Skip Navigation


Human Molecular Genetics Advance Access originally published online on April 27, 2007
Human Molecular Genetics 2007 16(13):1557-1568; doi:10.1093/hmg/ddm104
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
16/13/1557    most recent
ddm104v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (1)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Luo, X.
Right arrow Articles by Gelernter, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Luo, X.
Right arrow Articles by Gelernter, J.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2007. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

CHRM2 variation predisposes to personality traits of agreeableness and conscientiousness

Xingguang Luo1,2,*, Henry R. Kranzler3, Lingjun Zuo1,2, Huiping Zhang1,2, Shuang Wang4 and Joel Gelernter1,2

1 Department of Psychiatry, Yale University School of Medicine, New Haven, CT, USA, 2 VA Connecticut Healthcare System, West Haven Campus, CT, USA, 3 Department of Psychiatry, Alcohol Research Center, University of Connecticut School of Medicine, Farmington, CT, USA and 4 Department of Biostatistics, Mailman School of Public Health, Columbia University, New York, NY, USA

* To whom correspondence should be addressed at: Yale University School of Medicine, VA Psychiatry 116A2, 950 Campbell Avenue, West Haven, CT 06516, USA. Tel: +1 2039325711 ext. 3590; Fax: +1 2039374741; Email: xingguang.luo{at}yale.edu

Received December 18, 2006; Accepted April 13, 2007


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
Personality traits are among the most complex quantitative traits. Certain personality traits have been postulated to be part of the inherited component of substance dependence (SD) risk. Association between the M2 cholinergic receptor gene (CHRM2) and SD has recently been reported and replicated (Wang et al. Hum. Mol. Genet. (2004);13:1903–1911; Luo et al. Hum. Mol. Genet. 2005;14:2421–2434). In this study, we investigated the relationship between CHRM2 variation and personality traits in two American populations. We assessed dimensions of the five-factor model of personality, and genotyped six CHRM2 markers and 38 unlinked ancestry-informative markers in 239 subjects with SD [173 European-Americans (EAs) and 66 African-Americans (AAs)] and 275 healthy subjects (237 EAs and 38 AAs). The relationships between CHRM2 markers and personality traits were examined using multivariate analysis of covariance, controlling for marker–marker interaction effects and potential confounders. Associations were decomposed by Roy Bargmann stepdown analysis of covariance. Generally, substance-dependent patients, older individuals, males, and AAs scored higher on Neuroticism and lower on other personality factors. Diplotype CTCAAA/CTCGTT (P = 0.005) and the interaction between its two haplotypes (CTCAAA x CTCGTT) (P = 0.003) were associated with lower Conscientiousness scores. Haplotype CTCGAT (P = 0.006) and its interaction with haplotype TCAAAT (P = 0.002) were associated with higher Agreeableness scores. The trait-influencing variant site in CHRM2 for Agreeableness was close to marker rs1824024 (SNP3) (P = 0.002). CHRM2 variation may contribute to the genetic component of variation in personality traits. Personality traits might substantially underlie the heritable component of SD.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
Quantitative traits, including biological measures such as height and blood pressure, personality traits, and the severity of symptoms and their reduction by treatment, are common phenotypes of interest in genetics. Such traits are genetically complex, usually governed by several genetic loci (i.e. they are polygenic), and are also influenced by environmental factors. Humans have hundreds of personality traits (1), and every personality trait is quantitative and genetically complex. Because personality traits may have both main and interactive effects, they are among the most complex quantitative traits. A commonly used approach to the measurement of human personality traits is based on the well-accepted five-factor model (1), in which the five basic dimensions of personality are Extraversion, Agreeableness, Conscientiousness, Neuroticism and Openness to Experience (i.e. Factors I–V). Although these five major domains of personality describe personality from different perspectives, they are significantly correlated when taken in pairs (all P < 0.001). Neuroticism is negatively correlated with the other four personality factors (r < 0), which among themselves are positively correlated (r > 0) (2). The five-factor model can be operationalized via the NEO Five-Factor inventory (NEO-FFI) or NEO Personality Inventory-Revised (NEO PI-R) (3).

Many studies have provided evidence that personality characteristics may play a central role in the development of substance dependence (SD) [which in those studies include alcohol dependence (AD) and opioid dependence (OD)]. Many other studies suggest that common genetic factors may underlie some portion of the association between personality traits and SD (mainly AD) (reviewed in 2). We previously demonstrated that variation in the gene encoding alcohol dehydrogenase 4 (ADH4) predisposed to both SD [which includes AD, OD and cocaine dependence (CD)] and personality factors (2). These findings highlight the potential role of personality as a genetically determined risk factor for SD.

Genetic factors have been consistently implicated as contributing to individual differences in major dimensions of personality traits per se (i.e. independent of a genetic component of SD), with about one-third to one-half of the variation in personality typically being attributed to such factors (4–8). Twin studies have shown substantial heritability (median = 48%) for personality traits (9–11); additive genetic effects accounted for 25–65% of the reliable specific variance of the facet-level personality traits (12). A genomewide linkage scan mapped five trait-influencing sites for Neuroticism at chromosomes 1q, 4q, 7p, 12q and 13q (13). Varying degrees of direct evidence have also been found regarding the association between personality traits and specific genes, including SLC6A4, HTR2C, HTR2A, MAOA, DRD2, DRD3, DRD4, COMT, AP2B1, ESR1 and ADH4 (reviewed in 2). This evidence supports the hypothesis that personality traits are multigenic and that different personality traits may be determined by different genes. However, these findings have not always been confirmed (14), one reason for which might be that commonly used analytic methods [(e.g. t-test and analysis of variance (ANOVA)] are not sufficiently powerful and reliable (see Materials and Methods for further discussion of this issue). On the basis of the hypotheses of a strong link between personality and SD and the role of cholinergic activity in both SD and personality, the present study, using a series of novel, powerful, and reliable analytic approaches, focused on exploring the role of CHRM2 in determining personality traits.

Choline (CH), a ß-fatty acid, is primarily involved in the production of neurotransmitters in the brain that regulate mood, appetite, behavior and memory. Studies indicate that it can improve cognitive performance. Acetylcholine (ACh) is synthesized from choline and acetate, a reaction catalyzed by the enzymes acetyl coenzyme A (Acetyl-CoA) and choline acetyltransferase (ChAT); it is catabolized into choline and acetate by acetylcholinesterase (AChE). Acetylcholine receptors (AChR) mediate the response to ACh, which includes a series of intracellular and intranuclear events. These cholinergic events could directly or indirectly (via dopaminergic or adrenergic effects, or cognition) underlie some of the behavioral effects of drugs of abuse, thereby contributing to the risk of SD (including CD, OD and/or AD in the present study), and certain personality traits.

