Variation of global DNA methylation levels with age and in autistic children
Tsang et al. Human Genomics
Variation of global DNA methylation levels with age and in autistic children
Shui-Ying Tsang 1 2
Tanveer Ahmad 1 2
Flora W. K. Mat 2
Cunyou Zhao 2 4
Shifu Xiao 0
Kun Xia 3
Hong Xue 2
0 Department of Geriatric Psychiatry, Shanghai Mental Health Center, Shanghai Jiaotong University School of Medicine , Shanghai 200030 , China
1 Equal contributors
2 Division of Life Science, Applied Genomics Centre and Centre for Statistical Science, Hong Kong University of Science and Technology , Clear Water Bay, Hong Kong , China
3 The State Key Laboratory of Medical Genetics, Central South University , Changsha, Hunan 410078 , China
4 Department of Medical Genetics, School of Basic Medical Science, Southern Medical University , Guangzhou, Guangdong 510515 , China
Background: The change in epigenetic signatures, in particular DNA methylation, has been proposed as risk markers for various age-related diseases. However, the course of variation in methylation levels with age, the difference in methylation between genders, and methylation-disease association at the whole genome level is unclear. In the present study, genome-wide methylation levels in DNA extracted from peripheral blood for 2116 healthy Chinese in the 2-97 age range and 280 autistic trios were examined using the fluorescence polarizationbased genome-wide DNA methylation quantification method developed by us. Results: Genome-wide or global DNA methylation levels proceeded through multiple phases of variation with age, consisting of a steady increase from age 2 to 25 (r = 0.382) and another rise from age 41 to 55 to reach a peak level of ~80 % (r = 0.265), followed by a sharp decrease to ~40 % in the mid-1970s (age 56 to 75; r = −0.395) and leveling off thereafter. Significant gender effect in methylation levels was observed only for the 41-55 age group in which methylation in females was significantly higher than in males (p = 0.010). In addition, global methylation level was significantly higher in autistic children than in age-matched healthy children (p < 0.001). Conclusions: The multiphasic nature of changes in global methylation levels with age was delineated, and investigation into the factors underlying this profile will be essential to a proper understanding of the aging process. Furthermore, this first report of global hypermethylation in autistic children also illustrates the importance of agematched controls in characterization of disease-associated variations in DNA methylation.
Aging; Autism spectrum disorder; CpG methylation; Developmental epigenetics; Genome-wide methylation quantification
Genetic changes can alter the genomic DNA sequence
through point mutations, insertions, deletions, copy
number variations, and chromosomal rearrangements while
epigenetic modifications can modulate phenotype and
gene expressions. DNA methylation is the most common
epigenetic modification that plays an essential role in the
regulation of tissue-specific gene expression, cellular
differentiation, chromosome stabilization, genomic imprinting,
and suppression of transposable element mobility [
DNA methylation through DNA methyltransferases
convert cytosine to 5-methycytocine, with the majority of the
conversions occurring at CpG islands found in gene
promoter regions. Aberrant DNA methylation patterns have
long been associated with various human diseases
including cancers, cardiovascular diseases, psychotic disorders,
and autism [
Changes in epigenetics signatures, and in particular
DNA methylation, have been reported to occur in
normal physiological development and aging, and
alterations in DNA methylation associated with the signaling
and regulation of transcription have been demonstrated
in some genes [
]. Aging is the gradual deterioration
of various body functions and represents an important
risk factor for various age-related diseases such as
cancer, neurodegenerative disorders, cardiovascular diseases,
and type 2 diabetes mellitus . Several studies have
examined DNA methylation changes in old age as
disease risk factor, focusing mostly on CpG islands and
promoter regions in specific gene [
]. However, the
characterization of lifelong age-related changes in DNA
methylation at the whole genome level has remained a
largely unexplored area.
During the past decades, various HPLC-based,
sequencing-based, (e.g., bisulfite-sequencing and
methylated DNA immunoprecipitation) and microarray-based
methods have been introduced to quantitate genomic
DNA methylation [
]. Although these methods enable
high-resolution and detailed methylation profiles of
individual genes, they are time-consuming and incapable of
measuring whole genome methylation levels accurately.
