Association of the Dio2 Gene Single Nucleotide Polymorphisms with Recurrent Depressive Disorder*

Genetic factors may play a role in the etiology of depres-sive disorder. The type 2 iodothyronine deiodinase gene (DIO2) encoding the enzyme catalyzing the conversion of T4 to T3 is suggested to play a role in the recurrent depressive disorder (rDD). The current study investigates whether a specific single nucleotide polymorphism (SNP) of the DIO2 gene, Thr92Ala (T/C); rs 225014 or ORFa-Gly3Asp (C/T); rs 12885300, correlate with the risk for recurrent depression. Genotypes for these two single nucleotide polymorphisms (SNPs) were determined in 179 patients meeting the ICD-10 criteria for rDD group and in 152 healthy individuals (control group) using a poly-merase chain reaction (PCR) based method. The specific variant of the DIO2 gene, namely the CC genotype of the Thr92Ala polymorphism, was more frequently found in healthy subjects than in patients with depression, what suggests that it could potentially serve as a marker of a lower risk for recurrent depressive disorder. The distribution of four haplotypes was also significantly different between the two study groups with the TC (Thr-Gly) haplotype more frequently detected in patients with depression. In conclusion, data generated from this study suggest for the first time that DIO2 gene may play a role in the etiology of the disease, and thus should be further investigated.


INTRODUCTION
Depressive disorder is one of the most common psychiatric diseases (Whiteford et al., 2013).The existing evidence suggests a heterogenic etiology with a possible genetic background (Belmaker & Agam, 2008).One of the hypotheses postulates a deregulation of the hypothalamic-pituitary-thyroid (HPT) axis in depression.In adults, thyroid diseases can lead to various clinical manifestations (Bauer, 2008).For example, hypothyroidism causes fatigue, psycho-motor speed, attention and concentration and memory impairment (Samuels, 2008;Bonnin et al., 2010;Almandoz & Gharib, 2012).Hypothyroidism is also associated with bipolar affective disorders, depression, or loss of cognitive functions, especially in the elderly (Bonnin et al., 2010;Bauer et al., 2002;Fountoulakis et al., 2006;Bunevicius & Prange, 2010).The limbic system, where thyroid hormone (TH) receptors play a particularly essential role is implicated in the pathogen-esis of depression (Murray et al., 2011;Williams, 2008).Changes in TH levels associated with depression include an increase in thyroxine (T4) concentrations and elevated levels of reversed triiodothyronine (rT3) in the cerebrospinal fluid (CSF) (Kirkegaard & Faber, 1991), as well as elevated levels of circulating T4 (Williams, 2008), and lower levels of circulating T3 (Stipcević et al., 2008).
TH levels may also be affected by pharmacological treatment including antidepressants (Bauer et al., 2008).In the rat, treatment with various antidepressants results primarily in changes of local D2-activities and to a lesser extent of D3-activities (Eravci et al., 2000).Additionally, treatment with antidepressants results in an increase of T3 in the myelin fraction of homogenates of the amygdala, an essential structure implicated in emotion and fear regulation (Pinna et al., 2003).Various hormones of the thyroid axis, including T3 have been used to treat depression as mono-therapy or, more commonly, in combination with standard antidepressants (Cooper-Kazaz et al., 2009;Joffe, 2011).Interestingly, treatment with antidepressants, including a selective serotonin reuptake inhibitor, results in the induction of D2 (Baumgartner et al., 1994).
The D2 protein is mainly expressed in glial cells of various regions of the central nervous system (CNS) and plays an important role in mediating TH action both during CNS development and in the adult brain (Bauer et al., 2008).It is also suggested that D2 protects the thyroid status of the brain under conditions of TH deficiency (Galton et al., 2007).
There are results showing that single nucleotide polymorphisms (SNPs) in DIO2 gene may be associated with TH levels (Peeters et al., 2005) and that there is an association between the single nucleotide polymorphism (SNP) within the DIO2 gene, the D2 protein level and its enzymatic activity.For example, allele T of the Thr92Ala (T/C) variant was found to be related with higher D2 activity/TH levels, while C allele of the ORFa-Gly3Asp (C/T) polymorphism resulted in lower D2 activity/TH levels (Bauer et al., 2008;Peeters et al., 2005;Canani et al., 2005).The DIO2 gene polymorphism is linked to a bipolar disorder, mental retardation and well-being (Guo et al., 2004;He et al., 2009;Panicker et al., 2009).
Considering the possible changes in TH levels in depression, and the role of the D2 enzyme in maintaining the active TH levels, the current study examined a potential genetic contribution of the DIO2 gene to the etiology of recurrent depressive disorders (rDD).
The aim of the study was to investigate whether two common SNPs in the gene, Thr92Ala and ORFa-Gly3Asp, linked to expression/stability of the D2 protein (Peeters et al., 2005;Canani et al., 2005) are associated with rDD.

