Neurobiology of Aging 36 (2015) 2443.e1e2443.

e7

Contents lists available at ScienceDirect

Neurobiology of Aging
journal homepage: www.elsevier.com/locate/neuaging

NOTCH3 variants in patients with subcortical vascular cognitive

impairment: a comparison with typical CADASIL patients
Cindy W. Yoon a,1, Young-Eun Kim b,1, Sang Won Seo c, d, e, f, g, h, *, Chang-Seok Ki b, **,
Seong Hye Choi a, Jong-Won Kim b, Duk L. Na c
a
Department of Neurology, Inha University School of Medicine, Incheon, Korea
b
Department of Laboratory Medicine and Genetics, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea
c
Department of Neurology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea
d
Neuroscience Center, Samsung Medical Center, Seoul, Korea
e
Department of Clinical Research Design and Evaluation, SAIHST, Sungkyunkwan University, Seoul, Korea
f
Department of Neurology, Memory and Aging Center, University of California, San Francisco, CA, USA
g
Helen Wills Neuroscience Institute, University of California, Berkeley, CA, USA
h
Lawrence Berkeley National Laboratory, Berkeley, CA, USA

a r t i c l e i n f o a b s t r a c t

Article history: Although cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy
Received 6 December 2014 (CADASIL) is thought to be a common form of hereditary subcortical vascular cognitive impairment
Received in revised form 24 February 2015 (SVCI), there is little data on the frequency of NOTCH3 variants in SVCI patients. We prospectively
Accepted 17 April 2015
screened for NOTCH3 variants in consecutive SVCI patients who underwent brain magnetic resonance
Available online 25 April 2015
imaging and amyloid positron emission tomography as well as sequence analysis for mutational hotspots
in the NOTCH3 gene. Among 117 patients with SVCI, 16 patients had either known mutations or variants
Keywords:
of unknown significance in the NOTCH3 gene. There were no differences in clinical and neuroimaging
CADASIL
NOTCH3
features between SVCI patients with and without NOTCH3 variants, only except for a higher number of
Subcortical vascular cognitive impairment deep microbleeds in SVCI patients with NOTCH3 variants. Our findings suggest that there is a phenotypic
(SVCI) entity of NOTCH3 variant that is similar to that of sporadic SVCI but not of typical CADASIL. Notably,
2 SVCI patients with NOTCH3 mutations showed significant amyloid burden, which challenges the
prevailing concept that CADASIL represents the genetic model of pure small vessel disease.
Ó 2015 Elsevier Inc. All rights reserved.

