NEUROLOGICAL REVIEW

Understanding Neurological Disease Mechanisms

in the Era of Epigenetics
Irfan A. Qureshi, MD; Mark F. Mehler, MD

T
he burgeoning field of epigenetics is making a significant impact on our understand-
ing of brain evolution, development, and function. In fact, it is now clear that epigen-
etic mechanisms promote seminal neurobiological processes, ranging from neural stem
cell maintenance and differentiation to learning and memory. At the molecular level,
epigenetic mechanisms regulate the structure and activity of the genome in response to intracel-
lular and environmental cues, including the deployment of cell type–specific gene networks and
those underlying synaptic plasticity. Pharmacological and genetic manipulation of epigenetic fac-
tors can, in turn, induce remarkable changes in neural cell identity and cognitive and behavioral
phenotypes. Not surprisingly, it is also becoming apparent that epigenetics is intimately involved
in neurological disease pathogenesis. Herein, we highlight emerging paradigms for linking epi-
genetic machinery and processes with neurological disease states, including how (1) mutations in
genes encoding epigenetic factors cause disease, (2) genetic variation in genes encoding epigen-
etic factors modify disease risk, (3) abnormalities in epigenetic factor expression, localization, or
function are involved in disease pathophysiology, (4) epigenetic mechanisms regulate disease-
associated genomic loci, gene products, and cellular pathways, and (5) differential epigenetic pro-
files are present in patient-derived central and peripheral tissues.
JAMA Neurol. 2013;70(6):703-710.
Published online April 9, 2013. doi:10.1001/jamaneurol.2013.1443

The hallmarks of the human brain are its
extraordinary degree of cellular diversity,
capacity for synaptic and neural network
connectivity and plasticity, and intellec-
tual abilities. Ongoing efforts have sought
to better understand this hierarchical or-
ganization and the molecular, cellular, and
environmental mechanisms responsible for
generating it. The completion of the Hu-
man Genome Project and the continuing
characterization of functional genomic ele-
ments (tissue-specific promoters, enhanc-
Author Affiliations: Roslyn and Leslie Goldstein Laboratory for Stem Cell Biology
and Regenerative Medicine (Drs Qureshi and Mehler), Institute for Brain Disorders
and Neural Regeneration (Drs Qureshi and Mehler), Departments of Neurology
(Drs Qureshi and Mehler), Neuroscience (Dr Mehler), and Psychiatry and
Behavioral Sciences (Dr Mehler), Rose F. Kennedy Center for Research on
Intellectual and Developmental Disabilities (Drs Qureshi and Mehler), Einstein
Cancer Center (Dr Mehler), Ruth L. and David S. Gottesman Institute for Stem
Cell Biology and Regenerative Medicine Research (Dr Mehler), Center for
Epigenomics (Dr Mehler), and Institute for Aging Research (Dr Mehler), Albert
Einstein College of Medicine, Bronx, New York, New York.
ers, and alternative exons) represent prime
advances toward this goal.1,2 This postge-
nomic era has been defined by the rise of
epigenetics—the scientific discipline fo-
cused on interrogating how genomic pro-
cesses, such as gene transcription and DNA
replication and repair, are mediated in dif-
ferent cellular contexts.
Epigenetics promises to provide insights
thatwillhelpanswerseminalquestionsabout
the human brain. How did it evolve? How
does the human genome encode neural cel-
lular diversity? How do genetic factors and
environmental stimuli interact to promote
synaptic and neural connectivity and plas-
ticity?Howdocognitiveandbehavioraltraits
emerge? Most importantly, what mecha-
nisms are responsible for the pathogenesis
of complex neurological diseases?
Further, the rapidly emerging era of
highly personalized epigenetic and epi-

JAMA NEUROL/ VOL 70 (NO. 6), JUNE 2013 WWW.JAMANEURO.COM

703

©2013 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ by a PUC/RS User on 12/26/2021
 Environmental cues/neuromodulators/synaptic activity/
metabolic signals/stress responses

ADAR 1/2 miRNAs ESC

A RNA degradation
APOBECs Translational repression NSC
I

C
mRNA/ncRNA
U

Nucleosomes
Histones

DNA strand

N1 N2 N3 OL AS
me

P/Me/Ac/Ub
5′ ...NCGN...
...NGCN... 5′ H2A
me

H3 H4 SWI/SNF
H2B Adult neurogenesis

MBDs DG
DNMTs
HMTs/HDMs
HATs/HADCs

LTP
PcG/TrxG
ΔLTD/ΔLTP

ANRIL/HOTAIR
LTD

Figure 1. Environmental cues, neuromodulators, synaptic activity, metabolic signals, and stress responses lead to the activation of diverse epigenetic
mechanisms, including DNA methylation, histone modifications and chromatin remodeling, noncoding RNA (ncRNA) expression, and RNA editing. In turn, these
processes mediate embryonic stem cell (ESC) and neural stem cell (NSC) maintenance and maturation, adult neurogenesis, neural network formation, and
synaptic plasticity. Ac indicates acetylation; ADAR, adenosine deaminases that act on RNA enzymes; APOBECs, apolipoprotein B editing catalytic subunit enzymes;
AS, astrocyte; C, cytosine; DG, dentate gyrus; DNMTs, DNA methyltransferases; G, guanine; HATs, histone acetylases; HDACs, histone deacetylases; HDMs,
histone demethylases; HMTs, histone methyltransferases; LTD, long-term depression; LTP, long-term potentiation; MBDs, methyl-CpG-binding domain proteins;
me, methylation; miRNA, microRNA; mRNA, messenger RNA; N, any nucleotide; N1-3, neuronal subtypes 1-3; OL, oligodendrocyte; P, phosphorylation;
PcG, Polycomb group proteins; TrxG, Trithorax group proteins; and Ub, ubiquitination.