Recently, a number of methods have been developed to
render possible the measurement of whole genome
methylation levels, including the LUminometric Methylation
Assay (LUMA) method [
], the ELISA-based approach
], and the fluorescence polarization DNA methylation
(FPDM) method developed by us [
The objective of the present study was to analyze
whole genome DNA methylation in the normal
population in order to establish the quantitative relationship
between global DNA methylation levels and age using the
simple and accurate FPDM method, as well as delineate
any gender differences. The DNA methylation-aging curve
obtained for the normal population will provide a useful
reference to facilitate an improved understanding of the
regulation of DNA methylation in aging. Moreover,
autismassociated changes in genome-wide methylation were
investigated, which also served to demonstrate the
importance of using age-matched controls in methylation-disease
The main study population in this study were enrolled
from Beijing, Shanghai, Changsha, and Hong Kong and
consisted of 2116 healthy Chinese subjects including
1108 (52.36 %) males and 1008 (47.64 %) females. The
subjects spanned a wide age range from 2 to 97 years.
The 280 autistic children (age 2-13 years) and their
parents (n = 552; age 24–62 years) were recruited at Central
South University in Changsha. Samples from the parents
but not those from the children were included in the
main study set for age-methylation analysis.
Genomic DNA extraction
Leukocytes were isolated from 5-ml peripheral blood
samples. DNA was prepared by phenol extraction and
chloroform extraction followed by isopropanol
precipitation, washed with ethanol, and air-dried. Tris-EDTA
buffer pH 8.0 was used to dissolve the final genomic
Whole genome DNA methylation analysis by FPDM
To determine genome-wide or “global” DNA methylation
by fluorescence polarization DNA methylation
measurement, ~100 ng DNA sample was first subjected to separate
restriction-enzyme digestions by HpaII and MspI as
]; the methylation-sensitive HpaII cut only
un-methylated 5′CCGG-3′ sites, whereas the
methylationinsensitive MspI cut both methylated and un-methylated
5′-CCGG-3′ sites. After completion of the restriction
reactions, both digests were subjected to a one-label-extension
reaction through incubation with fluorescent
(5-propargylamino-dCTP-5/6-carboxytetra-methylrhodamine, Jena Bioscience) and Taq DNA polymerase.
Measurements of fluorescence polarization on the two
digests following the extension reaction yielded the
percentile global methylation in the DNA sample. The global
DNA methylation level in each instance was thus expressed
in terms of the global percentage of CpG sites in genomic
DNA that were methylated based on the average of
Statistical data analysis
Statistical analysis of data was performed using SPSS
19.0. Percentile methylation of each DNA sample
represented the average of three independent measurements.
To assess the relationship between global DNA
methylation and age, methylation levels of samples in every
5year age range were first analyzed for correlation with age
using Pearson’s correlation test. Based on these results,
the samples were further grouped into five age ranges to
represent different phases of methylation change with age
and again analyzed using Pearson’s correlation to yield an
overall correlation coefficient for each age range.
Differences in methylation levels between males and females
were analyzed for all samples using independent sample t
test as well as for each group of samples in the five age
ranges using multivariable linear regression. Independent
sample t test was used to analyze the methylation
difference between autistic children and parents and
between autistic children and age-matched healthy children.
A p value <0.05 was regarded as statistically significant.
The methylation data is given in Additional file 1: Table S1.
Although the global DNA methylation data determined
using the FPDM method displayed large standard
deviations, when the subjects were divided into 5-year age
groups and analyzed for within-group correlations with age,
positive correlations were detected in the 16–20 and 51–55
groups, and a negative correlation was detected in the 56–
60 group (Table 1). Based on the within-group correlations
and the methylation-age plot (Fig. 1), multiple phases of
change in methylation levels with age were discerned
including a steady increase from year 2 to year 25 and
another rise from year 40 onward to reach a peak level at
year 55, followed by a sharp decrease up to year 75 and
leveling off thereafter. Quantitatively, the increase from
age 2 up to the age of 25 was significant to p < 0.001 with
Pearson’s correlation coefficient r = 0.382. From 26 to
40 years of age, there was no significant change in
methylation (r = 0.028; p = 0.459). However, the DNA
methylation levels again significantly increased with age
between 41 to 55 years (r = 0.265; p < 0.001). From 56 to
75 years of age, there was a significant decrease in global
DNA methylation (r = −0.395; p < 0.001) showing an
inverse relationship between methylation and age, and no
significant change in methylation levels was observed
between 75 and 97 years of age (r = −0.061; p = 0.486). The
correlation data for these different age groups are given in
Additional file 2: Table S2.
With respect to the genders (Table 2), a statistically
significant gender effect was observed in the 41 to 55
age group (beta = 0.136; p = 0.010), where the average
methylation levels in males (73.49 ± 15.26) was higher
than that in females (77.52 ± 13.44). There was no
significant gender effect in any of the other age groups or
when all age groups were combined.
