Association of polymorphisms in the BDNF, DRD1 and DRD3 genes with tobacco smoking in schizophrenia

Introduction

Recent evidence points to the involvement of the DRD1-BDNF-DRD3 cluster in nicotine addiction (Comings et al., 1997; Kenny et al., 2000; Lang et al., 2007; Le Foll et al., 2007; Wang et al., 2007; Huang et al., 2008a) (see Le Foll et al. (2009) for a review). It is also becoming evident that the mesocorticolimbic dopaminergic system plays an important role in nicotine addiction (Balfour et al., 2000; Balfour, 2009; Le Foll et al., 2009). Furthermore, this system mediates both the positive and negative symptoms of schizophrenia (Lieberman et al., 1994; Seeman, 2006); hence, it may form the basis of the observed comorbidity between schizophrenia and nicotine addiction (Dalack et al., 1998).

The reinforcing/rewarding effects of nicotine are mediated by the release of dopamine (DA) within the nucleus accumbens (Corrigall et al., 1992; Pontieri et al., 1996). However, the dopaminergic response of individuals to nicotine shows great variability. This variability has a strong genetic component, and a number of gene variants, which contribute to the interindividual variation in dopaminergic response, have been identified. These include variants of DA receptors and other genes within the dopaminergic pathway (Brody et al., 2006; Novak et al., 2009).

Gene variants that alter dopaminergic response to nicotine likely contribute to different patterns of smoking behaviour between individuals. In particular, the dopamine D3 receptor (DRD3) polymorphism Ser9Gly (rs6280) (Lundstrom & Turpin, 1996) is thought to confer enhanced affinity for endogenous DA (Jeanneteau et al., 2006) and thus higher responsiveness to DA effects. A recent genetic study found significant association of this polymorphism with nicotine dependence (Huang et al., 2008a). The involvement of the DRD3 gene is also supported by behavioural studies in animals, demonstrating a role of the dopamine D3 receptor in nicotine conditioning (Le Foll et al., 2003a, 2005a).

The dopamine D1 receptor (DRD1) is another gene implicated in mediating the reinforcing effects of nicotine (Spina et al., 2006). To date, only one association study of DRD1 variants with nicotine addiction has been performed. Huang et al. (2008b) found a significant association of two SNPs (rs686 and rs4532) with nicotine dependence, with the rs686 functional polymorphism likely affecting the expression of the dopamine D1 receptor.

The dopaminergic response to nicotine is also influenced by other factors, such as the brain-derived neurotrophic factor (BDNF), which controls the expression of the dopamine D3 receptor via a mechanism involving stimulation of the dopamine D1 receptor (Guillin et al., 2001). Repeated nicotine administration results in upregulation of BDNF and triggers a mechanism that modulates nicotine-dependent DA release (Kenny et al., 2000), as well as DRD3 upregulation (Le Foll et al., 2003a). Polymorphisms within the BDNF gene were found to be associated with smoking, in particular the Met/Met genotype and the Met allele of the Val66Met (rs6265) variant (Lang et al., 2007; Wang et al., 2007), but only a limited number of association studies has been performed.

We have, therefore, selected 42 SNPs within these genes for analysis of association with the risk of smoking and with the degree of smoking. We have genotyped 34 SNP variants spanning the DRD3 gene, as well as two SNPs at each end of the DRD1 gene, and six SNPs across the BDNF gene. The DRD1 SNPs selected were already shown to be significantly associated with nicotine addiction (Huang et al., 2008b).

Materials and Methods

Participants were from a community-based sample of 341 unrelated schizophrenia patients, which included 217 smokers (S) and 124 non-smokers (NS), who were recruited through the Centre for Addiction and Mental Health (CAMH) in Toronto, Canada. For association of polymorphisms with the presence of smoking, the BDNF polymorphisms were analysed in the complete cohort of 341 individuals (217 S, 124 NS). For DRD1 and DRD3 polymorphism analysis, we were able to genotype 199 samples from this cohort (124 S, 75 NS). The association analysis of SNP variants with the degree of smoking required information stating the number of cigarettes consumed per day, which was not available for all subjects in the original cohort. Hence the analysis of SNP association with the degree of smoking within the BDNF gene was performed using 327 individuals (204 S, 123 NS), and dopamine receptor (DR) SNP analysis was performed using 191 individuals (116 S, 75 NS).

The procedures were approved and monitored by the CAMH Research Ethics Board. Written informed consent was obtained from all participants. The diagnosis was established by best estimate procedures and the clinical information was collected through a questionnaire addressing smoking status and quantity smoked.

Genomic DNA was extracted from white blood cells using a high salt extraction method and the markers were genotyped using a microarray on an Illumina platform (Enoch et al., 2006). Results for qualitative (smoking vs. non-smoking), allelic and haplotypic analyses were obtained using the Haploview statistical software (Barrett et al., 2005). Haplotype blocks (Fig. S1 – Supplementary material) were determined using the Haploview statistical software, which was also used for permutation analysis. Quantitative (degree of smoking) analyses were assessed using the UNPHASED 3.0 statistical software (Dudbridge, 2003).

