ISSN: 1524-4628
Copyright © 2006 American Heart Association. All rights reserved. Print ISSN: 0039-2499. Online
Stroke is published by the American Heart Association. 7272 Greenville Avenue, Dallas, TX 72514
DOI: 10.1161/01.STR.0000195177.61184.49
2006;37;248-255; originally published online Dec 8, 2005; Stroke
Gabrielle G. Leblanc, James F. Meschia, Donald T. Stuss and Vladimir Hachinski
Challenges
Genetics of Vascular Cognitive Impairment: The Opportunity and the
http://stroke.ahajournals.org/cgi/content/full/37/1/248
located on the World Wide Web at:
The online version of this article, along with updated information and services, is
http://www.lww.com/reprints
Reprints: Information about reprints can be found online at
journalpermissions@lww.com
410-528-8550. E-mail:
Fax:Kluwer Health, 351 West Camden Street, Baltimore, MD 21202-2436. Phone: 410-528-4050.
Permissions: Permissions & Rights Desk, Lippincott Williams & Wilkins, a division of Wolters
http://stroke.ahajournals.org/subscriptions/
Subscriptions: Information about subscribing to Stroke is online at
by on November 3, 2010 stroke.ahajournals.orgDownloaded from
Genetics of Vascular Cognitive Impairment
The Opportunity and the Challenges
Gabrielle G. Leblanc, PhD; James F. Meschia, MD; Donald T. Stuss, PhD; Vladimir Hachinski, MD
Background and Purpose—This review considers the current state of knowledge of genetic factors underlying vascular
cognitive impairment (VCI).
Summary of Review—We argue here that genes conferring susceptibility to VCI must be of 2 nonmutually exclusive
classes: (1) genes that confer susceptibility to cerebrovascular disease, and (2) genes that determine brain tissue
responses to cerebrovascular disease (ie, render parenchymal tissue more or less susceptible to injury or able to repair
itself after injury). Although some progress has been made in identifying genes of the first class, little has been done
to explore genes of the second class. Evidence for the existence of such genes is presented. We discuss the advantages
and disadvantages of different forms of cerebrovascular disease for studying these genes, and different study designs that
might be used.
Conclusion—The most critical challenge for genetic studies of VCI is to identify quantifiable phenotypes that can be
reliably and effectively determined in large samples of subjects. (Stroke. 2006;37:248-255.)
Key Words: cerebrovascular disorders � cognition � dementia � ischemia
Cognitive impairment attributable to cerebrovascular dis-ease is a rapidly escalating public health problem. For
example, up to one third of all stroke survivors exhibit
dementia within 3 months after their stroke.1–3 In addition,
postmortem pathological studies4–9 indicate that 15% to 34%
of dementia cases (of which there are currently �4 million in
the United States) show significant vascular pathology, either
alone or in combination with Alzheimer disease (AD) pathol-
ogy. However, dementia represents only a portion of the
burden of cognitive dysfunction associated with cerebrovas-
cular disease. In addition to patients who develop dementia,
there are those who develop cognitive impairment that does
not fulfill traditional criteria for dementia but that nonetheless
has a significant impact on quality of life and ability to carry
out activities of daily living. As a result, the older term
“vascular dementia” is being replaced with a new one:
“vascular cognitive impairment” (VCI), in which frank de-
mentia may or may not be a feature.4–6 Recent studies
indicate that the prevalence of VCI without dementia is equal
to that of VCI with dementia,7,8 suggesting that the total
prevalence of VCI (with or without dementia) could be as
high as 3 million cases in the United States.
The Opportunity: Genetics of VCI So
Far Unexplored
Although the prevalence of VCI approaches that of AD,
research on VCI has lagged considerably behind that on AD,
particularly with regard to pathogenic mechanisms. Our un-
derstanding of the pathobiology of AD vaulted forward with
the discovery of genes that produce monogenic forms of the
illness or contribute to polygenic forms. Similarly, the iden-
tification of genes contributing to VCI would no doubt
provide insight into the cellular and molecular basis of VCI.