Cocaine, opioids and alcohol may directly affect cholinergic activity via influencing AChE activity or modulating ACh release, which may be involved in the development of SD. AChE inhibitors suppressed both cocaine- and morphine-induced conditioned place preference and blocked the induction and persistence of cocaine-evoked hyperlocomotion, preventing long-lasting behavioral abnormalities associated with cocaine and morphine addiction by potentiating the actions of ACh released from nucleus accumbens (NAc) cholinergic neurons (15). Opioid receptors stimulated by agonists can depress ACh release and cause a reduction of cholinergic activity (16), which can be reversed by the corresponding antagonists. For example, the non-selective µ-opioid receptor agonist, morphine, inhibits ACh release, which is reversed by low concentrations of the non-selective antagonists, naloxone and naltrexone (17). Further, the {kappa}-opioid receptor agonist ethylketocyclazocine inhibits ACh release, which is reversed by the selective antagonist, Mr2266 (18). Additionally, the {delta}-opioid receptor agonists [Leu5]enkephalin and [D-Ala2, D-Leu5]enkephalin selectively inhibit ACh release (19). At high concentration, eseroline, an AChE antagonist, had both opioid and cholinomimetic activities (20). The cholinergic pathway is also considered to be important in the development of opioid tolerance and physical dependence (21). Furthermore, cholinergic activity may indirectly modulate a variety of behaviors related to SD, by altering cholinergic–adrenergic balance and cholinergic–dopaminergic balance. For example, an increase of ACh in the NAc causing an alteration of the dopaminergic-ACh balance may be involved in the ethanol withdrawal state (22); conversely, elevated mesolimbic dopaminergic activity has been hypothesized to serve as a final common pathway for the hedonic properties of a variety of abused substances.

Cholinergic activity may also be directly involved in the modulation of personality. For example, there is a relationship between some sensation seeking subscales and AChE activity (23). Greater behavioral and cardiovascular sensitivity to the cholinomimetic physostigmine was associated with more ‘neurotic’ traits (i.e. less efficient, habitually passive, ‘helpless’ strategies in stress coping) (24), ‘depressive’ personality traits (i.e. low life contentment, high irritability, feeling stressed, high emotional lability) (25) and traits of affective instability (i.e. affective instability per se, unstable relationships, identity disturbance and chronic feelings of emptiness and boredom) (26). The balance between Ach and both norepinephrine and dopamine may modulate personality as well. For example, the cholinergic–adrenergic balance appears to be related to shyness and low extraversion (27), spontaneous aggressiveness (28,29) and depressive attitudes (i.e. low life contentment, high irritability, feeling stressed, high emotional lability) (24,25). Enhanced dopaminergic activity is hypothesized to be related to (i) positive emotion, social engagement and life satisfaction, which are characteristics of extraversion (30) and (ii) novelty-seeking, which is characteristic of openness (31).

The effects of cholinergic activity on SD and personality might also be mediated, at least in part, by impaired cognition. Cholinergic pathways, in general, have been widely implicated in cognition and memory (32,33). At some point in their course, many neuropsychiatric disorders such as Alzheimer's disease, major depression, AD or drug dependence (DD) include an impairment of cognition. Further, certain personality traits have been known to contribute significantly to cognitive function, explaining 2–7% of the variance in ability across cognitive domains (34).

The muscarinic-2 cholinergic receptor (CHRM2), present in neurons throughout the central and peripheral nervous systems, cardiac and smooth muscles and a variety of exocrine glands, is predominantly a presynaptic autoreceptor, helping to inhibit ACh release from cholinergic terminals and control cAMP regulation, and is responsible for ACh-mediated inhibition of adenyl cyclase activity (35–38). CHRM2, the gene encoding the CHRM2 receptor, is located on chromosome 7q31–35. CHRM2 variation may influence transcription levels, mRNA stability, translational efficiency, density, ligand binding, affinity or other properties of the receptor. For example, some experimentally modeled functional variants within the coding sequence can lead to amino acid substitutions that may alter the properties of the receptor, including Tyr403Phe, which affects the receptor's ligand binding affinities of the receptor, and four other amino acid substitutions at Val385, Thr386, Ile389 and Leu390, which are essential for G-protein coupling specificity and G-protein activation (reviewed in 39). Genetic variation at CHRM2 may thus be expected to influence cholinergic activity and result in phenotypic variation.

Some of the individual genetic variability in susceptibility to the development of SD may be due to polymorphism at CHRM2, as demonstrated by two recent studies (39,40). We (39) initially examined six CHRM2 markers (Table 1) in 871 subjects, including 333 healthy controls [287 European-Americans (EAs) and 46 African-Americans (AAs)] and 538 SD subjects (415 with AD and 346 with DD; 382 EAs and 156 AAs). These CHRM2 markers were located in several haplotype blocks, i.e. SNPs 1–3 in the first block, and SNPs 4 and 5 in the second block. Genotype frequency distributions of all markers were in Hardy–Weinberg Equilibrium (HWE) in healthy controls; however, SNP3 (rs1824024) was in significant Hardy–Weinberg Disequilibrium (HWD) in cases (P = 0.010 in EA DD, P = 0.010 in AA AD and P = 0.008 in AA DD) even after correction for multiple tests, which indicates an association of SD with SNP3. Regression analysis showed that the alleles and genotypes of SNPs 1 and 2, which were in linkage disequilibrium (LD) with SNP3, and some specific haplotypes and diplotypes that contained SNP3, were also significantly associated with AD and/or DD.


View this table:
[in this window]
[in a new window]

  		Table 1. CHRM2 marker information

 
DD is one of the most common diseases comorbid with AD. Many studies have demonstrated that DD has a number of features in common with AD, including clinical symptomatology, neuropsychological impairments, hypothesized pathogenetic mechanisms and response to specific treatments (41). Further, DD has been reported to share some susceptibility genes with AD; for example, in our previous report that the CHRM2 gene affected risk for both (39). On the basis of these findings, in the present study, we examined AD and DD jointly as one phenotype.

Personality traits may plausibly also be associated with CHRM2, due to their strong physiologic (and in some cases, genetic) link to SD (reviewed earlier), and due to their strong link to the cholinergic signaling system (reviewed earlier), although there is no report about the direct links between personality traits and CHRM2 polymorphisms so far. In the context of this theoretical framework, one of the aims of the present study was to fine-map the contributory loci for personality factors and to investigate whether variation in CHRM2 is a common genetic factor underlying both personality traits and SD; a second aim was to see whether the gene-personality association is stronger than the gene-disease association in this sample.

