A family-based and case-control association study of the NOTCH4 gene and schizophrenia

Article metrics


Recently a strong positive association between schizophrenia and Notch4 has been reported.1 Both individual markers and haplotypes showed association with the disease, with five markers (three microsatellites and two SNPs) being tested. In order to test this finding we genotyped these markers in the Han Chinese population using a sample of 544 cases and 621 controls as well as >300 trios. Analysis of allele, genotype and haplotype frequencies in both samples showed no association between the markers and the disease. Our results would indicate that a significant role for the Notch4 gene in schizophrenia can be ruled out in the Han Chinese. However, similar studies are necessary in the Caucasian population as linkage disequilibrium arrangements and founder effects may differ between these two populations.


Schizophrenia is a severe and common complex illness with a large genetic component, however major common risk loci have yet to be found.2 Wei and Hemmings1 found that Notch4 was strongly associated with schizophrenia in a sample of 80 British trios. They analysed three microsatellites and two SNPs in and around the gene (on 6p23) and found marker and haplotype associations. The Notch gene family encodes large transmembrane receptors and may regulate embry- onic cell migration, vascular morphogenesis and remodelling.3 Evidence from neuropathology and epidemiology suggests a significant proportion of schizophrenia cases have neurodevelopmental aetiology.4 Additionally, the 6p21–6p23 region has previously been implicated by linkage studies.2 We have attempted to replicate the finding using two Han Chinese samples: (a) 544 cases and 621 controls; and (b) 327 trios. We analysed the five loci chosen by Wei and Hemmings and found negative results with both markers and haplotypes. The physical positions of the markers are shown in Figure 1.

Figure 1

Physical locations of the five markers used in the study. The marks on the scale bar are equivalent to approximately 10 kb. Exons are indicated by triangles in the direction of transcription.

In our case-control analysis 544 schizophrenics were genotyped and compared with a set of 621 controls. Table 1 gives the marker genotype and allele frequencies. The allele and genotype frequencies were in Hardy–Weinburg disequilibrium and did not differ by sex, age or region of origin. Likewise, there was little evidence of an overall difference in distribution in allele or genotype frequencies between cases and controls. Only the fifth marker, which is a slightly unstable TTAT repeat, gave slightly positive P-values of 0.03 overall and 0.01 for female sex, before correction for multiple testing (at least 50 tests were performed). Some difficulty was encountered in analysing haplotype frequencies and associations due to the large number of possible haplotypes (>2000), individuals (>1000) and marker combinations to be analysed. However, using an experimental version of EH Plus called FPEH (supplied by J Zhao), we analysed the haplotype frequencies for all markers and found no differences overall (P = 0.39) or for different combinations (data not shown). Thus there was no evidence from this study to support the association of Notch 4 with schizophrenia.

Table 1 Distribution of alleles in schizophrenia patients and healthy controls

However, one of the most common criticisms of case-control association studies is the possibility of population stratification. Thus we conducted a parallel study of 327 trios for all five markers. The results of transmission disequilibrium test (TDT) analysis were negative for all markers (see Table 2; analysis by region of origin, sex and age is also negative, data not shown).5 The multiple marker haplotypes were analysed by TRANSMIT v2.5 (see Table 3) and no differences were found.

Table 2 Transmission disequilibrium test (TDT) results for five polymorphisms of NOTCH4 in total samples
Table 3 Haplotype analysis of trios families for association of NOTCH4 variation with schizophrenia

Thus, in a very large sample of schizophrenic trios, cases and controls, we found no evidence for association of schizophrenia with NOTCH4 either by case-control analysis or the family-based TDT method. There are several possible reasons for this finding. (a) The initial result may be real. The Han Chinese population examined differs from the British sample used in the initial report, and it is possible that the differing genetic background in these groups accounts for the differences although this is not the case for common risk loci such as ApoE e4 in Alzheimer's disease. Likewise our finding may be a chance result. (b) Our result may be real and the initial report may be a chance result. The initial sample size was small compared to our sample (80 trios vs 327 trios, 544 cases and 621 controls) and did not report results from a replication sample. Likewise, it should be pointed out that haplotype analysis of a positive association could show positive results merely due to linkage disequilibrium between markers, although it may also result from increased informativeness due to the increased number of markers and haplotypes examined. Possibly the real test of haplotype analysis would be if it identified a marker which on its own showed a stronger association with the disease that the initial marker tested.

In conclusion, we can exclude a major role for the Notch4 gene in schizophrenia in the Chinese population and probably in schizophrenia as a whole. Both case-control and family-based studies in the Han Chinese failed to show any association. However, a similar study in a Caucasian population will be necessary before definitively excluding the locus from any involvement in schizophrenia.