Subjects.
The study enrolled 179 patients diagnosed and treated for rDD (110 females -61.45% and 69 males -38.55%).The diagnosis was established according to ICD-10 (1992) criteria (F33.0-F33.8).A medical history was obtained and assessed using the standardized Composite International Diagnostic Interview (CIDI) form (Patten 1999).The Hamilton Depression Rating Scale (HDRS) was used to assess the level of depressive symptoms.Next, the number of depressive episodes, the duration of disease and the age of patient at disease onset were recorded for each individual.The control group (CG) consisted of 152 healthy subjects (87 females -57.24% and 65 males -42.76%) with a negative family history for psychiatric disorders.Healthy controls constituted of healthy community volunteers, enrolled in the study on the basis of a CIDI psychiatric interview (Patten, 1999).
Individuals (both patients and CG) with other psychiatric diagnoses within the axis I and II disorders were excluded from the current study.Severe or chronic diseases with confirmed inflammatory or autoimmune etiology served as an additional exclusion criteria.All study subjects (patients and CG) were unrelated individuals from central Poland.To avoid a population stratification effect, genotypes were determined only in individuals of Polish origin, i.e., all four grandparents identified themselves to be of Polish origin.The study protocol had earlier been approved by the Local Bioethics Committee.
Genotyping of SNPs.Peripheral blood was collected and genomic DNA was extracted using a commercial isolation kit according to the manufactures' protocol (A&A Biotechnology, Gdańsk, Poland).
The rs225014 SNP is a polymorphism at nucleotide 674 of the D2 sequence predicting a threonine (Thr) to Alanine (Ala) substitution at codon 92 (Thr92Ala).The rs12885300, C/T polymorphism is in the most up- stream short open reading frame (ORFa-Gly3Asp) of the 5'untranslated region of DIO2.The region containing the Thr92Ala and the ORFa-Gly3Asppolymorphism was amplified by PCR-based method as described by He and coworkers (2009) and Dora and coworkers (2010) with some modifications.Statistical analysis.The results are reported as percentages (%) or means with standard deviations (± S.D.).In order to determine the association between SNPs within the DIO2 gene and recurrent depression disorder rDD, χ 2 test was used.A Post-hoc power analysis was performed with the use of non-central χ 2 distribution.The analysis of association was based on 95% confidence interval (CI) for the disease odds ratio (OR dis ), calculated with the use of logistic regression model including sex and age as covariates.Deviations from the Hardy-Weinbergs equilibrium were determined by comparison of observed genotype prevalence rates with the expected ones.The Hardy-Weinberg equilibrium for genotype frequencies in rDD group was calculated using c 2 tests.In all the analyses, p ≤ 0.05 was accepted as the level of statistical significance.

