Article Text

Download PDFPDF

NOTCH3 mutations and clinical features in 33 mainland Chinese families with CADASIL
  1. Zhaoxia Wang1,
  2. Yun Yuan1,
  3. Wei Zhang1,
  4. He Lv1,
  5. Daojun Hong1,
  6. Bin Chen1,
  7. Yang Liu1,
  8. Xinghua Luan1,
  9. Sheng Xie2,
  10. Shiwen Wu3
  1. 1Department of Neurology, Peking University First Hospital, Beijing, PR China
  2. 2Department of Radiology, Peking University First Hospital, Beijing, PR China
  3. 3Department of Neurology, The General Hospital of the Chinese Armed Police Force, Beijing, PR China
  1. Correspondence to Professor Yun Yuan, Department of Neurology, Peking University First Hospital, Beijing, 100034, PR China; yuanyun2002{at}


Background Cerebral autosomal dominant arteriopathy with subcortical infarcts and leucoencephalopathy (CADASIL) is an inherited small-vessel disease caused by mutations in NOTCH3. Although CADASIL cases have been identified worldwide, the data from mainland China are still limited.

Objective To identify NOTCH3 mutations and analyse the clinical and MRI findings in a large series of CADASIL patients from mainland China.

Methods Direct sequencing of NOTCH3 and/or skin or sural nerve biopsies were performed on 48 unrelated suspected CADASIL cases of Chinese descent. The clinical manifestations and MRI features were retrospectively collected and analysed in the 33 index patients with confirmed diagnosis and their available affected family members.

Results 20 different NOTCH3 mutations were identified in 33 families, including seven novel mutations. The highest mutation frequency was in exons 4 (55%) and 3 (30%). Granular osmiophilic material in smooth muscle cells was found in 30 cases who were biopsied. Clinical presentation included TIA/stroke in 82%, cognitive decline in 60%, and migraine with aura in only 5% of 57 symptomatic cases. MRI detected multiple lacunar infarcts and leucoaraiosis in all symptomatic cases, brainstem lesions in 64% of symptomatic cases and white-matter lesions in the temporal pole in 46% of affected members.

Conclusions The mutational spectrum and primary clinical features of patients with CADASIL from mainland China are similar to those in Caucasians. However, migraine with aura and abnormal white matter in the temporal pole are less common than among Caucasians, while brainstem involvement is more common than among Caucasians.

  • NOTCH3, mutation
  • migraine
  • brainstem
  • clinical neurology, genetics

Statistics from


Cerebral autosomal dominant arteriopathy with subcortical infarcts and leucoencephalopathy (CADASIL) is a cerebral vascular disorder with an autosomal dominant inheritance pattern. The causative gene for the disease is NOTCH3, which maps to chromosome 19p13.1 The age of onset of this disease usually ranges from the third to the sixth decade of life. Clinically, recurrent ischaemic stroke, migraine with aura, psychiatric symptoms, cognitive decline and dementia are frequently described in CADASIL patients.2 3 MRI reveals leucoaraiosis with multiple lacunar infarcts in the deep white matter, predominantly in the frontal and temporal regions extending to the temporal poles.4–6 The diagnosis of CADASIL depends on the accumulation of granular osmiophilic material (GOM) in degenerating vascular smooth muscle cells (VSMCs) as well as a mutation in the NOTCH3 gene.5

Although GOM has been considered specifically diagnostic for CADASIL, reports on the sensitivity of detecting GOM in patients' skin biopsies have been contradictory, ranging from 44%5 to nearly 100%.7 Diagnosis of CADASIL is therefore confirmed by identification of mutations in the NOTCH3 gene.

NOTCH3 is composed of 33 exons and encodes a 300 kDa transmembrane protein with a receptor and cell signal transduction function.8 Most mutations result in a gain or loss of a cysteine residue in one of the 34 epidermal growth factor-like repeats in the extracellular domain of the NOTCH3 protein. To date, over 170 NOTCH3 mutations have been reported, from exons 2 to 26.7 Because of its broad mutational spectrum, complete NOTCH3 gene mutation screening is costly and time-consuming. Fortunately, mutations tend to cluster in geographically localised populations,9 facilitating the quick diagnosis of suspected cases.