The genes underlying VCI must be of 2 nonmutually
exclusive classes: (1) genes that predispose individuals to
cerebrovascular disease, and (2) genes that determine
tissue responses to cerebrovascular disease (eg, genes
conveying ischemic tolerance or susceptibility, or the
ability to recover from ischemic insult). With regard to the
first class of genes, some progress has been made in
the past few years in identifying genes that confer suscep-
tibility to hypertension and stroke.9 –14 In addition, several
monogenic forms of cerebrovascular disease have been
identified. The 2 best studied of these are cerebral autosomal
dominant arteriopathy with subcortical infarcts and leucoen-
cephalopathy (CADASIL) and hereditary cerebral hemor-
rhage with amyloidosis-Dutch type (HCHWA-D). CADASIL
is a syndrome of subcortical small vessel disease accompa-
nied by lacunar strokes, migraine, and dementia.15 The
disease results from mutations in the Notch 3 gene,16,17 which
is normally expressed in vascular smooth muscle cells and
pericytes (including those of the cerebral vasculature).18,19
The gene appears to be involved in directing smooth muscle
Received May 11, 2005; final revision received September 2, 2005; accepted September 27, 2005.
From the Neurogenetics Group, National Institute of Neurological Disorders and Stroke (G.G.L.), Bethesda, Md; Department of Neurology (J.F.M.),
Mayo Clinic, Jacksonville, Fla; Baycrest Centre for Geriatric Care (D.T.S.), The Rotman Research Institute, University of Toronto, Ontario, Canada;
London Health Sciences Center (V.H.), University of Western Ontario, London, Ontario, Canada.
Correspondence to Gabrielle G. Leblanc, PhD, National Institute for Neurological Disorders and Stroke, 6001 Executive Blvd, Room 2114, Bethesda,
MD 20892. E-mail leblancg@ninds.nih.gov
© 2005 American Heart Association, Inc.
Stroke is available at http://www.strokeaha.org DOI: 10.1161/01.STR.0000195177.61184.49
248
Comments, Opinions, and Reviews
by on November 3, 2010 stroke.ahajournals.orgDownloaded from
cell proliferation and differentiation.20–24 HCHWA-D is a
syndrome of primarily hemorrhagic strokes and dementia.25,26
It is caused by a mutation in the gene for amyloid precursor
protein (APP) that causes abnormal deposition of amyloid in
the walls of leptomeningeal arteries and cortical arterioles (a
pathological condition known as cerebral amyloid angiopathy
[CAA]).27,28 Mouse models have been developed for CADA-
SIL28,29 and HCHWA-D28–30 and have contributed critical
insights into the cell biology of the pathogenic processes
underlying them.
In contrast, little attention has been paid to the second class
of genes: those that render the brain more or less susceptible
to injury in response to cerebrovascular disease. Evidence for
the existence of such response genes is that patients with
apparently similar loads of vascular pathology (with regard
to lesion type, number, and location) may range from no
cognitive impairment to severely cognitively impaired.31–35
Such differences could be attributable to either genetic or
environmental factors. Direct evidence for a role of genetic
factors comes from studies of the heritability of white matter
hyperintensities (WMH). Studies in older male twins,36 in a
family-based sample of middle-aged and older men and
women (the Framingham Study cohort37), and in hypertensive
sibships38 have shown that WMH volume is a highly heritable
trait, indicating a large genetic contribution to development of
this condition. Does the heritability of WMH volume simply
reflect the heritability of underlying cerebrovascular disease?
Two lines of evidence suggest not. First, in the Framingham
study group, high heritability of WMH volume was seen even
in younger patients in whom the prevalence of cerebrovas-
cular injury was relatively low. Second, in the study of
hypertensive sibships, levels of hypertension were controlled
for. Thus, these studies suggest that there are genetic factors
affecting cellular responses to cerebrovascular pathology that
are different from those that cause cerebrovascular pathology
itself.
One class of genes that must influence tissue responses to
cerebrovascular disease are the AD genes. As was first shown
in the Nun Study,39 there is an additive or synergistic
interaction between AD and cerebrovascular pathologies,
such that individuals with both of these pathologies show
greater cognitive impairment than those exhibiting either
pathology alone. This finding has been replicated.40,41 In
addition, at least 3 sets of genes in the AD pathway, the
presenilins, APP, and apolipoprotein E (apoE), are known to
interact with the VCI disease pathway. The presenilins,
mutations of which cause AD, have been shown to interact
directly with Notch proteins, including Notch 3 (mutations of
which cause CADASIL).42 Mutations in the APP gene can
lead either to AD or to hemorrhagic stroke and dementia (as
in HCHWA-D) depending on the site of the mutation and the
subsequent cellular site of amyloid accumulation.28,43,44 Vari-
ants of the apoE gene appear to affect not only susceptibility
to cerebrovascular disease but also recuperative responses to
it (see below). Thus, there appear to be links in the biochem-
ical pathways underlying VCI and AD pathologies, which
could be responsible for the observed interactive effects of
these pathologies on cognitive function.