The global methylation levels for autistic children
(n = 280; mean age = 4.7) were compared to both those of
their healthy parents (n = 552; mean age = 33.8) and those
of age-matched healthy children (n = 236; mean age = 5.3).
No significant difference (p = 0.872) was observed between
patients (65.18 ± 16.69) and parents (66.01 ± 19.98)
(Additional file 3: Table S3); but the difference between
patients and age-matched controls (54.35 ± 21.37) was highly
significant (p < 0.001), with increased methylation in the
patients (Fig. 2). There was no significant difference in
methylation between autistic children and either their fathers or
mothers separately (Additional file 3: Table S3).
The purpose of the present study has been to analyze,
using the FPDM method, the global DNA methylation
p < 0.05 is shown in italic font
ap values using gender as a variable in multivariable linear regression analyses
bp value for between-gender comparison using Student’s t test
levels in leukocytes as a function of age in order to establish
a continuous methylation-age curve for the population that
could serve as a basis for describing phenotypic changes
associated with aging and as an age-dependent standard for
the detection of any significant deviation caused by disease.
In providing a systematic characterization of the
dependence of global DNA methylation on age, the study revealed
that global DNA methylation as a genomic parameter of
age was distinctly multiphasic in character. The global
DNA methylation-age curve displayed evident increases
over the adolescent age group of 10–20 and late
middleage age group of 40–50 and a sharp decrease over the age
group of 55–70 to reach a hypomethylation level from age
71 onwards as a hallmark of old age. Previous studies on
age-associated global DNA methylation have reported
agedependent decrease in methylation based on adult to old
age populations. In addition, there are also reports of
ageassociated increase in methylation [
] and methylation
increases in the early years of life [
]. Moreover, while many
loci such as intergenic CpGs outside of CpG islands
display decreased methylation in later life, other loci
such as promoter-associated CpG islands show increased
methylation with age throughout the lifespan [
Therefore, in general global methylation decreases in old
age, it has not been established that this decrease
occurs continuously throughout life. Indeed, a general
increase in DNA methylation with age before
adulthood, followed by stabilization and an eventual
decrease in old age, has emerged from studies on different
age ranges using different methods [
]. The lifelong profile
obtained in the present study is consistent with this
general description, with the more comprehensive time
curve revealing a second period of methylation increase in
late middle-age prior to the methylation decline in old
Significant difference between male and female
subjects was observed only in the late middle-age age group,
suggesting that gender-related factors may contribute to
this second period of methylation increase. The profile
of global methylation variation between genders is
somewhat unclear with previous reports of significant
difference based on methylation levels in LINE-1 repeat
elements for a 45–75 age group [
] but no difference
for a 19–80 age group [
] and no significant difference
based on the LUMA method for CCGG sites for a group
with mean age of 24 [
]. The different results suggest
that gender differences in “global” methylation levels are
dependent on age as well as the subset of methylation
sites examined in the quantification method.
DNA hypomethylation at old age has been reported in
studies focused on a relatively limited number of gene loci
and narrow age ranges, suggesting the possible
association between DNA methylation and age-related
9, 20, 21
]. Indeed, the loss of global DNA
methylation is one of the first epigenetic abnormalities in
], and advanced age represents a potent risk
factor for human epithelial cancers with cancer incidence
increasing sharply from age 60 onward, especially in
], which is in accord with the sharp
decrease in DNA methylation over this age range shown
in Fig. 1. Likewise, evidence of age-associated loss of
DNA methylation in brain tissue suggests the
significant role of DNA hypomethylation at old age in the
pathogenesis of Alzheimer’s disease [
]. The onset of
Alzheimer’s disease at age 65 is also in accord with
the sharp decrease in global DNA methylation
between the ages 55–70 shown in Fig. 1.
