@k*::s*i".oepigenetics"l;

a

Epigenetics

ISSN: 1 559-2294 (Print) 1 559-2308 (Online) Journal homepage:

4a - r- - a -::fi'rfi,1 n

Aberrant signature methylome by DNMTl hot spotmutation in hereditary sensory and autonomicneuropathy 1E

Zhifu Sun, Yanhong Wu, Tamas Ordog Saurabh Baheti,Jinfu Nie, XiaohuiDuan, Kaori Hojo, Jean-Pierre Kochet PeterJ Dyck & ChristopherJ Klein

Io cite this article: Zhifu Sun, Yanhong Wu, Tarnas Ordog Saurabh Baheti,Jinfu Nie, XiaohuiDuan, Kaori Hojo,Jean-Pierre Kocher, PeterJ Dyck & ChristopherJ Klein QA14) Aberrantsignature methylome by DNMT1 hot spot mutation in hereditary sensory and autonomicneu ropathy 1 E, Epigenetics, 9:8, 1 1 8+1 1 93, DOI: 1 A.41 61 I epi.2967 6

To I in k to th is article: httol I dx.doi.ar zl 1 O.41 61 I epi.2967 6

ft ,,"* supplementary matedal f4 ffi trUtirned online: 07Jul 2014.

G' runrn,,yourarticletothisjournal [l lrltl nrti.t" uiews: 263

S ,,"* related articles fi Q vi.* crossmark data f4

frl a,r,ng "rticles:

5 View citing articles f4

Full Terms & Conditions of access and use can be found athttp://www.tandfonline.com/action/journallnformation?journalCode=kepi20

Download by: [Mayo Clinic Scottsdale] Date: i 5 April 20'16, At 09:50

\o

Nkg

ra

Orno\a

Cd

q)

o

oa

>'c(l

a

()ECd

tr*n

RESEARCH PAPER

Aberrant signature methylomeby DNMTI hot spot mutation in hereditary

sensory and autonomic neuropathy l EZhifu Sunr'2'3't, Yanhong Wua't, Tamas Ordog2's, Saurabh Bahetit, Jinfu Niet, Xiaohui Duan6, Kaori Hojd, Jean-pierre Kocherr,3,

Peter J DyclC, and Christopher J (lgipao,a,x

'Division of Biomedical Statistics and lnformatics; Mayo Clinic; Rochester, MN USA; ,Epigenomics Translational program, Mayo Clinic Center for lndividualized Medicine:Rochester, MN USA;3Bioinformatics Program; Mayo Clinic Center for lndividualized Medicine; Rochester, MN USA;3Department of Laboratory Medicine and pathology; MayoClinic; Rochester, MN USA; sDepanment of Physiology and Biomedical Engineering; Mayo Clinic; Rochester, MN USA; 6Department of Neurology; Mayo Clinic; Rochester, MN

U5A; THarima Sanatorium; Division of Neuropsychiatry; Hyogo, Japan; sDepartment of Medical Genetics; Mayo Clinic; Rochester, MN USA

,These authors contributed equally to this work.

Keywords: DNA methylation, neurodegeneration, OMIM 6t4Ll6, OMIM 6o4l2l,whole genome bisulfite sequencing, HSAN1E

Abbreviations: \fGBS, whole genome bisulfite sequencing; HSANIE, Hereditary Sensory and Autonomic Neuropathy lE; DMCs,differentially methylated CpGs; DMRs, diflerentially methylated regions; CGI, CpG island

DNA methyltransferase 1 (DNMTI) is essential for DNA methylation, gene regulation and chromatin stability. Wepreviously discovered DNfff7 mutations cause hereditary sensory and autonomic neuropathy type 1 with dementiaand hearing loss (HSANIE; OMIM 614116). HSANIE is the first adult-onset neurodegenerative disorder caused by a defectin a methyltransferase gene. HSANIE patients appear clinically normal until young adulthood, then begin developingthe characteristic symptoms involving central and peripheral nervous systems. Some HSAN1E patients also developnarcolepsy and it has recently been suggested that HSANIE is allelic to autosomal dominant cerebellar ataxia, deafness,with narcolepsy (ADCA-DN; OMIM 604121), which is also caused by mutations in DNMT|. A hotspot mutation y495Cwithin the targeting sequence domain of DNMT1 has been identified among HSANIE patients. The mutant DNMTIprotein shows premature degradation and reduced DNA methyltransferase activity. Herein, we investigate genome-wide DNA methylation at single-base resolution through whole-genome bisulfite sequencing of germline DNA in 3 pairsof HSAN'IE patients and their gender- and age-matched siblings. Over 1 billion 75-bp single-end reads were generatedfor each sample. ln the 3 affected siblings, overall methylation loss was consistently found in all chromosomes with X and18 being most affected. Paired sample analysis identified 564,218 differentially methylated CpG sites (DMCs; p < 0.05), ofwhich 300134 were intergenic and 264084 genic CpGs. Hypomethylation was predominant in both genic and intergenicregions, including promoters, exons, most CpG islands, L1, L2, Alu, and satellite repeats and simple repeat sequences.ln some CpG islands, hypermethylated CpGs outnumbered hypomethylated CpGs. ln 201 imprinted genes, there weremore DMCs than in non-imprinted genes and most were hypomethylated. Differentially methylated region (DMR)analysis identified 5649 hypomethylated and't872 hypermethylated regions. lmportantly, pathway analysis revealed1693 genes associated with the identified DMRs were highly associated in diverse neurological disorders and NAD+/NADH metabolism pathways is implicated in the pathogenesis. Our results provide novel insights into the epigeneticmechanism of neurodegeneration arising from a hotspot DNMT1 mutation and reveal pathways potentially important ina broad category of neurological and psychological disorders.

lntroduction

DNA methylation afiFects many cellular funcions includingtranscription regulation, cell differendation, gene imprinting andgenome stability.l'2 Mammalian DNA merhyladon is catalyzed byDNMTI, DNMT3A, and DNMT3B. DNA methyltransferasel(DNMTI), locates at chromosome 19p13.3-p13.2, is the

maintenance methylrransferase and essential for maintainingproper epigenetic inheritance. DNMT3A and DNMT3Bestablish new methylation patterns and are considered de novomethyltransferases. The line between maintenance and de novomethyltransferases has become increasingly blurred as studieshave linked DNMTI to de novo methylation3 and DNMT3A/Bto maintenance methylation.a Aberrant DNA methylation

