CRY2 Genetic Variants Associate with Dysthymia

People with mood disorders often have disruptions in their circadian rhythms. Recent molecular genetics has linked circadian clock genes to mood disorders. Our objective was to study two core circadian clock genes, CRY1 and CRY2 as well as TTC1 that interacts with CRY2, in relation to depressive and anxiety disorders. Of these three genes, 48 single-nucleotide polymorphisms (SNPs) whose selection was based on the linkage disequilibrium and potential functionality were genotyped in 5910 individuals from a nationwide population-based sample. The diagnoses of major depressive disorder, dysthymia and anxiety disorders were assessed with a structured interview (M-CIDI). In addition, the participants filled in self-report questionnaires on depressive and anxiety symptoms. Logistic and linear regression models were used to analyze the associations of the SNPs with the phenotypes. Four CRY2 genetic variants (rs10838524, rs7121611, rs7945565, rs1401419) associated significantly with dysthymia (false discovery rate q<0.05). This finding together with earlier CRY2 associations with winter depression and with bipolar type 1 disorder supports the view that CRY2 gene has a role in mood disorders.


Introduction
Circadian rhythms are about 24-hour oscillations in physiological processes that allow humans to anticipate routine changes in the environment. The rhythms are generated by the internal clock, the principal one being located in the suprachiasmatic nucleus of the anterior hypothalamus in the brain. At the molecular level, the transcription of a network of genes is switched on and off following the approximate 24-hour pattern. A feedback loop is initiated by binding of CLOCK protein [1] with aryl hydrocarbon receptor nuclear translocatorlike (ARNTL) protein [2]. The heterodimer then activates the transcription of a large set of genes and among them the period (PER1, PER2, PER3) and cryptochrome (CRY1, CRY2) genes. Again, the encoded proteins form heterodimers (PER-with-CRY dimers) which translocate back to the nucleus and inhibit the CLOCK-with-ARNTL mediated transcription, thereby repressing their own transcription [3]. In addition, nuclear receptors together with their co-activators and co-repressors feed back to ARNTL [4].
The cryptochrome genes may be of key importance here, because of their unique role in the core of these feedback loops [3,5,6]. The nuclear ratio of CRY1 to CRY2 proteins controls the period of the circadian clock [5] and the effects of CRY2 on circadian regulation are sensitive to the dosage of CRY1 [7]. Findings from knock-out mice suggest that without CRY2 the circadian period is lengthened similar to ''evening owls'', whereas without CRY1 it is shortened [8,9]. Furthermore, CRY2 appears to play a critical role in tuning the suprachiasmatic nucleus to 24 hour cycle by opposing and titrating the decelerating action of CRY1 [7]. ''Evening owls'' are known to have increased risk for anxiety and depressive symptoms [10,11] that may develop into depressive disorders [12,13]. Furthermore, dysfunction of signaling pathways involving CRY1 and CRY2 might lead into depressive behaviors as follows. Interaction of CRY1 and CRY2 proteins directly with the Gsa subunit of heterotrimeric G protein [14] and with adenylyl cyclase [15] reduces the accumulation of cyclic adenosine monophosphate in response to G protein-coupled receptor activation and inhibits downstream reactions [15,16]. If this inhibition is not appropriate, the inhibitory tone on cyclic adenosine monophosphate signaling is overruled, leading to the activation of cascades downstream, such as cyclic adenosine monophosphate response element-binding protein activity, and to the increase in depressive-like behaviors [17].
Disrupted or misaligned circadian rhythms and circadian gene expression are often found in major depressive disorder and bipolar disorder, with or without a seasonal pattern [18,19]. Since the original finding [20], there is a growing evidence that genetic variations in the core circadian clock genes associate with mood and anxiety disorders [21]. In addition, these genetic variations associate with sleep and metabolic disorders [22] that are often comorbid conditions with mood and anxiety disorders. With the focus on the cryptochrome genes, earlier studies have indicated CRY1 genetic variants in major depressive disorder [23] and CRY2 genetic variants in seasonal affective disorder, winter depression in particular, and in bipolar type 1 disorder [24,25].
The aim of the current paper is to study the associations of CRY1, CRY2 and TTC1 genes in depressive and anxiety disorders. In addition to the two cryptochrome genes, TTC1 was selected in order to analyze epistasis, if any, between the TTC1 and CRY2 genes, because the TTC1 protein interacts physically with CRY2 [26]. The present work extends the earlier studies by deeper linkage disequilibrium (LD) coverage of the genes, by including potential functional variants, by using a large sample derived randomly from the general population, by assessment based on both a diagnostic interview and self-reports, and by analysis of diverse phenotype information. Here we report a significant association of CRY2 genetic variants with dysthymia.