Unlike neurodegenerative and aging-related disorders,
neurodevelopmental disorders such as autism and Down
syndrome affect subject groups on the opposite end of the
age spectrum. Autism is highly heritable and affects
information processing in the brain, leading to symptoms that
include impairments in social interaction and
communication, restricted interest, and repetitive behavior [
characteristic symptoms become apparent in early
childhood, typically before the age of three. Although the
etiology of autism is mainly ascribed to genetic variations
including single nucleotide polymorphisms and copy
number variations [
], epigenetic mechanisms have been
invoked to affect the environmental influences [
such, DNA methylation has been associated with
dysregulation of biological pathways in autistic brains, with both
hypomethylated and hypermethylated genomic regions
being identified . Recently, in peripheral blood analysis,
the OXTR promoter was shown to be hypomethylated in
autism cases [
]. Here, we have demonstrated that the
global methylation in autistic children was increased
compared to healthy children (Fig. 2) with respect to the
overall effect across all CCGG sites recognized by the HpaII/
MspI enzymes, encompassing both hypermethylated and
hypomethylated sites as well as unchanged sites. The
overall increase suggests that hypermethylated regions were
more extensive than hypomethylated regions in the
autistic genome. Moreover, in comparison with the time profile
for methylation, the higher methylation level is that
expected of young to middle-aged adults and this could be
interpreted to suggest an abnormally advanced methylome
in autistic children. This is reflected in that no significant
difference in methylation was found between autistic
children and their parents. From another point of view, since
a general increase in methylation takes place from young
to middle age (Fig. 1), the comparison between children
and parents is confounded by the age factor, and the result
demonstrates the importance of using age-matched
controls in analyzing methylation differences. Notably,
environmental and nutritional factors may also affect
methylation, and application of the FPDM method will
facilitate the in-depth analysis of the quantitative effects of
such external factors.
In conclusion, global DNA methylation measurements
in the present study on leukocyte DNA have shown a
multiphasic variation with age that leads to depression of
methylation at old age to half its level at middle age,
thereby providing strong evidence for DNA
hypomethylation at old age as a potent risk marker for various
agerelated disorders such as cancers, cardiovascular and
neurodegenerative disorders, and type 2 diabetes. In contrast,
hypermethylation was observed for autism, a
neurodevelopmental condition. These measurements,
readily performed with the FPDM method, provide a
simple and quantitative approach to investigate the
multiple genetic and environment factors that
determine global DNA methylation. A delineation of
global methylation changes will complement studies on
gene-specific methylation changes to yield an
increasingly comprehensive understanding of the regulation of
DNA methylation and the roles of DNA methylation in
Additional file 1: Table S1. Sample information and methylation data.
(XLSX 118 kb)
Additional file 2: Table S2. Correlation of global DNA methylation with
age. (DOCX 13 kb)
Additional file 3: Table S3. Global DNA methylation levels of autistic
children and their parents. (DOCX 12 kb)
FPDM: Fluorescence polarization DNA methylation
We are grateful to the Hong Kong Red Cross and to Professor Yunlong Tan
for their contributions to blood sample collection.
This study was supported by a research grant to HX from the University Grants
Committee, Hong Kong SAR, for Knowledge Transfer (UGCOSE PCF.010.12/13).
This work was also supported by the National Basic Research Program of China
(2012CB517902). The funding bodies had no other roles in the study.
Availability of data and materials
The dataset supporting the conclusions of this article is included within the
article and its additional files.
HX, SYT, and CZ conceived and designed the experiments; TA and FM
performed the experiments; TA and SYT analyzed the data; SX and KX
contributed to the clinical samples; TA, SYT, and HX wrote the manuscript.
All authors read and approved the final manuscript.
The authors declare that they have no competing interests.
Consent for publication
Ethics approval and consent to participate
Written informed consent was obtained from each participant or from
parents/guardians of the participating children. Subject recruitment and
sample collection were approved by the research ethics review boards of
the Hong Kong University of Science and Technology; Southern Medical
University, Guangzhou; Beijing Huilongguan Hospital; Shanghai Jiaotong
University School of Medicine; and Central South University, Changsha.
Submit your next manuscript to BioMed Central
and we will help you at every step:
1. Gibney ER , Nolan CM . Epigenetics and gene expression . Heredity (Edinb) . 2010 ; 105 : 4 - 13 .
2. Smith ZD , Meissner A . DNA methylation: roles in mammalian development . Nat Rev Genet . 2013 ; 14 : 204 - 20 .
3. Hamidi T , Singh AK , Chen T. Genetic alterations of DNA methylation machinery in human diseases . Epigenomics . 2015 ; 7 : 247 - 65 .
4. Ladd-Acosta C , Hansen KD , Briem E , Fallin MD , Kaufmann WE , Feinberg AP . Common DNA methylation alterations in multiple brain regions in autism . Mol Psychiatry . 2014 ; 19 : 862 - 71 .
5. Nardone S , Sams DS , Reuveni E , Getselter D , Oron O , Karpuj M , et al. DNA methylation analysis of the autistic brain reveals multiple dysregulated biological pathways . Transl Psychiatry . 2014 ; 4 : e433 .
6. Petronis A . Epigenetics as a unifying principle in the aetiology of complex traits and diseases . Nature . 2010 ; 465 : 721 - 7 .