genomic medicine is poised to radically transform diag-
nostic and therapeutic strategies for neurological dis-
eases and to deliver innovative treatments to promote
neural protection and repair.
THE ERA OF EPIGENETICS
Revolutionary Insights
Among the most important insights to have emerged is
an appreciation for chromatin organization in regulat-
ing genomic function and establishing cellular memory
states. Chromatin refers to the packaging of the genome
within the cell nucleus. DNA is wrapped around a his-
tone protein octamer, forming a fundamental chroma-
tin structure, the nucleosome. Chromatin states play cen-
tral roles in coordinating the accessibility of DNA
sequences to factors in the nucleus mediating critical cel-
lular processes, including gene transcription. Nucleosome-
free regions represent DNA actively engaged in regula-
tory and other functions. These regions can be identified,
experimentally, by their relative hypersensitivity to nucle-
ases (DNase I).3 Higher-order chromatin exists as rela-
tively open (euchromatic) or highly condensed (hetero-
chromatic) structures. Euchromatin is generally associated
with active transcription, whereas heterochromatin is typi-
cally found in inactive regions, such as repressed genes
and structural components of chromosomes (centro-
meres and telomeres). Chromatin structure is dynamic
and subject to local modification at the level of indi-
vidual nucleotides, histone proteins, and nucleosomes
and genome-wide by higher-order chromatin remodel-
ing. Protein complexes mediate these processes by the
capacity to “read,” “erase,” and “write” specific chroma-
tin states (“marks”). Inhibiting specific chromatin-
modifying enzymes is a powerful tool for modulating gene
expression programs and a strategy approved by the Food
and Drug Administration for select disease indications
and now in preclinical and clinical trials for cancer and
neurodegenerative diseases.
Chromatin states are intimately linked to the estab-
lishment and maintenance of cell identity (Figure 1).
Chromatin exists in a relatively open conformation in em-
bryonic stem cells, reflecting their pluripotent state. As

