The present study sought to identify genomic loci that confer risk for antisocial behavior, depression, and impulsive substance use by conducting linkage scans of three scales from the MMPI-2 that were constructed to measure these behaviors, the Antisocial Practices content scale (ASP), the Depression content scale (DEP), and the revised MacAndrew Alcoholism scale (MAC-R), respectively. Suggestive evidence of linkage was observed to three genomic regions that proved independent of alcohol and cannabis dependence diagnostic status. Specifically, the ASP scale showed evidence of linkage to chromosome 13 at 11 cM, the MAC-R scale showed evidence of linkage to chromosome 15 at 47 cM, and all 3 scales showed evidence of linkage to chromosome 17 at 57–58 cM. Of note, previous linkage scans of alcohol and cannabis dependence conducted in the SFFAM failed to detect evidence of linkage to any of these regions (Gizer et al., 2011
and Ehlers et al., 2010b
, respectively). This lack of overlap in findings is somewhat surprising given the evidence from twin studies suggesting that common genetic influences underlie to some extent the phenotypic correlations between the dependence diagnoses and antisocial behavior, impulsive behavior, and even depression (e.g., Hicks et al., 2004
; Kendler et al., 1993
; Young et al., 2006
). It is possible that that the lack of overlap in findings between the diagnoses and MMPI-2 scales could be due to the reduced power of variance components linkage analysis for dichotomous traits, especially those with lower prevalence rates such as cannabis dependence, relative to quantitative traits (Duggirala et al., 1997
). Nonetheless, the uniformly low LOD scores for the diagnostic variables in those linkage regions identified for the MMPI-2 scales (all LODs <1.0) suggest that the discrepant results are not simply the result of low power. Further, the relatively modest correlations between the dependence diagnoses and the examined MMPI-2 scales in the UCSF sample (see ) suggest that these traits are largely distinct, thus providing a compelling explanation for the lack of overlapping findings.
As described, the locus on chromosome 17 at 57–58 cM appeared to be linked to scores on all three MMPI-2 scales that were evaluated. Further, the evidence for linkage was strongest when the 3 scales were included in a multivariate analysis. It should be noted that these scales do exhibit modest correlations in clinic-referred samples despite minimal item overlap (Greene, 2000
; see for correlations in the UCSF Family Sample). Such correlations clearly demonstrate overlap in what each scale is measuring, but they also demonstrate that each scale is capturing unique information. Given the observed increase in LOD scores when all 3 scales were considered in the analysis, it appears that the unique information provided by each scale is contributing to the observed linkage signal. This would suggest that the identified region on chromosome 17 is influencing psychopathology in some general fashion rather than substance dependence, antisocial behavior, or depression specifically. As an example, twin studies suggest the presence of genetic influences that jointly predispose an individual to externalizing (e.g., antisocial behavior, substance dependence) and internalizing disorders (e.g., major depressive disorder) (Cosgrove et al., 2011
; O'Conner et al., 1998
). Low effortful control, which can manifest as high levels of reward seeking behavior as well as high levels of negative emotionality, has been suggested by some researchers as a broad risk factor for psychopathology that could account for the shared genetic influences common to externalizing and internalizing disorders (MacDonald, 2008
; Nigg, 2006
). Supporting the conclusion that this region might confer a broad risk for psychopathology, studies have reported evidence of linkage to this region with a range of psychiatric disorders including major depressive disorder (Holmans et al., 2007
; Middeldorp et al., 2009
), alcohol dependence (Hill et al., 2004
), substance dependence (Gelernter et al., 2006
), and conduct disorder (Stallings et al., 2005
; Stallings et al., 2003
This chromosome 17 locus is located in a relatively gene-rich region and includes several candidate genes that have been widely studied in relation to psychiatric disorders and related phenotypes. For example, the gene encoding for the serotonin transporter (SLC6A4
) is one of the most widely studied genes in the psychiatric genetics literature and polymorphisms of this gene have shown replicable evidence of association with disorders such as depression (Lopez-Leon et al., 2008
), alcohol dependence (McHugh et al., 2010
), ADHD (Gizer et al., 2009
), and schizophrenia (Allen et al., 2008
) in meta-analytic reviews. Additional candidate genes in the region are involved in dopaminergic (PPP1R1B/DARPP32), noradrenergic (PNMT), and cholinergic (VAT1) neurotransmission, neuronal development (MAPT, NEUROD2), and the stress response hormone system (CRHR1). Given the diversity of psychiatric phenotypes that have been related to this genomic region as well as the diversity in neurobiological systems represented by potential candidate genes in the region, a substantial amount of work will be required to refine our understanding of the observed relations; however, preliminary findings such as those described in the present report as well as those reviewed above suggest that this chromosomal region may either harbor several genetic variants that each confer risk to a specific disorder, and/or genetic variants that confer risk to psychopathology more broadly.
