/journal.pone.0156361 341. Sharp, G. C. et al. Maternal pre-pregnancy BMI and gestational weight gain , offspring DNA
methylation and later offspring adiposity : findings from the Avon Longitudinal Study of Parents and Children. 1288–1304 (2015). doi:10.1093/ije/dyv042
139
342. Almgren, M., Schlinzig, T., Gomez-cabrero, D., Norman, M. & Ekstro, T. J. Cesarean delivery and hematopoietic stem cell epigenetics in the newborn infant : implications for future health ? Am. J. Obstet. Gynecol. 211, 502.e1-502.e8 (2014).
343. Yousefi, P. et al. Sex differences in DNA methylation assessed by 450 K BeadChip in newborns. BMC Genomics 1–12 (2015). doi:10.1186/s12864-015-2034-y
344. Mozhui, K., Smith, A. K. & Tylavsky, F. A. Ancestry Dependent DNA Methylation and Influence of Maternal Nutrition. 1–17 (2015). doi:10.1371/journal.pone.0118466
345. Armstrong, D. A. et al. Maternal smoking during pregnancy is associated with mitochondrial DNA methylation. 1–9 (2016). doi:10.1093/eep/dvw020
346. Austin, P. C. An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies. 399–424 (2011). doi:10.1080/00273171.2011.568786
347. Illumina. Infinium MethylationEPIC BeadChip. (2017). 348. Infinium ® HD Assay Methylation Protocol Guide. (2015). 349. De, S. et al. Aberration in DNA Methylation in B-Cell Lymphomas Has a Complex Origin and
Increases with Disease Severity. PLoS Genet. 9, (2013). 350. Harris, A. Illumina Infinium 450K Array. 351. Systems, I. Calculating and Interpreting the p-values for Functions , Pathways and Lists in IPA. 352. Sze, M. A., Abbasi, M., Hogg, J. C. & Sin, D. D. A Comparison between Droplet Digital and
Quantitative PCR in the Analysis of Bacterial 16S Load in Lung Tissue Samples from Control and COPD. 9, 1–8 (2016).
353. Manual, I. QX200 TM Droplet Reader and QuantaSoft TM Software. 354. Paul D. Thomas, Michael J. Campbell, Anish Kejariwal, Huaiyu Mi, Brian Karlak, Robin
Daverman, Karen Diemer, Anushya Muruganujan, A. N. PANTHER: a library of protein families and subfamilies indexed by function. Genome Res. 13, 2129–2141 (2003).
355. Choi, S. et al. Biochemical and Molecular Actions of Nutrients Vitamin B-12 Deficiency Induces Anomalies of Base Substitution and Methylation in the DNA of Rat Colonic Epithelium 1. 750–755 (2004).
356. Brunaud, L., Alberto, J., Ayav, A. & Gérard, P. Original Paper : Metabolism Vitamin B 12 Is a Strong Determinant of Low Methionine Synthase Activity and DNA Hypomethylation in Gastrectomized Rats. 133–140 (2003). doi:10.1159/000075307
357. Piyathilake, C. J. et al. Localized Folate and Vitamin B-12 Deficiency in Squamous Cell Lung Cancer Is Associated With Global DNA Hypomethylation. 37, 99–107 (2000).
358. Matthews, R. G. & Haywood, B. J. Inhibition of Pig Liver Methylenetetrahydrofolate Reductase by Dihydrofolate : Some Mechanistic and Regulatory Implications1 ". 18, (1985).
359. Kamen, B. A. et al. Lack of Dihydrofolate Reducíase in Human Tumor and Leukemia Cells In Vivo. 2, 133–138 (1985).
360. Sowers, R. et al. mRNA Expression Levels of E2F Transcription Factors Correlate with Dihydrofolate Reductase , Reduced Folate Carrier , and Thymidylate Synthase mRNA Expression in Osteosarcoma 1. 2, 535–541 (2003).
361. Nijhout, H. F., Reed, M. C., Budu, P. & Ulrich, C. M. A Mathematical Model of the Folate Cycle. 279, 55008–55016 (2004).
362. Gadgil, M. et al. Imbalance of folic acid and vitamin B12 is associated with birth outcome: an Indian pregnant women study. Eur. J. Clin. Nutr. 68, 726–9 (2014).