Genes that influence brain responses to cerebrovascular
disease do not appear to be limited to those within AD
pathway. First, it has been shown that VCI can occur in the
complete absence of AD pathology in sporadic VCI and in
hereditary forms.34,45,46 In addition, the cognitive sequelae of
pathogenic processes associated with VCI are different from
those seen in “pure” AD, in that executive function appears
more strongly affected in VCI than is memory.6,47,48–51 Consis-
tent with these observations, different brain regions seem differ-
entially affected in VCI and AD, with prefrontal circuits being
more affected in VCI and the hippocampus in AD.52–56
There is direct evidence from both human and animal
studies for specific non-AD genes that play a role in tis-
sue responses in ischemia. First, studies in humans suggest
that variants in the genes for platelet glycoprotein and
�-fibrinogen affect poststroke outcomes without affecting
stroke risk per se.57–61 Furthermore, studies in animal models
have demonstrated that a number of proteins outside the AD
pathway contribute to (or protect against) tissue injury after
ischemia. These include glutamate and �-aminobutyric acid
receptors, acid-sensing ion channels, proteases, growth fac-
tors and their receptors, and transcription factors.62–64
Genes that affect an individual’s premorbid level of cog-
nitive ability also seem likely to affect performance in the
wake of cerebrovascular disease. For example, several studies
have shown now that subjects who in their youth perform
better on measures of linguistic or mental ability are less
likely to develop cognitive impairment or dementia later in
life.65–67 Baseline cognitive function in “healthy” individuals
at all ages clearly has a strong genetic component.68–70
Indeed, the heritability of certain cognitive measures (includ-
ing measures of executive function) actually increases with
increasing age, raising the possibility that genetic influences
become even more important in later life.71
With respect to specific genes that may influence cognitive
function, candidate gene studies have indicated strong asso-
ciations between measures of prefrontal function and poly-
morphisms in the catechol-o-methyl transferase gene.70 Ge-
netic studies of attention deficit hyperactivity disorder
suggest statistically significant associations of this disorder
with genes in the dopaminergic and serotonergic neurotrans-
mitter systems;72 it remains to be determined to what extent
these findings are applicable to prefrontal function in aging
“normal” individuals. Studies in genetically engineered ani-
mal models have also identified specific genes that may be
involved in human cognition,68 although much of the focus in
those studies has been on hippocampal memory formation;
executive function has been far less studied.
Finally, some of the genes underlying VCI are likely to
affect both susceptibility to cerebrovascular disease and
the response of the brain to ischemic or hemorrhagic
injury. ApoE appears to be an example of gene that affects
disease incidence and disease responses. On the disease
incidence side, apoE genotype influences risk of intracerebral
hemorrhage. In the Greater Cincinnati/Northern Kentucky
population, about one third of all cases of lobar intracerebral
hemorrhage are attributable to the possession of the e2 or e4
allele.73 E4 allele carriers also have a nearly 4-fold higher risk
of lobar warfarin-related intracerebral hemorrhage.74 On the
Leblanc et al Genetics of Vascular Cognitive Impairment 249
by on November 3, 2010 stroke.ahajournals.orgDownloaded from
injury response side, the presence of the e4 allele has been
associated with reduced survival in intracerebral hemor-
rhage.58,75 Conversely, an increasing dose of the e4 allele has
been associated with improved survival in patients after
ischemic stroke, even after adjusting for baseline severity of
neurological impairment.76
Although a useful conceptual model, differentiating inci-
dence from injury-response genes may be difficult experi-
mentally. This is especially true for genes encoding for or
affecting the expression of vascular growth factors such as
ephrins, which have overlapping effects with neural growth
factors.77 Because there are gene products that can have
vasculotrophic and neurotropic effects, human studies should
encompass not only information on the presence or absence
of stroke (whether symptomatic or not) but also on markers of
neural injury or response to injury (such as measures of
neurological deficit or cognition).