*Correspondence to: Christopher J Klein; Email: klein.christopherpmayo.eduSubmitted: 05/i412014; Revised: 06/1612014; Accepred:.O6/20/2O14;Published Online: 07/O7nOAhttp://dx.d oi.o rgl 1 0.41 61 / e pi.2967 6

Epigenetics Volume 9 lssue 81 184

\oot\tL&rn

o\

={)dq

oaI

U

d<{zs{)!(!

oo

and other epigenetic defects by either Mendelian or polygenicpathogenesis are increasingly associated with diverse forms ofneurological diseases.l'5'6 DNMT3B mutation leading to thepartial inactivation of its catalytic domain has been linked toa human developmental disorder, ICFI (immunodeficiency,

centromere instabiliry and facial anomalies) syndrome,T

DNMT3B mutations cause DNA hypomethylation in satellites

2 and 3 and ct repetitive sequences, a major part of constitutiveheterochromatin. Recently, we identifi ed heterozygous mutationsin DNMT1 targeting sequence (TS) domain causing hereditarysensory and autonomic neuropathy type 1 with dementia

and hearing loss (HSANIE; OMIM 614116).8 Subsequently,

additional DNMT1 TS domain mutations were found underlyingautosomal dominant cerebellar ataxia, deafness and narcolepsy(ADCA-DN, OMIM 604121).e It has been suggested that these

two aduleonset neurodegenerative disorders can be considered

in the differential of each other as HSAN1E and ADCA-DNpatients may share a similar set of phenotypes with varied severity

and onset age.10 More HSANIE families have been recognized

and a hot spot mutation at amino acid 495 was identified as

Y495C mutation is the most common causal mutation foundamong HSAN1E patients.8'rl The DNMTI Y495C mutationleads to DNMT1 protein misfolding and decreased enzymaticactiviry DNMTI is indispensable for development, cellularfunctions and cell cycle regulation. Genomic deletion ofDNMTL in mice leads to embryonic death,rz and complete

deletion of DNMTI in human cancer cells results in significantloss of global methylation, pronounced chromosomal defects

and apoptosis.l3 Dynamic changes in DNA methylation are also

linked to neuronal synaptic stimulation in the brain.ta DNMTLmutations in HSAN1E do not lead to malignancy; instead, theyresult in the neurodegeneration ofperipheral and central nervous

systems.s'15

The aim of the present study was to investigate the genome-

wide DNA methylation changes in HSAN1E patients arisingfrom the hotspot DNMT1 mutation Y495C. There are -30million cytosines preceding guanine nucleotides (CpGs)16

in the human genome and the ma;'ority of them (-80o/o) are

methylated,2 especially at repetitive elements and centromericsatellite repeats, which comprise approximately half of human

genome. A small percentage of CpG dinucleotides are clustered

within gene promoters as CpG islands (CGI), but normally only3o/o of CpG islands are methylated.tT The pathogenic role ofpromoter methylation has been extensively studied especially incancer, but there is limited understanding on how methylationchanges in intergenic regions relate to human disease.

'We analyzed CpG methylation at single-base resolutionin 3 affected individuals and their age-, gender-matched (<5

y difference) unaffected siblings by whole genome bisulfitesequencing (\[GBS). This paired-sample study design allowedus to perform statistical analysis of differential methylation whilecontrolling for entraneous factors, minimizing inter-familialand environmental influences. Furthermore, we investigated thepathways downstream of consequent aberrant DNA methylationthat are involved in HSANIE pathogenesis in relation to otherneurological and psychological disorders.

Figure 1. Study design and analysis workflow. Three pairs of affected(with DNMT1 mutation and HSANI) and unaffected siblings (\.rithoutDNMTI mutation and HSANI) were selected from three different fami-lies of the same kinship.The sibling pairs were matched for age (<5 y)

and gender. WGBS was performed on peripheral lymphocyte cells andDMCs and DMR5 were identified by paired analyses. Genes with DMRs

5 kb from theTSS were used for pathway enrichment analysis. A1-A3,

affected individuals with DNMTI mutation; U1-U3, unaffected siblingswithout DNMTI mutation.

Results

Genomic DNA was entracted from peripheral blood B cells

of three affected patients carrying the heteroaTgous Y495CDNMT1 mutation and their unaffected siblings and analyzed

by .W'GBS and bioinformatics approaches as illustrated in

Figure 1. Our sequencing results generated approximately one

billion single-end reads of 75$p length for each individualsample. Alignment efificiency was high wrth 94-95a/o of these

rea& mapped to the human reference genome (hg19; Table 1).

Our sequencing results covered -24 of the -30 million CpGspresent in the human genome. The bisulfite conversion rate

was assessed using all non-CpG sites with > 10X coverage inthe genome, and found to be close to I (>0.999), indicatingalmost complete conversion. On average, 18-19 million of CpG

rytosines were obtained with >10X coverage for each individualgenome and. L2035253 CpG sites had >l0X coverage across all 6samples (Fig. 2,{. and B), of which 6051917 w€re in genic (5kb

upstream or inside gene body) regions and 5983334 were inimergenic (outside of genic regions) regions. The genic CpGs hadhigher overall methyladon (figi 2C) than the intergenic CpGs(FB. 2D). Notably, the affected patients consistently had lowermethylation than their paired siblings, particulady in intergenicregions. The methylation of CpG islands (CGI) was low (-30olo)

across all samples, whereas methylation in repeat regions (Alu,Ll, L2, Satellite repeat, and simple repeat) was high {>70%)(Fig. 2E). All the repeat regions had loq'er methylation in theaffected patients than their unaffected siblings except for CGIs(Fis.2F).

f Affr"t"d-l

l'u*fi""t"d I

Pairl

m[I]

