|Home | About | Journals | Submit | Contact Us | Français|
The α7 neuronal nicotinic acetylcholine receptor subunit gene (CHRNA7) is localized in a chromosomal region (15q14) linked to schizophrenia in multiple independent studies. CHRNA7 was selected as the best candidate gene in the region for a well-documented endophenotype of schizophrenia, the P50 sensory processing deficit, by genetic linkage and biochemical studies.
Subjects included Caucasian-Non Hispanic and African-American case-control subjects collected in Denver, and schizophrenic subjects from families in the NIMH Genetics Initiative on Schizophrenia. Thirty-five single nucleotide polymorphisms (SNPs) in the 5′-upstream regulatory region of CHRNA7 were genotyped for association with schizophrenia, and for smoking in schizophrenia.
The rs3087454 SNP, located at position −1831 bp in the upstream regulatory region of CHRNA7, was significantly associated with schizophrenia in the case-control samples after multiple-testing correction (P = 0.0009, African American; P = 0.013, Caucasian-Non Hispanic); the association was supported in family members. There was nominal association of this SNP with smoking in schizophrenia.
The data support association of regulatory region polymorphisms in the CHRNA7 gene with schizophrenia.
Schizophrenia is a complex neuropsychiatric disorder partially characterized by sensory processing and cognitive deficits (Leonard et al., 2001; Light and Braff, 2003; George et al., 2006; Adams and Stevens, 2007; Martin and Freedman, 2007). Sensory processing deficits are normalized in schizophrenic patients and first-degree relatives by smoking (Adler et al., 1993, 1998; Olincy et al., 1998; Light and Braff, 2003; Leonard et al., 2007b), and cognitive deficits are improved by nicotine administration (Levin et al., 1998, 2006; Rezvani and Levin, 2001). These discoveries led to a hypothesis of self-medication wherein schizophrenics smoke to correct an underlying biological deficit (Leonard, 2003; Leonard et al., 2007a; Adler et al., 1998; Kumari and Postma, 2005). The hypothesis is supported by studies demonstrating that smoking alters gene expression in normal individuals and differentially regulates gene expression in schizophrenics (Mexal et al., 2005, 2008; Kuehn, 2006).
Nicotine exerts its effect through neuronal nicotinic acetylcholine receptors expressed in the brain and periphery (Leonard and Bertrand, 2001; Gotti et al., 2006). The α7 neuronal nicotinic receptor subunit gene (CHRNA7), localized at 15q14, was first genetically linked to the P50 auditory sensory processing deficit in schizophrenia (Freedman et al., 1997), and then to schizophrenia as a disease in multiple independent studies. These reports included cohorts of European American, African American, South African Bantu, Azorean, and Han Chinese ancestry, indicating that the linkage is valid across ethnicities (Coon et al.,1994; Kaufmann et al., 1998; Leonard et al., 1998; Riley et al., 2000; Freedman et al., 2001; Freedman and Leonard, 2001; Gejman et al., 2001; Liu et al., 2001; Tsuang et al., 2001; Xu et al., 2001; Fallin et al., 2003). Activation of the homopentameric α7* receptor results in the influx of Ca++ and neurotransmitter release (Vijayaraghavan et al., 1992; Aramakis and Metherate, 1998; Berg and Conroy, 2002; Dajas-Bailador and Wonnacott, 2004). Binding studies with an α7* receptor antagonist (α-bungarotoxin) show there are 50% fewer receptors in postmortem hippocampus of individuals with schizophrenia compared to control subjects (Freedman et al., 1995). Decreased expression has also been found in cortex (Guan et al., 1999; Marutle et al., 2001) and in the reticular thalamic nucleus (Court et al., 1999). Low levels of the α7* receptor may have downstream consequences for multiple neuro-transmitter systems, altering the balance of neurotransmitter release and activation (Leonard, 2003). The CHRNA7 gene is now considered one of the important candidate genes for schizophrenia (Harrison and Weinberger, 2005), and α7* receptor agonists are ranked as the most promising targets for development of a drug to treat cognitive impairments in the disorder (Psychiatric News, 2006). Recent Phase I and Phase II studies of an α7* receptor partial agonist, DMXB-A (Martin et al., 2004) resulted in improvements of both sensory processing and attention in non-smoking schizophrenics (Olincy et al., 2006; Freedman et al., 2008).