Materials and methods


The 327 trios were composed of 170 trio families from Shanghai, 71 trio families from Shanan'xi province, and 86 trio families from Jilin province. There were 179 male and 148 female probands with a mean age of 25.8 ± 7.7 years. In addition to a family-based design, we also incorporated a case-control designed testing. The case sample included 544 Han Chinese (53% male), mean age 42, SD = 16.5. The control sample included 621 healthy Han Chinese (50% male) mean age 41.5 ± 16.1 years. Clinical diagnosis was made according to DSM-IIIR; an independent clinician using the same criteria reviewed all diagnoses, and all the subjects gave informed consent. All subjects were Han Chinese in origin.


Two SNP polymorphisms and three microsatellite markers were genotyped in up to 327 family trios by PCR using standard agarose gel PCR-RFLP analysis or fluorescent microsatellite marker analysis on an ABI 377 sequencer (PE Applied Biosystems, USA). Of the five markers, four pairs of primers [(TAA)n, SNP1, SNP2 and (TTAT)n] were synthesized and genotyped using sequences provided by J Wei (personal communication). However, the primers for (CTG)n polymorphism were re-designed by us. Parents and probands were checked for Mendelian inheritance. The exact sequences were as follows: SNP1 5′-TGCT GGCTCACGGGCTTCC-3′ and 5′-TGGATTGCAGTGGC ACGACC-3′; SNP2 5′-AAACAGCAGGGCTGGGACTG-3′ and 5′-ACCTCTGGGTCTGACCACTG-3′; (CTG)n 5′FAM-GAAACAGCTCAGACGTGAGG-3′ and 5′-ACC TCTGGGTCTGACCACTG-3′; (TAA)n 5′HEX-TCATGA CCAGCAACATAGGG-3′ and 5′-TACACACTACCATT CCTGGG-3′ and (TTAT)n 5′FAM-TGAATACACCCTTC CTCCTC-3′ and 5′-ACAGACTGGGACTCCATCTC-3′.

Statistical analysis

Association of individual markers with schizophrenia was examined by: (i) comparing cases and controls; and (ii) using family controls in an extension of the transmission disequilibrium test. Allele frequencies in different groups of subjects were compared using the CLUMP program for the microsatellites and EpiInfo v6.0 for the SNPs (WHO/CDC, 1999).6

For the family-based analysis, a conditional logistic regression test was performed using the ETDT program, which can analyse multi-allele marker loci.7 For analysis of multiple marker haplotype transmission, we used the program TRANSMIT v2.5, which can deal with the transmission of multi-locus haplotypes even if phase is unknown and some parental genotypes are missing.8,9 TRANSMIT requires parameters to be entered for minimum haplotype frequencies and ambiguity of haplotypes. For the present analysis minimum haplotype frequencies were set at 0.03 and ambiguity set at 3, ie haplotypes which occurred with a frequency of less than 3% were excluded and if there are more than three possible parental haplotypes the family was excluded.10


  1. 1

    Wei J, Hemmings GP . The NOTCH4 locus is associated with susceptibility to schizophrenia Nat Genet 2000 25: 376–377

  2. 2

    Riley BP, McGuffin P . Linkage and associated studies of schizophrenia Am J Med Genet 2000 97: 23–44

  3. 3

    Joutel A, Corpechot C, Ducros A, Vahedi K, Chabriat H, Mouton P et al. Notch3 mutations in CADASIL, a hereditary adult-onset condition causing stroke and dementia Nature 1996 24: 707–710

  4. 4

    Chua SE, Murray RM . The neurodevelopmental theory of schizophrenia: evidence concerning structure and neuropsychology Ann Med 1996 6: 547–555

  5. 5

    Spielman RS, McGinnis RE, Ewens WJ . Transmission test for linkage disequilibrium: the insulin gene region and insulin-dependent diabetes mellitus (IDDM) Am J Hum Genet 1993 52: 506–516

  6. 6

    Sham P, Curtis D . Monte-Carlo tests for associations between disease and alleles at highly polymorphic loci Ann Hum Genet 1995 59: 97–105

  7. 7

    Sham PC, Curtis D . An extended transmission/disequilibrium test (TDT) for multi-allele marker loci Ann Hum Genet 1995 59: 323–336

  8. 8

    Chiano MN, Clayton DG . Fine genetic mapping using haplotype analysis and the missing data problem Ann Hum Genet 1998 62: 55–60

  9. 9

    Clayton D . A generalization of the transmission/disequilibrium test for uncertain haplotype transmission Am J Hum Genet 1999 65: 1170–1177

  10. 10

    Li T, Ball D, Zhao J, Murray RM, Liu X, Sham PC et al. Family-based linkage disequilibrium mapping using SNP marker haplotypes: application to a potential locus for schizophrenia at chromosome 22q11 Mol Psychiatry 2000 5: 77–84

Download references


Dr S Li, and Professor B Lu for statistical and other help. This work was supported by grants from the 973 Project, the National Natural Science Foundation of China and the Royal Society of the UK, Qiu Shi Science & Technologies Foundation, and Shanghai Municipal Commission for Science and Technology.

Author information

Correspondence to L He.

Rights and permissions

Reprints and Permissions

About this article


  • schizophrenia
  • Han
  • Chinese
  • family trios
  • Notch4
  • haplotype
  • TDT

Further reading