CADASIL was initially thought to be a rare disorder, but increasing numbers of families have been identified worldwide, including in Europe, America, Australia, South Africa and Asia.10–15 Most studies indicate that the mutation hotspots are located in exons 3 and 4 of the NOTCH3 gene, although some studies in geographically localised populations have revealed other distributions.16 Lee et al reported that R544C in exon 11 was a mutational hotspot in Taiwanese patients.17 Until now, there have been limited data published on mainland Chinese CADASIL patients. Here, we report the mutation pattern of the NOTCH3 gene in 33 unrelated Chinese families with CADASIL from mainland China, as well as a retrospective analysis of their clinical features and MRI findings.

Materials and methods


From March 2003 to May 2010, 48 unrelated index cases of Chinese descent (45 Han, two Mongol and one Man Chinese) from 12 provinces in mainland China, were referred to the department of Neurology, Peking University First Hospital, for confirmation of a diagnosis of CADASIL. All the suspected cases had a cranial MRI scan that showed marked leucoaraiosis and at least one of the following: young age of onset of lacunar infarction or transient ischaemic attacks (TIA), migraine with aura, a family history of ischaemic stroke or cognitive decline or psychiatric disorders. After informed consent was obtained, NOTCH3 gene mutation screening and skin/sural nerve biopsies were performed. All the index cases gave their consent for genetic analysis, and 42 gave consent for biopsy.

Mutation screening of the NOTCH3 gene

Genomic DNA was extracted from peripheral blood using standard protocols. Exons 2–24 (starting with exons 4 and 3) and their flanking sequences of the NOTCH3 gene were amplified by PCR. The 25 μl PCR reaction contained 0.2 μmol/l each primer, 100 μmol/l each dNTP, 10 mmol/l Tris-HCl (pH 8.3), 50 mmol/l KCl, 1.5 mmol/l Mg2+, 0.1% Triton X-100, 1 unit Taq DNA polymerase and 50–100 ng of genomic DNA. PCR products were directly sequenced after purification using BigDye (Applied Biosystems, Foster City, California) on an automated sequencer (ABI 3730; Applied Biosystems). Sequencing in the opposite direction was performed whenever abnormal sequence was found.

Pathological examination

Sural nerve and/or skin biopsies were performed on 42 index patients. The samples were initially fixed in 2.5% glutaraldehyde, then in 1% buffered osmium tetroxide, dehydrated in ascending grades of ethanol and embedded in Epon. Semithin sections of skin or sural nerve were cut and stained with Toluidine Blue for selecting arterioles for thin sectioning. Thin sections were double-stained with uranyl acetate and lead citrate, and then examined under an electron microscope.

Clinical assessment and neuroimaging

We retrospectively collected and analysed the clinical and cranial MRI features of all confirmed index cases following the molecular and/or pathological studies. All index patients and their available affected family members were interviewed and examined by at least two neurologists (YY, WZ, HL, YL, SW or ZW). Because we previously noted a lower frequency of migraine with aura in Chinese versus Caucasian CADASIL patients,18 19 the symptom of migraine with aura was emphasised when patient histories were taken. Brain MRIs of all index patients and some of their family members were retrospectively analysed for the location of ischaemic lesions as well as for hyperintensity in the white matter by a radiologist (SX) who was blinded to the clinical and genetic status of the patients.


Mutations in the NOTCH3 gene

Twenty different mutations were detected among the 33 index cases (table 1). Of those, seven were novel: six missense mutations (C49R and G53C in exon 2, C117R, H170C and C222S in exon 4, T1098S in exon 20) and one deletion (C76-L78del in exon 3). The other 13 mutations have been reported elsewhere: R75P, R90C, C108R and R110C in exon 3; C117F, R141C, C144Y, C146R, C162S, R169C, R182C and R207C in exon 4; and R607C in exon 11. The mutations in family 3 and family 33 do not involve a cysteine residue. In family 3, c.224G→C causes substitution of an arginine for a proline in codon 75, and in family 33, c.3292A→T causes the substitution of a threonine for a serine in codon 1098. Since cysteine-sparing mutations have rarely been reported, we screened R75P and T1098S mutation among 250 unrelated, healthy control subjects, and no one was found to be carrying these mutations. Although the R75P mutation has been found in Korean patients with CADASIL, R75P does not segregate with the disease phenotype in the paper from Kim et al.20 Therefore, it is still open to discussion as to whether these two cysteine-sparing mutations are pathogenic.