Personality may develop with age (in fact, certain changes are predictable, on a group level), and it is also sex-related (42). Allele frequency is known to be population-specific; and the disease-related allele frequency may be phenotype-specific. In previous studies, we demonstrated that the associations between ADH4, OPRM1 and personality traits could be modified by affection status, age, sex, and/or ethnicity (2,43). Generally, SD patients, older individuals, AAs and males had significantly lower scores on Extraversion, Agreeableness, Conscientiousness, Openness to Experience and/or significantly higher Neuroticism scores (2). In the present study, we therefore included SD patients and healthy subjects, older and younger adults, EAs and AAs and men and women in our sample, and took into account statistically the influence of these factors, which included the confounding effects of age and sex, and population stratification and admixture effects on the association between CHRM2 variation and personality traits. We predicted the presence of different gene effects on personality scores in different subgroups based on sex, age, SD diagnosis or population.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
There were nominally significant differences in genotype and/or allele frequencies between affected and healthy subjects for many markers: in AAs SNP2 (pg = 0.035 for genotypewise analysis), SNP4 (pa=0.014 for allelewise analysis), and SNP6 (pa=0.018). After correction for multiple testing using SNPSpD (a Bonferroni-type correction that takes marker correlation into account (44)), the genotype and allele frequencies of these markers were only suggestively different between affected and healthy subjects (Table 2). There were also significant differences in haplotype and/or diplotype frequencies between patients and healthy subjects (data not shown). These findings were basically consistent with our previous report using a larger sample (n = 871) (39) from which the present sample was drawn, where there were suggestively significant differences in genotype and/or allele frequencies between cases and healthy subjects for more markers, including SNP1 (Pg=0.049) and SNP2 (Pa=0.023) in AAs for SD and SNP3 (Pa=0.018, Pg=0.027) in EAs for drug dependence-only (i.e. in the absence of AD) and there were significant differences in haplotype and diplotype frequencies between SD and controls in both populations.


View this table:
[in this window]
[in a new window]

  		Table 2. Genotype and allele frequencies of CHRM2 markers

 
No markers were in significant HWD in controls (except for SNP4 with nominal P = 0.029). There were many markers in nominally-significant HWD in cases (which could indicate valid marker-disease associations): SNP1 (P = 0.019) in EAs, and SNP2 (P = 0.015) and SNP3 (P = 0.013) in AAs. After correction for multiple testing using SNPSpD, these markers were only suggestive of HWD (Table 2), consistent with our previous report, where more markers were in significant HWD (39; also see Introduction).

Two ancestries, i.e. European and African, were detected in our sample, and the ancestry proportions for each individual were estimated, as reported previously (39). The estimate of the extent of admixture is a function of the exact populations used in the STRUCTURE analysis (45,46), and of the marker set used, as well as of ‘actual’ ancestry proportion.

t-test and ANOVA showed nominally significant differences in individual personality factor scores by population, age, sex, and/or affection status (data not shown), completely consistent with our previous study (2). There were also significant differences in individual personality factor scores in subjects with varying alleles or genotypes of many markers, in the whole sample and in subgroups divided by population, age, sex or affection status (Table 3).


View this table:
[in this window]
[in a new window]

  		Table 3. P-values for associations between personality factors and markers (from ANOVAs or t-tests)

 
MANCOVA indicated that personality traits were related to affection status, age, ancestry and sex; and the multi-locus haplotypes and diplotypes and single-locus markers had main or interaction effects on personality traits, with some effects modified by age, affection status or sex (Table 4).


View this table:
[in this window]
[in a new window]

  		Table 4. P-values for MANCOVAs on the relationships between composite personality trait and CHRM2 gene

 
Haplotypewise, diplotypewise, allelewise and genotypewise MANCOVAs showed significant associations between the personality traits and age, affection status, ancestry and sex (1.5 x 0–65≤P ≤ 1.8 x 0–4; Table 4).

Haplotypewise MANCOVA showed that the haplotypes TCAAAT (P = 0.017) and CTCAAA (P = 0.022) had significant main effects on the composite personality trait. Age x CTCAAA (P = 0.035), diagnosis x CTCGAT (P = 0.044), TCAAAA x TCAAAT (P = 0.042), TCAAAT x CTCGAT (P = 0.004), CTCGTT x TCAAAT (P = 0.032), CTCGTT x CTCAAA (P = 0.012) and TCAGTT x CTCAAT (P = 0.026) had significant interaction effects on the composite personality trait. Diplotypewise MANCOVA showed that the diplotype CTCAAA/CTCGTT (P = 0.006) had significant main effects on the composite personality trait. Most of the contributory haplotypes and diplotypes contained the haplotypes TCAAAT (f = 0.026) and CTCAAA (f = 0.045), and some contained CTCGAT (f = 0.016) and CTCGTT (f = 0.173). Allelewise MANCOVA showed that the alleles of SNP3 had significant effects on the composite personality trait (P = 0.025), modified by sex. Genotypewise MANCOVA showed no significant effects for any marker.

Roy Bargmann Stepdown Analyses showed that different personality traits were related to affection status, age, ancestry, and/or sex, and that the haplotypes, diplotypes and markers had main effects and/or interaction effects on different personality factors (Table 5).


View this table:
[in this window]
[in a new window]

  		Table 5. Roy Bargmann Stepdown ANCOVAs on the relationships between every personality factor and predictor variables

 
The haplotypewise, diplotypewise, allelewise and genotypewise Roy Bargmann Stepdown Analyses showed that different personality traits were related to affection status, age, ancestry and/or sex. Agreeableness, Conscientiousness and Openness scores were significantly lower (ß < 0), and Neuroticism scores were significantly higher (ß > 0) in patients than in healthy subjects (4.1 x 10–71≤P ≤ 0.003); Agreeableness (4.7 x 10–7≤P ≤ 0.006) and Neuroticism (2.1 x 10–9≤P ≤ 1.1 x 10–4) were significantly lower in males than in females; Extraversion, Neuroticism and Openness to Experience were significantly inversely (ß < 0), but Agreeableness was significantly positively (ß > 0) associated with age (5.0 x 10–16≤P ≤ 0.011); Extraversion, Agreeableness, Neuroticism, and Openness to Experience were positively related to European ancestry (ß > 0; 1.0 x 10–6≤P ≤ 0.010). These findings are completely consistent with our previous findings (2).

Haplotypewise Roy Bargmann Stepdown Analyses showed that the haplotypes TCAAAT and CTCGAT had a significant interaction effect on Agreeableness scores (ß > 0, P = 0.006), and CTCAAA and CTCGTT had a significant interaction effect on Conscientiousness scores (ß < 0, P = 0.003); the haplotype CTCGAT had significant effects on Agreeableness scores (P = 0.002), which was modified by affection status; that is, this haplotype increased Agreeableness scores in patients (ß > 0, P = 0.008), but decreased them in healthy subjects (ß < 0, P = 0.018). Diplotypewise Roy Bargmann Stepdown Analyses showed that the diplotype CTCAAA/CTCGTT significantly decreased Conscientiousness scores (ß < 0, P = 0.005), consistent with the above interaction effect between its two haplotypes.

Allelewise Roy Bargmann Stepdown analyses showed that SNP3 (rs1824024) had a significant effect on Agreeableness scores (P = 0.002), modified by sex; that is, the A-allele of this marker was associated with higher Agreeableness scores in males (ß > 0, P = 0.005) and lower Agreeableness scores in females (ß < 0, P = 0.030). Genotypewise Roy Bargmann Stepdown analyses showed no association between genotypes and personality factors.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
MATERIALS AND METHODS
 REFERENCES
 
The findings in the present study suggest that CHRM2 variation may play an important role in the development of personality traits. Further, personality traits might be a substantial heritable component of SD, and the two phenotypes could have intrinsically-related neurobiological mechanisms.