Pair?E[E

Pair3

[r][q

www.landesbioscience.com Epigenetics

WGBS -Bhol€ g€rcme bisulfrte*quercing

DMC - ditforstially o1€thylated CpGDMR - difle€itisiry methyhled rogior

'I t85

Aes0o

=40EXso(Ja20E

Ero(E()0

B

0.75Ju1A1 U2tr U3A3

1X

5X

10x

q(q

=EoE

0.70

41 U1

\oOc'l

^.

od

(Ja()

Uo

ruiiliml;iililrlli-: rrllrl i*i :lI nrurnzuzagug i nrutA2uzngus

i

ie 1 - iFoo, itorAlii

sI#ffi:= s::'tIItI=t,I E---EF"'I

""'*"oS*'B*e i

Figure 2. Genome coverage and overall methylation patterns. (A) Captured Cs in CpG context at different depths of coverage. Almost all CpGs in the

genome were covered; however, CpGs with at least 10X coverage accounted for about 1/3 of -30million CpGs in the genome. (B) Overall mean methyla-

tion was reduced by 0.3-3olo in affected individuals (A1-3) relative to unaffected controls (Ul-3r. (C) Methylation distribution of genic CpGs by sample.

The affected had slightly lower methylation than their unaffected siblings. The horizontal bar within each box is the median methylation for the sample.

(D) Methylation distribution of intergenic CpGs by sample. The affected patients had lower methylation than unaffected individuals and the differences

were greater than in genic regions. The horizontal bar within each box is the median methylation for the sample. (E) CGI and repeat region methyla-

tion. CGI methylation was low (-30%) across all samples with little difference between samples while methylation in repeat regions (Alu, 11,12, satel-

lite, and simple repeats) was higher (>70oi6). The affected patients consistently had lower methylation (l-5% reduction) than their unaffected siblings'(F) Methylation difference between sibling pairs in CGls and repeat elements. All repeat elements but not CGls had lower methylation in the affected\r / !vrstrt/rsUvr,

individuals with Pair 1 (P1) displaying the smallest differences.

Genome-wide CpG methylation patterns reflect DNMTImutation status and kinship

Previous studies have shown that meth,vlation is associated

with gender, age, and individual genetic background, as wellas environmental influences. Therefore, we designed our study

to compare age- and gender-matched sibling pairs to minimizethe impact of potentially confounding variables. To vaiidate this

approach, we performed unsupervised clustering using the top126 000 CpGs in which methylation was most varied across the 6

samples (CpGs with the highest standard deviations). W'e found

that the affected and the unaffected samples formed rwo distinct

clusrers (Fig. 3A), suggesting a strong effect of DNMTI mutationon DNA methylation. Interestingly, when we used the top 10 000mostvaried CpGs across the 6 samples for unsupervised clustering,the samples clustered by relationship (FiS. 3B), demonstrating an

effect ofthe overall genetic background. Our results from cluster

analysis confirm that it is important to compare the methylationchange between gender and age-matched individuals within a

family, ideally under similar environmental influence. Also, itindicates that our paired-samples design will likely increase the

accuracy in interrogating the impact of DNMTL mutation on

methylation changes of different genomic regions.

Epiqenetics Volume 9 lssue B

AE,l o

ot?o

=oO'o c!TO

o-:

tr)IJo

c,

f:o

Eo#o

ao

\o

e.l

o

Or

oi

o(n

ao()ao

Uo

a

()

o

l-.1

Figure 3. (A) Unsupervised clustering using the top 126000 varied Cpcs shows the affectedand unaffected samples form different clusters, suggesting the similar methylation pat-

terns by mutation status. (B) Unsupervised clustering using top 10000 highly varied CpGs

shows samples are clustered by family, indicating the strong genetic influence on DNA

methylation.

Differentially methylated CpGs (DMCs) are mosdyhlpomethylated

.We performed paired t-statistics to identifr differentiallymethylated CpGs or DMCs between 3 affected and 3 unaffectedsiblings after removing 190 349 CpGs without any yariance across

samples (constant methylation, non-informative). The mean

methylation difference was used to measure the scale of a change

along with the differential ? value. At P < 0.05, 564218 DMCswere identified, of which 308962 had mean ratio differencegreater than 0.1 (>100/o methylation difference). Noted is thateven with paired analysis for increased power, the small sample

size would only detect the CpGs with a large difference and lowvariability. Therefore, we compared the 2 value disuibutionfrom the permutation test with randomly assigned siblingpairs and found that the Pvalues from the true pairs were non-uniform with lower Pvalues shi&ed to the left side of significantdifference (Fig. SlA), whereas the Pvalues from the permutationdeviated toward the non-significant side (Fig. SfB). These results

indicate that the significant CpGs were not likely random. Themajority of the DMCs were hypomethylated in all chromosomes(Fig. 4A). Interestingly, chromosome X (3145 hyper- and 10104

hypo-methylated CpGs) and chromosome 18 (4342 hyper- and1L346 hypo-methylated CpGs) had the largest methylationreduction (Fig. 48 and C), suggesting the impact of DNMTImutation is chromosome- and region-specific.

Ofthe 308 962 DMCs with mean methyladonchange greater than 10o/o, therewere significantlymore in intergenic than in genic regions (182766vs. 126L96; chi Square test P <2.2e-16).In genicregions, there were 2-fold more hypomethylatedCpGs than hypermethylated CpGs; in intergenicregions, there were 3-fold more hypomethylatedCpGs than hypermethylated CpGs (Fig. 5A).These results demonstrate higher prevalence

of hypomethylation in the intergenic regions.

Although the overall, genome-wide reductionin DNA methylation was moderate berweenaffected and unaffected siblings, a grearerreduction (5-l0o/o) was observed aroundtranscription start sites (TSS) (Fig. 58).

'W'hen we assessed the DMC disuibution inthe genome according to the genomic features

such us exons, CGIs, Alu elements, Ll, L2,satellite, and simple repeats, we found all but the

CGIs to be significantly hypomethylated in bothgenic (Fig. 5C) and intergenic regions (FiS. 5D).

Differentially methylated regioas (DMRS)

are mostly intergenic and hypomethylatedDNA methylation often occurs in clusters and

it is important to evaluate how the differentiallymethylated regions are scattered across thegenome. For DMR analysis, we included CpGscaptured in all 6 samples and used a smoothingapproachrs to stabilize the CpG sites with lowerthan 10X coverage.