1. Introduction (WMH), especially in the anterior-temporal lobe and external

Cerebral autosomal dominant arteriopathy with subcortical
infarcts and leukoencephalopathy (CADASIL) is an autosomal
dominant disorder of cerebral small vessels caused by mutations in
the NOTCH3 gene on chromosome 19 (Joutel et al., 1996). CADASIL is
characterized by young or middle age-onset, migraine with aura,
stroke episodes, and cognitive impairment (Chabriat et al., 1995;
Dichgans et al., 1998). CADASIL patients usually have a family his-
tory of stroke or dementia but no past history of vascular risk fac-
tors (Chabriat et al., 2009). Magnetic resonance imaging (MRI) of
these patients reveals diffuse white matter hyperintensities
* Corresponding author at: Department of Neurology, Samsung Medical Center,
Sungkyunkwan University School of Medicine, Irwon-dong, Gangnam-gu, Seoul,
Korea. Tel.: þ82 2 3410 1233; fax: þ82 2 3410 0052.
** Alternate corresponding author at: Department of Laboratory Medicine and
Genetics, Samsung Medical Center, Sungkyunkwan University, School of Medicine,
Irwon-dong, Gangnam-gu, Seoul, Korea. Tel.: þ82 2 3410 2709; fax: þ82 2 3410 2719.
E-mail addresses: sangwonseo@empal.com (S.W. Seo), changski@skku.edu
(C.-S. Ki).
1
The first 2 authors contributed equally to this work.
0197-4580/$ e see front matter Ó 2015 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.neurobiolaging.2015.04.009
capsule (Chabriat et al., 1998; O’sullivan et al., 2001).
Previous studies have shown that CADASIL is very rarely
observed in the general population (0.001%e0.002%) (Markus et al.,
2002; Razvi et al., 2005). However, patients with NOTCH3 muta-
tions may not exhibit every feature of CADASIL, instead only
showing part of the clinical spectrum. Moreover, de novo mutations
have also been reported (Joutel et al., 2000). When these pheno-
typic and genotypic variations are considered, it could be possible
that CADASIL patients might be clinically underdiagnosed.
Vascular cognitive impairment (VCI) is a term for all degrees of
cognitive impairment because of cerebrovascular disease, which in-
cludes vascular dementia (VaD) and vascular mild cognitive impair-
ment (Bowler and Hachinski, 1995; O’Brien et al., 2003). Subcortical
VCI (SVCI) is one of the most common subtypes of VCI, especially in
East Asia (O’Brien et al., 2003). Vascular risk factors produce cerebral
small vessel disease (CVD), manifested on MRIs as WMH, lacunes, and
microbleeds resulting in SVCI (Román et al., 2002).
Although CADASIL has been considered as a common form
of hereditary SVCI, to the best of our knowledge, there has been no
2443.e2 C.W. Yoon et al. / Neurobiology of Aging 36 (2015) 2443.e1e2443.e7
Fig. 1. Subject selection and comparison groups *From a previous study (Kim et al., 2014). Abbreviations: CADASIL, cerebral autosomal dominant arteriopathy with subcortical
infarcts and leukoencephalopathy; SVCI, subcortical vascular cognitive impairment.
study systematically investigating the frequency of NOTCH3 vari-
ants in SVCI patients. In this study, we prospectively screened for
NOTCH3 variants in SVCI patients and investigated genotypic
and phenotypic characteristics of SVCI patients with NOTCH3 vari-
ants. In addition, we compared the amyloid and CVD burdens
between SVCI patients with and without NOTCH3 variants.
2. Methods
2.1. Participants
We prospectively recruited 137 Korean patients with SVCI who
underwent Pittsburgh compound-B (PiB) positron emission
tomography (PET) and structural brain MRI between September
2008 and August 2011. This included 70 patients with subcortical
VaD (SVaD) and 67 patients with subcortical vascular mild cognitive
impairment (svMCI). Patients with SVaD met the diagnostic criteria
for VaD as determined by the Diagnostic and Statistical Manual of
Mental Disorders Fourth Edition as well as the imaging criteria for
SVaD proposed by Erkinjuntti et al. (Inzitari et al., 2000). All svMCI
patients met Petersen’s criteria for MCI with modifications as pre-
viously described (Seo et al., 2009). All SVCI patients had evidence
of signficant ischemia on their MRI scans, defined as a cap or band
10 mm and a deep white matter lesion 25 mm, as modified from
the Fazekas ischemia criteria (Fazekas et al., 1987; Lee et al., 2011).
Among 137 SVCI patients, no patients met probable CADASIL
criteria proposed by Davous et al. (Davous, 1998). We excluded
patients with territorial infarctions and those with high signal
abnormalities on MRI due to radiation injury, multiple sclerosis,
vasculitis, or leukodystrophy. All patients completed a clinical
interview and neurological examination, as described previously
(Lee et al., 2011).
Among the 137 SVCI patients, 20 patients were excluded
because they or their caregivers chose not to participate in the
study. The results comparing included and excluded patients are
shown in Supplementary Table 1. There was a difference in Mini-
Mental State Examination scores between these 2 groups. Finally,
117 SVCI patients, consisting of 54 SVaD patients and 63 svMCI
patients, were included.
For comparison with SVCI patients with NOTCH3 variants,
24 typical CADASIL patients who were clinically diagnosed and then
genetically confirmed in our previous study were included in this
study (Kim et al., 2014) (Fig. 1). In the previous study, we used the
criteria for probable CADASIL (Davous, 1998) with some modifica-
tions. The inclusion criteria were marked leukoaraiosis on brain
MRI and at least one of the following: young age (<50) at onset of
stroke or cognitive decline, or a family history of stroke or dementia
suggesting autosomal dominant inheritance.
The Institutional Review Board of Samsung Medical Center
approved the study protocol and written consent was obtained
from each patient.
2.2. Molecular genetic analysis
Peripheral blood specimens were collected after obtaining
informed consent. Genomic DNA was extracted using the Wizard
Genomic DNA purification kit according to the manufacturer’s
instructions (Promega, Madison, WI, USA). Mutational hotspots of
the NOTCH3 gene including exons 2e6, 8, 11, 18, 19, and 22 were
sequenced. All tested exons and their exon-intron boundaries in
the NOTCH3 gene were amplified by polymerase chain reaction, as
described previously (Kim et al., 2014). Cycle sequencing was
performed with a BigDye Terminator Cycle Sequencing Ready
Reaction kit (Applied Biosystems, Foster City, CA, USA) on an ABI
3130xl Genetic Analyzer (Applied Biosystems). The nucleotides of
NOTCH3 complementary DNA were numbered according to a
reference sequence, GenBank accession number NM_000435.2.
Sorting Intolerant From Tolerant (Ng and Henikoff, 2001) and
Polymorphism Phenotyping (PolyPhen-2 v2.1) (Ramensky et al.,
2002) servers were used to predict the effect of a non-
synonymous variants of unknown significance (VUS) on protein
structure, function, phenotype, sequence conservation and/or
protein structure. In addition, 358 age- and sex-matched healthy
Korean controls were screened for novel VUS in the NOTCH3 gene
 C.W. Yoon et al. / Neurobiology of Aging 36 (2015) 2443.e1e2443.e7 2443.e3