JAMA NEUROL/ VOL 70 (NO. 6), JUNE 2013 WWW.JAMANEURO.COM

704

©2013 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ by a PUC/RS User on 12/26/2021
 cellular differentiation proceeds, chromatin becomes more
closed, with marks indicative of gene repression, par-
ticularly in chromatin associated with pluripotency genes.
Conversely, chromatin associated with lineage-specific
genes selectively opens and becomes decorated with marks
of active gene transcription. Differences in chromatin
states explain how the genome gives rise to the plethora
of organismal cell types, including the extraordinary di-
versity of neurons and glia of the human brain. Agents
targeting epigenetic enzymes have, therefore, been used
to reprogram cell identity and function and may also pro-
mote neural regeneration and repair.
Another key insight is that the entire human genome
is actively transcribed, including regions that code for pro-
teins (⬍2%) as well as those that do not. Such noncoding
regions give rise to numerous classes of noncoding RNAs
(ncRNAs)—a previously unexplored layer within the ge-
nome—that are interlaced with each other and with protein-
coding messenger RNAs (mRNAs). The identification of
large numbers of these ncRNAs embedded within the ge-
nome and recognition of their important biological roles
directly challenges the central dogma of molecular biol-
ogy (DNA makes RNA and RNA makes protein).
Noncoding RNAs are far more abundant than protein-
coding mRNAs in human cells, particularly in neural cells.
They engage in complex interactions with DNA, RNA,
and proteins and exhibit a spectrum of sophisticated func-
tions. Certain classes of ncRNAs are implicated in gene
silencing whereas others are involved in regulating the
activity of retrotransposons. Notably, the expression of
ncRNAs is most prominent in the brain and subject to
intricate temporal and spatial regulation during neural
development, homeostasis, and plasticity, highlighting
their significance.4 In fact, it is becoming clear that non-
coding DNA sequences and associated ncRNAs serve as
substrates for evolutionary innovations in the human
brain. Human accelerated regions represent elements in
the human genome that have evolved rapidly since di-
vergence from our common ancestor with chimpan-
zees. One of hundreds of such regions, HAR1, gives rise
to an ncRNA selectively expressed in Cajal-Retzius cells
in the developing fetus and implicated in neuronal sub-
type specification and cell migration in the neocortex.
Epigenetic Mechanisms
The major epigenetic mechanisms include DNA meth-
ylation, histone modifications and nucleosome and higher-
order chromatin remodeling, ncRNAs, and RNA editing
(Table 1). DNA methylation refers to covalent modifi-
cation of cytosine residues to form 5-methylcytosine
(5mC). 5-Methylcytosine levels are dynamically regu-
lated during development and adulthood. Enzymes that
catalyze DNA methylation are DNA methyltransferases;
those that catalyze DNA demethylation include DNA ex-
cision repair, cytidine deaminase, and Gadd45 proteins.
The presence of 5mC at specific gene loci was previ-
ously thought to promote transcriptional repression and
long-term gene silencing. However, high-resolution ge-
nome-wide studies interrogating DNA methylation re-
veal that 5mC is distributed throughout the genome (gene
regulatory regions, transcriptional start sites, gene bod-
JAMA NEUROL/ VOL 70 (NO. 6), JUNE 2013
705
©2013 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ by a PUC/RS User on 12/26/2021
Table 1. Principal Epigenetic Mechanisms and Factors
Epigenetic Mechanism Epigenetic Factors
DNA (de)methylation and DNA methyltransferase enzymes
hydroxymethylation Methyl-CpG-binding domain proteins
DNA excision repair enzymes
Cytidine deaminase enzymes
Gadd45 proteins
Ten-Eleven Translocation enzymes
Histone and chromatin Histone-modifying (histone [de]acetylase
modifications and [de]methylase) enzymes
SWI/SNF nucleosome remodeling
complexes
Polycomb group proteins
Trithorax group proteins
RE1-silencing transcription factor
CoREST
Noncoding RNAs MicroRNAs
Small nucleolar RNAs
Endogenous short-interfering RNAs
PIWI-interacting RNAs
Long noncoding RNAs
RNA editing Adenosine deaminases that act on RNA
enzymes
Apolipoprotein B editing catalytic subunit
enzymes
Abbreviations: CoREST, corepressor for element-1–silencing transcription
factor; RE1, repressor for element-1.
ies, and repetitive elements) and DNA methylation seems
to be context specific. The effects of DNA methylation
are mediated by methyl-CpG-binding domain proteins,
including methyl-CpG-binding protein 2 (MECP2), which
bind to 5mC and recruit additional regulatory factors to
methylated loci. 5-Hydroxymethylcytosine, another modi-
fied cytosine residue generated by the Ten-Eleven Trans-
location family of enzymes that oxidizes 5mC, appears
to counteract 5mC by inhibiting the function of methyl-
CpG-binding domain proteins.
Each nucleosome contains 2 of each core histone pro-
tein (H2A, H2B, H3, and H4). Specific classes of histone
deacetylases (HDACs) as well as histone acetyltransfer-
ases, histone demethylases, histone methyltransferases, and
others catalyze histone posttranslational modifications.
These enzymes serve as “writers” of posttranslational modi-
fications, thereby altering nucleosome structure and func-
tion by modifying interactions with other proteins that
“read,” “erase,” and “write” epigenetic marks. Macromo-
lecular complexes containing various proteins with the abil-
ity to simultaneously read, erase, and write epigenetic marks
promote nucleosome (SWI/SNF) and higher-order chro-
matin (Polycomb group and Trithorax group proteins) re-
modeling. Epigenetic regulatory complexes (repressor for
element-1 [RE1]–silencing transcription factor [REST] and
co-RE1–silencing transcription factor [CoREST] com-
plexes) can also be highly modular and have functions that
range from recognizing specific nucleotide sequences and
DNA methylation states to promoting local histone and
higher-order chromatin modifications.
Newly recognized classes of ncRNAs are classified as