-We designated a DMR

as the minimum of 5 CpGs with the mean

difference >100/o berween the 3 pairs of affected

and unaffected siblings. Across the genome, 7521 significandychanged DMRs were found, ofwhich 5649 werehypomethylatedand 1872 were hypermethylated in the afFected siblings. Themajority of these DMRS (4890, 650/o) were intergenic and 2631(350lo) were within genic regions. Additionally, over 82o/o ofthe 4890 intergenic DMRs and 620/o of the 2631 genic DMRswere hypomethylated, indicating most DMRs showed loss

of methylation in the affected siblings (Fig. 6A). The DMRanalysis further confirmed more prevalent hypomethylation inthe intergenic regions in the affected siblings.

DMR-associated genes are enriched in neurological disordersTo investigate whether the methylation changes resulted from

DNMTI mutation are directly associated with genes enriched ina particular pathway, functional class or disease, we conducted apathway enrichment analysis for the genes that have identifiedDMRs within 5kb of their TSS using Ingenuiry PathwayAnalysis(IPA; see Methods for details). IPA constructs the networks thatoptimize for both interconnectivity and number of focus genes

under the constraint of maximal network size.le The top threesignificantly enriched gene interaction networks were "hereditary

disorder, metabolic disease, cancer," "hereditary disorder,neurological disease, cellular assembly, and organization," and"cancer, cellular development, tumor morphology." Next, we

applied canonical pathy.ay analysis using the IPA library of

www.landesbioscience.com Epigenetics 1187

EGi)Eq(Jo

=o

1.0A3A1

c@o

o,E ,^gd:-c tf,oo

c\lCi

oo

\o

alL&

t.)

rn(}'

(g

q)

roO

t)a(.)

oxda>'€o(t

Bo

Chr X methylation

Ul A2 U2

Sample

canonical pathways (part of the Ingenuity Pathways KnowledgeBase; IPKB). Eleven canonical pathways were significantlyenriched (P < 0.05; Table S1). Of particular interest, several

of the significant pathways are related to NAD+/NADHmetabolism which has been implicated in neurodegeneration.2o'2r

Finally we applied disease association functional analysis, whichidentifies the diseases that are most significant to the gene set.

\7e found that the top three significantly enriched diseases were

"neurological diseases" witfi enrichment p-value between l.5lE-04 to 5E-02 (Table S2), 'psychological disorders" (P = 1.88 to04*4.988-02; Table S3), and "skeletal and muscular disorders"(?= 1.8E-04 to 4.778-02; Table 54) Some ofthe top neurological

12

Chr

ChrX methylation density

0.0 0.4 0.6

Methylratio

diseases were "progressive neuropathy," "Parkinsont disease,"

"attention deficit disorders," "attention deficit hyperactivirydisorder," "narcolepsy," and "dementia." All these findings are

consistent with the qypical HSAN1E phenotypes which includeprogressive neuropathy, dementia, sleep disorder, and personalitydisorders, indicating that the gene set we identified is associated

with pathogenesis of central and peripheral neurodegeneration,

and may have important implication in other neurodegenerativedisorders.

CpG methylation in imprinted gcnesGenomic imprinting is an epigenetic regulatory mechanism

that turns offexpression from either the paternal or the maternal

3" g

o frHiiiilii: ;,ri;iii

:ii:tl--L--l-----:-.-, i

t* tL f,J Jt* l! -u JL_it--tl.-iL-tt-.tLlI tr r|-rr t+t I: i

ltl:li:'li:l,i

Ig!UI:;H8EC5c6g"3

I

E

oBE

i

i

=;

I

i

s

o

EH

Ii

I

i

TEEI

o

8

Ii

:

:

i

IIo

o

I

_.:_

-l

Io

20

Figure 4. Methylation by chromosome. (A) DMCs are plotted for each chromosome separately. All show the dominant hypo-DMCs with chromosomeX changed the most. Y-axis: The distribution of methylation mean difference between affected and unaffected individuals. (B) Chromosome X CpG

methylation distribution. Affected patients (red) have more obvious reduced methylation than the unaffected (green) for all pairs. (C) Density plot ofCpG methylation for chromosome X for each sample. Affected patients have left-shifted (decreased) methylation ratio curves. Solid line: patients withmutation; Dashed lines: sibling controls without mutation. Red lines: A1, A2; Green lines: A2, U2; Blue lines: A3, U3.

Epigenetics Volume 9lssue 8

A

B C

T

ITr=L]l. ,. : : ItJ

1 188

ilB-oq

.906aqC'=(o

€oOFEo

-4000 -2000 0 2000 4000

Oistance kom TSS

r hyper* hypo

C 1oo%

80%

Genic Region

__l

ltrLpc6OEE66

lillll€€gi5sEXEoE.

c,(J

=I 60%r9

oo

40%co()oo'

zo,h

sill

l

L__I

iD lntergenic Region

AFiGEIoEtE5lItl

Figure 5. DMC distribution in different genomic regions. (A) DMCs occur across genomic regions; however, intergenic regions have higher percentage

of hypomethylated CpGs than genic regions. (B) Hypomethylation around transcription start site (TSS). All genes are aligned around TSS and average

methylation is displayed separately for affected (red) and unaffected (green) subjects. (C) DMCs in different genomic features of genic regions. All

regions except CpG islands have more hypo-methylated DMCs in the affected individuals. (D) DMCs in different genomic features of intergenic regions.

All have more hypo-methylated DMCs in the affected individuals. X-axis, genomic region; y-axis, percentage of hypermethylated (purple bars) and

hypomethylated DMCs (green bars).

r hyPerr hypol00,ta

86:

glrlllBi=5

t__

www.landesbioscience.com

\oaa:l

d

ooa.)