using matrix-assisted laser desorption/ionization time-of-flight Table 1

mass spectrometry, as described previously (Supplementary Known mutations of the NOTCH3 gene identified in 10 SVCI and 20 typical CADASIL
patients
Table 2) (Song et al., 2012).
Exon Nucleotide change Amino acid No. of No. of mutations
change index cases per exon (%)
2.3. MRI techniques
SVCI patients with NOTCH3 mutation
3 c.224G>C p.R75P 1 1 (10)
T2, T1, 3-dimensional fluid-attenuated inversion recovery 5 c.709G>A p.V237M 2 2 (20)
(FLAIR), and T2 Fast Field Echo-MR images were acquired using the 11 c.1630C>T p.R544C 5 7 (70)
same 3.0T MRI scanner (Philips 3.0T Achieva). c.1759C>T p.R587C 2
Typical CADASIL patientsa
2 c.160C>T p.R54C 2 2 (10)
2.4. Measurement of WMH volume 3 c.224G>C p.R75P 6 7 (35)
c.268C>T p.R90C 1
We quantified WMH volume (in mL) on FLAIR images using an 4 c.397C>T p.R133C 1 1 (5)
5 c.698G>A p.C233Y 1 1 (5)
automated method as previously described (Seo et al., 2009).
6 c.994C>T p.R332C 2 2 (10)
11 c.1630C>T p.R544C 6 7 (35)
2.5. Assessment of the presence of hyperintense lesions in the c.1732C>T p.R578C 1
anterior-temporal lobe white matter and the external capsule Key: CADASIL, cerebral autosomal dominant arteriopathy with subcortical infarcts
and leukoencephalopathy; SVCI, subcortical vascular cognitive impairment.
a
The presence of WMH in the anterior-temporal lobe white From a previous study (Kim et al., 2014).

matter and the external capsule were defined as a Fazekas score of 2

(11e25 mm) or 3 (>25 mm) (Fazekas et al., 1987). Two experienced
neurologists, who were blinded to patient data, independently
evaluated WMH in the anterior-temporal lobe and the external
capsule on the T2-weighted and FLAIR images. The kappa values for
the agreement between the 2 neurologists regarding the presence
of WMH in the anterior-temporal lobe and the external capsule
were 0.97 and 0.99, respectively. Consensus was reached in any case
of discrepancy.
2.6. Assessment of lacunes and microbleeds on MRI
Lacunes were defined as small lesions (15 mm and 3 mm in
diameter) with low signal on T1-weighted images, high signal on
T2-weighted images, and a perilesional halo on 80 axial slices from
FLAIR images. The kappa value between the 2 neurologists for the
presence of lacunas was 0.78, and a consensus was reached in any
case of discrepancy.
Microbleeds were defined as 10 mm in diameter using criteria
proposed by Greenberg et al. (2009) on 20 axial slices of T2*
gradient echo-MR images. Two experienced neurologists, who
were blinded to other patients’ data, reviewed the number and
location of the lacunes and microbleeds. Microbleeds were counted
in 4 lobar regions (frontal, temporal, parietal, and occipital) and
deep brain regions. The lobar regions were defined as 10 mm from
the brain surface. The kappa value for the agreement between the 2
neurologists for the presence of microbleeds was 0.92, and a
consensus was reached in any case of discrepancy.
2.7. [11C] PiB-PET
All patients completed a standardized [11C] PiB-PET scan at
Samsung or Asan Medical Center using a Discovery STE PET/CT
scanner (GE Medical Systems, Milwaukee, WI, USA) to minimize
any variance due to scanner differences. The detailed radio-
chemistry profiles, scanning protocol, and data analysis method
were described in a previous study (Lee et al., 2011; Park et al.,
2013). Briefly, we calculated the PiB uptake ratio of each voxel
using the cerebellum as a reference region in the analysis. The
global cortical PiB uptake ratio was determined by combining the
bilateral frontal, parietal, and temporal cortices, and posterior
cingulate gyrus. Patients were considered PiB-positive if their
global PiB uptake ratio was more than 2 standard deviations
(PiB retention ratio 1.5) from the mean of the normal controls
(Lee et al., 2011).
2.8. Neuropsychological testing
All patients underwent neuropsychological testing using the Seoul
Neuropsychological Screening Battery (Kang, 1998; Yoon et al., 2013).
2.9. Statistical analysis
To analyze the differences between SVCI patients with and
without NOTCH3 variants, we performed the c2 test or the Fisher’s
exact test for categorical variables, and the Mann-Whitney U-test or
Student t test for continuous variables.
3. Results
3.1. Frequency of NOTCH3 variants in SVCI patients
Among 117 patients with SVCI, 16 (13.7%) patients had either
known mutations or VUS. Four known mutations were found in
10 patients, including p.R544C (n ¼ 5), p.V237M (n ¼ 2), p.R587C
(n ¼ 2), and p.R75P (n ¼ 1) (Table 1). In addition, 7 VUS were
identified in 6 patients: 5 missense variants (p.P167S, p.C542R,
p.P572L, p.S947I, and p.R1175W), 1 nonsense variant (p.S567*), and
1 frameshift variant (p.Glu990Argfs*282). Two VUS (p.P572L and
p.R1175W) were found in 1 patient. A control study of 716 chro-
mosomes identified 2 variants: p.P167S (n ¼ 3, 0.42%) and
p.R1175W (n ¼ 8, 1.12%) (Table 2). These 2 variants might be poly-
morphisms rather than pathogenic mutations. All VUS were
deemed to affect protein structure according to in silico analysis
with Polymorphism Phenotyping or Sorting Intolerant From
Tolerant, with the exception of p.P167S, which was predicted to be a
benign amino acid change. One SVCI patient with the p.P167S
variant that was also identified in the control study was excluded
from comparative analysis. However, 1 SVCI patient with the
p.R1175W variant was not excluded because the patient also had
another variant, p.P572L, which was not found in the control study.
Detailed characteristics of SVCI patients with NOTCH3 variants are
shown in Supplementary Table 3.
3.2. Comparison of clinical features between SVCI patients with and
without NOTCH3 variants
No significant differences in demographic and clinical charac-
teristics were seen between SVCI patients with and without
2443.e4 C.W. Yoon et al. / Neurobiology of Aging 36 (2015) 2443.e1e2443.e7