short or long. Salient classes of short ncRNAs include mi-
croRNAs (miRNAs), endogenous short-interfering RNAs,
PIWI-interacting RNAs, and small nucleolar RNAs. Each
WWW.JAMANEURO.COM
 class is associated with specific biogenesis pathways,
mechanisms of action, and functions. MicroRNAs are the
best-characterized class of short ncRNAs. They engage
in posttranscriptional gene regulation. Specifically, single-
stranded, 20- to 23-nucleotide miRNAs bind to protein-
coding mRNAs in their 3' untranslated regions through
complementary sequence-specific interactions and pre-
vent translation of these target mRNAs. Importantly, a
single miRNA molecule can repress a network of mRNAs.
Long ncRNAs (lncRNAs) are the most abundant but least
well-characterized class of ncRNAs. The most common
strategy for analyzing lncRNAs is to study their genomic
context relative to protein-coding genes. Some lncRNAs
are derived from intergenic regions (long intergenic/
intervening ncRNAs). Other lncRNAs are transcribed in
antisense or overlapping orientations relative to protein-
coding genes, and these transcripts are often function-
ally interrelated. For example, expression of a protein-
coding mRNA might be specifically regulated by an
antisense lncRNA from the same genomic locus. Never-
theless, lncRNAs are highly versatile with a spectrum of
novel activities. These include roles in recruiting non-
selective transcriptional and epigenetic regulators to spe-
cific loci distributed throughout the genome, forming
nuclear bodies, modulating nuclear-cytoplasmic trans-
port, and controlling local protein synthesis at synapses.
RNA editing is a mechanism for modifying RNA mol-
ecules. Adenosine-to-inosine editing is catalyzed by aden-
osine deaminases that act on RNA and cytidine-to-
uridine editing is catalyzed by apolipoprotein B editing
catalytic subunit enzymes. Intriguingly, cytidine deami-
nases also act on DNA and are implicated in DNA demeth-
ylation pathways (see earlier). RNA editing modifies the
information content of RNA molecules and affects their
intracellular dynamics. For example, RNA editing can
change codons in mRNAs, leading to expression of pro-
tein molecules different from those encoded in the ge-
nome. This process occurs primarily in proteins involved
in neural excitability, such as ion channels and neurotrans-
mitter receptors, and is thought to fine-tune synaptic re-
sponsiveness to environmental stimuli. RNA editing also
plays a role in controlling the subcellular localization of
transcripts. RNAs subject to high levels of editing can be
sequestered in nuclear bodies and released in response to
specific signals including cellular differentiation, viral in-
fection, and stress. While RNA editing has been studied
primarily in protein-coding mRNAs, the majority of RNA
editing occurs in noncoding regions. Lastly, RNA editing
is much more pronounced in the human brain than in other
organ systems and other species, highlighting its respon-
siveness to environmental stimuli.
EVOLVING EPIGENETIC LANDSCAPE OF BRAIN
A broad range of studies has focused on elucidating how
epigenetic mechanisms together coordinate brain devel-
opment and functioning throughout the life span, includ-
ing neural stem cell maintenance and differentiation and
synaptic plasticity underlying learning and memory.
Each epigenetic mechanism contributes to the estab-
lishment and maintenance of neural cell identity. In fact,
the appropriate expression and integrated functional
JAMA NEUROL/ VOL 70 (NO. 6), JUNE 2013
706
©2013 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ by a PUC/RS User on 12/26/2021
repertoires of DNA methyltransferases, methyl-CpG-
binding domain proteins, histone- and chromatin-
modifying enzymes, and ncRNA biogenesis factors are
necessary for mediating neurogenesis and gliogenesis.
Moreover, as embryonic stem cells differentiate into neu-
ral lineage-committed progenitors, hundreds of gene pro-
moter regions are methylated, including those associ-
ated with pluripotency and alternative lineage genes. These
DNA methylation events are responsible for silencing these
genes, thereby promoting neural lineage commitment.
Histone modifications and chromatin remodeling are simi-
larly involved in establishing neural cell identity. For ex-
ample, the developmental stage-specific deployment of
REST and CoREST complexes modulates neural lineage
elaboration, including neuronal and glial subtype speci-
fication, and terminal differentiation. In terms of ncRNAs,
both miRNAs and lncRNAs regulate neural stem cell self-
renewal and differentiation programs. The balance be-
tween specific miRNAs is implicated in the execution of
these programs. For example, miR-184 promotes neural
stem cell proliferation and inhibits neuronal differentia-
tion, whereas miR-9 and miR-137 inhibit neural stem cell
proliferation and promote neural differentiation. Other
miRNAs, including miR-125b, miR-128, miR-132, miR-
134, and miR-138, are associated with neuronal matu-
ration and neural network integration. Long ncRNAs also
influence the establishment of neural cell identity. For
example, mice lacking the Dlx6 antisense RNA 1 (Dlx6as1)
lncRNA have reduced numbers of GABAergic interneu-
rons selectively in the postnatal hippocampus and den-
tate gyrus, because Dlx6as1 regulates the expression of
DLX5 and DLX6, key transcription factors responsible
for the development of GABAergic interneurons.5
Epigenetic mechanisms also underpin synaptic plas-
ticity (long-term potentiation and depression) and learn-
ing and memory. Each epigenetic mechanism is respon-
sible for regulating genes critical for learning and memory,
such as brain-derived neurotrophic factor and cAMP re-
sponsive element-binding protein. In parallel, these
mechanisms can themselves be deployed in response to
synaptic activity. Indeed, DNA methylation levels, his-
tone posttranslational modification profiles, ncRNA ex-
pression patterns, and RNA editing in hippocampal neu-
rons all vary with memory consolidation and storage.
Animal models have also demonstrated the importance