U

-o

c=C

\

:t

aI

1 189

lAI

I

4.890 intergenic * 2631 genic&.p""

t-B H2az

the regions of 11 histone modifications and variants (H2L.Z,H3K4mel, H3K4me2, H3K4me3, H3K9ac, H3K9me3,H3K27 ac, I{3K27me3, H3K36me3, F{3K79 me3, H4K20mel)reported in a lymphoblastoid cell line of a healthy donor(GM12878) by the ENCODE project (https:l/genome.ucsc.edu/ENCODE/). In addition, we also compared the CTCF(CCCTC-Binding factor) binding sites defined in rwo replicatesof the same cell line (CTCF.repl and CTCF.rep2). We foundthat CpG methylation was reduced in the affected individualsin the regions normally occupied by the repressive histone marksH3K27me3 (4.1o/o median reduccion) and H3K9me3 (1.2o/o

median reduction), as well as CTCF (2.3o/o median reduction)(FiS. 68). Increasing evidence indicates that DNA methylationand histone modification pathways work synergistically toregulate chromatin structure and gene expression. DNAmethylation may serve as a template for histone modifications,and decreased methylation in the repressive histone marks may'

lead to interruption of proper heterochromatin state.

CpG methylation in genes associated with mitochondrialfunction

Mitochondrial genes perform the important function inenergy generation for cells. The fact that HSAN1E patientsmanifest some clinical presentations similar to mitochondrialdiseases or the mutation leads to the loss of high energydemanding neurons prompted us to investigate the methylationprofile of mitochondrial genes.'We identified 666 mitochondrialfunction-related genes encoded either in the nuclear or themitochondrial genome (Table S5), and 654 were captured in thedata set with 10x coverage. Comparing the mean methylationof these genes berween the affected and unaffected samples,

we did not see noticeable difference (0.83 vs. 0.84). About4.1o/o of CpGs in genes related to mitochondrial function were

differentially methylated vs. 4.5o/o methylated CpGs found ingenes unrelated to mitochondrial function. Of the differentiallymethylated genes annotated to mitochondrial funcrions, 54.8o/o

was hypomethylated.

Discussion

Accumulating evidence suggests DNMTI plays multiplecritical functions in the nervous system,2a but the exactmolecular mechanism(s) remain unclear. HSAN1E is the firstidentified adult-onset, monogenic neurodegenerative disease

caused by DNMTI mutations. The disease-causal DNMTImutation provides a unique window through which we caninvestigate specific methylation defects underlying an adult-onset neurodegenerative disorder at the molecular level. Therole of DNMT1 is also of fundamental biological interest, as itis the only essential methyltransferase to ensure the fideliry ofepigenetic inheritance. Through whole genome methylationevaluation with single base resolution, our study reveals

a signature pattern of methylation alteration in this adultonset neurodegenerative disease, characterized by a prevalenthypomethylation in intergenic regions, a significant methylationdecrease at surrounding regions oftranscription start sites (TSS),

\o

(\l

a4

rno\

o(d

q

oa()

>'z>'s{.)

oH

H3k27me3

H4k20me'1 H3k36me3

allele to achieve mono-allelic gene er<pression through DNAmethylation and histone modulation. As a DNA methylationmaintenance enzyme, DNMTI has been shown to be the enzyme

responsible for maintaining the imprinting process duringand after embryogenesis.22'23 Any defect of this enzyme likelyleads to reduced methylation in imprinted genes. 'We a*alyzed,the list of 201 well-characterized imprinted genes obtainedfrom the geneimprint web resource (http://wwwgeneimprint.orgi). Collectively, the CpGs in imprinted genes were more

differentially methylated than in non-imprinted genes (5.1olo vs.

4.5o/oCpGs, respectively, Chi-square testP= 9.01 x 10-12). Therewere more hypomethylated CpGs in the imprinted genes than inother genes (72o/o vs. 660/o hypomethylated CpGs, Chi-squaretest P = 9.01 x 10-"), suggesting DNMTI mutation has a higherimpact on the imprinted genes.

CpG methylation in histone modification mark sitesTo examine whether differential DNA methyladon

preferentially occurred in regions normally occupied by specifichistone marks and variants, we compared the overall CpGmethylation between the affected and unaffected siblings in

Figure 6. DMR distribution in the genome and DNA methylation inhistone marks. (A) Most DMRs are hypomethylated for both genic andintergenic regions; however, this is much more dramatic in intergenicregions (82% vs. 62%). (B) Median methylation difference in differenthistone mark region between the affected and unaffected individuals.Hypomethylation is observed in H3K27me3, H3K9me3, and CTCF sites.

Epigenetics Volume 9lssue 8

H3k4me1

H3kgac*Affected

- - Unaffecled

1 i90

\oONk

\o

r.)oioG'

it)GI

ao(.)v)utr

U

a>.,g()(!o

o

and marked methylation reduction on chromosomes X and

18 (5-100/o). Additionally, DNMT1 mutation leads to morehypermethylated DMC in some CpG islands.

Our results provide an interesting contrast yet uniqueresemblance to a recent study that also used whole genome

bisulfite sequencing to assess the methylation profile change ina ICF syndrome patient wkh DNMT3B mutation.25 This studyshowed DNMTiB mutations cause severe hypomethylationacross the whole genome (-25o/o overall reduction) and profoundlosses in chromosome X (630/0). ICF syndrome patients have

autosomal recessive inheritance and an average life expectancy

of only 8 y. In contrast, our whole genome bisulfite sequencing

results showed that the impact of DNMTI heterorygous

mutation on global genome methylation is only modestly

changed in 3 affected HSANIE patients relative to their age-

and gender-matched siblings. However, the strong impact ofDNMTI heterozygous mutation is rather exerted in intergenicregions, TSS regions and chromosomes l8 and X. These regions

have between 5-l0o/o reducdons in methylation, suggesting a

targeted and insidious effect of DNMTI methylation defect.

HSANIE pati€nts have autosomal dominant inheritance; they

are normal from bimh to early teenager years, but then start todevelop progressive sensory neuropathy, sensorineural hearingloss and memory loss in young adulthood. The life expectancy

of HSAN1E is approximately 50 y and the cause of death is

usually related to dementia complications. The ICF syndromet

early-onset and severe phenotypes are consistent with the drastic

methylation decrease while HSANIE's adult-onset feature,

unique phenotypes and progressive course are linked with region-

specific methylation reduction. Our findings suggest DNMTIheterozygous mutation has minimal effect during development,

but has deleterious accumulating impact on the nervous system

over a long period of dme and specifically on memory formationand sensory neurons.