Gene-gene interaction analysis between the DRD1, DRD3 and BDNF genes was performed using HelixTree interactive tree analysis software (HelixTree Genetics Analysis Software, Golden Helix Inc. Bozeman, MT, USA; e.g., Zai et al., 2008). The adjusted p-value was used as the p-value threshold.

Corrections for multiple marker testing in the qualitative analysis were performed using 10,000 permutation analysis (Haploview statistical software). Corrections for multiple marker testing in the quantitative analysis were performed using the Single Nucleotide Polymorphism Spectral Decomposition (SNPSpD) method of Nyholt (Nyholt, 2004). In order to maximise the stringency of the Nyholt test, redundant SNPs were removed using tagger, as recommended by Nyholt (Nyholt, 2005).

The following analyses were performed in order to account for possible confounding factors. We have performed an analysis of smoking in a Caucasian population, selecting only individuals with all 4 grandparents of Caucasian origin. The cohort consisted of 245 individuals (166 S, 79 NS) for BDNF analysis and of 112 S and 57 NS for DRD1 and DRD3 analyses. For the analysis of sex effects on degree of smoking, our sample consisted of 63 females (26 S, 37 NS) and 134 males (93 S, 41 NS). To account for possible confounding factors due to schizophrenia, we have compared our combined schizophrenia group to 139 controls genotyped for the BDNF markers, of which 107 were genotyped for the DRD1 and DRD3 markers. The control subjects were screened for any significant mental health disorders and were recruited in the same area as individuals with schizophrenia.

The Odds Ratios and the 95% confidence intervals (Table 1) were determined using the RELRISK software developed by J. Ott (Ott, 1991).

Results

Association with Degree of Smoking

A quantitative analysis testing the association of the SNPs with the degree of smoking, expressed as cigarettes smoked per day (CPD), showed that the DRD3 marker rs1025398 is strongly associated with the number of cigarettes smoked (p = 0.002). This association remained significant after correction for multiple testing using the Nyholt method (Table 1) (Nyholt threshold = 0.003) .

We have also identified a significant association of both DRD1 SNPs (rs686, p = 0.005 and rs4532, p = 0.01) with the number of cigarettes smoked, which remained significant after correction for multiple testing (Nyholt threshold value = 0.05) (Table 1).

When CPD consumption was compared among individuals, individuals homozygous for the allele G of the DRD1 rs4532 marker smoked 36% more than individuals homozygous for the A allele (16.8 vs. 22.9 CPD; p = 0.03, t-Test). Individuals homozygous for the G allele of the DRD1 marker rs686 had 32% higher CPD consumption than individuals homozygous for the A allele (17 vs. 23, p = 0.05, t-Test). The same 32% increase was observed in individuals homozygous for the G allele in the DRD3 marker rs1025398 (18 vs. 23 CPD; p = 0.02, t-Test) (Fig. 1).

For association with degree of smoking, we performed a power calculation using the Quanto statistical software, version 1.2.3 (Gauderman & Morrison, 2006). The calculations were based on critical value of alpha = 0.05. For BDNF, we have 80% power to detect a between-genotype difference of 3.1 cigarettes, if the minor allele frequency is 20%. For DRD1 or DRD3, we have over 80% power to detect a between-genotype difference of 4.5 cigarettes, if the minor allele frequency is 20%.

Discussion

The most noteworthy result of this study is the strong association of both DRD1 markers tested (rs4532 and rs686) with the degree of smoking in individuals with schizophrenia. We also show an association of the DRD3 marker rs1025398 with the number of cigarettes smoked per day in the same cohort. Furthermore, we have identified that the BDNF gene haplotype ACCG formed by the markers Val66Met (rs6265), rs11030104, rs2049045, and rs7103411 is associated with the risk of smoking in this population (Table 1).

We were able to show that individuals homozygous for the G alleles of the DRD1 markers rs4532 and rs686, as well as the DRD3 marker rs1025398, have higher CPD consumption (36%, 32% and 32% higher CPD consumption, respectively) than individuals homozygous for the A alleles of these markers (Fig. 1).

The mesolimbic brain DA system mediates the psychoactive and reinforcing effects of nicotine and other addictive drugs. Within the mesolimbic DA pathway, nicotine stimulates acetylcholine nicotinic receptors located in the ventral tegmental area (VTA), resulting in the release of DA in the nucleus accumbens (NAc) shell (Picciotto et al., 1998; Tapper et al., 2004; Brody et al., 2006; Balfour, 2009). In the nucleus accumbens, this DA signal subsequently activates a number of receptors, in particular the DRD1 and DRD3 receptors (Le Foll et al., 2003a,b; Marcellino et al., 2008; Balfour, 2009).

