dna methylation in alzheimer’s and down syndrome · 2011-04-25 · down syndrome (trisomy 21)...
TRANSCRIPT
DNA Methylation in Alzheimer’s and Down syndrome
B T k MD PhDBen Tycko MD, PhDTaub Institute for Research on Alzheimer’s Disease
and the Aging Brainand the Aging BrainColumbia University Medical Center
Why study DNA methylation?Why study DNA methylation?
• Promoter CpG methylation silences gene expression;Promoter CpG methylation silences gene expression; methylation elsewhere can activate genes.
• Methylation is perturbed in disease – best studied inMethylation is perturbed in disease best studied in cancer but now of interest (NIH Roadmap) in many other diseases – especially neurological ones.p y g
• Methylation can be manipulated by drugs and diet (5‐aza‐dC; folate); )
• Mapping CpG methylation can highlight genetic‐epigenetic interactions – help to find disease risk loci.ep ge et c te act o s e p to d d sease s oc
DNA methylation: faithfully transmitted toDNA methylation: faithfully transmitted to daughter cells in proliferating tissues
CpG
G C
me
CpG
G C
me
CpG
meGpC
me
GpC
DNMT1 re‐methylates daughterGpC
me me
DNMT1 re‐methylates daughter strands in S‐phase
CpG
GpC
CpG
GpC
meme
Changing DNA methylation:demethylation and de novo methylation
• Pre‐existing methylation can be lost when DNA replicates in the absence of DNMT1p
• de novo gain of methylation at previously unmethylated sequences is catalyzed by theunmethylated sequences is catalyzed by the other methyltransferase enzymes – DNMT3A and DNMT3B (DNMT3A is most important inand DNMT3B (DNMT3A is most important in the brain)
Could DNA methylation be y“involved in AD”?
By Braak stage 2, tau tangles have accumulated further and have caused some neurons to burst apart and die. However, mental testing still shows minimal impairment. Tangles at this level or worse are found in the brains of about 60% of over‐65s. At this stage the tau tangles are much more extensive in the transitional entorhinal region (highlighted in yellow) and have begun to kill neurons here. This region is a "relay station" where sensory input is y ) g g y y pfiltered before being stored in the memory. Meanwhile, in the brain's hippocampus (pink) and neo‐cortex, tau protein is also beginning to aggregate but has not yet formed tangles.
W k b t t iWe know about toxic Abeta aggregates and tau tangles – why do
At Braak stage 2, tau tangles are causing cell death (yellow) in the transitional entorhinal region, a part of the brain that filters sensory
input before passing it on to the memory. Tau protein is aggregating in th hi ( i k) hi h i i l f d i th t
we need another hypothesis?
the hippocampus (pink) which is crucial for memory, and in the cortex(blue) which is associated with conscious thought.
Methods for profiling CpGp g pmethylation
• MSNP and related methods (methyl‐sensitive restriction enzymes; microarrays)y ; y )
• MeDIP/MIRA (affinity capture of methylated DNA; microarrays or Nextgen sequencing)DNA; microarrays or Nextgen sequencing)
• Illumina Infinium (bisulfite conversion; primer extension on Beadarrays)extension on Beadarrays)
• Nextgen bisulfite sequencing of targeted h l isequences or whole epigenomes
Promoter meth: major differences among tissues
Y‐axis: 2,300 genes with methylation fraction >.5 in at least 20% of the samples
Promoter meth: few differences in AD vs. control
Promoter meth: BA37 (temporal cortex)AD vs controlAD vs. control
Need to use lenient criteria to find anything:criteria to find anything:>10% absolute diff in methylation at a given gene;p<.05 by simple ANOVA (not Bonferoni (corrected)
AD vs. control whole temporal cortex:h h l d i i dhypermethylated promoter regions associated with decreased mRNA expression in published
d (AZ S H l h )data (AZ Sun Health group)
mRNA foldMethyl CpGfractional
gene symbolBrainregions
mRNA P‐value
mRNA fold change AD vs. control
fractional change AD vs.