Ethics Statement
The ethics committees of the National Public Health Institute and the Helsinki and Uusimaa Hospital District accepted the study protocol and its ethical approval, and all participants provided a written informed consent.

Subjects
This study was part of a nationwide health interview and examination survey, Health 2000. The survey included a health status examination and the Munich-Composite International Diagnostic Interview (M-CIDI) [27]. M-CIDI is a valid and reliable instrument for the assessment of depressive, anxiety and alcohol use disorders, yielding diagnoses according to Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV). In addition, participants gave venous blood samples for DNA extraction and filled in two sets of questionnaires. The sample for our current study from the Health 2000 participants aged 30 years and older (n = 8028) included 5910 individuals (3283 women, 2627 men) who had given blood samples, taken part to the M-CIDI interview and filled in the self-report on seasonal changes in mood and behavior. The study design and methods have been described in detail elsewhere (http://www.terveys2000. fi/indexe.html).

Phenotypes
Altogether 12 phenotypes related to mental well-being (diagnoses according to DSM-IV and psychometric scales) were analyzed (Table 1). Mental disorders analyzed in the current study included major depressive disorder, dysthymia, depressive disorders (major depressive disorder, dysthymia), panic disorder, anxiety disorders (panic disorder w/o agoraphobia, generalized anxiety disorder, social phobia, agoraphobia), and the comorbid conditions of depressive, anxiety and alcohol use disorders (alcohol dependence and abuse), as assessed with the M-CIDI. The controls did not have any diagnosis of mental disorders nor met any sub-threshold criteria as assessed with the M-CIDI.

Gene and SNP Selection
Three genes of interest were selected: CRY1, CRY2 and TTC1. The SNP selection was based on phase 3 data of the HapMap database (http://www.hapmap.org/) and done using the Tagger program included in the Haploview 4.1 software [34]. Information about the LD within the chosen areas of the genome was used to select an optimal set of SNPs capturing most of the genetic variation. Areas of the genome covered were the genes and 10 kb of their 59 and 39 flanking regions. For CRY1, this area of interest was 122 kb (chr12:105,899-106,021 kb, NCBI36/hg18 assembly) in size, whereas for CRY2 it was 56 kb (chr11:45,815-45,871 kb) and for TTC1 77 kb (chr5:159,358-159,435 kb).
The HapMap database included 21, 34 and 15 SNPs having a minor allele frequency (MAF) .5% in the European population (CEU and TSI) for the areas containing the CRY1, CRY2 and TTC1 genes, respectively. The aim was to capture all these SNPs by setting the limit for the pair-wise r 2 to $0.9. Based on this criterion, ten CRY1 SNPs, ten CRY2 SNPs and nine TTC1 SNPs Table 1. Number of subjects in mental health related phenotypes (DSM-IV based diagnoses and quantitative scores on psychometric scales) analyzed in the study.  were decided to be genotyped and all, except one (TTC1 rs3733868), were successfully included in the genotyping multiplexes. One SNP (CRY1 rs11113153) failed in the genotyping phase.
In addition to the aforementioned SNPs, potentially functional variants, as many as possible, were included in the genotyping multiplexes. These 21 additional SNPs (8 for CRY2, 11 for CRY1 and 2 for TTC1) were selected using Pupasuite [35], Variowatch [36], dbSMR [37] and miRNASNP [38] databases. Table 2 presents all the 48 SNPs that were successfully genotyped in this study and their selection criteria.  Genotyping Genomic DNA was isolated from the whole blood according to standard procedures. DNA samples were available from 5910 subjects for this study. The SNPs were genotyped using the Sequenom MassARRAY system and the iPLEX Gold Single Base Extension chemistry (Sequenom, San Diego, CA, USA) in a multiplex format [39]. This method has excellent success (.95%) and accuracy (100%) rates [40]. For the genotyping quality control purposes, both positive (CEPH) and negative water controls were included in each 384-plate. Genotyping was performed blind to phenotypic information.
173 individuals were removed from the statistical analyses due to a high missing genotype rate (i.e. .0.1). The total genotyping rate in the remaining individuals was 0.9987. In addition, the following five SNPs were removed because of the minor allele frequency of ,0.01: rs41275305, rs12520927, rs3747548, rs35488012, rs7294758. Finally, there were 5737 individuals and 43 SNPs for the statistical analyses.