Table 1

Clinical, MRI, pathological and molecular data from 33 unrelated index patients with cerebral autosomal dominant arteriopathy with subcortical infarcts and leucoencephalopathy in mainland China

Eighteen cases (55%) had a mutation in exon 4, 10 cases (30%) had a mutation in exon 3, two cases (6%) had a mutation in exon 2, two cases (6%) had a mutation in exon 11, and one case (3%) had a mutation in exon 20. R90C and R182C were detected in four unrelated families, R169C was present in three unrelated families, and R110C, C144Y and R607C were detected in two unrelated families.

Pathological examination

GOM in the basal layer of the smooth muscle cells was found in 30 index cases who underwent a sural or skin biopsy (figure 1). The GOM appeared mainly in the arterioles, and occasionally also in the veins. The endothelial cells of the small vessels were intact. All of these 30 cases were found to carry a pathogenic NOTCH3 mutation (table 1). GOM was not detected in skin or sural biopsies from the remaining 12 patients, who were also negative for NOTCH3 mutation.

Figure 1

Pathological changes in the sural nerve of index patient 3. Electron microscopy showed granular osmiophilic material in the smooth muscle cells, with marked atrophy

Clinical and MRI features

The clinical and MRI findings of the 33 unrelated index cases with confirmed CADASIL, as well as their 30 affected family members (24 symptomatic and six asymptomatic mutation-carriers), are summarised in tables 1, 2. Twenty-nine families showed autosomal dominant inheritance. Among all 57 symptomatic cases, the onset age of neurological symptoms was 26–65 years (42.74±9.12 years, mean±SD), with the most common onset symptom being TIA/stroke. With clinical progression, the main symptoms included TIA/stroke in 82% (47/57) and cognitive decline in 60% (34/57). Migraine with aura, mental and/or psychiatric manifestations such as depression, anxiety and mania, however, were noted in only 5% (3/57) and 7% (4/57), respectively. Some rare symptoms were also reported in this patient series, including alopecia in one case, tinnitus and deafness in two cases.

Table 2

Clinical, MRI and molecular data from 30 affected family members (other than the index patients) in 12 families with cerebral autosomal dominant arteriopathy with subcortical infarcts and leucoencephalopathy in mainland China

Brain MRIs were available for 48 members, including all 33 index cases and 15 of their mutation-carrying relatives (nine symptomatic and six asymptomatic). Multifocal hyperintensity in the cerebral white matter on T2 weighted images, as well as multiple lacunar infarcts in the basal ganglia or deep white matter, could be detected for all of the symptomatic cases. White-matter lesions in the temporal pole were noted in 46% (22/48) of all affected subjects. Infarct lesions in the brainstem were found in 64% (27/42) of the symptomatic cases (figure 2). In the six asymptomatic cases carrying a NOTCH 3 mutation, abnormal signals in the white matter could also be found by MRI in all of them, and, in one, lacunar infarction in the pons was noteworthy.

Figure 2

Cranial MRI of index patient 31. Lacunar infarcts can be seen in the cerebral hemisphere and brainstem as well as hyperintensity in the white matter of the bilateral temporal lobes.


It is well known that the great heterogeneity of clinical manifestations and the substantially non-specific pattern of cranial MRI changes hinder the clinical diagnosis of CADASIL. To our knowledge, only four mainland Chinese CADASIL families have been reported by other groups.15 21 22 Compared with the large number of CADASIL patients reported among the French, German, Finnish and Italian populations, this disease seemed rather rare in mainland China. Here, we present a substantial number of unrelated mainland Chinese cases with CADASIL who have been collected in the past 7 years by us, thus suggesting that this disease is far more common than previously thought.