The α7 nicotinic receptor subunit gene, CHRNA7, was found to be partially duplicated, with both loci mapping to the 15q14 region (Gault et al., 1998; Riley et al., 2002). The partially duplicated gene (CHRFAM7A) is located 1.6 Mb centromeric to the full-length gene; its function remains unknown. Mutation screening of the coding region and intron/exon borders of CHRNA7 and of CHRFAM7A identified 33 polymorphisms (Gault et al., 2003). Three of the polymorphisms were non-synonymous and mapped to the full length CHRNA7 coding region. These polymorphisms, however, were very rare and were not associated with either schizophrenia or the P50 gating deficit (Gault et al., 2003). Mutation screening of the core promoter in the CHRNA7 gene identified a large number of polymorphisms (Leonard et al., 2002). Functional analysis of polymorphisms in the 231 base pairs upstream of the translation initiation site, the core promoter region, demonstrated that most decrease transcription. These polymorphisms were also found to be statistically more prevalent in schizophrenics than in control subjects (P=0.007) (Leonard et al., 2002). Further, the presence of a promoter polymorphism in non-schizophrenic controls was associated with a P50 gating deficit (P<0.0001) (Leonard et al., 2002). Reporter gene assays with fragments 1.0 kb and 2.6 kb proximal to the translation initiation site exhibit less activity than the core promoter, suggesting that repressor elements may lie upstream of the core promoter.
Utilizing overlapping genomic fragments, we have identified 35 SNPs in the 2 kb upstream regulatory region of CHRNA7. SNPs were genotyped utilizing a combination of heteroduplex analysis by denaturing high-performance liquid chromatography and sequencing, and were analyzed for association with schizophrenia in Caucasian-Non Hispanic and African-American subjects. Smoking history was considered as a secondary outcome. The results show significant association of a specific SNP, rs3087454 (−1831 bp), in the 5′ upstream regulatory region of the CHRNA7 gene with schizophrenia.
Characteristics of study participants are shown in Tables 1a and 1b. DNA samples from both case-control subjects collected in Denver, and schizophrenic subjects from NIMH families were included in the association studies. For the family-based study, a total of 329 African-Americans (47 nuclear families) and Caucasian-Non Hispanic subjects (73 nuclear families) from the NIMH Schizophrenia Genetics Initiative were chosen based on a diagnosis of schizophrenia. This cohort has previously shown positive LOD scores and association to markers at the 15q14 locus in both African-American and Caucasian individuals (Kaufmann et al., 1998; Leonard et al., 1998; Freedman et al., 2001). Detailed information on the NIMH family structure is available from the web site (http://www.nimhgenetics.org).
Case-control association studies of 612 African-American and Caucasian-Non Hispanic schizophrenics and controls included individuals collected in our laboratory, and schizophrenic subjects from the NIMH families. Ethnicities of case-control subjects were recorded from self-report or family interview. The Caucasian-Non Hispanic case-control sample was comprised of 307 schizophrenic patients, 56 childhood onset schizophrenics, and 99 controls. The African-American case-control sample included 103 schizophrenics, 2 childhood onset schizophrenics, and 45 controls. P50 data were available for 89 Caucasian-Non Hispanic controls subjects and 43 African-American controls. Adult case subjects were chosen based on a diagnosis of schizophrenia, utilizing a Structured Clinical Interview for DSM-IV Axis I Disorders (First et al., 1996b). Control subjects were interviewed and found to have no evidence for current or past psychosis, using a Structured Clinical Interview for DSM-IV Axis I Disorders for non-patients (First et al., 1996a). Childhood onset schizophrenia cases were diagnosed utilizing the K-SADS-PL (Kaufman et al., 1997). Smoking history was determined on local subjects utilizing a detailed questionnaire (Breese et al., 1997; Breese et al., 2000) and for the NIMH Genetics Initiative subjects, from the web site. P50 auditory evoked potentials were recorded on controls, using published methods (Freedman et al., 1991).