Table 2
Variants of unknown significance (VUS) of the NOTCH3 gene identified in 6 SVCI and 4 typical CADASIL patients

Exon Nucleotide change Amino acid No. of No. of alleles In silico analysis rs number
change index cases in control
PolyPhen-2 SIFT
chromosomes (%)
(probabilistic score) (tolerance index)
SVCI patients with NOTCH3 variants
4 c.499C>Ta p.P167S 1 3/716 (0.42) Benign (0.00) Tolerable (0.39) rs202157633
11 c.1624T>C p.C542R 1 0/716 (0) Probably damaging (1.00) Not tolerable (0.02) NA
c.1700C>G p.S567* 1 0/716 (0) NA NA NA
c.1715C>Tb p.P572L 1 0/716 (0) Probably damaging (0.99) Tolerable (0.12) NA
18 c.2840G>T p.S947I 1 0/716 (0) Probably damaging (1.00) Tolerable (0.11) NA
c.2968delG p.E990Rfs*282 1 0/716 (0) NA NA NA
a,b
22 c.3523C>T p.R1175W 1 8/716 (1.12) Probably damaging (0.994) Not tolerable (0.02) rs200504060
Typical CADASIL patientsc
7 c.1078T>C p.C360R 1 0/716 (0) Probably damaging (1.00) Not tolerable (0.00) NA
8 c.1241C>G p.S414C 2 0/716 (0) Probably damaging (1.00) Not tolerable (0.00) NA
11 c.1624T>C p.C542R 1 0/716 (0) Probably damaging (1.00) Not tolerable (0.02) NA

Key: CADASIL, cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy; NA, not applicable; PolyPhen, polymorphism phenotyping; SIFT,
sorting intolerant from tolerant; SVCI, subcortical vascular cognitive impairment.
a
These variants were excluded from comparative analysis.
b
From a previous study (Kim et al., 2014).
c
These 2 VUS were identified in 1 patient.

NOTCH3 variants, except for a higher prevalence of hyperlipidemia
in SVCI patients with NOTCH3 variants (Table 3).
3.3. Comparison of clinical features between SVCI patients with
NOTCH3 variants and typical CADASIL patients
Compared with typical CADASIL patients from our previous
study (Kim et al., 2014), SVCI patients with NOTCH3 variants were of
more advanced age (71.7 years vs. 53.2 years) and had a higher
prevalence of hypertension (86.7% vs. 29.2%) and hyperlipidemia
(60.0% vs. 25.0%), whereas they had less frequent family history of
stroke (26.7% vs. 66.7%) and anterior-temporal involvement
(6.7% vs. 54.2%) (Table 3).
deep microbleeds in SVCI patients with NOTCH3 variants
(5 [interquartile range 1e11] vs. 1 [interquartile range 0e3],
p ¼ 0.034) (Table 4). There were no differences in the frequency of
the anterior-temporal and external capsule involvement, total and
regional volumes of WMH, and number of lacunes and lobar
microbleeds.
Thirty-two (31.7%) of the 101 patients without NOTCH3 variants
tested positive for PiB retention, whereas 2 (13.3%) of the
15 patients with NOTCH3 variants tested positive. There were no
significant differences in the PiB-positive rate or PiB retention ratio
between the 2 groups (Table 4). The 2 instances of PiB retention in
PiB-positive SVCI patients with NOTCH3 variants were distributed
to the frontal, temporal, and parietal regions (Fig. 2).

3.4. Comparison of neuroimaging features between SVCI patients 4. Discussion

with and without NOTCH3 variants
In our consecutive SVCI patients, NOTCH3 variants are not
No significant differences were seen in CVD markers between the uncommon. Specifically, among 117 SVCI patients, there were
2 groups, with the only exception being a higher median number of 10 (8.5%) patients with known mutations, 5 (4.3%) with probably

Table 3
Demographic and clinical characteristics of subjects

SVCI pa Typical CADASILb

Total NOTCH3 variants () NOTCH3 variants (þ)

Number 117 101 15c 24
Demographics
Age, median (IQR) 75 (69e78) 75 (69e78) 74 (68e78) 0.461 53 (48e60)
Gender, female, N (%) 72 (61.5) 62 (61.4) 9 (60.0) 0.565 16 (66.7)
Education, median (IQR) 9 (6e12) 9 (6e12) 6 (6e11) 0.474
Vascular risk factor, N (%)
Hypertension 91 (77.8) 77 (76.2) 13 (86.7) 0.296 7 (29.2)
Diabetes 29 (24.8) 26 (25.7) 3 (20.0) 0.453 4 (16.7)
Hyperlipidemia 40 (34.2) 30 (29.7) 9 (60.0) 0.024 6 (25.0)
ApoE4 carrier, N (%) 31 (26.5) 28 (27.7) 2 (13.3) 0.195
Clinical history, N (%)
Ischemic TIA/stroke 27 (23.1) 23 (22.8) 4 (26.7) 0.480 14 (58.3)
Headache 11 (9.4) 8 (7.9) 3 (20.0) 0.152 9 (37.5)
Family history, N (%)
Dementia 23 (19.7) 20 (19.8) 2 (13.3) 0.426 3 (12.5)
Ischemic TIA/stroke 38 (32.5) 33 (32.7) 4 (26.7) 0.444 16 (66.7)
White matter hyperintensities, N (%)
Anterior temporal 3 (2.6) 2 (2.0) 1 (6.7) 0.342 13 (54.2)
External capsule 99 (84.6) 83 (82.2) 15 (100) 0.066 21 (87.5)