Among the 48 suspected CADASIL patients, 30 were confirmed both genetically and pathologically, and three were confirmed genetically. Among these 33 index patients and their available affected family members, most patients showed clinical pictures typical of CADASIL, including recurrent ischaemic stroke, cognitive decline, psychiatric changes, white matter abnormalities and multiple lacunar infarcts revealed by cranial MRI, as well as an autosomal dominant family history.3 23 However, there were distinct features in our patient cohort, which were different from those in previous reports. First, migraine with aura, encountered in more than one-third of Caucasian patients,24 25 was present in only 5% of our patients. Similarly, migraine was infrequently observed in other Asian CADASIL patients, such as in Japan,26 Thailand,13 Korea12 or Chinese Taiwanese,17 although recent data from Japan revealed that 40% of index patients suffered from migraines.27 We speculate that modifying genes or environmental factors may underlie the low frequency of migraine among Asian CADASIL patients. In support of this, epidemiological studies in the general population have shown a lower prevalence of migraine in Asia than in Europe.28 29 Second, index case 2 had alopecia and absence of a family history, but a diagnosis of CADASIL was made after both pathological and molecular examinations. In the Taiwan report by Lee et al,17 approximately 40% of the index patients did not have a dominantly inherited family history. This suggests CADASIL may show marked clinical variability, and that diagnosis cannot be made judging only on the classically described phenotypes. Third, in our patients, MRI revealed different lesion distributions. The sensitivity of T2-weighted MRI in detecting anterior temporal abnormalities in our patient cohort was only 46%, which is similar to the 42.9% reported for the Chinese Taiwanese, but notably lower than the 89–95% reported for Caucasian populations.5 30 Conversely, lacunar infarction in the brainstem is more common among Asian than among Caucasian CADASIL patients31 (64% vs 45%).

Until now, the largest number of CADASIL patients of Chinese descent was reported by Lee et al, who reported that in 21 unrelated Taiwanese families with CADASIL, 47.6% of NOTCH3 mutations were in exon 11. R544C in exon 11 was very common, but only 19% of mutations were in exon 4.17 In the present study, however, exons 4 and 3, not exon 11, of the NOTCH3 gene were responsible for 85% of the patients with CADASIL of mainland Chinese descent. Our results are quite different from those of Taiwanese patients, but are in agreement with the Caucasian population.9 These data support the finding that the spectrum of NOTCH3 mutation is different not only from country to country, but also from region to region in the same population.16 Because we identified 20 different mutations in 33 unrelated cases in our patient series, we speculate that CADASIL mutations might have more than one ancestor in mainland China. More cases and further research are needed to confirm this.

All but one of the mutations we identified were missense, and six of them were novel: two in exon 2, three in exon 4 and one in exon 20 (table 1). Most mutations involved the loss or gain of a conserved cysteine residue in an epidermal growth factor-like domain, but the mutations in families 3 and 33 did not involve a cysteine residue. In addition, family 4 carried a novel deletion in exon 3 of the NOTCH3 gene, causing loss of a cysteine residue. All of these novel mutations have added to the complexity of the NOTCH3 gene variation database.7

If CADASIL is suspected in a patient, confirmation of the presence of GOM is an important step32 in addition to NOTCH3 gene analysis. Of the 42 cases who underwent skin or sural biopsies, GOM was identified in the VSMCs in 30 cases. NOTCH3 mutation analysis revealed pathogenic mutations in all 30 cases with GOM. In the 12 cases without GOM, however, exhaustive gene mutation screening of NOTCH3 did not show any pathogenic mutation. Our results were consistent between the pathology and molecular genetic test, which is in accordance with previous results.7 Although biopsy is an invasive examination, and genetic testing has become widely available, we still consider pathological examination to be essential in the diagnosis of CADASIL. The presence or absence of GOM can influence the decision about whether to undertake exhaustive NOTCH3 mutation screening, especially for those patients who do not have a mutation within a hotspot. Furthermore, Peters et al reported that in 125 unrelated biopsy-proven CADASIL patients, genetic testing was associated with a non-negligible proportion of false-negative results despite performing direct sequencing of exons 2–24 of the NOTCH3 gene, indicating the crucial role of biopsy in the diagnosis of CADASIL.33 However, the sensitivity of detecting GOM in patients' skin biopsies varies greatly from report to report.5 7 In our study, we found that GOM rarely appeared in the venule wall and never appeared in the capillary wall. We suggest that the arterioles should be targeted when looking for GOM, to increase the sensitivity.


In conclusion, in reviewing a cohort of 33 unrelated mainland Chinese patients with CADASIL, we found that the characteristics and prevalence of the main clinical features, including TIA, stroke and cognitive decline, were similar to those seen in Western countries, while migraine with aura was rare. Because not every patient shows the complete CADASIL phenotype, pathological examination and genetic analysis of the NOTCH3 gene should be performed in all suspected cases. In suspected cases of CADASIL in mainland China, we advise that exons 4 and 3 should be screened first, followed by exons 2 and 11.


We thank all the families with CADASIL (and their referring physicians) for participating in the study.



  • Funding This study was supported by grants 30470595 and 30971006 from the National Natural Science Foundation of China.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval The study was approved by the Medical Ethics Committee of Peking University First Hospital.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.