Nucleated cells were obtained from anticoagulated blood (EDTA) via lysis with a high sucrose solution (0.3 M sucrose, 0.01 M Tris–HCl, pH 7.5, 0.005 M MgCl2, and 1% Triton X-100) and subsequent centrifugation. DNA was then isolated from these cells as previously described (Miller et al., 1988). DNA samples for the NIMH Genetics Initiative subjects were obtained from the NIMH collection at Rutgers (http://www.rucdr.org/ASHG/Schizophrenia.htm).
Eleven amplicons were designed to screen 2.0 kb of DNA in the 5′-upstream regulatory region of the CHRNA7 gene. The Polymerase Chain Reaction (PCR) with AmpliTaq Gold™ and GeneAmp® PCR System 9600 (Perkin-Elmer, Foster City, CA) was used for fragment amplification; thermal cycler conditions unique to each amplicon are shown in Table 2. Fragments were characterized using temperature modulated heteroduplex analysis with Denaturing High-Performance Liquid Chromatography (DHPLC; Transgenomic WAVE™) (Transgenomic Inc., San Jose, California) as previously described (Leonard et al., 2002; Gault et al., 2003). Samples were “spiked” with control DNA to ensure that a homozygous variant that might migrate as a single peak would be detected. When a sample resulted an unusual WAVE™ pattern, automated DNA sequencing on an Applied Biosystems 3100 Avant DNA Sequencer (Applied Biosystems, Foster City, CA) was employed for genotyping. For most SNPs, a specific and recognizable WAVE™ pattern was generated.
The human neuroblastoma cell lines SH-SY5Y (a gift from Dr. June Biedler) (Biedler et al., 1978), and SK-N-BE (ATCC number CRL-2271; ATCC, Manassas, VA) were grown in 1:1 Ham F12 Dulbecco Eagle medium supplemented with 15% fetal bovine serum, 100 μg/ml streptomycin, 100 units/ml penicillin, and 2 mM L-glutamine (Invitrogen, Carlsbad, CA) at 37 °C in a 5% CO2 incubator.
Isolation and characterization of 2.6 kb of 5′-flanking sequence in the human CHRNA7 subunit gene was described previously (Gault et al., 1998; Leonard et al., 2002). The promoter-luciferase construct Pr2B contains a CHRNA7 5′-upstream regulatory region fragment from −2004 bp to the translation start site at +1 bp inserted upstream of the luciferase gene in the pGL3-Basic vector. Pr2B was generated by first TA cloning a 2004 bp PCR fragment from the previously cloned 2.6 kb promoter into pDrive, according to manufacturer’s instructions (Promega, Madison, WI). Restriction enzymes (HindIII and KpnI) were then used to isolate the 2004 bp fragment and insert it into the pGL3 reporter plasmid. The promoter-luciferase fusions were confirmed by DNA sequencing.
Pr2B contains the (C) allele at −1831 bp. The −1831(A) polymorphism was introduced into Pr2B by PCR, utilizing the QuikChange II XL Site-Directed Mutagenesis Kit, according to the manufacturer’s protocols (Stratagene). The mutagenesis primers were: a7Prom-1831A forward (5′-gcca-tacatactccagaaaaaatAaataaattcccttggccc-3′) and a7Prom-1831A reverse (5′-gggccaagggaatttattTattttttctggagtatg-tatggc-3′); the mutant construct was confirmed by DNA sequencing and named Pr2B-1831A.