363. Visentin, C. E. et al. Maternal Choline Status , but Not Fetal Genotype , Influences Cord Plasma Choline. J. Nutr. 145, 1491–7 (2015).
364. Casamar, E. P. et al. DNA methylation signature of human fetal alcohol spectrum disorder. Epigenetics Chromatin 1–20 (2016). doi:10.1186/s13072-016-0074-4
365. Demetriou, C. A., Veldhoven, K. Van, Relton, C., Stringhini, S. & Kyriacou, K. Biological
140
embedding of early-life exposures and disease risk in humans : a role for DNA methylation. Eur J Clin Nutr 45, 303–332 (2015).
366. Szeto, I. M. Y. et al. High multivitamin intake by Wistar rats during pregnancy results in increased food intake and components of the metabolic syndrome in male offspring. 575–582 (2008). doi:10.1152/ajpregu.90354.2008.
367. Szeto, I. M. Y., Das, P. J., Aziz, A. & Anderson, G. H. Multivitamin supplementation of Wistar rats during pregnancy accelerates the development of obesity in offspring fed an obesogenic diet. 364–372 (2009). doi:10.1038/ijo.2008.281
368. Rogne, T. Europe PMC Funders Group Maternal vitamin B12 in pregnancy and risk of preterm birth and low birth weight : A systematic review and individual participant data meta-analysis. 185, 212–223 (2017).
369. Lm, D., Ac, F., Dowswell, T. & Jp, P. Effects and safety of periconceptional folate supplementation for preventing birth defects ( Review ). (2010).
370. Khadem, N. et al. Relationship between Low Birth Weight Neonate and Maternal Serum Zinc Concentration. 14, 240–244 (2012).
371. Reuther, G. W. et al. Leukemia-associated Rho guanine nucleotide exchange factor, a Dbl family protein found mutated in leukemia, causes transformation by activation of RhoA. J. Biol. Chem. 276, 27145–27151 (2001).
372. Wang, Q. et al. Thrombin and lysophosphatidic acid receptors utilize distinct rhoGEFs in prostate cancer cells. J. Biol. Chem. 279, 28831–28834 (2004).
373. Bourguignon, L. Y. W., Gilad, E., Brightman, A., Diedrich, F. & Singleton, P. Hyaluronan-CD44 interaction with leukemia-associated RhoGEF and epidermal growth factor receptor promotes Rho/Ras co-activation, phospholipase C epsilon-Ca2+ signaling, and cytoskeleton modification in head and neck squamous cell carcinoma cells. J. Biol. Chem. 281, 14026–14040 (2006).
374. Klarlund, J. K. et al. Signaling complexes of the FERM domain-containing protein GRSP1 bound to ARF exchange factor GRP1. J. Biol. Chem. 276, 40065–40070 (2001).
375. Gutierrez-Juarez, R. et al. Critical role of stearoyl-CoA desaturase-1 (SCD1) in the onset of diet-induced hepatic insulin resistance. J. Clin. Invest. 116, 1686–1695 (2006).
376. Attie, A. D. et al. Relationship between stearoyl-CoA desaturase activity and plasma triglycerides in human and mouse hypertriglyceridemia. J. Lipid Res. 43, 1899–1907 (2002).
141
Appendix
Appendix A. Analysis of blood biomarkers
Aliquots of maternal and cord blood stored in biorepository were used to analyse blood biomarkers of the
PREFORM study folate and B12 measurements.