Statistical Analyses
Statistical analyses were performed with the PLINK software [41], release v1.07. The additive model was calculated using logistic and linear regression models controlling for age and sex. In addition, dominant and recessive models were calculated and the analyses were performed in both sexes separately. For continuous phenotypes 10,000 permutations were used to produce empirical p-values in order to relax the assumption of normality. The results were corrected for multiple testing across all the tests (SNPs, phenotypes, gender, genetic models) by calculating false discovery rate (FDR) q-values [42] using R software (http://www.r-project.org/). The q-values of ,0.05 were considered significant. Haploview software [34] was used to define haplotype blocks by the confidence interval method. Each haplotype in the formed blocks was analyzed by PLINK software using linear and logistic regression, additive model and controlling for age and sex. Sexes were also analyzed separately. Continuous phenotypes were permuted 10,000 times. Interactions between CRY2 and TTC1 (SNP 6 SNP epistasis) were tested and calculated using the PLINK software.

Results
The genotype and allele frequencies and Hardy-Weinberg equilibrium (HWE) estimates are shown in Table 2 (see Table S1 for the genotype counts by phenotype). The results of the association analyses where q,0.15 are presented in Table 3 and the remaining results of the association analyses in Table S2. Four CRY2 SNPs showed evidence of association with dysthymia using the additive model (rs10838524 risk allele G, p = 0.000020, OR = 1.75; rs7121611 risk allele A, p = 0.000022, OR = 1.74; rs7945565 risk allele G, p = 0.000039, OR = 1.71; rs1401419 risk allele G, p = 0.000041, OR = 1.71). CRY2 SNP rs10838524 showed evidence of association also using the dominant model (p = 0.000059, OR = 3.27) and the additive model for women (p = 0.000026, OR = 1.97). The CRY1 and TTC1 genetic variants studied showed no evidence of association with any of the phenotypes analyzed. No epistasis between CRY2 and TTC1 SNPs was observed.
Haploview produced two haplotype blocks for CRY1 (Figure 1), one for CRY2 ( Figure 2) and one for TTC1 (Figure 3). The haplotype analyses supported the association of CRY2 with dysthymia, as the ATTCGCGGTGGCACG haplotype containing the risk alleles A, G, G and G of SNPs rs7121611, rs10838524, rs7945565 and rs1401419, respectively, showed association with dysthymia in all cases (p = 0.00034, OR = 1.58) and among women (p = 0.0011, OR = 1.66). Table 4 presents the haplotype associations with p-values (or empirical p-values) of ,0.010.