SH-SY5Y and SK-N-BE cells grown to subconfluency were transfected with the ProFection Mammalian Transfection Calcium Phosphate System (Promega). A pGL3-Control construct in which the luciferase gene is regulated by the SV40 promoter and enhancer was used as a positive control to measure the maximum reporter gene activity. A pRL-TK construct (Promega) with the Renilla luciferase gene, controlled by the HSV thymidine kinase promoter, served as an internal control for normalization of transfection efficiency. 2×105 cells/35 mm plate were incubated with molar equivalents of Pr2B, Pr2B-1831A, pGL3-Control, or pGL3-Basic luciferase reporter constructs and 1 μg of pRL-TK vector as an internal control for the transfection efficiency. Cells were harvested after 48 h and luciferase activity was determined with the Dual-Luciferase Reporter Assay System according to the manufacturer’s instructions (Promega). Each transfection was performed using eight separate plates in triplicate.
The case group of schizophrenic individuals included one randomlychosen schizophrenic individual per family. For each marker, allele frequencies were estimated and conformance of genotype frequencies with Hardy–Weinberg Equilibrium (HWE) expected proportions was tested via a 1 df chi-squared goodness-of-fit test using Haploview (Barrett et al., 2005). Linkage disequilibrium parameters D′ and r2 were calculated for each pair of markers and tests for differences in allele frequencies between cases and controls were computed using a chi-squared test (1 df) for each SNP in Haploview. Since several of the SNPs were rare, empirical P-values were utilized rather than relying on the large-sample approximation. Further, since several SNPs were tested for each outcome, empirical P-values were generated via permutation that were corrected for the number of SNPs tested. For each permutation, the minimum P-value over all SNPs was compared to the minimum P-value over all SNPs in the original data. Chi-squared tests of association (2 df) between each SNP and schizophrenia under a genotypic model were calculated using the UNPHASED software version 3.0.13 (Dudbridge, 2003); a similar permutation procedure was applied. Corrected P-values are indicated with an (*). Odds ratio estimates (OR) and 95% confidence intervals (CI) for comparing relevant genotype or allele classes were computed with SPSS 16 software (SPSS Inc., Chicago IL).
Analysis of the P50 gating deficit in controls was conducted using simple linear regression models that tested the mean differences in P50 ratios under both additive and genotypic models of association separately for each SNP. In the absence of evidence for ethnic-specific effects, ethnicities were combined for most SNPs in order to gain power to detect association, and ethnicity was included as a covariate. All analyses were performed using SAS version 9.1 software (SAS Institute Inc.: Cary, NC, 2004).
Power estimates (at a significance level of 5%) for case-control analyses were obtained with the Quanto software (Gauderman, 2002a,b). Results indicated that the Caucasian schizophrenic/control sample size had power between 0.74 and 0.99 to detect a genotype relative risk (GRR) ≥2.0 over a range of marker allele frequencies between 0.10 and 0.40. At GRRs <2.0, power was substantially lower (0.12–0.88) for most allele frequencies. The African-American schizophrenic/control sample size had less power to detect a GRR ≥2.0 (0.38–0.87) over a range of marker allele frequencies between 0.10 and 0.40. Similarly, at GRRs <2.0, power was considerably lower (0.07–0.49) for all allele frequencies.
For both ethnicities, case-control samples were evaluated for the presence of admixture by genotyping 176 ancestry-informative SNP markers (Hodgkinson et al., 2008). Within each ethnic group, relative frequencies of alleles between the cases and controls were compared using a chi-squared test (Pritchard and Rosenberg, 1999). There were no significant differences from what was expected, indicating a similar degree of admixture between cases and controls in both ethnic groups. These conclusions were corroborated by estimating individual admixture proportions within each ethnic group using structure (Pritchard et al., 2000). These analyses indicated no significant differences in average admixture proportions between cases and controls within each ethnic group.
Prior to data analysis, the PEDCHECK software was utilized to detect marker allele errors in Mendelian inheritance (O’Connell and Weeks, 1998). Pedigree Disequilibrium Test(s) (PDT) of association at SNP markers were implemented in the UNPHASED software version 3.0.13 (Dudbridge, 2003). Similar to the case-control analyses, UNPHASED was used to generate an empirical P-value corrected for the number of SNPs.