Key: CADASIL, cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy; IQR, interquartile range; SVCI, subcortical vascular cognitive
impairment; TIA, transient ischemic attack.
a
NOTCH3 variant() versus NOTCH3 variant(þ).
b
From a previous study (Kim et al., 2014).
c
One SVCI patient with a variant that was identified in a control study was excluded from comparative analysis.
 C.W. Yoon et al. / Neurobiology of Aging 36 (2015) 2443.e1e2443.e7 2443.e5

Table 4
Comparisons of amyloid and cerebrovascular disease burdens between SVCI patients with and without NOTCH3 variants

NOTCH3 variants () (n ¼ 101) NOTCH3 variants (þ) (n ¼ 15a) p

PiB retention
PiB positivity frequency, N (%) 32 (31.7) 2 (13.3) 0.225
PiB uptake ratio, median (IQR) 1.27 (1.17e1.65) 1.22 (1.12e1.34) 0.156
Frontal 1.22 (1.11e1.66) 1.13 (1.06e1.24) 0.073
Parietal 1.33 (1.21e1.64) 1.29 (1.23e1.38) 0.325
Temporal 1.46 (1.33e1.99) 1.34 (1.30e1.50) 0.231
Cingulate 1.31 (1.20e1.67) 1.25 (1.18e1.38) 0.154
WMH
Volume (mm3), median (IQR)
Total 33,203.71 (24,345.22e46,822.66) 38,763.43 (21,058.46e51,935.76) 0.787
Frontal 20,680.17 (16,027.50e30,718.16) 21,230.00 (15,684.07e32,015.50) 0.734
Parietal 8571.31 (5231.47e12,487.55) 9771.03 (4847.59e15,612.90) 0.654
Temporal 3517.78 (2217.39e4824.02) 4355.16 (2462.80e5516.02) 0.818
Occipital 288.04 (108.34e502.33) 186.45 (4.84e370.03) 0.395
Involvement of AT, N (%) 2 (2.0) 1 (6.7) 0.342
Involvement of external capsule, N (%) 83 (82.2) 15 (100.0) 0.066
Lacunes
Frequency, N (%) 91 (90.1) 14 (93.3) 1.000
Total, median (IQR) 6 (2e14) 10 (7e18) 0.078
White matter, median (IQR) 3 (0e10) 6 (4e11) 0.089
Basal ganglia and thalamus, median (IQR) 2 (0e4) 3 (1e7) 0.250
Microbleeds
Frequency, N (%) 66 (65.3) 11 (73.3) 0.771
Total, median (IQR) 2 (0e6) 6 (1e15) 0.141
Lobar, median (IQR) 0 (0e3) 1 (0e3) 0.705
Deep, median (IQR) 1 (0e3) 5 (1e11) 0.034

Key: AT, anterior temporal; IQR, interquartile range; PiB, Pittsburgh compound-B; SVCI, subcortical vascular cognitive impairment; WMH, white matter hyperintensities.
a
One SVCI patient with a variant that was identified in a control study was excluded from comparative analysis.