control
TERF2IP HC 2 1E‐06 ‐2 44 0 15TERF2IP HC 2.1E‐06 ‐2.44 0.15
KCNC3 EC 7.9E‐03 ‐3.34 0.12
SUMO3 HC 6.1E‐03 ‐1.63 0.10
MGMT MTC 1.5E‐04 ‐2.24 0.10
C16orf24 EC 4.4E‐04 ‐2.66 0.10
NPTX2 MTC 1.1E‐05 ‐4.68 0.10
Separating neuronal vs. glial cellsSeparating neuronal vs. glial cells
Dounce frozen (autopsy) brain tissue
Isolate nuclei using density digradient
Label neuronal nuclei with NeuNtib dantibody
Separate neuronal and glial nuclei using FACS
DNA extraction Determine differential methylation patterns using microarrays, y p g y ,
Nextgen sequencing, etc.
Sorted neurons from autopsy brainsSorted neurons from autopsy brains
NeuN+
NeuNNeuN-
Sorted neurons: very few differentially methylated genes in AD vs. controls
• Infinium 27K “first generation” Beadarrays; mostly analyzing promoter regions; not other genomicanalyzing promoter regions; not other genomic regions.
• AD and control neurons from BA9 (frontal) and BA37AD and control neurons from BA9 (frontal) and BA37 (temporal) cortex.
• Zero differentially methylated CpG’s in AD vs. control y y psorted neurons, even with lenient criteria.
…but there are clear differences in neuronal DNA methylation between
different neocortical regionsdifferent neocortical regions
ANOVA p 05;ANOVA p.05; require >15 % difference in methylation.methylation.BA9 control vs. BA37 control neurons.Compare samples on controls only; add AD and DLBD to the clustering.
Results from Infinium 27KResults from Infinium 27K• In whole grey matter a few interesting genes do have mildly
l d h l laltered promoter CpG methylation in AD vs. control.
• In purified neurons we find almost no examples of genes with altered promoter methylation in AD vs. control.altered promoter methylation in AD vs. control.
• However, in purified neurons, promoter methylation varies substantially across different neocortical regions. Some of the differentially methylated genes are interesting.
Conclusion/hypothesis: patterns of promoter region DNAConclusion/hypothesis: patterns of promoter region DNA methylation likely help to specify neuronal identities in the human cerebral cortex but, except for a few interesting genes, these methylation patterns are not strongly altered in the earlythese methylation patterns are not strongly altered in the early to middle stages of AD.
Assessing global methylation: anti‐me‐C
T111-5 control D t t
T111-5 control CA1
T111-5 control LGN
Assessing global methylation: anti me C
DentateCA1 LGN
10x
40x
Globally reduced me‐C in CA1 in AD?
T247 -5 ADT247-5 AD T247-5 AD
Globally reduced me C in CA1 in AD?
DentateCA1 LGN
10x
40x
IHC using anti‐OHmeCT99 Control
IHC using anti OHmeC
DentateCA1 LGN
10x
40x
Globally reduced OH‐meC in CA1 in AD?
T247 AD
Globally reduced OH meC in CA1 in AD?
DentateCA1 LGN
10x
40x
Global meC and OH‐meC: work is still in progress
• Careful checking of technical issues (antigen accessibility) using other nuclear antigensy) g g
• Examining other brain regions in AD and control brainscontrol brains
• Careful correlations with neuropathology including plaques and tanglesincluding plaques and tangles
DNA methylation and neuronal specification: what happens in Down
syndrome?syndrome?