Thirty-five polymorphisms were identified in the 2 kb of CHRNA7 5′-upstream regulatory sequence, screened in a multi-ethnic cohort in this study. Many of these SNPs reside in transcription factor binding sites (Fig. 1). Twenty-six of these SNPs were present in Caucasian-Non Hispanic subjects and/or in African-American subjects. The minor allele frequencies (MAF) for the 26 SNPs are shown in Table 3. Genotype frequencies of all SNPs were consistent with HWE (all P>0.05; data not shown). Only SNPs with MAFs ≥0.05 in at least one ethnic population were considered for statistical analyses.
Of note, the minor alleles for rs3087454 (−1831 bp) were different in the African-American and Caucasian-Non Hispanic samples. The frequency of the rs3087454 (−1831 bp) “C” allele was 0.360 in the Caucasian-Non Hispanic sample, and 0.350 for the “A” allele in the African-American sample.
Fig. 2 shows the LD plot for common SNPs in the core promoter of CHRNA7 (Leonard et al., 2002) and in the upstream regulatory region of the gene. LD parameter estimates were generated using the case-control data. Many of the 26 polymorphisms are rare and were not included in the LD plots. In the Caucasian sample, markers rs3087454 (−1831 bp) and rs6494165 (−1512 bp) were designated as a haplotype block using the Gabriel definition (Gabriel et al., 2002). Two SNP marker pairs, rs3087454 (−1831 bp)/rs3826029 (−1313 bp) (D′ =0.89), and rs6494165 (−1512 bp)/rs3826029 (−1313 bp) (D′=1.0) were in particularly strong LD as they are all contained within a 500 bp region (see Table 3). In the African-American sample, SNP marker rs3826029 (−1313 bp) was in strong LD with marker −1252 bp (D′=1.0), as well as two core promoter polymorphisms [−190 bp (D′=1.0) and −46 bp (D′=1.0)]. Additionally, in both sample populations, SNP markers at −905 bp/−768 bp were in LD with one another (D′=1.0) and with a core promoter polymorphism at −191 bp (D′=1.0).
Eight regulatory region SNPs, −46 bp, −86 bp, −178 bp, −190 bp, −1252 bp, rs3826029 (−1313 bp), rs6494165 (−1512 bp), and rs3087454 (−1831 bp), were analyzed for association with schizophrenia and smoking in schizophrenia. Other SNPs were too rare for analysis. P-values were adjusted by generating empirical P-values via permutation that were corrected for the number of SNPs tested and are indicated by an (*) (Table 4). Genotype and allele counts for the association studies are shown in Supplementary Tables 1–6.
A significant difference was found in the genotype distribution of the rs3087454 (−1831 bp C/A) polymorphism between control and schizophrenic subjects (Table 4). In the Caucasian-Non Hispanic sample, the “C/C” genotype was associated with schizophrenia (P=0.0009*) (OR=1.57; 95% CI=1.31–2.19) and in the African-American sample, the “A/A” genotype was associated with schizophrenia (P=0.013*) (OR=2.0; 95% CI=1.19–3.35). Under an allelic model, the “C” allele was only nominally associated with schizophrenia (P=0.005) in the Caucasian-Non Hispanic sample, and the “A” allele was nominally associated with schizophrenia in the African-American sample (P=0.005) (Table 5). rs3087454 (−1831 bp C/A) was not significantly associated with the outcome of smoking status in either ethnic population.
Fisher’s method (Fisher, 1948) was employed to combine P-values for the ethnic groups, to determine the overall significance of the finding (Table 6). As expected, this meta-analysis of the rs3087454 (−1831 bp) SNP indicated a significant association with schizophrenia (P =0.0001) under a genotypic model, but not under an allelic model (P=0.124).
The 58 childhood-onset subjects included in this study were analyzed separately for association with rs3087454 (−1831 bp), but there was no evidence for association (P=0.560*). COS subjects were excluded from the Caucasian-Non Hispanic case-control sample (a majority of COS subjects were Caucasian), the P-value for an allelic model of association decreased from P=0.135* to P=0.033*. Similarly, the P-value for a genotypic model of association decreased from P=0.0009* to P=0.0002*.