In addition to the locus on chromosome 17, suggestive evidence for linkage for the ASP scale to chromosome 13 at 11 cM and for the MAC-R scale to chromosome 15 at 47 cM was identified and proved to be independent of alcohol and cannabis dependence diagnostic status. Both of these loci have shown evidence of linkage to phenotypically similar traits in previous studies. For example, the locus on chromosome 13, which was linked to antisocial behavior as measured by the ASP scale, has previously shown evidence of linkage to externalizing behavior in a Native American population (Ehlers et al., 2008
) as well as tobacco use in the same Native American population (Ehlers and Wilhelmsen, 2006
) and the COGA sample (Saccone et al., 2003
). Additionally, recent genome-wide association studies have reported suggestive association signals in this region with several psychiatric disorders including mood disorders (Huang et al., 2010
), attention-deficit hyperactivity disorder (ADHD) (Anney et al., 2008
; Lasky-Su et al., 2010
; Sonuga-Barke et al., 2008
), and schizophrenia (Sullivan et al., 2008
) suggesting a relation between this region and psychiatric phenotypes. Despite this convergence of findings to the described region, it is a relatively gene-poor region without any obvious candidate genes, and the top hits reported in several of the GWAS’s described above come from intergenic regions or genes that would appear to lack a direct relation with psychiatric disorders.
The locus on chromosome 15, which was linked to impulsive substance use as assessed by the MAC-R scale in the present study, has also shown previous evidence of linkage to psychiatric phenotypes. For example, COGA has reported evidence of linkage between this region and alcohol misuse related to heightened anxiety (Dick et al., 2002
) and tobacco use (Bierut et al., 2004
), and evidence of linkage to this region for alcohol withdrawal was reported in a Native American sample (Ehlers et al., 2004
). Further, a locus approximately 50 cM telomeric has shown evidence of linkage to depression phenotypes in four previous studies (Holmans et al., 2007
; McGuffin et al., 2005
; Middeldorp et al., 2009
; Zubenko et al., 2002
). Notably, Terracciano and colleagues (2010)
recently reported that a SNP in the RORA
gene located within the linkage region reported in the present study represented the top hit in their GWAS of trait depression. This gene was also implicated in bipolar disorder using a functional genomics approach integrating data from human and animal studies (Le-Niculescu et al., 2009
). Thus, it may be that the linkage signal reported for the MAC-R in the present study is related to cyclic changes in mood that co-occur with substance misuse. Additional candidate genes in the region include two genes that encode for the glutamate receptor, ionotropic, N-methyl D-aspartate-like 1A (GCOM1, GRINL1A) and CYP19A1, which encodes for the aromatase enzyme responsible for the conversion of androgen to estrogen.
Despite the potential importance of these findings, there are limitations of the present study that should be considered. First, families in the UCSF Family Study sample were selected to be enriched for the alcohol dependence diagnosis, and as a result, it is not clear how findings from the present study will generalize to other populations. Second, it should be noted that multiple phenotypes were evaluated in the present study, though corrections were not made for multiple testing due to the generally low power of linkage analysis for complex traits. Nonetheless, we attempted to limit spurious evidence for linkage by restricting follow-up analyses (i.e., those including covariates and the multivariate analyses) to those regions of interest identified in the initial univariate analyses. Third, the scales from the MMPI-2 included in the present study, while well-studied, have not been previously included in molecular genetic studies of psychiatric disorders. Thus, replication of the reported results is needed. Nonetheless, the convergence between the reported results and studies using alternative assessment methods is promising.
In summary, the present study attempted to identify genetic loci that confer risk to antisocial behavior, depressive symptoms, and impulsive substance use and reward-seeking behavior. Suggestive evidence for linkage that was independent of alcohol and cannabis dependence diagnostic status was observed between antisocial behavior and chromosome 13 at 11 cM and between impulsive substance use and chromosome 15 at 47 cM. In addition, a region on chromosome 17 at 57–58 cM showed evidence of linkage in a multivariate analysis to antisocial behavior, depressive symptoms, and impulsive substance use. These regions have shown prior evidence of linkage and association to substance dependence and externalizing behavior as well as other psychiatric disorders such as depression suggesting potentially broad relations with psychopathology.