Our observation that the DRD1 markers (rs4532, rs686) are associated with the number of cigarettes smoked per day is in agreement with the results of Huang et al., who have identified the same DRD1 markers as being strongly associated with nicotine addiction (Huang et al., 2008b). Furthermore, the DRD1 rs686 marker has been identified by Huang et al. as a functional polymorphism, likely affecting the expression of this gene (Huang et al., 2008b). The involvement of these receptors in nicotine addiction is supported by behavioural studies indicating that nicotinic receptors as well as dopamine D1 receptors within the VTA mediate the rewarding effects of nicotine in mice (David et al., 2006) and that dopamine D1 receptors located in the nucleus accumbens shell play a role in nicotine-induced place conditioning in rats (Spina et al., 2006). Taken together, these studies suggest an influence of the D1 dopamine receptor on the reinforcing effects of nicotine.

In addition to a strong association of the DRD3 marker rs1025398 with the number of cigarettes smoked, we have identified a trend for association of this marker and of the DRD3 markers Ser9Gly (rs6280), rs167770 and rs7616367 with the risk of smoking. It is interesting to note that both the positive association with nicotine addiction of Ser9Gly (rs6280) and a trend by rs167770 have also been identified by Huang et al. (their study did not include the markers rs1025398 and rs7616367) (Huang et al., 2008a). The involvement of the dopamine D3 receptor in nicotine addiction is also supported by behavioural research showing that blockade of the dopamine D3 receptor attenuates reactivity to nicotine-associated stimuli (Le Foll et al., 2003b, 2005a; Pak et al., 2006; Khaled et al., 2010). Antagonism of DRD3 also blocks the expression of nicotine-induced place preference (Le Foll et al., 2005a; Pak et al., 2006) and the reinstatement of nicotine seeking (Andreoli et al., 2003; Khaled et al., 2010). It is also possible that the DRD3 variant affects dopamine D1 receptor function. The dopamine D3 receptor is expressed in the most ventral parts of the striatal complex (NAc) (Sokoloff et al. 1990; Bouthenet et al., 1991), where it is colocalised with the dopamine D1 receptor (Schwartz et al., 1998; Marcellino et al., 2008). These two receptors form heteromers and recent evidence suggests that dopamine D3 receptor stimulation potentiates dopamine D1 receptor-mediated behavioural effects (Marcellino et al., 2008).

Our observations that the ACCG haplotype of the BDNF gene (markers Val66Met (rs6265), rs11030104, rs2049045 and rs7103411) is associated with the risk of smoking parallels observations by (Lang et al., 2007). The haplotype association implies that the functional variants are in LD with one or more of the SNPs in the block.

BDNF is a protein responsible for supporting existing neurons and encouraging growth and differentiation of new neurons and synapses. It has generated much interest due to its possible role in depression and other severe mental health disorders (Rybakowski, 2008). The Val66Met (rs6265) polymorphism is associated with abnormal intracellular trafficking and secretion of BDNF (Egan et al., 2003). In human subjects, the Met allele was shown to negatively affect memory-related brain activity (Hariri et al., 2003).

BDNF is synthesised in dopaminergic neurons, which project from the VTA to the NAc where dopamine D3 receptors are primarily expressed (Guillin et al., 2001). BDNF appears to be involved in drug reward and relapse (Lu et al., 2004; Graham et al., 2007; Schoenbaum et al., 2007). In fact, the expression of dopamine D3 receptor in the nucleus accumbens in response to various drugs seems to be controlled by BDNF (Sokoloff et al., 2002; Le Foll et al., 2002, 2005b). Hence, our results showing that the risk of smoking is influenced by polymorphisms within the BDNF gene provide some validity to the hypothesis that BDNF is a contributor to drug addiction in humans.

Because many different factors play a role in smoking, smokers from the general population would, therefore, be expected to form a very heterogeneous group. By focusing on individuals with schizophrenia, we were aiming to reduce this heterogeneity. Since predisposition to smoking may be related to the same dopaminergic pathways involved in schizophrenia, we were hoping that such a group would be enriched in alleles of these genes playing a role in smoking, allowing us to detect alleles which may be more difficult to detect in the general population. Because of the heterogeneity of the general population, identifying these alleles may require a large sample size. In addition, since heavy smoking is much less prevalent in the general population, recruiting a group containing a large proportion of heavy smokers may pose a challenge, but would be an important confirmation of the role of these alleles in heavy smoking in the general population.

In summary, functional variants of these genes are likely to lead to phenotypic changes, which have the capacity to affect the response of individuals to nicotine (Brody et al., 2006). Our study provides an important confirmation of the involvement of polymorphisms within the DRD1-BDNF-DRD3 gene cluster in nicotine addiction, especially pertaining to the schizophrenia population.

Acknowledgements

This project has been funded in part by Fellowship awarded to GN by the Strategic Training Program in Tobacco Use in Special Populations (TUSP) – CIHR.

Competing Interests

The authors declare no competing interests.