of specific epigenetic factors in mediating cognitive and
behavioral phenotypes. These exciting studies suggest that
modulating the expression and function of epigenetic fac-
tors can modulate these processes. For example, Dnmt1
and Dnmt3a double knockout mice exhibit decreases in
DNA methylation selectively leading to abnormalities in
plasticity in the hippocampal CA1 region and impair-
ments in learning and memory. Furthermore, the
DNMT3A2 isoform is involved in age-associated cogni-
tive decline, and upregulation of DNMT3A2 seems to re-
store cognitive function in mice.6 Similarly, abnormali-
ties in HDAC2-mediated acetylation of histones associated
with genes involved in learning and memory contribute
to cognitive symptoms in neurodegenerative diseases, such
as Alzheimer disease, and they can be rescued by inhib-
iting HDAC2 activity.7
WWW.JAMANEURO.COM
 PARADIGMS LINKING EPIGENETICS
WITH NEUROLOGICAL DISEASE PROCESSES
Here, we describe several distinct but nonmutually ex-
clusive paradigms linking epigenetic machinery and
mechanisms with neurological disease states (Table 2
and Figure 2). Mutations in genes encoding epigenetic
factors can cause neurological disease. Among the most
salient examples is the MECP2 gene, which encodes a mul-
tifunctional methyl-CpG-binding domain family pro-
tein. MECP2 mutations, including missense, nonsense,
frameshift, and large deletion mutations, are the pri-
mary causes of Rett syndrome, an autism spectrum dis-
order. MECP2 mutations are also linked to a severe neo-
natal encephalopathy, syndromic forms of intellectual and
developmental disability, and an Angelman-like syn-
drome. Mutations in other epigenetic factors are simi-
larly associated with neurodevelopmental disorders. For
example, mutations in an increasing array of genes en-
coding proteins that directly and indirectly regulate chro-
matin structure cause syndromic and nonsyndromic forms
of intellectual and developmental disability.8 These emerg-
ing observations suggest that disruption of epigenetic net-
works is one of the principal pathogenic mechanisms
underlying intellectual and developmental disability. Mu-
tations in genes encoding epigenetic factors also pro-
mote the onset and progression of various forms of can-
cer, including primary brain tumors. Recent studies have
detected a strong association between mutations in the
epigenetic machinery—particularly in factors mediat-
ing histone lysine (de)methylation—and medulloblas-
toma.9 These findings are consistent with the observa-
tion that the cellular functions of lysine methylation,
including roles in regulating cell identity, DNA repair,
cell cycle, stress responses, and transcription, are all
known to be deregulated in cancer, in general, and me-
dulloblastoma, in particular. These studies implicate many
novel genes in medulloblastoma pathogenesis and also
correlate specific gene mutations with subgroups of tu-
mors with distinct histological and clinical features, sug-
gesting significant diagnostic, prognostic, and therapeu-
tic value in further characterizing these epigenetic targets.
Genetic variation in genes encoding epigenetic fac-
tors can modify the risk of neurological disease onset and
progression. For example, in patients with multiple scle-
rosis, single-nucleotide polymorphisms (SNPs) in 3 HDAC
gene loci—rs2522129 (SIRT4), rs2675231 (HDAC11), and
rs2389963 (HDAC9)—are correlated with brain volume
measurements on neuroimaging, markers of disease se-
verity. 1 0 In another prominent example, a large
genome-wide association study performed by the Inter-
national Stroke Genetics Consortium demonstrated that
a SNP in an intron of the HDAC9 gene (rs11984041) on
chromosome 7p21.1 is associated with risk of large-
vessel ischemic stroke (odds ratio, 1.38; 95% CI,
1.22-1.57).11 Although the precise pathogenic molecu-
lar mechanism is still unclear, HDAC9 is known to play
numerous roles in developmental programs and adap-
tive responses across many organ systems involved in the
pathophysiology of atherosclerosis and stroke. Given its
intronic location, the effect of this SNP might not simply
manifest via the protein-coding gene in which it is em-
JAMA NEUROL/ VOL 70 (NO. 6), JUNE 2013
707
©2013 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ by a PUC/RS User on 12/26/2021
Table 2. Emerging Paradigms Linking Epigenetic
Factors and Mechanisms With Diverse Neurological
Disease Processes
Paradigms
Mutations in genes encoding epigenetic factors cause disease
Genetic variations in genes encoding epigenetic factors modify
disease risk
Abnormalities in epigenetic factor expression, localization, or function
are involved in disease pathophysiology
Epigenetic mechanisms regulate disease-associated genomic loci,
gene products, and cellular pathways
Differential epigenetic profiles are present in patient-derived central
and peripheral tissues
bedded. Rather, the risk of large-vessel ischemic stroke
might be modified by the effect of this SNP on 1 or more
ncRNAs derived from the same locus. Indeed, it is well
documented that SNPs within disease-linked genomic loci
harboring both protein-coding genes and ncRNA genes
can modify disease risk via effects on ncRNA genes. For
example, CDKN2B-AS1, also known as ANRIL, is an an-
tisense lncRNA transcribed from the CDKN2B-
CDKN2A gene cluster on chromosome 9p21.3. A case-
control study demonstrated that SNPs at this locus,
including those that preferentially influence expression
of ANRIL transcripts, modify the risk of large-vessel is-
chemic and hemorrhagic stroke and, independently, the
risk of recurrent stroke and cardiovascular mortality.12
Emerging data suggest that ANRIL interacts with other
epigenetic regulatory factors, including PRC2, and can
modulate gene expression programs underpinning di-
verse cellular processes (proliferation, apoptosis, extra-
cellular matrix remodeling, and inflammation). Nota-
bly, other genome-wide association studies have found
that the ANRIL locus is a genomic hot spot influencing
susceptibility to other disease states affecting the ner-
vous system and beyond.13 These observations highlight