deleterious variants, and 1 (0.9%) with a probably benign variant.
Previously, sequencing of the NOTCH3 gene in 218 patients with
lacunar stroke showed a mutation in only 1 patient (Dong et al.,
2003). Another study based on patients with all types of acute
ischemic stroke showed that the frequency of NOTCH3 mutation
was 4% (Choi et al., 2013). A previous study investigating the
frequency of NOTCH3 variants in elderly individuals with coalescent
white matter lesions showed that the NOTCH3 gene is highly vari-
able in age-related CVD; however, deleterious variants were
observed in only 3.1% of subjects (Schmidt et al., 2011). On the
contrary, ˃10% of SVCI patients had either known mutations or VUS
in the present study, which is a higher frequency than in other
previous settings such as consecutive ischemic stroke patients or
Fig. 2. PiB-PET images of 2 SVCI patients with NOTCH3 variants. Abbreviations: PiB,
Pittsburgh compound-B; PET, positron emission tomography; SVCI, subcortical
vascular cognitive impairment.
subjects with incidental white matter changes on MRI. It is unclear
why the frequency of NOTCH3 variants in this study is higher than
in other studies. However, the discrepancy may be explained by our
study population, which only consisted of cognitively impaired
patients with severe WMH.
We found 2 probably deleterious NOTCH3 variants, including a
nonsense (c.1700C > G; p.S567*) and a frameshift (c.2968delG;
p.E990Rfs*282) variant. These variants are predicted to produce
truncated proteins lacking C-terminal parts of the Notch3 protein,
which may be crucial for protein function and interactions
(Chabriat et al., 2009). Aberrant transcripts might fail to encode
truncated proteins due to nonsense-mediated decay (Chang et al.,
2007). However, it remains unclear whether a loss of function
variant in the NOTCH3 gene would be a neutral variant or a caus-
ative mutation. A few studies have reported deletions or insertions
in the cysteine residue of the NOTCH3 gene and have suggested that
these variants are underlying causes of CADASIL (Bianchi et al.,
2013; Dichgans et al., 2000; Dotti et al., 2004; Wang et al., 2011).
In contrast, Rutten et al. reported that a nonsense variant (p.R103*)
and a large intragenic heterozygous frameshift deletion of exons
3e16, did not lead to a CADASIL phenotype (Rutten et al., 2013).
Furthermore, evaluation is needed to address the effect of loss of
function variants in the NOTCH3 gene. We could not confirm these
variants using skin biopsy. Skin biopsy might be helpful in pre-
dicting the pathogenicity of these variants.
Another finding in this study was that no differences in clinical
characteristics were observed between SVCI patients with and
without NOTCH3 variants. On the contrary, compared with typical
CADASIL patients, SVCI patients with NOTCH3 variants were of more
advanced age and more frequently had a history of hypertension,
but were less likely to have a family history of stroke, and anterior-
temporal involvement. There was a case report showing that a
NOTCH3 mutation was observed in a patient presenting with spo-
radic VaD without CADASIL-specific clinical features (Pradotto et al.,
2008), which is consistent with our findings. A previous study
performing whole exome sequencing in a Turkish patient clinically
2443.e6 C.W. Yoon et al. / Neurobiology of Aging 36 (2015) 2443.e1e2443.e7
diagnosed with Alzheimer’s disease identified a mutation in
NOTCH3 (p.R1231C), previously described as causing CADASIL
(Guerreiro et al., 2012). Along with previous reports, our findings
suggest that some patients with NOTCH3 variants are clinically
underdiagnosed because they lack typical characteristics of
CADASIL.
Notably, 2 SVCI patients with NOTCH3 variants showed signifi-
cant amyloid burden, which challenges the prevailing concept that
CADASIL represents the genetic model of pure small vessel disease
(Chabriat et al., 2009; Dichgans et al., 2008). Previous studies have
found that 20%e30% of cognitively normal elderly individuals have
amyloid pathology (Buchman and Bennett, 2011; Sojkova et al.,
2011). Considering that the mean age of SVCI patients with
NOTCH3 variants is ˃70 years in our study, it is possible that SVCI
patients with NOTCH3 variants may incidentally have amyloid
burden. Furthermore, 1 patient was an APOE4 carrier. However, to
the best of our knowledge, this is the first study directly showing
that some SVCI patients with NOTCH3 variants may also have
combined Alzheimer’s disease pathology.
There were no significant differences in CVD burden, possibly
because the effects of old age and hypertension on ischemia over-
ride the effects of NOTCH3 variants in SVCI patients. However, we
found that SVCI patients with NOTCH3 variants had a marginally
higher number of deep microbleeds than those without NOTCH3