• Methylation profiling and detailed studiesMethylation profiling and detailed studies comparing normal vs. Down syndrome cells– Blood cells (PBL T‐cells)Blood cells (PBL, T cells)
– Cortical grey matter and purified neurons
Down Syndrome (trisomy 21)
Cognitive disability
Down Syndrome (trisomy 21)
Cognitive disabilityCongenital heart defectsEarly‐onset Alzheimer’s disease (1.5X APP)Childhood leukemiasAutoimmune diseases (50‐fold increased)Recurrent infectionsRecurrent infectionsAltered lymphocyte and NK cell function
Altered patterns of CpG methylation in PBL from adults with DS
AML1
AML2
AML3
DS PBL Control PBL DS PBL Control PBL
Illumina Infinium 27KAffymetrix 250K MSNP
Validations: the TMEM131 locus
BstUI
U
M
DS PBL Normal PBL
HpyCH4IV
M
U
M
DS PBL (TMEM131 short ) Normal PBL (TMEM131 short)
240 KbTMEM131 long
TMEM131 short
0.5 Kbrs6760008
CGI
Validations: the PLD6 and CD3Z genes
DS PBL Normal PBL DS PBL Normal PBL
g
S o a S o a
PLD6 (LOC201164)
Infinium probe CD247 (CD3Z)
Infinium probe
200bp 200bpCGI
MS-Pyroseq
DS PBLNormal PBLNormal cord bloodNormal PBMCN l PMNNormal PMN
Altered mRNA expression of the differentially th l t d i DS PBL
essi
on .
10 DS av=5.40Control av=1.68T t t <0 0001
40 DS av=2.92
Control av=16.5T t t 0 00012
methylated genes in DS PBL
M131-
shor
t exp
re
3
5
8
T-test p<0.0001
10
20
30
CF7
expr
essi
on
T-test p= 0.00012
TMEM
DSN=30
ControlN=20
0
3
DSN=24
ControlN=23
0
10TC
250 Kb (additional exons)TMEM131 long
CGI TMEM131 short
rs67600081 kb
TCF7
rs756699
5 kb 23 kb
CGI 1 kb
The altered methylation patterns are functionally relevant: gene activation
JurkatPrimary PBMC,
0 40.60.81.01.21.4
shor
t mR
NA
-sho
rt m
RN
A
15202530
cells expanded with IL-150
0.20.4
TMEM131-
s
0 0.25 0.55aza-dC (M) TM
EM131-
0510
0 0.25 0.55aza-dC (M)
56
1012
012345
TCF7
mR
NA Primary
PBMC, expanded with PHA
02468
TCF7
mR
NA
Jurkatcells
00 0.25 0.5
5aza-dC (mM)
00 0.25 0.5
5aza-dC (mM)
A programmed epigenetic response to a
Altered patterns of
simple chromosomal aneuploidy in humans
Trisomy 21DNA methylation
Altered
Self‐renewing tissue:
Cellular selection
regulatory networks
ghematopoietic system
Outgrowth of clones of cells with stereotypical epigenetic abnormalities. Genes are involved in lymphocyte function – candidates for producing immune deficits and increased autoimmunity in Down syndrome and in the general population, i.e. Down syndrome as an “experiment of nature”.
Recent publications from our groupgroup
Infinium 27K Methylation DS vsnormal cerebellum; ANOVA p.05; 1.2fold diff; 0 1 absolute diffdiff; 0.1 absolute diff
Infinium 27K Methylation; DS vscontrol cerebral cortex; 1.2fold; 0 05diff; ANOVA p 050.05diff; ANOVA p.05
ESR1: preliminary validations of differential p ymethylation in DS vs. normal brains
ESR140 kb
ESR1
cg20627916
5965DS BRAIN
5983CONTROL BRAIN
5980DS BRAIN
ESR1 Methyl: .31 .26 .09
ESR1 is a risk locus for AD in DS:potential convergence of genetic and epigenetic data
Current Research GroupCurrent Research Group
Ri h d M• Huferesh Darbary
• Martha Salas
• Richard Mayeux
• Nicole Schupf
• Sandra Barral
• Maite Mendioroz
• John Crary
• J‐P Vonsattel
• Alexis Temkin
• Kristi Kerkel
• Jerzy Wegiel
• Wayne SilvermanKristi Kerkel• Warren Zigman
• Ed JenkinsGrant Support from the NIA d NICHDNIA and NICHD