The simplest and most straightforward analyses in the present study, t-test and ANOVA, showed that the personality factors were significantly associated with affection status, age, sex and population (consistent with the findings of hundreds of previous studies, including our own (2)), and nominally with alleles and genotypes of at least five CHRM2 markers. However, these commonly used analyses have some disadvantages that are not present in MANCOVA. Although given these advantages, MANCOVA should theoretically be able to generate reliable and powerful results; for genetically complex traits like personality, the power of MANCOVA is also limited. Personality is a quantitative trait that is genetically complex and multigenic—different genes produce additive or epistatic effects on phenotype, but each risk gene usually exerts minor effects. In the present study, hundreds of personality traits were represented by five personality factors, which were combined into a composite phenotype using MANCOVA. It is unlikely that this composite trait is strongly determined by a single gene. Nonetheless, we did observe significant associations in the present study, and the modest effects (e.g. 0.01 < P < 0.05) obtained with MANCOVA are reasonable and sufficient to suggest an underlying genetic factor for personality traits. Additionally, although the main effect from each haplotype is small, the interaction effect of two haplotypes could be greater, since they may exert multiplicative effects, as was observed in the present study.

To decompose the composite personality trait set into separate personality factors, MANCOVA was followed by Roy Bargmann Stepdown ANCOVA. Although the results obtained using the Roy Bargmann Stepdown ANCOVA differed from those obtained with t-test and ANCOVA (Table 3 versus Table 5), this analysis has most of the advantages of MANCOVA, so that the findings from Roy Bargmann Stepdown ANCOVA are also considered to be reliable. Additionally, this analysis yields regression coefficients, the sign of which makes it possible to judge the nature of the association of the gene variants with the personality factors. One of the disadvantages of this analysis is that the five personality factors are tested separately, which inflates the Type I error rate. We therefore set {alpha} = 0.01 for these ANCOVAs.

Given their analytic advantages, we draw conclusions mainly from MANCOVA and Roy Bargmann Stepdown ANCOVA. MANCOVA demonstrated that the personality traits were very strongly related to age, affection status, ancestry and sex. Decomposition using ANCOVA showed that, generally, SD patients, older individuals, AAs and males had significantly lower scores on Extraversion, Agreeableness, Conscientiousness, Openness to Experience and/or significantly higher scores on Neuroticism. Similar findings were demonstrated by other analytic approaches (such as t-test and ANOVA) in the present study. Haplotypewise and diplotypewise MANCOVAs (at the ‘whole gene’ level) showed that (i) some specific haplotypes and diplotypes, most of which contained the haplotypes TCAAAT and CTCAAA and some of which contained CTCGAT and CTCGTT, had main or interaction effects on personality (see Table 4). Decomposition using ANCOVA showed that the interaction between haplotypes TCAAAT and CTCGAT increased Agreeableness scores, but the interaction between haplotypes CTCAAA and CTCGTT decreased Conscientiousness scores; the haplotype CTCGAT increased Agreeableness scores in SD patients, but decreased them in healthy subjects; (ii) the diplotype CTCAAA/CTCGTT significantly decreased Conscientiousness scores. Diplotypewise analysis provided much more significant results than haplotypewise analysis, probably because diplotypes incorporated interaction between haplotypes, leading to more sensitive detection of the sources of variance. This also supports our prior findings that diplotypewise analysis is more powerful than haplotypewise analysis especially in the absence of HWE for markers (39,47); (iii) the inter-haplotype interaction effects were much stronger than the main effects of these haplotypes; the strongest interaction effect was between two haplotypes (i.e. TCAAAT and CTCGAT; P = 0.004). These findings suggest that CHRM2 might harbor contributory loci for personality traits and might be involved in their development (mainly for Agreeableness and Conscientiousness).

To fine-map the contributory loci within CHRM2, we performed allelewise and genotypewise MANCOVAs (at the ‘single-point’ level). These analyses showed that the contributory locus was close to SNP3 (rs1824024) and was modified by sex. Because single-locus analysis is less powerful than multi-locus haplotype or diplotype analysis, the P-values from the single-locus analysis yielded lower levels of statistical significance. Decomposition using ANCOVA showed that allele A of SNP3 was associated with higher Agreeableness scores in males (ß > 0, P = 0.005) and lower Agreeableness scores in females (ß < 0, P = 0.030). We posit that this is because higher Agreeableness scores in males are close to lower Agreeableness scores in females; both reflecting a specific range of Agreeableness scores. Additionally, the Roy Bargmann Stepdown ANCOVA demonstrated that Neuroticism was negatively correlated with the four other personality factors, which were all positively intercorrelated, consistent with results of pairwise correlation analysis (see Introduction).

From the gene-disease association analyses (conventional case–control comparison, regression analysis and HWE test) in the previous and present studies, some specific CHRM2 haplotypes, diplotypes and single markers (especially, SNP3) were associated with SD. Wang et al. (40) found that SNP3 was significantly associated with AD (P = 0.007; case–control comparison); SNP3 contributed most to the risk and protective haplotypes. In our initial study (39), case–control comparison showed a suggestively significant association between SNP3 and EA DD-only (Pa=0.018, Pg=0.027), which was the most significant of all of the markers. In that study, SNP3 was in significant HWD in cases (P = 0.010 in EA DD, P = 0.010 in AA AD and P = 0.008 in AA DD). The present study also showed that SNP3 was in LD with SNP1 (rs978437) and SNP2 (rs1455858), both of which are also risk markers; and SNP3 was in significant HWD in AA SD patients (P = 0.013, which was the most significant of all of the markers). In the present study, the contributory locus for personality traits was close to SNP3; we therefore conclude that personality traits and SD have a partially shared genetic basis, consistent with the extant literature, including our previous study (2). This is plausible (i) because personality traits might have some underlying neurobiological mechanisms that are intrinsically related to those influencing risk for SD, mediated by cholinergic activity (as discussed in the Introduction), or (ii) because, as postulated by many other investigators, personality is part of the heritable component in SD. In other words, certain personality traits are inherited and increase the likelihood that subjects who have them will develop SD or SD-related disorders; thus, personality features may be more clearly genetic in their origin than SD per se. Some opposite personality characteristics predispose to SD and depression, e.g. subjects with high Extraversion or Openness scores are more likely to develop SD (authors' unpublished data), but subjects with low Extraversion and Openness scores are more likely to develop depression, which could partially explain why the opposite phases of CHRM2 variation predispose to SD and depression (39,40). Similarly, personality features in subjects who have a psychiatric disorder may be associated with any gene related to that psychiatric disorder. For example, Neuroticism in subjects with major depression was associated with a promoter polymorphism (5-HTTLPR) in the serotonin transporter gene (SLC6A4), which was also associated with major depression per se (48). On the basis of these findings, we postulate that personality traits are a substantial heritable core for some psychiatric disorders, including SD.