Discussion
Our main finding was that CRY2 genetic variants associated with dysthymia in our random sample derived from the general population aged 30 years and older. The associations with dysthymia were stronger in the additive than in the dominant model and became weaker when we analyzed only women (n = 88). No association was observed in men. The sample size for dysthymia was limited (n = 128) and the lack of association in men could be due to a small sample size (n = 40). We did not find any significant association of CRY2 with major depressive disorder, whether with a single or recurrent episode, or with anxiety disorders, and there was no significant association of CRY1 or TTC1 with any of the phenotypes assessed.
Of the individual CRY2 SNPs, rs7121611 (upstream), rs10838524 (intron 1), rs7945565 (intron 2) and rs1401419 (intron 2) showed evidence of association with dysthymia, with the A, G, G and G alleles being predisposing, respectively. These associations are supported by the haplotype analysis, as the CRY2 ATTCGCGGTGGCACG haplotype including the individual risk alleles associated with the phenotype. The associated haplotype of CRY2 spans almost the whole gene and the four associated SNPs are located in the first half of the haplotype block ( Figure 2). According to transcription factor CHiPSeq from ENCODE (http://genome.ucsc.edu/ENCODE/) SNP rs10838524 is located in a regulatory region and according to Pupasuite software [35] rs7945565 and rs1401419 were predicted to be putative triplex disrupting SNPs which may also be part of regulatory regions for controlling gene expression [43].
Of the four associated CRY2 SNPs, only rs10838524 has previously been linked to mental health disorders. Concerning this SNP, our current results are partly in line with the findings of Lavebratt et al. [24] who found the same risk allele for winter depression in the Finnish subpopulation of their study. The allele frequencies in cases and controls were similar in these two sets, as expected for the controls, since they were drawn from the same Health 2000 cohort. The cases, on the other hand, had been recruited from outpatient services in the capital area [24]. Another Finnish study, using a population based sample also from the Health 2000 cohort, found the same minor allele to weakly associate with depressive disorders in women and in women having depressive disorders with early morning awakening, and the association was supported with haplotype analyses but did not hold after correcting for multiple testing [44]. We also analyzed depressive disorders in women and found similar crude p-value (p = 0.035, Table S2). In the study by Utge et al., the statistical analyses had been done separately in men and women. In our analyses, the associations were stronger when using the dominant model in both sexes together (p = 0.0010), but still the association to depressive disorders did not reach statistical significance after correcting for multiple testing. For the Swedish subpopulation of the winter depression study [24] as well as for a Swedish bipolar study [25], the allele frequencies for CRY2 rs10838524 were reversed and the other allele, i.e. A, was found as the risk allele. Earlier, the A allele has also been associated with a greater degree of chronic course of depressive symptoms in a sample of 35 individuals with major depression or bipolar disorder [45].
Concerning SNP rs3824872, located in the 39 region of CRY2 (but also in the 59 region of MAPK8IP1 gene), our study suggests (q = 0.08) that the T allele is the protective one against dysthymia, and a similar (though non-significant) tendency is seen in the Finnish subpopulation of the winter depression study [24], while in the Swedish subpopulation of the same study, the other allele, i.e. A, was the risk allele. With regard to CRY2 rs7123390, the SNP was earlier associated with winter depression in the Finnish subpopulation [24], but in the current study, we did not observe any significant association with this intronic SNP. Moreover, there was no significant association with the 39UTR CRY2 rs10838527, similar to earlier findings where no association with winter depression was observed among the Finnish subpopulation, whereas this SNP did associate with winter depression among the Swedish subpopulation [24].
Although preliminary, our analysis has demonstrated that CRY2 associates robustly with dysthymia which is characterized by a chronic course of illness where a depressive episode lasts for two years or longer and often deepens into a major depressive episode. Earlier, CRY2 genetic variants have been associated with winter recurrent depressive disorder [24], and with the rapid cycling of bipolar type 1 disorder where depressive episodes dominate the clinical course of illness [25]. A single CRY1 variant has been associated with major depressive disorder earlier [23] but there is no replication of the finding [21]. Furthermore, it has been previously observed that following the antidepressant sleep deprivation, the expression of CRY2 mRNA increased in controls, whereas no change or response to the treatment was observed in patients with bipolar disorder whose CRY2 mRNA levels were markedly lower than in controls [24]. Moreover in an animal study, CRY2 mRNA expression was abnormal when inbred-strain mice with the intrinsic level of high anxiety were deprived of sleep [46]. Currently, it is however not known, whether the knock-out or the knock-down of CRY2 in the whole organism or specific to a tissue produces any change in anxiety-like or depressive-like behaviors, or how mutated CRY2 proteins influence mood-related behaviors. CRY2 and CRY1 proteins are functionally different [7] which could impact their circadian outputs differently and make a difference in the consequent phenotype. This difference could explain why CRY2, not CRY1, associated with dysthymia in our study. Moreover, it might be that CRY2 variants play a leading role in circadian hierarchy [3] and thereby in the pathogenesis of mood disorders as well. Some circadian clock gene variants seem to have a role in neuropsychiatric disorders in general [47,48]. Elucidating the mechanisms of action with these genetic variants might therefore benefit the development of treatments for and beyond mood and anxiety disorders [18,49,50].
A limitation in our current epidemiological health examination study is that the seasonal pattern, whether in major depressive disorder or bipolar disorder, was not assessed at all. Another limitation is that the diagnoses of bipolar disorders were considered not adequate for analysis. However, it is of note that bipolar type 1 disorder is markedly less frequent in Finland compared with the prevalence in many Western-type societies [51].
To conclude, CRY2 genetic variants associated with dysthymia. This new data, together with the earlier findings, reinforces the involvement of CRY2 in mood disorders.