the interrelated nature of different epigenetic regulatory
processes, their impact on critical cellular programs, and
their relevance to disease.
Abnormalities in epigenetic factor expression, local-
ization, or function can be involved in the molecular patho-
physiology of neurological disease. Convergence be-
tween epigenetic factors and disease-associated pathways
and proteins, including those that explicitly cause dis-
ease when mutated, results in this kind of deregulation.
For example, mutations in the huntingtin protein (HTT),
which cause Huntington disease, lead to abnormal pro-
files of REST expression, localization, and function. RE1-
silencing transcription factor is upregulated in Hunting-
ton disease and clearly contributes to the transcriptional
dysregulation that is a hallmark of the disease. Wild-
type HTT modulates the translocation of REST into the
neuronal cell nucleus by forming a complex with other
factors. Mutant HTT alters the conformation of this com-
plex, leading to aberrant nuclear accumulation of REST.14
In turn, REST-regulated genes, including protein-
coding and ncRNA genes, are deregulated in Hunting-
ton disease.15 Other proteins that cause disease are simi-
larly involved in modulating the nuclear-cytoplasmic
shuttling of epigenetic factors. The ataxia telangiectasia
WWW.JAMANEURO.COM
 A Epigenetic gene mutations
Missense Nonsense Frameshift
C Abnormal epigenetic factor expression, localization, and function
E Abnormal central and peripheral epigenetic profiles
Donor cell
Figure 2. Emerging paradigms linking epigenetic factors and mechanisms with diverse neurological disease processes. A indicates adenosine; C, cytosine;
CSF, cerebrospinal fluid; G, guanine; me, methylation; RISC, RNA-induced silencing complex; and SNP, single-nucleotide polymorphism.
mutated protein indirectly regulates the subcellular lo-
calization of HDAC4.16 Deficiency of ataxia telangiec-
tasia mutated leads to neurodegeneration as a conse-
quence of HDAC4 accumulation in the neuronal cell
nucleus. Moreover, inhibiting HDAC4 activity or block-
ing its nuclear localization can rescue the pathological
phenotype. Intriguingly, perturbations of the subcellu-
lar localization of HDAC1 that occur in response to in-
flammatory insults disrupt axonal integrity in animal mod-
els of multiple sclerosis, and HDAC1 is mislocalized in
regions of demyelination in tissues derived from pa-
tients with multiple sclerosis.17 These observations raise
the possibility that abnormalities in nuclear-cytoplas-
mic shuttling of epigenetic factors represent a common
pathogenic mechanism for neuronal damage and loss. In
addition to interactions with chromatin regulatory fac-
tors, disease-associated pathways and proteins also in-
fluence ncRNAs. Mutations in leucine-rich repeat ki-
nase 2 (LRRK2) that cause familial and sporadic forms
of Parkinson disease compromise the integrity of miRNA-
mediated translational repression, resulting in deregu-
lation of pathways implicated in Parkinson disease patho-
genesis.18 Inhibiting the expression and function of let-7
and miR-184* reproduces these effects in wild-type LRRK2
cells and increasing let-7 and miR-184* expression
levels alleviates mutant LRRK2-induced pathology, sug-
gesting that these miRNAs are key players in LRRK2-
mediated Parkinson disease pathogenesis. Other disease-
associated factors, including fragile X mental retardation
protein (FMRP), also act cooperatively with the miRNA
machinery. FMRP is an RNA-binding protein with a va-
JAMA NEUROL/ VOL 70 (NO. 6), JUNE 2013
©2013 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ by a PUC/RS User on 12/26/2021
B Epigenetic gene variations
Deletion SNP SNP
D Epigenetic regulation of disease-associated genes and pathways
GAAn
me
me
CG CG
CSF
AAA
Bloodstream
RISC
Lymphatics
Recipient cell
riety of functions including the ability to serve as a trans-
lational repressor for target neuronal mRNAs by inter-
acting with select miRNAs and miRNA effector pathways
(RNA-induced silencing complex). These pathways are
perturbed in fragile X syndrome and related disorders
caused by mutations in the gene encoding FMRP. In ad-
dition, disease-linked factors affect lncRNAs. TAR DNA-
binding protein (TARDBP/TDP-43) is an RNA-binding
protein that forms inclusions in frontotemporal lobar de-
generation and amyotrophic lateral sclerosis. Interest-
ingly, interrogating the profiles of RNAs bound by TDP-43
reveals that the most significant increases in binding in
frontotemporal lobar degeneration–TDP are to nuclear
paraspeckle assembly transcript 1 (NEAT1) and metastasis-
associated lung adenocarcinoma transcript 1 (MALAT1;
also known as NEAT2).19 These lncRNAs localize to
nuclear bodies referred to as speckles (NEAT2) and para-
speckles (NEAT1) that coordinate the execution of post-
transcriptional programs within the nuclear compart-
ment. Speckles contain factors implicated in splicing. By
contrast, paraspeckles seem to serve as reservoirs for hy-
per adenosine-to-inosine edited RNAs that are rapidly ex-
ported into the cytoplasm in response to cellular stress.
The relationship between TDP-43 and these lncRNAs high-
lights the importance of TDP-43 for modulating post-
transcriptional RNA processing and transport in fronto-
temporal lobar degeneration and amyotrophic lateral
sclerosis.
Epigenetic mechanisms can regulate disease-asso-
ciated genomic loci, gene products, and cellular path-
ways. By definition, epigenetic mechanisms modulate the
WWW.JAMANEURO.COM
708
 entire genome, which includes loci involved in the patho-
genesis of every neurological disease with a genetic com-
ponent. Imprinting is the classic example of this phenom-
enon. It refers to monoallelic gene expression and is
mediated by the effects of DNA methylation, histone and
chromatin modifications, and ncRNAs that occur at the
imprinted locus in a parent-of-origin–dependent man-
ner. Clinical phenotypes associated with perturbations in
imprinting on chromosome 15q11-13 include Prader-
Willi and Angelman syndromes. Additionally, a rapidly in-