We previously reported a gene-personality association study that was more robust than the corresponding gene-disease association study, in that significant effects were detected even with only half the sample size (2). The present study leads to the same conclusion. In addition to the reasons described earlier, this may be because the study of quantitative traits (e.g. personality) yields more power than the study of qualitative traits (e.g. affection status) in association studies. On the other hand, the strong gene-personality associations in the present study support the gene-disease associations in our initial study (39), although our initial findings were only suggestive (P-values close to 0.05). Based on this finding, in the future, the more sensitive gene-personality association study designs might reasonably be used to detect and predict gene-disease associations, in which personality serves as an intermediate phenotype, or a factor that influences risk of the disease in question.

Genetic markers are consistent throughout a person's life. Personality traits are also relatively invariant over the adult life span (or they show predictable patterns of change, e.g. certain features tend to become more pronounced with age). It is, therefore, reasonable to consider that an association may exist between genetic markers and personality traits, and that genetic variation may underlie variation in personality traits. CHRM2 variation may affect the properties of the CHRM2 receptor (reviewed earlier), and thus influence cholinergic activity and cholinergic–dopaminergic and cholinergic–adrenergic balance, which might contribute to the development of personality (as discussed in Introduction).

In conclusion, based on findings from the present study, CHRM2 appears to contribute to the genetic component of variation in personality traits. Personality traits might be a substantial heritable component of SD. Common neurobiological mechanisms for the contributions of CHRM2 variation to both phenotypes are presumed to be related to cholinergic activity, and the balance of ACh with dopamine and norepinephrine. Validation of these findings is required in subsequent studies, preferably of large study samples that permit separate examination of these findings in groups differentiated by age, sex, population and affection status.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
REFERENCES
 
Subjects
Two hundred and thirty-nine subjects with SD (173 EAs and 66 AAs) and 275 healthy subjects (237 EAs and 38 AAs) were included in the present study. This is a subsample of that used in the study by Luo et al. (39), and is the same sample as that used in the study by Luo et al. (2). The patients (138 males; 101 females) met lifetime DSM-III-R criteria (49) for a diagnosis of AD (n = 188) and/or DD (n = 137 with CD; n = 92 with OD). Diagnoses were made using the Structured Clinical Interview for DSM-III-R (SCID) (50), the computerized Diagnostic Interview Schedule for DSM-III-R (C-DIS-R) (51) or a checklist comprising DSM-III-R symptoms. The healthy subjects (101 males; 174 females) were screened using the SCID or the C-DIS-R to exclude major Axis I disorders, including AD or DD, psychotic disorders (including schizophrenia or schizophrenia-like disorders), mood disorders and major anxiety disorders. The average ages were 38.0 ± 9.3 years for patients and 27.8 ± 8.5 years for healthy subjects. Population assignment was done empirically by ancestry proportions estimated on the basis of ancestry-informative marker (AIM) genotypes (see in what follows).

Subjects were recruited at the University of Connecticut Health Center. All subjects gave informed consent before participating in the study, which was approved by the Institutional Review Board.

Marker inclusion
Following extraction of genomic DNA from peripheral blood by standard methods, six markers mapping within and flanking the CHRM2 coding sequence (see Table 1) and 38 AIMs unlinked to CHRM2, including 37 short tandem repeats (STRs) and one Duffy antigen gene (FY) marker (rs2814778) (39,47,52) were included. The FY marker is highly informative for distinguishing between European and African ancestry. The FY(–) allele frequency has been reported to be close to 0 in Europeans, and close to 1.0 in sub-Saharan Africans (53). These markers were genotyped previously (39). Six CHRM2 markers were located in intron 3, intron 4, intron 5 and the 3'-UTR, with an average inter-marker distance of 15 kb. The markers were selected because, at the time of genotyping, they were all available as validated assays from Applied Biosystem, Inc. (ABI, Foster City, CA, USA), and most of these SNPs are tagSNPs in the HapMap database (r2=1, MAF > 0.2).

Assessment of personality
The NEO-FFI (3) was used to assess personality dimensions in patients and healthy individuals. The NEO-FFI is a 60-item self-report questionnaire that is rated on a five-point scale to yield scores in five major domains of personality (Extraversion, Agreeableness, Conscientiousness, Neuroticism and Openness to Experience) and requires a sixth-grade reading level. Internal consistency values (Cronbach's alpha coefficient) range from 0.74 to 0.89 (3). The personality scores for different factors in our sample were: Extraversion (4–48, mean 28.7 ± 6.9), Agreeableness (10–46, mean 31.0 ± 6.6), Conscientiousness (0–48, mean 31.8 ± 7.8), Neuroticism (2–48, mean 20.4 ± 9.4) and Openness to Experience (9–45, mean 28.4 ± 6.5). These five factors are significantly intercorrelated. Every personality factor and the linear combination of all personality factors are normally distributed (2).

Data analysis
Gene-disease association analysis
The allele and genotype frequencies in different populations are shown in Table 2. Associations between alleles, genotypes, haplotypes, diplotypes and disease were analyzed by comparing the frequency distributions between affected and healthy subjects (within EAs and AAs) with exact tests implemented in the program PowerMarker (54) (for allelewise and genotypewise analyses) or logistic regression analysis (for haplotypewise and diplotypewise analyses) [details described by Luo et al. (2)].

HWE of the genotype frequency distribution for each marker was tested within each population, and separately in patients and healthy subjects, via the application of the program PowerMarker. HWD in cases, i.e. deviation from HWE expectations, can indicate a valid gene-disease association and sometimes can be more powerful than case–control comparison (41,55). Further, the HWD test is a prerequisite for deciding which software programs can be used to reconstruct haplotypes (because some approaches require the assumption of HWE).

Ancestry proportion estimation
To detect admixture and measure its magnitude in an admixed population, we used the ancestry information content from a set of 38 AIMs (39,47,52). The optimal number of ancestries and the ancestry proportions for an individual can be estimated by the program STRUCTURE (45,46) using these 38 AIMs. We have shown that this marker set is sufficient to the task (52). The ancestry proportion scores were entered as covariates into the following general linear models (including MANCOVA and ANCOVA) to exclude population stratification and admixture effects in analyses in which EAs and AAs were considered together.

Individual haplotype and diplotype probability estimation
The program PHASE (56,57), which is not subject to the HWE assumption, was used to reconstruct haplotypes and estimate the probabilities of all likely pairs of haplotypes (i.e. diplotypes) for every individual. The individual phased diplotypes that can be unambiguously inferred by PHASE have a probability of 1.0; the unphased diplotypes that were ambiguously inferred by PHASE have probabilities between 0.0 and 1.0. These haplotype and diplotype probabilities were entered into the general linear model described in what follows, for analysis. Haplotypes were reconstructed within the ‘genetic’ EAs (European ancestry proportion > 0.5) and AAs (African ancestry proportion > 0.5), respectively, rather than within the self-reported EAs and AAs. Haplotype and diplotype frequencies are shown in Table 6.


