No Evidence for Linkage of the CHRNA7 Gene Region in Canadian Schizophrenia Families

Abstract

Schizophrenia patients demonstrate a deficiency in the filtering of sensory information, and one specific measure involves a response to the second of a pair of auditory stimuli. A neurophysiological measure of this consists of the electroencephalographic response to pairs of auditory signals, emitted fractions of a second apart. Schizophrenic patients and some of their unaffected relatives show a failure of inhibition of a second tone if it occurs 50 msec after the first. A recent genome scan indicated that the gating defect is linked to the alpha 7 neuronal nicotinic acetyl choline receptor gene on chromosome 15. We genotyped 5 schizophrenia families with a total of 96 subjects with a dinucleotide polymorphic marker located less than 120 kb from the first exon of the alpha 7 neuronal nicotinic acetylcholine receptor gene. Linkage analysis was undertaken using parametric and nonparametric statistical methods. The results of the parametric analysis showed negative lod scores under both narrow and broad diagnosis (lod = −3.6 and −4.8, respectively, at θ = 0), and dominant and recessive modes of transmission of the disease. Non-parametric analysis using GENEHUNTER produced nonsignificant NPL scores (NPL = −0.4 and −0.3 for broad and narrow diagnoses, respectively). In summary, we did not find any evidence that the α7 neuronal nicotinic acetylcholine receptor gene (CHRNA7) is linked to schizophrenia. However, we have not been able to assess the P50 measures in these families.

INTRODUCTION

A genome scan found evidence for linkage of markers near the CHRNA7 gene and a neuronal mechanism that regulates the response to auditory stimuli [Freedman et al., 1997]. An abnormality in filtering auditory stimuli observed in schizophrenia patients has been measured by a test that involves the evoked response, the P50, to the second of a pair of stimuli. Schizophrenic patients and some of their unaffected relatives demonstrate a deficit in the inhibition of the P50 response produced 500 msec after the first one, and this trait is an autosomal-dominant trait in these families. The abnormal P50 was used as the phenotype in nine schizophrenia families, and genome-scanning found linkage to D15S1360 located less than 120 kb from the first exon of CHRNA7 on chromosome 15 [Freedman et al., 1997]. In view of the evidence that nicotine administration normalizes the P50 defect in schizophrenia patients, Freedman et al. [1997] suggest that this provides further rationale for examining the CHRNA7 gene. Linkage analysis was performed in the nine families, using the diagnosis of either schizophrenia or schizoaffective disorder as the phenotype. A maximum lod score of 1.33 for schizophrenia at θ = 0.07 was obtained. Here, we report on an attempt to replicate the linkage of D15S1360 with schizophrenia in five eastern Canadian families.

SUBJECTS AND METHODS

Five multiplex families from the eastern provinces of Canada had family members assessed with the Structured Clinical Interview of DSMIII-R (SCID-I and SCID-II) and review of medical records [Bassett et al., 1993; Bassett and Honer, 1994]. The total number of subjects of the five families studied was 122 members with 36 members (7.2 per family on average) diagnosed with schizophrenia or schizoaffective disorder (narrow diagnosis), and an additional 14 (2.8 per family on average) with a broader spectrum of disorders such as schizotypal personality disorder, psychosis NOS, schizoid personality disorder, or paranoid personality disorder (broad diagnosis). DNA was available for 93 subjects in these families.

Genomic DNA was extracted from Epstein-Barr-transformed cell lines using standard high-salt methods. The primer sequences for the CA repeat polymorphism of D15S1360 were obtained from the Genome Data Base (GDB, accession no. 1105081). One primer was end-labelled with γ-32ATP and PCR was performed, as described in the GDB. The PCR product was electrophoresed on 6% polyacrylamide gels, which were then exposed to X-ray film for a few hours or overnight.

Parametric linkage analysis was performed by the program MLINK, assuming both dominant and recessive modes of transmission using the following parameters: autosomal-dominant (disease gene frequency = 0.85%, fAA = fAa = 70%, faa = 0.1%); autosomal-recessive (disease gene frequency = 12%, FAA = 70%, fAa = faa = 0.1%). Single-point nonparametric linkage analysis was performed using GENEHUNTER NPL, with the options “skip large off” and “discard on” [Kruglyak et al., 1996].

RESULTS AND DISCUSSION

No positive lod scores were obtained at θ = 0.00 for D15S1360 (broad dominant, −4.95; broad recessive, −5.79; narrow dominant, −3.04; narrow recessive, −3.25); see Table I. These lod scores allow us to exclude a region of 10 cM on either side of the marker. Single-point nonparametric linkage analysis, performed using GENEHUNTER, found no significant NPL scores. We obtained an NPL = −0.4 (P = 0.6) for broad diagnosis category. The narrow diagnosis yielded an NPL of −0.3. The heterozygosity of D15S1360 was 0.52 in the families analyzed.

TABLE I.

Two-Point Results for D15S1360 at θ = 0.0

Families	MLINK				GENEHUNTER
	Narrow/dominant	Narrow/recessive	Broad/dominant	Broad/recessive	NPL broad	P
1	−1.73	−0.11	−1.14	−1.34	−0.36	0.55
2	−0.08	−0.12	0.13	0.07	0.19	0.31
3	−0.16	−1.36	−0.10	−1.12	−0.22	0.44
4	0.06	−0.31	−2.66	−3.13	−0.34	0.55
5	−1.13	−1.36	−1.18	−0.28	−0.22	0.52
Total	−3.04	−3.25	−4.95	−5.79	−0.43	0.62

CONCLUSIONS

We find no evidence for linkage in five large schizophrenia families of a microsatellite marker D15S1360, which is located 120 kb from the CHRNA7. The heterozygosity of the marker is high enough to clearly exclude our parametric hypothesis, and further marker information is not required. Greenberg et al. [1996] suggested that in cases like this where traits are complex or for which the mode of inheritance is unclear, the use of both the parametric dominant and recessive model is appropriate. The exclusion region included more than 1 cM on either side of D15S1360, thus effectively excluding the CHRNA7 locus under the specified models. Also, nonparametric testing with GENEHUNTER provided no support for linkage. Therefore, we were not able to replicate the evidence for linkage to schizophrenia reported by Freedman et al. [1997]. Previous studies on this set of families investigated specific genetic hypotheses, with no evidence for linkage thus far [Sidenberg et al., 1993; Vicente et al., 1997; Petronis et al., 1996; King et al., 1997; O’Malley et al., 1996]. The power to detect linkage may be low if this locus confers only minimal to moderate relative risk for schizophrenia [Risch and Merikangas, 1996]. Alternatively, the negative parametric and nonparametric values obtained in our study might be a consequence of a high degree of genetic heterogeneity of schizophrenia, with this locus not conferring susceptibility in these particular families. It is of note that none of the markers flanking D15S1360 in the initial study showed lod scores greater than 1.0, using schizophrenia as the phenotype [Freedman et al., 1997]. One limitation of the present study is that we have not been able to assess the P50 measures in the families studied; thus, our results only address the hypothesis of a role for CHRNA7 in schizophrenia per se.