variants. It could be possible that NOTCH3 variants may aggravate
leakage from vessels in SVCI patients. However, further studies will
be needed to validate this finding.
This study is the first to systematically screen for NOTCH3 vari-
ants in SVCI patients. However, several limitations in our study
must be acknowledged. First, there was a relatively small sample of
SVCI patients with NOTCH3 variants in our study; therefore, the
statistical power might be relatively low. Second, we did not
confirm 5 SVCI patients with VUS using skin biopsy, although these
VUS were not found in 716 control chromosomes. Finally, excluded
SVCI patients had a lower average Mini-Mental State Examination
score than included patients, which could be a source of bias.
Disclosure statement
The authors have no actual or potential conflicts of interest.
Acknowledgements
This study was supported by the Basic Research Program
through the National Research Foundation of Korea (NRF) funded
by the Ministry of Education (NRF-2013R1A1A2065365 and NRF-
2013R1A1A2009756), the Korean Healthcare Technology R and D
Project, the Ministry for Health and Welfare Affairs (HI10C2020 and
HI12C0713), a Samsung Medical Center Clinical Research Develop-
ment Program Grant (CRL-10801 and CRS110-14-1), an Inha Uni-
versity Research Grant (INHA-49307), the Inha University Hospital
Research Grant, and the Converging Research Center Program
through the Ministry of Science, ICT, and Future Planning, Korea
(2013K000338).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.neurobiolaging.
2015.04.009.
References
Bianchi, S., Dotti, M.T., Gallus, G.N., D’Eramo, C., Di Donato, I., Bernardi, L., Maletta, R.,
Puccio, G., Bruni, A.C., Federico, A., 2013. The first deep intronic mutation in the
NOTCH3 gene in a family with late-onset CADASIL. Neurobiol. Aging 34,
2234.e9e2234.e12.
Bowler, J., Hachinski, V., 1995. Vascular cognitive impairment: a new approach to
vascular dementia. Bailliere’s Clin. Neurol. 4, 357e376.
Buchman, A.S., Bennett, D.A., 2011. Amyloid pathology in persons with “normal”
cognition. Neurology 78, 228e229.
Chabriat, H., Joutel, A., Dichgans, M., Tournier-Lasserve, E., Bousser, M.-G., 2009.
Cadasil. Lancet Neurol. 8, 643e653.
Chabriat, H., Levy, C., Taillia, H., Iba-Zizen, M.T., Vahedi, K., Joutel, A., Tournier-
Lasserve, E., Bousser, M.G., 1998. Patterns of MRI lesions in CADASIL. Neurology
51, 452e457.
Chabriat, H., Vahedi, K., Bousser, M., Iba-Zizen, M., Joutel, A., Nibbio, A., Nagy, T.,
Tournier Lasserve, E., Krebs, M., Julien, J., 1995. Clinical spectrum of CADASIL: a
study of 7 families. Lancet 346, 934e939.
Chang, Y.F., Imam, J.S., Wilkinson, M.F., 2007. The nonsense-mediated decay RNA
surveillance pathway. Annu. Rev. Biochem. 76, 51e74.
Choi, J.C., Lee, K.H., Song, S.K., Lee, J.S., Kang, S.Y., Kang, J.H., 2013. Screening for
NOTCH3 gene mutations among 151 consecutive Korean patients with acute
ischemic stroke. J. stroke Cerebrovasc. Dis. 22, 608e614.
Davous, P., 1998. CADASIL: a review with proposed diagnostic criteria. Eur. J. Neurol.
5, 219e233.
Dichgans, M., Ludwig, H., Muller-Hocker, J., Messerschmidt, A., Gasser, T., 2000.
Small in-frame deletions and missense mutations in CADASIL: 3D models
predict misfolding of Notch3 EGF-like repeat domains. Eur. J. Hum. Genet. 8,
280e285.
Dichgans, M., Markus, H.S., Salloway, S., Verkkoniemi, A., Moline, M., Wang, Q.,
Posner, H., Chabriat, H.S., 2008. Donepezil in patients with subcortical vascular
cognitive impairment: a randomised double-blind trial in CADASIL. Lancet
Neurol. 7, 310e318.
Dichgans, M., Mayer, M., Uttner, I., Brüning, R., Müller-Höcker, J., Rungger, G.,
Ebke, M., Klockgether, T., Gasser, T., 1998. The phenotypic spectrum of CADASIL:
clinical findings in 102 cases. Ann. Neurol. 44, 731e739.
Dong, Y., Hassan, A., Zhang, Z., Huber, D., Dalageorgou, C., Markus, H.S., 2003. Yield
of screening for CADASIL mutations in lacunar stroke and leukoaraiosis. Stroke
34, 203e206.
Dotti, M.T., De Stefano, N., Bianchi, S., Malandrini, A., Battisti, C., Cardaioli, E.,
Federico, A., 2004. A novel NOTCH3 frameshift deletion and
mitochondrial abnormalities in a patient with CADASIL. Arch. Neurol. 61,
942e945.
Fazekas, F., Chawluk, J.B., Alavi, A., Hurtig, H., Zimmerman, R., 1987. MR signal ab-
normalities at 1.5 T in Alzheimer’s dementia and normal aging. Am. J. Roent-
genology 149, 351e356.
Greenberg, S.M., Vernooij, M.W., Cordonnier, C., Viswanathan, A., Al-Shahi
Salman, R., Warach, S., Launer, L.J., Van Buchem, M.A., Breteler, M.M., 2009.
Cerebral microbleeds: a guide to detection and interpretation. Lancet Neurol. 8,
165e174.
Guerreiro, R.J., Lohmann, E., Kinsella, E., Brás, J.M., Luu, N., Gurunlian, N., Dursun, B.,
Bilgic, B., Santana, I., Hanagasi, H., 2012. Exome sequencing reveals an unex-
pected genetic cause of disease: NOTCH3 mutation in a Turkish family with