View this table:
[in this window]
[in a new window]

  		Table 6. Frequencies of common haplotypes and diplotypes

 
t-test and one-way ANOVA
To test the difference on the scores for every personality factor between different alleles or genotypes of every marker, t-tests (for allelewise analysis) and one-way ANOVA (for genotypewise analysis) were applied, in the whole sample and within each subgroup divided by population, sex, age and affection status. As in our previous study (2), age categories were defined as ‘older’ (>32 years, the mean age in our sample) or ‘younger’ (≤32 years). The diagnostic phenotypes include affected and healthy individuals.

There are limitations of analyses based on t-test and ANOVA, which are likely to result in information loss and bias in the findings. These limitations result from the fact that, in the whole sample (or in each subgroup divided by covariates), the correlation among personality factors is ignored, the interactions between genetic markers (many markers are in strong LD) are ignored, the confounding effects of all of the covariates (or other covariates) including population, sex, age and affection status are ignored, and the transformation of ancestry and age from continuous to categorical traits is arbitrary. Given these disadvantages, in the present study, t-test and ANOVA analyses are considered exploratory.

Stepwise multivariate analysis of covariance
Multivariate analysis of covariance (MANCOVA), implemented in SPSS 14.0, is the core analytic method used in this study to test associations between genes and personality traits. MANCOVA is the preferred analytic approach for a number of reasons [for more details see Luo et al. (2)]. First, considering all of the personality factors together as one observation in a single MANCOVA model accounts for the correlations between different personality factors, incorporates the overlap among them and requires fewer repetitive tests than a series of ANCOVAs. This reduces the familywise error rate ({alpha}FW, an {alpha} inflation or a cumulative Type I error that increases with the number of separate ANCOVAs) and thus offers extra power and reliability. Second, data from subjects with different affection status, age, ethnicity and sex can be included in a single model to increase the sample size and statistical power. Confounding effects from these variables can be controlled. Different predictor variables can be entered into a single model, to avoid multiple tests and information loss. Third, the interactions between genetic markers, the analysis of which has been demonstrated to be a more powerful approach than single locus analysis (58), can be taken into account in MANCOVA. Fourth, MANCOVA allows diplotype phase to be uncertain (in the present study, the proportion of individuals with unambiguous diplotypes, i.e. probability = 1) was 28.3%. A multi-locus haplotype or diplotype usually is more informative and closer to representing a ‘whole gene’ than any single marker, because haplotypes and diplotypes contain the information of multiple markers and probably unknown markers as well. Additionally, the probability of unphased haplotypes and diplotypes usually preserves more information than categorical data. Furthermore, haplotypes or diplotypes incorporate the information on LD among markers; therefore, the multiple-way interaction effects between markers can be incorporated in haplotypewise and diplotypewise MANCOVAs. Fifth, MANCOVA can control for population stratification and admixture effects on association analysis. Sixth, diplotypewise MANCOVA allows analysis even in the presence of deviation from HWE (when diplotypes are reconstructed by the programs which obviate the need for HWE).

In the haplotypewise and diplotypewise MANCOVA models, five correlated personality factors served as one composite dependent variable; haplotype or diplotype probabilities served as predictor variables; affection status, age, ancestry proportions and sex served as covariates. Two-way interactions between the predictor variables and covariates and between any two haplotypes were also considered as independent predictor variables in this model. Statistical significance was evaluated using the Pillai's Trace statistic (59). When positive findings were obtained from the haplotypewise or diplotypewise analysis (which suggested that the contributory loci were harbored within the haplotype block), allelewise and genotypewise MANCOVA was performed (using allele or genotype data) to fine-map the specific loci contributing to personality. Alleles or genotypes of each marker, instead of haplotype or diplotype probabilities, served as predictor variables in these analyses, in which two-way interactions between the predictor variables and covariates and between any two markers were also considered as independent predictor variables. All of the MANCOVAs were run as backward stepwise analyses with only statistically significant variables (i.e. P < 0.05) retained in the final equations.

Stepwise Roy Bargmann stepdown analysis (60)
A series of univariate ANCOVA tests was employed to assess the unique contribution of each personality factor to the associations involving covariates and markers. In these models, each personality factor served as one dependent variable in a univariate ANCOVA test in which predictor variables, covariates and interaction variables were the same as those in the MANCOVA models discussed earlier. These ANCOVA tests prioritized the theoretical importance of the five personality factors in the associations with CHRM2 and resulted in the same priority order as in our previous study (which included sample from which the present sample was drawn), which provided the criteria for deciding the priority (2): Neuroticism > Agreeableness > Conscientiousness > Extraversion > Openness.

When positive findings were obtained with MANCOVA, we used univariate ANCOVA to test each personality factor in the order of their priority, with all higher-priority personality factors as covariates. Other covariates and predictor variables were those retained in the final MANCOVA equations. This stepdown process decomposed the findings from MANCOVA without neglecting the correlations among the personality factors.

To consider the inflation of Type I error due to multiple tests using Roy Bargmann Stepdown Analysis, {alpha} was set at 0.01. For each Roy Bargmann ANCOVA, a backward stepwise process was applied, so that only the statistically significant (P < 0.01) variables were retained in the final equations.

We used the ‘parameter estimate’ option implemented in ANCOVA to calculate the regression coefficient (ß) for each variable in the final ANCOVA equations. From the signs of the ßs, it can be judged whether the variables are ‘risk’ or ‘protective’ factors, and which specific allele or genotype of a marker is the ‘risk allele (or genotype)’ for the personality factor. This decomposition analysis is preferable to t-test and ANOVA, because it controls for the confounding effects of other contributing variables when decomposing one of the contributing variables.


   ACKNOWLEDGEMENTS
 
Ann Marie Lacobelle provided excellent technical assistance. The constructive comments of two anonymous reviewers are highly appreciated. This work was supported in part by NIH grants R01-DA12849, R01-DA12690, R01-AA016015, K24-DA15105, R01-AA11330, P50-AA12870, K08-AA13732, K24-AA13736, K02-MH01387 and M01-RR06192 (University of Connecticut General Clinical Research Center), by funds from the U.S. Department of Veterans Affairs [the VA Medical Research Program, and the VA Connecticut–Massachusetts Mental Illness Research, Education and Clinical Center (MIRECC) and the VA Research Enhancement Award Program (REAP) research center] and Alcoholic Beverage Medical Research Foundation (ABMRF) grant award R06932 [GenBank] (X.L.).