In an attempt to further refine the schizophrenia phenotype, beyond the results shown in Tables 4 and and5,5, additional analyses were included under the assumption that schizophrenic smokers may be biologically distinct from schizophrenic non-smokers. These analyses included the following dependent variables; schizophrenic smokers versus control smokers, schizophrenic non-smokers versus control non-smokers, and schizophrenic smokers versus all controls. In the African-American sample, nominally significant P-values at rs3087454 (−1831 bp) were found with the outcomes of schizophrenic smokers versus control smokers (P=0.018, genotype; P=0.508, allele), schizophrenic non-smokers versus control non-smokers (P = 0.0025, genotype; P=0.008, allele), and schizophrenic smokers versus all controls (P =0.001 genotype; P = 0.606 allele). In the Caucasian-Non Hispanic sample, a nominally significant P-value at rs3087454 (−1831 bp) was found with the outcome of schizophrenic smokers versus all controls (P=0.041 genotype; P=0.179 allele).
The frequency of a polymorphism is an important factor in determining the power to detect the risk allele(s) in a given association study. Many of the CHRNA7 5′ regulatory region When polymorphisms are very rare with minor allele frequencies (MAF) <0.001. With this information and the recent hypothesis that the ‘common disease–rare alleles’ model might explain many cases of schizophrenia (McClellan et al., 2007), we examined the relationship between the rs3087454 (−1831 bp) “risk” genotype and rare proximal promoter polymorphisms. The number of individuals positive for both a proximal promoter polymorphism and the schizophrenia “risk” genotype at rs3087454 (−1831 bp) were compared against those with only proximal promoter polymorphisms (Table 7). Of the rare promoter SNPs in the Caucasian-Non Hispanic sample population, the SNP at −194 bp occurs more often in individuals with the “C” risk allele and the “CC” risk genotype. The D′ value for −194 bp and rs3087454 is 0.73, however, the confidence bounds are large (0.40–0.89). Of the rare SNPs in the African-American sample population, the SNP at −190 bp occurs more often in individuals with the “A” risk allele and the “AA” risk genotype. The D′ value is 0.20 for −190 bp and rs3087454, with confidence bounds of 0.02–0.75. Because these SNPs are rare and their D′ values are not well estimated, the confidence bounds are large. With many rare SNPs and relatively low D′ values, haplotype analyses would have limited power to detect a risk haplotype due to the presence of many rare haplotypes and the error associated with haplotype frequency estimation. However, observations provided by Table 7 suggest rare CHRNA7 proximal promoter polymorphisms may be relevant to schizophrenia as a haplotype with rs3087454 (−1831 bp) or as single markers.
Results of the family-based PDT analysis in a second cohort supported the majority of results obtained in case-control studies (not shown). At rs3087454 (−1831 bp) in the Caucasian-Non Hispanic sample, the “C” allele was found to be linked/associated with schizophrenia (nominal P=0.048). The rs3087454 SNP was not associated with schizophrenia in the African-American sample. The rs6494165 (−1512 bp) SNP was nominally associated with smoking (P=0.045) when Caucasian-Non Hispanic schizophrenic non-smokers and smokers were compared, as was the proximal promoter SNP −86 bp (P=0.026). These results were not replicated in the African-American sample.
Linear regression models were estimated separately for the two ethnic groups for the SNP at rs3087454 (−1831 bp) because different alleles were associated with schizophrenia in African-Americans (“A” allele) and Caucasian-Non Hispanics (“C” allele) at this biallelic marker. For all other SNPs, ethnic groups were combined in order to gain statistical power and ethnicity was included as a covariate. Two proximal promoter SNPs were nominally associated with higher P50 ratios in control subjects (SNP −86 bp, P=0.03; SNP −191 bp, P<0.01). This study did not identify significant association with the P50 deficit at rs3087454 (−1831 bp).
SNP rs3087454 (−1831 bp) was examined for effects on α7 nicotinic receptor gene expression in a luciferase reporter gene assay. Results of functional studies in two cell lines (SH-SY5Y and SK-N-BE) found no significant difference in the functional expression of either the “A” or “C” alleles for this SNP (data not shown).