Because patients with DNMTI mutations have a syndromicappearance similar to many mitochondrial disorders, we

investigated the DMC pattern of mitochondrial function-related

genes encoded either in the nuclear or the mitochondrial genome,

but did not find significant difference. However, our IPA analysis

demonstrated that the 1693 genes associated whh DMRs are

highly enriched in many neurological and psychological disorders.

Alzheimer disease, narcolepsy, and dementia were among the top

significantly associated diseases. The canonical pathway analysis

revealed NAD+/NADH metabolism pathways are implicatedin the pathogenesis. NAD+ is a coenryme involved in redox

reactions, expressed ubiquitously and has critical functions in the

neural cells. NAD+ has been shown not onlyimportant in energy

metabolism and mitochondrial functions, but also in caicium

homeostasis, aging, and cell death, In animal models, increased

nuclear NAD+ has neuroprotective function against the negative

effects of both mechanical and chemical insult.26 Emerging

evidence shows NAD+ as an important therapeutic target for the

treatment of age-related degenerative diseases, such as Alzheimer

disease.zz28 Our result indicates that DNMTI likely impacts a

specific set of pattrway involving NAD+/NADH metabolism

that are particularly imperative in degenerative neurologicaldisorders.

Increasingly epigenetic mechanisms are also being recognized

important in sporadic neurological diseases without apparent

genetic causes.5'2e The high fidelity DNMTI enzyme activity is

critical in epigenetic regulation. Ow DNMTI Mendelian disease

model provides a unique and valuable avenue for future research

in common varieties of human neurodegenerative disorders.

Udlizing a unique paired-sample design, our study showed

that even modest reductions in DNMT1 enzyme activiry and

preferential site-specific intergenic changes of DNA methylationcan lead to devastating neurological consequences. Llnderstandinghow DNMT1 mutation exerts its deleterious impact in a region-

and chromosome- specific manner and how the methylationchanges specifically lead to sensory nerve axonal loss and brainatrophy with cognitive decline will require a significant amountof future study. Our work provided first insights into a specific

neurodegenerative pathway by identifying a signature genome

methylation pattern potentially having broader implicationsbeyond HSAN lE.

Materials and Methods

Study subjects

Since mettrylation is associated wkh age and gender, we

identified three same-gender, similar aged (<5 y) sibling pain&om the large kindred of European origins (Caucasian) welongitudinally studied in our original publication for whole

genome methylation analysis. Three HSANlpatients withDNMT1 heterozygous Y495C $rutation were matched withtheir normal siblings without the mutation. Their genomic DNAsamples were collected from blood and whole geoome bisulfitesequenciog was performed. The clinical features are very similarbetween the affected persons in this kindred.s'ao'at They allstarted tn lose their sensory nerve action potentials during youngadulthood. By the 3r&4th decade, prominent sensorineural

hearing loss with cognitive and behavioral changes was notedwith progressive dedines, leading to the requirement of total care,

and death often occurred in the 4th-5th decades. Brain imagingand limited autopsy of the affected persons showed globalauophy and a reduced weight ofthe autopsied brains, spinal cord

posterior column loss ofaxons, and pansensory fiber loss in &stalneryes. Histopathologic analysis did not reveal any inclusion

bodies such as B-amyloid, as,TDP-43, and ct-synuclein.8'30 f-hestudy protocols were approved by Mayo Clinic Institute Review

Boards.DNA extracti,onDNA samples were €r$racted from patients' blood. Brieflp the

anticoagulated blood samples were centrifuged $ 2500 g for 10

min, and the separated bu& coat layers were collected. ThenGenome DNA was erfiract from bu& coat using Qiamp DNABlood Kit (Qiagen).

'Whole genome biculftte methylation sequencing10;rg genomic DNAnras used for bisulfite sequencing library

con$rucdon according to the standarrd Illumina Paired-Eod

www-landesbioscience.com Epigenetics 1 191

\o

6ILp

.4

<)rno\

()

r3oo()au

Uo

z>rpC;)

(Go

=o

protocol. Briefly, genomic DNA was fragmented and adaptorswere ligated. DNA samples were then subjected to the bisulfiteconversion using EZ DNA Methyladon-Gold kh (Zymo),

purified using AMPure beads (Agencourt) and amplified byl0 cycles PCR. Sequencing was performed on HiSeq2000(Illumina) with single-end reads on HiSeq2000 using TiuSeqSBS sequencing kit version 3 and HCS y2.0.12 data collectionsoftware.

\7GBS data analysis

The raw sequence fastq data were processed using our bisulfitesequence analytical pipeline, SAAP-BS (vl)3'z with alignmentto human reference genome version tg (hg19) to obtain CpGsite methylation including both methylated and unmethylatedcytosine counts and methylation ratio. All CpG sites wereextracted. To ensure data quality and reliability, we only includedCpG sites with >l0x coverage in the downstream differentialmethylation CpG analyses. Overall methylation was measured

by mean and median genome wide, by chromosome, or specialregions of the genome for each individual or comparison group.DMC detection between affected and unaffected siblings was

performed by paired t statistics. The constant CpGs across

sampies (the methylation ratio is either = 1 or = 0 in all six

samples) were first removed. CpGs with P < 0.05 in paired rtest were considered differentially methylated. The distributionof these DMCs in the genome was summarized against various

genomic regions including intergenic, genic regions, promoters,

repetitive sequences, histone modification, and imprinted genes.

Differential methylation region analysis

Differentially methylated regions (DMRs) berween affected

and unaffected siblings were identified using the R package

"bsseq" as previously described.t8 For this detection analysis,we retained all CpG sites that were sequenced in all six samples

regardless of coverage. A smoothing approach was used tostabilize the sites with a lower coverage since nearby CpGstend to have similar methylation patterns.33 The smoothingprocedures regresses these CpGs to reduce potential outliers atlow-covered CpG sites. T statistics was then performed for each

CpG independently and similarly changed CpGs were groupedinto DMRs according to specified minimum CpG numberand methylation difference berween affected and unafGctedindividuals (minimum of 5 CpGs and at least 10o/o meth,vlationdiffereoce in the DMR region were used).