Conflict of Interest statement. None declared.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 

  1. Digman J.M. Personality structure: emergence of the five-factor model. Annu. Rev. Psychol. (1990) 41:417–440.[ISI]

  2. Luo X., Kranzler H.R., Zuo L., Wang S., Gelernter J. Personality traits of agreeableness and extraversion are associated with ADH4 variation. Biol. Psychiatry (2007) 61:599–608.[CrossRef][ISI][Medline]

  3. Costa P.T., McCrae R.R. Stability and change in personality assessment: the revised NEO Personality Inventory in the year 2000. J. Pers. Assess. (1997) 68:86–94.[CrossRef][ISI][Medline]

  4. Eaves L.J., Eysenck H.J., Martin N.G. Genes, Culture and Personality: an Empirical Approach (1989) New York: Academic Press.

  5. Loehlin J.C. Genes and Environment in Personality Development (1992) 106. Newbury Park, CA: Sage. 266–279.

  6. Ebstein R.P., Benjamin J., Belmaker R.H. Personality and polymorphisms of genes involved in aminergic neurotransmission. Eur. J. Pharmacol. (2000) 410:205–214.[CrossRef][ISI][Medline]

  7. Noble E.P., Ozkaragoz T.Z., Ritchie T.L., Zhang X., Belin T.R., Sparkes R.S. D2 and D4 dopamine receptor polymorphisms and personality. Am. J. Med. Genet. (1998) 81:257–267.[CrossRef][ISI][Medline]

  8. Bouchard T.J., Loehlin J.C. Genes, evolution, and personality. Behav. Genet. (2001) 31:243–273.[CrossRef][ISI][Medline]

  9. Jang K.L., Livesley W.J., Vernon P.A. Heritability of the big five personality dimensions and their facets: a twin study. J. Pers. (1996) 64:577–591.[CrossRef][ISI][Medline]

  10. Jang K.L., Livesley W.J., Vernon P.A., Jackson D.N. Heritability of personality disorder traits: a twin study. Acta. Psychiatr. Scand. (1996) 94:438–444.[ISI][Medline]

  11. Torgersen S., Lygren S., Oien P.A., Skre I., Onstad S., Edvardsen J., Tambs K., Kringlen E. A twin study of personality disorders. Compr. Psychiatry (2000) 41:416–425.[CrossRef][ISI][Medline]

  12. Jang K.L., McCrae R.R., Angleitner A., Riemann R., Livesley W.J. Heritability of facet-level traits in a cross-cultural twin sample: support for a hierarchical model of personality. J. Pers. Soc. Psychol. (1998) 74:1556–1565.[CrossRef][ISI][Medline]

  13. Fullerton J., Cubin M., Tiwari H., Wang C., Bomhra A., Davidson S., Miller S., Fairburn C., Goodwin G., Neale M.C., et al. Linkage analysis of extremely discordant and concordant sibling pairs identifies quantitative-trait loci that influence variation in the human personality trait neuroticism. Am. J. Hum. Genet. (2003) 72:879–890.[CrossRef][ISI][Medline]

  14. Tochigi M., Umekage T., Kato C., Marui T., Otowa T., Hibino H., Otani T., Kohda K., Kato N., Sasaki T. Serotonin 2A receptor gene polymorphism and personality traits: no evidence for significant association. Psychiatr. Genet. (2005) 15:67–69.[CrossRef][ISI][Medline]

  15. Hikida T., Kitabatake Y., Pastan I., Nakanishi S. Acetylcholine enhancement in the nucleus accumbens prevents addictive behaviors of cocaine and morphine. Proc. Natl Acad. Sci. USA. (2003) 100:6169–6173.[Abstract/Free Full Text]

  16. Paton W.D. The action of morphine and related substances on contraction and on acetylcholine output of coaxially stimulated guinea-pig ileum. 1956. Br. J. Pharmacol. (1997) 120(Suppl. 4):123–131.[ISI][Medline]

  17. Beani L., Bianchi C., Siniscalchi A. The effect of naloxone on opioid-induced inhibition and facilitation of acetylcholine release in brain slices. Br. J. Pharmacol. (1982) 76:393–401.[ISI][Medline]

  18. Ahmed M.S., Schoof T., Zhou D.H., Quarles C. Kappa opioid receptors of human placental villi modulate acetylcholine release. Life Sci. (1989) 45:2383–2393.[CrossRef][ISI][Medline]

  19. Mulder A.H., Wardeh G., Hogenboom F., Frankhuyzen A.L. Kappa- and delta-opioid receptor agonists differentially inhibit striatal dopamine and acetylcholine release. Nature (1984) 308:278–280.[CrossRef][Medline]

  20. Galli A., Renzi G., Grazzini E., Bartolini R., Aiello-Malmberg P., Bartolini A. Reversible inhibition of acetylcholinesterase by eseroline, an opioid agonist structurally related to physostigmine (eserine) and morphine. Biochem. Pharmacol. (1982) 31:1233–1238.[CrossRef][ISI][Medline]

  21. Frederickson R.C. Morphine withdrawal response and central cholinergic activity. Nature (1975) 257:131–132.[CrossRef][Medline]

  22. Rada P., Johnson D.F., Lewis M.J., Hoebel B.G. In alcohol-treated rats, naloxone decreases extracellular dopamine and increases acetylcholine in the nucleus accumbens: evidence of opioid withdrawal. Pharmacol. Biochem. Behav. (2004) 79:599–605.[CrossRef][ISI][Medline]

  23. Arque J.M., Unzeta M., Torrubia R. Neurotransmitter systems and personality measurements: a study in psychosomatic patients and healthy subjects. Neuropsychobiology (1988) 19:149–157.[ISI][Medline]

  24. Fritze J., Sofic E., Muller T., Pfuller H., Lanczik M., Riederer P. Cholinergic-adrenergic balance: Part 2. Relationship between drug sensitivity and personality. Psychiatry Res. (1990) 34:271–279.[CrossRef][ISI][Medline]

  25. Fritze J., Lanczik M., Sofic E., Struck M., Riederer P. Cholinergic neurotransmission seems not to be involved in depression but possibly in personality. J. Psychiar. Neurosci. (1995) 20:39–48.

  26. Gurvits I.G., Koenigsberg H.W., Siever L.J. Neurotransmitter dysfunction in patients with borderline personality disorder. Psychiatr. Clin. North. Am. (2000) 23:27–40.[CrossRef][ISI][Medline]

  27. Stein M.B., Schork N.J., Gelernter J. A polymorphism of the ß1-adrenergic receptor is associated with shyness and low extraversion. Biol. Psychiatry (2004) 56:217–224.[CrossRef][ISI][Medline]

  28. Angst J., Clayton P. Premorbid personality of depressive, bipolar, and schizophrenic patients with special reference to suicidal issues. Compr. Psychiatry (1986) 27:511–532.[CrossRef][ISI][Medline]

  29. Moises H.W., Bering B., Muller W.E. Personality factors predisposing to depression correlate significantly negatively with M1-muscarinic and beta-adrenergic receptor densities on blood cells. Eur. Arch. Psychiatry Neurol. Sci. (1988) 237:209–217.[CrossRef][Medline]

  30. Ashby F.G., Isen A.M., Turken A.U. A neuropsychological theory of positive affect and its influence on cognition. Psychol. Rev. (1999) 106:529–550.[CrossRef][ISI][Medline]