7. Bell JT , Tsai PC , Yang TP , Pidsley R , Nisbet J , Glass D , et al. Epigenome-wide scans identify differentially methylated regions for age and age-related phenotypes in a healthy ageing population . PLoS Genet . 2012 ; 8 : e1002629 .
8. Jones MJ , Goodman SJ , Kobor MS . DNA methylation and healthy human aging . Aging Cell . 2015 ; 14 : 924 - 32 .
9. Lopez-Otin C , Blasco MA , Partridge L , Serrano M , Kroemer G. The hallmarks of aging . Cell . 2013 ; 153 : 1194 - 217 .
10. Finkel T , Serrano M , Blasco MA . The common biology of cancer and ageing . Nature . 2007 ; 448 : 767 - 74 .
11. Maegawa S , Gough SM , Watanabe-Okochi N , Lu Y , Zhang N , Castoro RJ , et al. Age-related epigenetic drift in the pathogenesis of MDS and AML . Genome Res . 2014 ; 24 : 580 - 91 .
12. Beck S , Rakyan VK . The methylome: approaches for global DNA methylation profiling . Trends Genet . 2008 ; 24 : 231 - 7 .
13. Karimi M , Johansson S , Stach D , Corcoran M , Grander D , Schalling M , et al. LUMA (LUminometric Methylation Assay ) -a high throughput method to the analysis of genomic DNA methylation . Exp Cell Res . 2006 ; 312 : 1989 - 95 .
14. Gay MS , Li Y , Xiong F , Lin T , Zhang L . Dexamethasone treatment of newborn rats decreases cardiomyocyte endowment in the developing heart through epigenetic modifications . PLoS One . 2015 ; 10 : e0125033 .
15. Zhao C , Xue H . A simple method for high-throughput quantification of genome-wide DNA methylation by fluorescence polarization . Epigenetics . 2012 ; 7 : 335 - 9 .
16. Numata S , Ye T , Hyde TM , Guitart-Navarro X , Tao R , Wininger M , et al. DNA methylation signatures in development and aging of the human prefrontal cortex . Am J Hum Genet . 2012 ; 90 : 260 - 72 .
17. Zhang FF , Cardarelli R , Carroll J , Fulda KG , Kaur M , Gonzalez K , et al. Significant differences in global genomic DNA methylation by gender and race/ethnicity in peripheral blood . Epigenetics . 2011 ; 6 : 623 - 9 .
18. Zhu ZZ , Hou L , Bollati V , Tarantini L , Marinelli B , Cantone L , et al. Predictors of global methylation levels in blood DNA of healthy subjects: a combined analysis . Int J Epidemiol . 2012 ; 41 : 126 - 39 .
19. El-Maarri O , Becker T , Junen J , Manzoor SS , Diaz-Lacava A , Schwaab R , et al. Gender specific differences in levels of DNA methylation at selected loci from human total blood: a tendency toward higher methylation levels in males . Hum Genet . 2007 ; 122 : 505 - 14 .
20. Pogribny IP , Beland FA . DNA hypomethylation in the origin and pathogenesis of human diseases . Cell Mol Life Sci . 2009 ; 66 : 2249 - 61 .
21. Pogribny IP , Vanyushin BF . Age-related genomic hypomethylation . In: Tollefsbol TO, editor. Epigenetics of ageing . New York: Springer; 2010 . p. 11 - 27 .
22. DePinho RA . The age of cancer . Nature . 2000 ; 408 : 248 - 54 .
23. Wang SC , Oelze B , Schumacher A . Age-specific epigenetic drift in late-onset Alzheimer's disease . PLoS One . 2008 ; 3 : e2698 .
24. Huguet G , Ey E , Bourgeron T. The genetic landscapes of autism spectrum disorders . Annu Rev Genomics Hum Genet . 2013 ; 14 : 191 - 213 .
25. State MW , Levitt P. The conundrums of understanding genetic risks for autism spectrum disorders . Nat Neurosci . 2011 ; 14 : 1499 - 506 .
26. Wong CC , Meaburn EL , Ronald A , Price TS , Jeffries AR , Schalkwyk LC , et al. Methylomic analysis of monozygotic twins discordant for autism spectrum disorder and related behavioural traits . Mol Psychiatry . 2014 ; 19 : 495 - 503 .
27. Elagoz Yuksel M , Yuceturk B , Karatas OF , Ozen M , Dogangun B. The altered promoter methylation of oxytocin receptor gene in autism . J Neurogenet 2016 : 1 - 5 . [Epub ahead of print].