creasing number of studies have focused on identifying
differential profiles of DNA methylation and histone modi-
fications associated with disease-linked gene loci (in cen-
tral and peripheral tissues) and correlating these patterns
with corresponding gene expression changes and clinical
phenotypes. For example, an expansion repeat mutation
in the frataxin (FXN) gene causes Friedreich ataxia, and a
provocative study analyzed DNA methylation profiles as-
sociated with the FXN gene locus in peripheral blood mono-
nuclear cells and buccal cells from patients with Fried-
reich ataxia. It found that DNA methylation upstream of
the expansion repeat is correlated with decreased FXN
mRNA expression and increased disease severity and also
that DNA methylation downstream of the expansion re-
peat is correlated with age at onset of disease.20 Comple-
mentary studies have identified distinct profiles of his-
tone modifications, which likely regulate FXN expression,
flanking the mutant FXN expansion repeat in blood as well
as the brain and heart, the primary tissues affected in Fried-
reich ataxia. Another recent study demonstrated that lev-
els of an miRNA, miR-886-3p, are elevated in cell lines and
blood samples derived from patients with Friedreich ataxia
and that this miRNA plays a role in silencing FXN expres-
sion.21 In addition, epigenetic mechanisms regulate genes
and pathways involved in disease processes arising from
environmental exposures or genetic and environmental in-
teractions. One study focusing on chronic pain syn-
dromes caused by peripheral nervous system injuries dem-
onstrated that C-fiber dysfunction is mediated by decreased
levels of sodium channel, voltage-gated, type X, alpha sub-
unit and opioid receptor, mu 1 expression in the dorsal
root ganglion and that REST is responsible for the down-
regulation of these genes in response to injury.22 Another
study using animal models of chronic pain found that per-
sistent painful stimuli led to downregulation of the glu-
tamate decarboxylase 2 gene (Gad2) in pain-modulating
neurons of the brainstem nucleus raphe magnus. Hypo-
acetylation of histones in the Gad2 gene promoter region
was responsible for the decreased expression of Gad2, and
HDAC inhibition increased GAD2 protein expression and
mitigated pain sensitization. In fact, a plethora of evi-
dence suggests that pharmacological agents, such as HDAC
inhibitors, can target genes involved in disease-linked pro-
cesses including apoptosis, blood-brain barrier function,
cell-cycle control, DNA repair, excitotoxicity, heat shock,
inflammation, oxidative stress, and neurogenesis and glio-
genesis. These drugs have significant potential to modify
disease pathogenesis and symptoms and are actively being
developed for these purposes.
Differential epigenetic profiles can be present in patient-
derived central and peripheral tissues. Indeed, a rapidly
increasing number of studies have identified differen-
JAMA NEUROL/ VOL 70 (NO. 6), JUNE 2013
709
©2013 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ by a PUC/RS User on 12/26/2021
tial levels of DNA methylation, histone and chromatin
modifications, and ncRNA expression in a spectrum of
neurological disorders using patient-derived central and
peripheral tissues and animal models. Although the con-
tribution of these aberrant epigenetic profiles to disease
processes is sometimes unclear, deregulation of these mul-
tilayered epigenetic processes is likely to be respon-
sible, either directly or indirectly, for disease pathophysi-
ology by mediating cross talk between genetic
susceptibilities and sex, environment, nutritional states,
and aging. These seemingly ubiquitously abnormal sig-
natures may, therefore, have clinical applications as in-
dicators of disease risk, onset, progression, and respon-
siveness to treatment. For example, the DNA methylation
status of genomic repetitive elements in blood corre-
lates with Alzheimer disease and Mini-Mental State Ex-
amination scores, stroke and total mortality, and other
clinical phenotypes. Moreover, in multiple sclerosis,
miRNA expression profiles have been measured in white
matter lesions, where they can discriminate between ac-
tive and inactive lesions, and blood, where they can dis-
criminate between relapsing-remitting and other forms
of the disease and patients treated with different thera-
pies (glatiramer acetate and natalizumab) and those who
are untreated.23 The secretion of microvesicles, includ-
ing exosomes, is one interesting mechanism that might
explain why peripheral levels of ncRNAs reflect central
disease processes. Microvesicles are secreted by donor cells
(neural, immune, and other cell types) into cerebrospi-
nal fluid, lymphatics, and blood.24 Microvesicles trans-
fer DNA, RNA, and proteins and can actively “repro-
gram” recipient cells. For example, transferred miRNAs
can silence target mRNAs in recipient cells. The major-
ity of miRNAs found in the peripheral circulation are de-
rived from microvesicles, suggesting that these factors are
not simply bystanders but key effectors of dynamic central-
peripheral signaling. Long ncRNAs are similarly found
in microvesicles and their expression patterns are also
deregulated in a range of neurological diseases,25 sug-
gesting that interrogating central and peripheral lncRNA
profiles may have diagnostic and prognostic value.
PERSPECTIVES
Epigenetics is a swiftly advancing, but nascent, field. Much
of the intricacy and nuance surrounding epigenetic mecha-
nisms is yet to be defined fully. Nevertheless, epigenetics
has already revolutionized our understanding of how the
human brain evolved and how neural cellular diversity,
connectivity and plasticity, and advanced intellectual abili-
ties emerge. There is also an evolving literature focused
on discovering whether epigenetics is involved in medi-
ating the risk, onset, progression, and treatment respon-
siveness of neurological diseases. It is becoming clear from
these studies that aberrations associated with epigenetic
factors, pathways, and profiles can almost always be un-