  31. Cloninger C.R., Svrakic D.M., Przybeck T.R. A psychobiological model of temperament and character. Arch. Gen. Psychiat. (1993) 50:975–990.[Abstract]

  32. Baxter M.G., Chiba A.A. Cognitive functions of the basal forebrain. Curr. Opin. Neurobiol. (1999) 9:178–183.[CrossRef][ISI][Medline]

  33. Everitt B.J., Robbins T.W. Central cholinergic systems and cognition. Annu. Rev. Psychol. (1997) 48:649–684. Review.[CrossRef][ISI][Medline]

  34. Booth J.E., Schinka J.A., Brown L.M., Mortimer J.A., Borenstein A.R. Five-factor personality dimensions, mood states, and cognitive performance in older adults. J. Clin. Exp. Neuropsychol. (2006) 28:676–683.[CrossRef][ISI][Medline]

  35. Mash D.C., Flynn D.D., Potter L.T. Loss of M2 muscarine receptors in the cerebral cortex in Alzheimer's disease and experimental cholinergic denervation. Science (1985) 228:1115–1117.[Abstract/Free Full Text]

  36. Wichmann T., Starke K. Modulation by muscarine and opioid receptors of acetylcholine release in slices from striato-striatal grafts in the rat. Brain. Res. (1990) 510:296–302.[CrossRef][ISI][Medline]

  37. Zhou C., Fryer A.D., Jacoby D.B. Structure of the human M(2) muscarinic acetylcholine receptor gene and its promoter. Gene (2001) 271:87–92.[CrossRef][ISI][Medline]

  38. Fenech A.G., Billington C.K., Swan C., Richards S., Hunter T., Ebejer M.J., Felice A.E., Ellul-Micallef R., Hall I.P. Novel polymorphisms influencing transcription of the human CHRM2 gene in airway smooth muscle. Am. J. Respir. Cell. Mol. Biol. (2004) 30:678–686.[Abstract/Free Full Text]

  39. Luo X., Kranzler H.R., Zuo L., Wang S., Blumberg H.P., Gelernter J. CHRM2 gene predisposes to alcohol dependence, drug dependence, and affective disorders: results from an extended case-control structured association study. Hum. Mol. Genet. (2005) 14:2421–2434.[Abstract/Free Full Text]

  40. Wang J.C., Hinrichs A.L., Stock H., Budde J., Allen R., Bertelsen S., Kwon J.M., Wu W., Dick D.M., Rice J., et al. Evidence of common and specific genetic effects: association of the muscarinic acetylcholine receptor M2 (CHRM2) gene with alcohol dependence and major depressive syndrome. Hum. Mol. Genet. (2004) 13:1903–1911.[Abstract/Free Full Text]

  41. Luo X., Kranzler H.R., Zuo L., Lappalainen J., Yang B.Z., Gelernter J. ADH4 gene variation is associated with alcohol dependence and drug dependence in European Americans: results from HWD tests and case–control association studies. Neuropsychopharmacology (2006) 31:1085–1095.[CrossRef][ISI][Medline]

  42. Martin C.S., Lynch K.G., Pollock N.K., Clark D.B. Gender differences and similarities in the personality correlates of adolescent alcohol problems. Psychol. Addict. Behav. (2000) 14:121–133.[CrossRef][ISI][Medline]

  43. Hernandez-Avila C.A., Covault J., Gelernter J., Kranzler H.R. Association study of personality factors and the Asn40Asp polymorphism at the mu-opioid receptor gene (OPRM1). Psychiatr. Genet. (2004) 14:89–92.[CrossRef][ISI][Medline]

  44. Nyholt D.R. A simple correction for multiple testing for single-nucleotide polymorphisms in linkage disequilibrium with each other. Am. J. Hum. Genet. (2004) 74:765–769.[CrossRef][ISI][Medline]

  45. Pritchard J.K., Stephens M., Donnelly P. Inference of population structure using multilocus genotype data. Genetics (2000) 155:945–959.[Abstract/Free Full Text]

  46. Falush D., Stephens M., Pritchard J.K. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics (2003) 164:1567–1587.[Abstract/Free Full Text]

  47. Luo X., Kranzler H.R., Zuo L., Yang B.Z., Lappalainen J., Gelernter J. ADH4 gene variation is associated with alcohol and drug dependence in European Americans: results from family-controlled and population-structured association studies. Pharmacogenetics Genomics (2005) 15:755–768.

  48. Munafo M.R., Clark T.G., Roberts K.H., Johnstone E.C. Neuroticism mediates the association of the serotonin transporter gene with lifetime major depression. Neuropsychobiology (2006) 53:1–8.[CrossRef][ISI][Medline]

  49. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders (1987) 3rd edn. revised. Washington, DC: American Psychiatric Press.

  50. Spitzer R.L., Williams J.B.W., Gibbon M., First M.B. The structured clinical interview for DSM-III-R (SCID), I: History, rationale, and description. Arch. Gen. Psychiat. (1992) 49:624–629.[Abstract]

  51. Blouin A.G., Perez E.L., Blouin J.H. Computerized administration of the diagnostic interview schedule. Psychiat. Res. (1988) 23:335–344.[CrossRef][ISI][Medline]

  52. Yang B.Z., Zhao H., Kranzler H.R., Gelernter J. Population group assignment: effects of markers and methods. Genet. Epidemiol. (2005) 28:302–312.[CrossRef][ISI][Medline]

  53. Parra E.J., Marcini A., Akey J., Martinson J., Batzer M.A., Cooper R., Forrester T., Allison D.B., Deka R., Ferrell R.E., et al. Estimating African American admixture proportions by use of population-specific alleles. Am. J. Hum. Genet. (1998) 63:1839–1851.[CrossRef][ISI][Medline]

  54. Liu K., Muse S. PowerMarkerand new genetic data analysis software. Version 3.0. Free program distributed by the author over the internet from http://www.powermarker.net (accessed October 1, 2004).

  55. Luo X., Kranzler H.R., Zuo L., Wang S., Lappalainen J., Schork N.J., Gelernter J. Diplotype trend regression (DTR) analysis of the ADH gene cluster and ALDH2 gene: multiple significant associations for alcohol dependence. Am. J. Hum. Genet. (2006) 78:973–987.[CrossRef][ISI][Medline]

  56. Stephens M., Smith N.J., Donnelly P. A new statistical method for haplotype reconstruction from population data. Am. J. Hum. Genet. (2001) 68:978–989.[CrossRef][ISI][Medline]

  57. Stephens M., Donnelly P. A comparison of bayesian methods for haplotype reconstruction from population genotype data. Am. J. Hum. Genet. (2003) 73:1162–1169.[CrossRef][ISI][Medline]

  58. Marchini J., Donnelly P., Cardon L.R. Genome-wide strategies for detecting multiple loci that influence complex diseases. Nat. Genet. (2005) 37:413–417.[CrossRef][ISI][Medline]