The current study expands our original data implicating core promoter polymorphisms in the α7 nicotinic acetylcholine receptor gene, CHRNA7, in susceptibility for schizophrenia. The rs3087454 (−1831 bp) polymorphism, located in a putative repressor region upstream of CHRNA7 was significantly associated with schizophrenia in both the Caucasian-Non Hispanic and African-American case-control samples. Nominal P-values for family-based analyses supported results in the case-control sample. Notably, different “risk” alleles were associated with schizophrenia in African-Americans (“A” allele) and Caucasian-Non Hispanics (“C” allele) at this biallelic marker. Associations of different alleles at the same locus with the same disease have been reported across different ethnic groups (Singleton et al., 2003; Tan et al., 2003) and within the same ethnic group (Park et al., 2002; Glatt et al., 2003; Wonodi et al., 2003; Chen et al., 2004; Sanders et al., 2005). This could be explained by population differences where heterogeneous effects of the same polymorphism are caused by differences in genetic background. Alternatively, in complex diseases such as schizophrenia, in which multiple loci act jointly to confer susceptibility for disease, associations of different alleles at the same locus could be attributed to a correlation with a causal polymorphism at another locus (Lin et al., 2007). The correlation can be due to either an interaction between the loci or to linkage disequilibrium. The finding that transcription levels were not changed in an in vitro reporter gene analysis of the rs3087454 (−1831 bp) polymorphism suggests that the results of our genetic association studies are likely due to linkage disequilibrium.
Attempts to refine the schizophrenia phenotype utilizing smoking history did not improve evidence for association, although nominal significance was noted in several comparisons. This indicates that the principal finding in this study is association with schizophrenia and not with smoking in the disorder.
Exploratory examination of CHRNA7 proximal promoter polymorphisms and their relationship with rs3087454 (−1831 bp) suggests that rare CHRNA7 promoter polymorphisms may occur more frequently in individuals with the rs3087454 (−1831 bp) risk genotype and could be relevant to schizophrenia as a haplotype with rs3087454 (−1831 bp) or as single markers in some individuals.
Exclusion of the COS subjects in the Caucasian case-control sample resulted in a more significant P-value for both allelic and genotypic models of association with schizophrenia. It is, therefore, likely that the rs3087454 “risk” allele is more relevant to the etiology of adult onset schizophrenia.
Results of this study continue to support the α7 nicotinic receptor gene (CHRNA7) as a candidate gene for schizophrenia. Future studies to investigate the CHRNA7 gene will attempt to capture all genetic variance in the linkage peak region by fine mapping with a combination of SNPs and microsatellite markers. Such studies could uncover a causal SNP, correlated with rs3087454, and reveal whether variance in the candidate gene CHRNA7, its upstream regulatory regions, or a relevant SNP in the partial duplication (CHRFAM7A) is associated with schizophrenia.
Role of funding source
This research was supported by National Institutes of Health grants DA09457 and MH081177 (SL), MH068582 (RF), and the Veterans Affairs Medical Research Service (SL, RF).
We acknowledge the valuable contributions of clinical information and blood by all the participants in the genetic study. We would like to thank Sarah Bulman, Philippa Kirby, Bernadette Sullivan, and Shannon Straub for technical assistance.
ContributorsDr. Leonard designed the study and edited the manuscript. Dr. Sarah Stephens wrote the first draft of the paper and Dr. Sharon Graw provided editorial assistance. Dr. Sarah Stephens, Alexis Franks, Judith Logel, Amanda Barton, Jane Dickenson, and Benjamin James performed the genotyping. Jessica Schultz and Margaret Short performed the functional studies. Drs. Tasha Fingerlin and Brandie Wagner provided statistical expertise. Dr. Colin Hodgkinson graciously shared his admixture markers. Drs. Randal Ross and Robert Freedman recruited the Denver patients participating in the study.
Confiict of interest
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.schres.2008.12.017.