Alzheimer’s disease. Neurobiol. Aging 33, 1008.e1017e1008.e1023.
Inzitari, D., Pantoni, L., Wallin, A., Scheltens, P., Rockwood, K., Roman, G.,
Chui, H., Desmond, D., 2000. Research criteria for subcortical vascular de-
mentia in clinical trials. In: Advances in Dementia Research. Springer,
Verlag/Wien, pp. 23e30.
Joutel, A., Corpechot, C., Ducros, A., Vahedi, K., Chabriat, H., Mouton, P.,
Alamowitch, S., Domenga, V., Cécillion, M., Maréchal, E., 1996. Notch3 mutations
in CADASIL, a hereditary adult-onset condition causing stroke and dementia.
Nature 383, 707e710.
Joutel, A., Dodick, D.D., Parisi, J.E., Cecillon, M., Tournier-Lasserve, E., Bousser, M.G.,
2000. De novo mutation in the Notch3 gene causing CADASIL. Ann. Neurol. 47,
388e391.
Kang, Y.W., 1998. Samsung Neuropsychological Screening Battery. Current Research
in Dementia. The Korean Dementia Association, Seoul, pp. 99e107.
Kim, Y.E., Yoon, C.W., Seo, S.W., Ki, C.S., Kim, Y.B., Kim, J.W., Bang, O.Y., Lee, K.H.,
Kim, G.M., Chung, C.S., 2014. Spectrum of NOTCH3 mutations in Korean patients
with clinically suspicious cerebral autosomal dominant arteriopathy with
subcortical infarcts and leukoencephalopathy. Neurobiol. Aging 35,
726.e721e726.e726.
Lee, J., Kim, S., Kim, G., Seo, S., Park, H., Oh, S., Kim, J., Cheong, H., Na, D., 2011.
Identification of pure subcortical vascular dementia using 11C-Pittsburgh
compound B. Neurology 77, 18e25.
Markus, H.S., Martin, R.J., Simpson, M.A., Dong, Y.B., Ali, N., Crosby, A.H., Powell, J.F.,
2002. Diagnostic strategies in CADASIL. Neurology 59, 1134e1138.
Ng, P.C., Henikoff, S., 2001. Predicting deleterious amino acid substitutions. Genome
Res. 11, 863e874.
O’Brien, J.T., Erkinjuntti, T., Reisberg, B., Roman, G., Sawada, T., Pantoni, L.,
Bowler, J.V., Ballard, C., DeCarli, C., Gorelick, P.B., 2003. Vascular cognitive
impairment. Lancet Neurol. 2, 89e98.
O’sullivan, M., Jarosz, J., Martin, R., Deasy, N., Powell, J., Markus, H., 2001. MRI
hyperintensities of the temporal lobe and external capsule in patients with
CADASIL. Neurology 56, 628e634.
Park, J.H., Seo, S.W., Kim, C., Kim, G.H., Noh, H.J., Kim, S.T., Kwak, K.C., Yoon, U.,
Lee, J.M., Lee, J.W., 2013. Pathogenesis of cerebral microbleeds: In vivo imaging
of amyloid and subcortical ischemic small vessel disease in 226 individuals with
cognitive impairment. Ann. Neurol. 73, 584e593.
 C.W. Yoon et al. / Neurobiology of Aging 36 (2015) 2443.e1e2443.e7
Pradotto, L., Azan, G., Doriguzzi, C., Valentini, C., Mauro, A., 2008. Sporadic vascular
dementia as clinical presentation of a new missense mutation within exon 7 of
NOTCH3 gene. J. Neurol. Sci. 271, 207e210.
Ramensky, V., Bork, P., Sunyaev, S., 2002. Human non-synonymous SNPs: server and
survey. Nucleic Acids Res. 30, 3894e3900.
Razvi, S., Davidson, R., Bone, I., Muir, K., 2005. The prevalence of cerebral autosomal
dominant arteriopathy with subcortical infarcts and leucoencephalopathy
(CADASIL) in the west of Scotland. J. Neurol. Neurosurg. Psychiatry 76, 739e741.
Román, G.C., Erkinjuntti, T., Wallin, A., Pantoni, L., Chui, H.C., 2002. Subcortical
ischaemic vascular dementia. Lancet Neurol. 1, 426e436.
Rutten, J.W., Boon, E.M., Liem, M.K., Dauwerse, J.G., Pont, M.J., Vollebregt, E., Maat-
Kievit, A.J., Ginjaar, H.B., Lakeman, P., van Duinen, S.G., Terwindt, G.M., Lesnik
Oberstein, S.A., 2013. Hypomorphic NOTCH3 alleles do not cause CADASIL in
humans. Hum. Mutat. 34, 1486e1489.
Schmidt, H., Zeginigg, M., Wiltgen, M., Freudenberger, P., Petrovic, K., Cavalieri, M.,
Gider, P., Enzinger, C., Fornage, M., Debette, S., Rotter, J.I., Ikram, M.A., Launer, L.J.,
Schmidt, R., 2011. Genetic variants of the NOTCH3 gene in the elderly and
2443.e7
magnetic resonance imaging correlates of age-related cerebral small vessel
disease. Brain 134, 3384e3397.
Seo, S.W., Cho, S.S., Park, A., Chin, J., Na, D.L., 2009. Subcortical vascular versus
amnestic mild cognitive impairment: comparison of cerebral glucose meta-
bolism. J. Neuroimaging 19, 213e219.
Sojkova, J., Zhou, Y., An, Y., Kraut, M.A., Ferrucci, L., Wong, D.F., Resnick, S.M., 2011.
Longitudinal patterns of beta-amyloid deposition in nondemented older adults.
Arch. Neurol. 68, 644e649.
Song, M.J., Lee, S.T., Lee, M.K., Ji, Y., Kim, J.W., Ki, C.S., 2012. Estimation of carrier
frequencies of six autosomal-recessive Mendelian disorders in the Korean
population. J. Hum. Genet. 57, 139e144.
Wang, Z., Yuan, Y., Zhang, W., Lv, H., Hong, D., Chen, B., Liu, Y., Luan, X., Xie, S., Wu, S.,
2011. NOTCH3 mutations and clinical features in 33 mainland Chinese families
with CADASIL. J. Neurol. Neurosurg. Psychiatry 82, 534e539.
Yoon, C.W., Shin, J.S., Kim, H.J., Cho, H., Noh, Y., Kim, G.H., Chin, J.H., Oh, S.J., Kim, J.S.,
Choe, Y.S., 2013. Cognitive deficits of pure subcortical vascular dementia vs
Alzheimer disease PiB-PETebased study. Neurology 80, 569e573.