Pathway/gene enrichment analyses by ingenuity pathwayanalysis

For the genes with DMRS within 5kb of TSS, we conducteda pathway, gene network and funcrional enrichment analysis

using IPA (http://www.ingenuity.com/, performed onSeptember 5, 2013), a widely used pathway and system biologyinterpretation toolle in order to identi!, relevant biological andmolecular networks. IPA uses a proprietary database created

from continuously updated, published, peer-reviewed scientificpublications stored in the Ingenuity Pathways Knowledge Base

(IPKB). It offers several function modules including Core, Tox,Metabolomics, Comparison, Biomarker Filter, and microRNATarget Analysis. The Core Analysis was used for this data,which inciudes the enrichment for canonical pathway, disease/

biological function, toxicological function and gene-gene

interaction nerworks. The enrichment Pvalue is calculated usingthe right-tailed Fisher Exact Test, representing the likelihood ofthe association between a set of focus genes and a given process

or pathway is due to random chance.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

This work was supported in part by the EpigenomicsTranslational Program and the Bioinformatics Program ofthe Mayo Clinic Center for Individualized Medicine and theNational Institutes of Health (K08 NS065007).

Supplemental Materials

Supplemental materials may be found here:www.landesbioscience.com/j ournals /epi ge netics I ar :j.cle I 29 67 6

t.

References

Qureshi IA, Mehler MF. Epigenetic mechanisms

governing the process of neurodegeneration. MolAspects Med 2013: 34:875-82: PM1Dt22782013-,http://dx.doi.org/10.1016/i.mam.2012.06.01 1

Bergman Y, Cedar H. DNA methylation dynmics inhealth md diseue. NatStructMol Biol 2013;20:274-81; PMID:23463312; htry:ildx.doi.org/10.1038/nsmb.2518

Jair KW', Bachman KE, Suzuki H, Ting AH, RheI, Yeo RW, Baylin SB, Schuebel KE. De novo CpGisland methylation in hmu rucr cells, Cucer Res

2OO6 ; 66 :682-92; PMID :\642i997 : htp://&.doi.orgl10.1 158/0008-5472.CAN-05-1980

Liang G, Chan MF, Tomigahual', Tui YC, Gonzales

FA, Li E, Laird PV, Jones PA. Coperativiqv bemenDNA methyltrmsferases in the maintenancemethylation of repetitive elements. Mol Cell Biol2002; 22:48O-91 : PMID:1\756544: http://dx.doi.org/10.I128/MCB.22,.2.480-491.2002

Marques SC, Oliveira CR, Pereira CM, OuteiroTF. Epigenetics in neurodegeneration: a no layer

of complexity. Prog Neuropsychophamacol BiolPrychiary 2011; 35:348-55: PMID:20736O41:http://dx.doi.org/ 10.1016ij.pnpbp.2010.08.008

Klein CJ, Beoarroch EE. Epigenetic regulationrbasic concepts ud relryance ro neurologicdisease. Neurology 2Ol4; 82:1813-40;PMID:24759839: http:lld.x.doi.orgll0.l2l2l\(iNL.0000000000000440Xu GL, Bestor TH, Bourc'his D, Hsieh CL,Tommerup N, Bugge M. Hulten M, Qu X, Russo

JJ, Viegas-Piquignot E. Chromosome instabiliw and

immunodeficiency smdrome caused by mutationsir a DNA methylransferoe gene. Nature 1999;

402:787-91; PMID:10647011; http;//dx,doi.orgl10.1038l46214

Klein CJ. Boruyan MV N0u Y, Ward CJ, NicholsonCA, Hmmos S, Hoio K, Ymmi$i H, KarpfAR, Vallae DC, et al. Mutations in DNMT1our hcreditary sensory neuropathy with dementiaand hearing loss. Nat Genet 2Ol1; 43:595-600:PMID :21,53257 2 ; hrp://dx.doi.org/ 10. 1038/ng.830

Winl<elmun J, Lir L, Schormair B, Kornum B&,Frnco J, Pluzi G, Melbelg A, Cornel-io F, UrbuAE, Pizza F, a al. Mutations in DNMTI caure

autosomal dominant erebellar atuia. deafness andnarcolepsy. Hum Mol Genet 2012; 2l:2205-10;PMID:22328086; hnp: I I dx.doi.orgl 1O.1093 lhmg/dds035

Moghadm KK, Pizza F, La Morgia C, FranceschiniC, Tonon C, Lodi R, Barboni B Seri M, Ferrui S,

Liguori R, et al. Narcolepsv is a common phenoq'pein HSAN IE md ADCA-DN. Brail 2014; r37tr643-55 : PMID : 247 27 570 ; http :/ldx.doi.orgl 10. 1093 /brain/awu069

Klein CJ, Bird T, Erckia-Imer N, Lincoh S, HjonhR, Wu Y, Kwok J, Mer G, Dvck PJ, Nicholson GA.DNMTI mutation hot spot causes varied phenotvpes

of HSANi with dementiaud hearing loss. Neurologv2013; 80:824-8; PMID :23365052: http://dx.doi.org1l0. 1212l1i7NL.0b013e318284076d

Li E, Bestor TH, Jaenisch R. Tugeted muration ofthe DNA methyltrmsferrc gcne results in embryoniclethaliry. Cell 1992; 69:915-26; PMID:1606615;http://dx.doi.org/ I O.rcrc I 009?-867 4 (92) 9061 1 -F

Chen T, Hqi S, Gay F, Trujimoto N, He T, ZhogB, Ueda Y, Li E. Complete inactivation of DNMT1leds to mirotic camtrophe in huma cucer ells.Nat Genet 2007; 39 :391-6 ; PMID :17 322882 ; hdp : I Idx.doi.olg/10.1038/ogl982

Ma DK, Guo JU, Ming GL, Song H. DNA excision

repair proeins and Gadd45 u moleculr players forective DNA demethylation. Cell Cycle 2009; 8:1526-31 ; PMID :19377 292 t http : I I dx.doi.oryl l0.416l /cc.8.10.8500

11.