covered, perhaps not surprisingly given the panoramic scope
of epigenetics. The contributions of epigenetic abnormali-
ties to neurological disease pathophysiology range from
being primary pathogenic mechanisms, protective re-
sponses, or simply passenger phenomena. Herein, we have
presented one framework for understanding and inter-
WWW.JAMANEURO.COM
 preting neurological disease mechanisms in the context
of epigenetics and for promoting the development of para-
digm-shifting clinical applications.
Accepted for Publication: October 17, 2012.
Published Online: April 9, 2013. doi:10.1001/jamaneurol
.2013.1443
Correspondence: Mark F. Mehler, MD, Albert Einstein
College of Medicine, Rose F. Kennedy Center, 1410 Pel-
ham Pkwy South, Room 401, Bronx, NY 10461 (mark
.mehler@einstein.yu.edu).
Author Contributions: Study concept and design: Qureshi
and Mehler. Acquisition of data: Qureshi and Mehler.
Analysis and interpretation of data: Qureshi and Mehler.
Drafting of the manuscript: Qureshi and Mehler. Critical
revision of the manuscript for important intellectual con-
tent: Qureshi and Mehler. Obtained funding: Mehler. Ad-
ministrative, technical, and material support: Mehler. Study
supervision: Mehler.
Conflict of Interest Disclosures: None reported.
Funding/Support: Dr Mehler is supported by grants
NS071571, HD071593, and MH66290 from the Na-
tional Institutes of Health, as well as by the F. M. Kirby,
Alpern Family, Mildred and Bernard H. Kayden, and Ro-
slyn and Leslie Goldstein foundations.
Additional Information: We regret that space con-
straints have prevented the citation of many relevant and
important references.
REFERENCES
1. Lander ES, Linton LM, Birren B, et al; International Human Genome Sequencing
Consortium. Initial sequencing and analysis of the human genome [published
correction appears in Nature. 2001;411(6838):720]. Nature. 2001;409(6822):
860-921.
2. Dunham I, Kundaje A, Aldred SF, et al; ENCODE Project Consortium. An inte-
grated encyclopedia of DNA elements in the human genome. Nature. 2012;
489(7414):57-74.
3. Thurman RE, Rynes E, Humbert R, et al. The accessible chromatin landscape of
the human genome. Nature. 2012;489(7414):75-82.
4. Qureshi IA, Mehler MF. Emerging roles of non-coding RNAs in brain evolution,
development, plasticity and disease. Nat Rev Neurosci. 2012;13(8):528-541.
5. Bond AM, Vangompel MJ, Sametsky EA, et al. Balanced gene regulation by an
embryonic brain ncRNA is critical for adult hippocampal GABA circuitry. Nat
Neurosci. 2009;12(8):1020-1027.
JAMA NEUROL/ VOL 70 (NO. 6), JUNE 2013
710
©2013 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ by a PUC/RS User on 12/26/2021
6. Oliveira AM, Hemstedt TJ, Bading H. Rescue of aging-associated decline in Dnmt3a2
expression restores cognitive abilities. Nat Neurosci. 2012;15(8):1111-1113.
7. Gräff J, Rei D, Guan JS, et al. An epigenetic blockade of cognitive functions in
the neurodegenerating brain. Nature. 2012;483(7388):222-226.
8. van Bokhoven H. Genetic and epigenetic networks in intellectual disabilities. Annu
Rev Genet. 2011;45:81-104.
9. Robinson G, Parker M, Kranenburg TA, et al. Novel mutations target distinct sub-
groups of medulloblastoma. Nature. 2012;488(7409):43-48.
10. Inkster B, Strijbis EM, Vounou M, et al. Histone deacetylase gene variants pre-
dict brain volume changes in multiple sclerosis. Neurobiol Aging. 2013;34
(1):238-247.
11. Bellenguez C, Bevan S, Gschwendtner A, et al; International Stroke Genetics Con-
sortium (ISGC); Wellcome Trust Case Control Consortium 2 (WTCCC2).
Genome-wide association study identifies a variant in HDAC9 associated with
large vessel ischemic stroke. Nat Genet. 2012;44(3):328-333.
12. Zhang W, Chen Y, Liu P, et al. Variants on chromosome 9p21.3 correlated with
ANRIL expression contribute to stroke risk and recurrence in a large prospec-
tive stroke population. Stroke. 2012;43(1):14-21.
13. Pasmant E, Sabbagh A, Vidaud M, Bièche I. ANRIL, a long, noncoding RNA, is
an unexpected major hotspot in GWAS. FASEB J. 2011;25(2):444-448.
14. Shimojo M. Huntingtin regulates RE1-silencing transcription factor/neuron-
restrictive silencer factor (REST/NRSF) nuclear trafficking indirectly through a
complex with REST/NRSF-interacting LIM domain protein (RILP) and dynactin
p150 Glued. J Biol Chem. 2008;283(50):34880-34886.
15. Zuccato C, Belyaev N, Conforti P, et al. Widespread disruption of repressor ele-
ment-1 silencing transcription factor/neuron-restrictive silencer factor occu-
pancy at its target genes in Huntington’s disease. J Neurosci. 2007;27(26):
6972-6983.
16. Li J, Chen J, Ricupero CL, et al. Nuclear accumulation of HDAC4 in ATM defi-
ciency promotes neurodegeneration in ataxia telangiectasia. Nat Med. 2012;
18(5):783-790.
17. Kim JY, Shen S, Dietz K, et al. HDAC1 nuclear export induced by pathological
conditions is essential for the onset of axonal damage. Nat Neurosci. 2010;
13(2):180-189.
18. Gehrke S, Imai Y, Sokol N, Lu B. Pathogenic LRRK2 negatively regulates microRNA-
mediated translational repression. Nature. 2010;466(7306):637-641.
19. Tollervey JR, Curk T, Rogelj B, et al. Characterizing the RNA targets and position-
dependent splicing regulation by TDP-43. Nat Neurosci. 2011;14(4):452-458.
20. Evans-Galea MV, Carrodus N, Rowley SM, et al. FXN methylation predicts ex-
pression and clinical outcome in Friedreich ataxia. Ann Neurol. 2012;71(4):
487-497.
21. Mahishi LH, Hart RP, Lynch DR, Ratan RR. miR-886-3p levels are elevated in
Friedreich ataxia. J Neurosci. 2012;32(27):9369-9373.
22. Uchida H, Ma L, Ueda H. Epigenetic gene silencing underlies C-fiber dysfunc-
tions in neuropathic pain. J Neurosci. 2010;30(13):4806-4814.
23. Thamilarasan M, Koczan D, Hecker M, Paap B, Zettl UK. MicroRNAs in multiple
sclerosis and experimental autoimmune encephalomyelitis. Autoimmun Rev. 2012;
11(3):174-179.
24. Lai CP, Breakefield XO. Role of exosomes/microvesicles in the nervous system
and use in emerging therapies. Front Physiol. 2012;3:228.
25. Qureshi IA, Mattick JS, Mehler MF. Long non-coding RNAs in nervous system
function and disease. Brain Res. 2010;1338:20-35.
WWW.JAMANEURO.COM