Maternal blood Umbilical cord blood Assay
Serum folate Serum folate Access Folate System (immunoassay)
RBC folate RBC folate Elecsys Folate Assay (immunoassay)
Plasma unmetabolized FA (UMFA)
Plasma UMFA LC-MS/MS
Plasma homocysteine Plasma homocysteine Synchron LX20/DxC System (immunoassay)
Serum B12
Serum B12
Access competitive-binding immunoenzymatic assay
Plasma Methylmalonic acid (MMA)
Plasma MMA LC-MS/MS
Plasma pyridoxal 5’ phosphate (PLP)
Plasma PLP (vitamin B6) Non-radioactive enzymatic
assay
Plasma choline, betaine, DMG, TMAO
Plasma choline, betaine, DMG, TMAO
LC-MS/MS
Plasma formate Plasma formate GC-MS
Global DNA methylation and hydroxymethylation
LC-MS/MS
38 SNPS in 26 one-carbon nutrient related genes
Genotyping
142
Appendix B. Final propensity score model in SAS code
proc logistic descending data = cutoffs; Model case = a_b6 infant_sex a_tchol mom_age caesarian preg_smoke preg_smoke*caesarian*a_b6; Output out = propensity2 p = p; run; Data propensity2; set propensity2; propensity_score = log(p/(1-p)); run; proc means data = propensity2 std; var propensity_score; output out=temp std=std; run; data temp; set temp; std1=0.2*std; run; proc print data= temp; run; %gmatch (data = propensity2, Group = case , Id = subject_id, Mvars = propensity_score a_bmi, Wts = 1 1, Dmaxk =0.39568 3, ncontls = 1, Seedca = 1234, Seedco = 5678, Print = y, Out = matches);
143
Appendix C. Histograms of CpG island and non-CpG island methylation
Illumina Infinium MethylationEPIC Array contains additional probes in non-CpG islands resulting in a bias
towards hypermethylated sites. This was controlled for in the analyses by separating the global DNA
methylation analysis by total, CpG island and non-CpG island sites. GenomeStudio was used to generate bar
graphs for all 44 subjects showing the distribution of DNA methylation depending on CpG site probe
location. The dataset with the total amount of CpG sites (Appendix Figure 1B) shows that in both cases
and controls, there is a high occurrence of both hypo- and hypermethylation. The level of hypermethylation
decreases when only accounting for CpG islands only (Appendix Figure 2B). In the non-CpG island
dataset (Appendix Figure 3B), the additional probes are clearly depicted with a high level of
hypermethylation like the total dataset.
Appendix Figure 1C. Total CpG site dataset including CpG and non-CpG islands. The x-axis indicates the level of DNA methylation and the y-axis indicates the number of occurrences. Cases are shown on the left and controls on the right.
144
Appendix Figure 2C. CpG island only dataset. The x-axis indicates the level of DNA methylation and the y-axis indicates the number of occurrences. Cases are shown on the top and controls on the bottom. Appendix Figure 3C. Non-CpG island only dataset. The x-axis indicates the level of DNA methylation and the y-axis indicates the number of occurrences. Cases are shown on the top and controls on the bottom.
145
Appendix D. IPA functional analysis summary table of all four datasets
Datasets
Mann-Whitney test T-test IPA Functions 371 Total 219 Hypo only 162 Hyper only 66 Total Diseases and
Disorders Cancer, GI, Endocrine,
Organ injury and
abnormalities
Cancer, GI, Endocrine, Organ
injury and abnormalities
Cancer, GI, Endocrine, Organ
injury and abnormalities,
Metabolic
Cancer, Neurological, Organ
injury and abnormalities
Molecular and Cellular
Functions
Cellular movement, Cell
cycle, growth and
proliferation, function and meabolism
Cell movementand cycle, Lipid
metabolism, Small molecule
biochemistry
Cell growth, development, cycle, death and survival, Gene expression
Carbohydrate and Lipid metabolism, Cell cycle, Small
molecule biochemistry
Physiological System
Development and Function
Hematological, Reproductive, Cardiovasular,
Tissue and Embryonic
Tissue, Organismal, Connective Tissue,
Cardiovasular, Skeletal
Hematological, Lymphoid,
Reproductive and Connective Tissue
Behaviour, Embryonic,
Endothelial, Hair, Skin, Nervous
Networks Cancer, Connective
Tissue disorders, Cell
signalling, Developmental
disorder
Cell death, Development,
Growth, Proliferation,
Movement, Cancer, Molecular Transport
Organismal Development,
Growth, Proliferation,
Cardiovascular, Cell assembly,
Organization, DNA replication and repair,
Cell cycle, Death, Survival, Cancer
Celll Death, Survival, Development,
Growth, Proliferation, Cancer,
Hematological, Cardiovascular, Cell
assembly, DNA replication and repair