1192 Epigenetics

t1

Volume 9lssue 8

\o

a:l

g

oi

()(n

()a(.)

()o

z>'

()

o

B

IJ

Klein CJ. DNMTl-Rclated Demotia, Deafnes, mdSensory Nzuropath,v. In: Pagon RA, Bird TD, DolmCR, Stephens K, Adm MB eds. GeneRwiews.Seade (VA), 1993.

Ehrlich M, Gma-Sosa MA, Huang LH, MidgemRM, Kuo KC, McCune RA, Gehrke C. Amount mddistribution of 5-methylcytosine io humm DNAfrom different q,pes of tisrues of cells. Nucleic AcidsRes 1982; 1O2709-21; PMlDt7O79l82; htcp:lldx.doi.org/ 10. 1093 I t* I 10.8.27 09

Bird A, Taggart M, Frommcr M, Miller OJ, MacleodD. A fraction of the mouse getrome that is derived

from islmds of nonmethylated, Cpc-rich DNA.Cell 1985; 40:91-9; PMID:2981636; http://dx.doi.org/ 10. l0 16/0092-8 67 4\85)9 0312-5

Horen M, Gerds TA. Nielsen OH, Seidelin JB,Troelsen JT, Olsen J. pcaGoPromoter--u R package

for biological md regulaton interpretation ofprincipal components in genome-wide gene expression

data. PLoS One 2012; 7:e32394tPMID:22384239;hnp://dx.doi.orglI0.l37l /journal.pone.003239 4

Thomas S, Bonchev D. A suney of orrcnt sofwuefor network malysis in molecular biology. HumCenomics 2010; 4:353-60 ; PMID:20650822; http : / Idx.doi.org/10.1 186 I 1479 ".1 36 4-4-5 -353

Araki T, Scaki Y, Milbrmdt J. Increased nuclearNAD biosvnthesis ud SIRTI activation prryeotaxonal degeneration. Science 2004; 305:1010-3: PMID:15310905; http://dx.doi.orgl10.1126/rcience.l098014

Raff MC, 'W'hitmorc AV, Finn JT. Axoaal self-

destruction md neurodegeneration. Science 2002;296:868a1l' PMID:11988563; hnp:l/dx.doi.org/10.1 126/science.l0686l3

Hircawa R, Chiba H, Kmeda M, Taiima S, Li E,

Jreoisch R, Suki H. Maernal ud rygotic Dtrmtlare necesss and rufficient for the maintenmce ofDNA mahylation imprinrs during preimplantationdwelopment. Genes Ds 20O8: 22:1607-161PMID:18559477; http:i/dx.doi.org/l0.ll0l/gad.i667008

Smith ZD, Chu MM, Miklelsen TS, Cu H, GnirkeA, Reg* A, Meissner A. A unique regulatory phu ofDNA methylation in the dy mmmaliao embryo.Nature 2012; 484:339-44; PMID:2245O10; hnp:l /dx.doi.org/10.1038/narurel0950

Lister R, Mukamel EA, NeryJR, Urich M, PuddifootCA, Johnsoa ND, Lucero J, Huang Y, Dwork AJ,

Schule MD, etal. Global epigoomic reonfigurationduring mammalim brain d*elopment. Science 2013;341:1237905: PMID:23828890; http:i/dx.doi.org/10. I 126lscience.l237 905

Heyn H, Vidal E, Sayols S, Sanchez-Mut JV,Moru S, Medina I, Smdoval J, Sim6-Riudalbas L,Szczesna K, Huertas D, et al. 'Whole-genome bisulfiteDNA sequencing of a DNMT3B orulant patienr.Epigenetics ?01); 7:542-50', PMID:22595875;Ircp://dx.doi.org/ [O.416l I epi.20523

Bedalov A, Simon JA. Neuroscience, NAf) ro therescue. Science 2004; 305:954-5; PMlD:15310883;lntpr//dx.doi.org/ lO.1126 I sqene.l 102497

Belenkv B Racette FG, Bogm KL, McClure JM,SmithJS, Brelner C. Nicotinamide riboside promotesSir2 silencing md *tends lifespan via Nrk and Urhl/Pnpl/Meul pathways to NAD+ . Cell 2007;1292473-84 I PMID r 174825 43 : http t / / dx.d.oi.orgl 10.7016 I j.cel|.20O7.03.O24

Belenky P, Bogan KL, Brenner C- NAD+metabolism in helth ud disec. Trends BiochemSci 2007; 32:12-9; PMID:l7 161604; http://dx.doi.orgi 10.10r6/j.tibs.2006.l 1.006

Qureshi lA. Mehler MF. Ad"mces in epigenericr

md epigenomics for neurodegmerativc direars.Curr Neurol Neurosci Rep 20lll' ll:464-73;PMID:21571i62; http://dx.doi,orgl10.1007/s1 r9l0-01 1-0210-2

Hojo K, ImamuraT, Takuashi N{, Ishii K, Sasaki 1r'1,

lmura S, Ozono R, Takatsuki Y, Takauchi S, MoriE. Hereditary rosory nilroparhy wirh deafoess uddementia: a clinical md neuoimaging sild,v. Eur J

Neurol 1999; 6357-61; PMID:10210919; hrp:/idx.doi.org/10.1046 I i.1468 -1331.1999.630357 x'Wright A, Dvck PJ. Hereditary sensory neuroparhvwith sensorineural deafness and early-onset dementia.Neurology 1995; 45:560-2; PM1D17898717r htrp://dx.doi-org/10.1212l\fNL.45.3.560Sun Z, Baheti S, Middha S, Kanrvu R. Zhang Y,

Li X, Beutler AS, Klee E, Asmann YW. ThompsonEA, et al. SAAP-RRBS: stremlined oal"sis mdmnotation pipeline for reduced represeoarionbisulfite sequencing. Bioinformatics 20121 28:2180-l; PMID:22689387; hnp://dx.doi.orgl10.1093ibioinformaticslbm3ST

Bird A. DNA methylarior patterns md epigenericmemory. Genes Dw 2002; 16 :6 -21 ;PMID :117 82440 t

http://dr.doi.org/10.1 101/gad.947 102

www.landesbioscience.com Epigenetics 1 193