Challenge 1: Choice of Subject Population
and Study Design
Acute Large Vessel Infarct
The cerebrovascular pathologies that cause VCI are hetero-
geneous. Hence, there are several different patient popula-
tions that could be used for genetic studies. The first of these
is patients with acute ischemic stroke enrolled at the time of
admission to hospital. Because only a subpopulation of these
patients will develop a clinically diagnosable cognitive im-
pairment subsequent to their stroke, one could compare the
genotypes of patients with different cognitive outcomes. Use
of this patient population would have distinct practical
advantages: patient collection would be straightforward and
would generate a cohort enriched with individuals destined
for cognitive decline.
There are several obvious confounders that would have
to be addressed with this patient population. Symptomatic
cerebral infarction is heterogeneous with regard to lesion
volume, location, and laterality. Hence, the degree of cogni-
tive impairment seen may reflect these factors as well as the
genetic factors that render neural tissue more or less resilient
to injury. However, these factors could be controlled for with
a sufficiently large sample size of patients with large vessel
strokes in similar brain regions. A second confounding issue
is that the cognitive status of the patients before stroke would
be variable and could not be measured directly at the time of
presentation with stroke. Using the Informant Questionnaire
on Cognitive Decline in the Elderly (IQCODE), investigators
estimate the rate of prestroke dementia at 12% to 16%.78,79
The confound of premorbid cognitive impairment might be
controlled for by testing cognition at �2 time points after the
stroke and using rates of change in cognitive ability as the
phenotype to be measured.
The timing of cognitive testing after stroke would de-
pend on what phase of the postischemic tissue injury re-
sponse one wished to analyze and what classes of genes one
thereby hoped to target. After a discrete ischemic event, there
is an acute phase of tissue injury, and it is typically during
this phase that the individual presents at the hospital with
clinical symptoms. Over the first 6 to 12 months after infarct,
a large proportion (up to 50%) of patients with early presen-
tation of cognitive impairment exhibit significant recov-
ery.80,81 However, over the subsequent months and years,
progressive deterioration is seen in the overall population of
stroke patients.82–84 This long-term cognitive deterioration
may reflect additional ischemic events (which may or may
not be attended by acute symptoms), the interaction of the
acquired cognitive impairment with normal aging changes,
progressive nonvascular dementias such as AD, or some
combination.
The likely occurrence of nonvascular cognitive impairment
in a subset of individuals complicates the analysis of long-
term cognitive outcomes considerably. There are neuropsy-
chometric tools that could potentially help identify patients
with significant AD; these are discussed in a later section of
this article. Insights could also be gained through serial MRI,
amyloid imaging, or testing for AD-related genes. However,
such long-term studies ideally would be performed over a
period of at least several years and would carry considerable
per-patient costs. An alternative solution would be to target
the initial phase of poststroke recovery. It is possible that the
degree of short-term recovery would be less influenced by
concomitant AD pathology than would long-term outcome.
This approach might also uncover genes that promote tissue
repair as well as ones that render it susceptible to damage.
Such a study could be hospital based and loss to follow-up
would be minimal.
Subcortical Small Vessel Disease
It has been suggested that subcortical small vessel disease
produces a more homogeneous set of cognitive deficits than
do large vessel infarcts,85 although the deficits may be milder
and slower in onset. In this regard, subcortical small vessel
disease could be particularly well suited to genetic studies.
The major challenge with this population would be subject
recruitment. Large-scale imaging studies have shown that
most subcortical small vessel disease does not produce acute
symptoms but rather an insidious decline in neurological
function or no symptoms at all.31,35,86 In theory, one could
screen a large number of “normal” subjects for MRI signs of
small vessel disease and then compare the genotypes of
subjects who do and do not subsequently develop cognitive
impairment. However, only�15% to 25% of elderly subjects
show MRI signs of small vessel disease,86–89 and only a
subpopulation of these then go on to develop cognitive
impairment. Because it has been estimated that about a
thousand cases and a thousand controls would be needed for
a case/control study design (John Hardy and Don Bowden,
personal communications, 2004), this kind of longitudinal
study design would be extremely costly.
CADASIL and CAA
CADASIL patients are another intriguing population for
studying the genetics of VCI. There are 2 potential advan-
tages in using this population. First, the population is rela-
tively homogeneous with respect to underlying vascular
disease (small artery angiopathy causing degeneration of the
smooth muscle cell layer) and type of infarct (primarily
subcortical lacunar). Second, despite this homogeneity in the
250 Stroke January 2006
by on November 3, 2010 stroke.ahajournals.orgDownloaded from
nature of the vascular disease, there is a high degree of
variability in penetrance and age of onset of stroke and
cogni