RESULTS
No significant differences were found between rDD patients and CG with respect to gender (p > 0.05).Groups were gender matched (p=0.44),but varied significantly with respect to the age distribution (p < 0.0001).The mean age was higher for rDD patients than for CG: 48.5y ± 10.8y vs 31.7y± 9.1 (Table 1).
No significant difference in the distribution of demographic and clinical characteristics for different genotypes was observed, except for the age difference for the Thr92Ala polymorphism between rDD patients and the control group (Kruskal-Wallis test; p=0.048).The observed differences in age and gender distribution between groups and certain genotypes caused necessity of age and sex adjustment using a logistic regression.
A distribution comparison between rDD patients and the control group with respect to genotypes/alleles of  Thr92Ala and ORFa-Gly3Asp polymorphisms did not reveal a significant difference except for the CC genotype of Thr92Ala polymorphism.The CC genotype of the Thr92Ala polymorphism was significantly less frequent in rDD patients than in the controls.A summary of results for genotypes and allele frequencies within the examined SNPs is presented in Table 2.
The results of the DIO2 gene haplotype analysis indicate a significant difference in the distribution of haplotypes for the combined SNPs between rDD patients and controls (Table 3).

DISCUSSION
This study is the first to test the hypothesis whether DIO2 polymorphisms are linked to depressive disorders.
Our results suggest that common variation in the DIO2 gene may be linked to a lower risk for rDD.The CC genotype is present in a small proportion of the population, approximately 4% of our study population.In a group of patients we found only 1 person with CC genotype.Interestingly, a relatively low prevalence of the CC genotype (16%) was reported by Panicker et al. (2009) for a larger group from Bristol, United Kingdom.Similarly, Babenko and coworkers (2012) observed an 8% presence of Ala/Ala (CC) homozygous in the Caucasian population.In the cohort consisting of 946 subjects Zevenbergen et al., (2014) found CC genotype also to be less frequent (16.6%).
The Thr92Ala genotype frequency differences were noted between the Caucasian and two other Asian and Indian populations (He et al., 2009;Nair et al., 2012).Ethnical and geographical differences are also suggested by Guerra et al. (2013) who investigated the Thr92Ala polymorphism in the Brazilian population.By contrast to the above effect, neither the ORFa-Gly3Asp polymorphism nor its genotype frequencies were significantly different between two study groups.However, the frequency distribution of ORFa-Gly3Asp polymorphism genotypes in our study remains in agreement with previous observations for the Caucasian population (Peeters et al., 2005;de Jong et al., 2007;Hoftijzer et al., 2011).
The obtained results show that the TC (Thr-Gly) haplotype is statistically more frequently detected in rDD patients.The ORFa-Gly3Asp is located in 5' untranslated region and is a functional variant related to the enzyme activity.Our findings of TC haplotype increasing rDD risk may possibly support both the fact that depressive disorder is a low T3 syndrome (Baumgartner et al., 1998;Premachandra et al., 2006) and the suggestion on the supplementation of antidepressant therapy, especially in some cases.
The DIO2 gene polymorphisms were linked to psychiatric disorders.For example, the DIO2 gene is linked with bipolar disorder in the Asian population (He et al., 2009).Activity of the D2 in the brain may be a determinant of well-being and neurocognitive function (Panicker et al., 2009).It is known that the CC genotype of the Thr92Ala polymorphism is relevant to TH metabolism, as it is linked to a greater therapeutic improvement in hypothyroid patients on T4/T3 combination hormone replacement as compared with T4 mono-therapy.It is also associated with poorer psychological well-being, as measured by General Health Questionnaire (GHQ) score (Panicker et al., 2009).
The possible effect of the DIO2 gene polymorphism on TH metabolism could be related to the effect of a small amino acid substitution in the D2 protein, thus influencing its level and/or activity.Studies on subclinical hypothyroidism show that subtle changes in TH bioavailability may have clear effects on well-being and neurocognitive function (Cooper, 2001;Toft, 2001).Previous studies have shown that Thr92Ala polymorphism is related to minimal effects on circulating levels of TH (Canani et al., 2005;Torlontano et al., 2008), whereas the ORFa-Gly3Asp polymorphism was associated with lower serum T4 and free T4 but unaltered TSH and T3 levels (Peeters et al., 2005;Canani et al., 2005;Peltsverger et al., 2012).It was also postulated that the Thr92Ala substitution may result in the instability loop in D2 protein.Such instability may affect ubiquitination, prompting D2 protein degradation and impairing the ability of D2 enzyme to increase its activity in the presence of low T4 levels, reducing the ability to maintain homeostasis and increasing dependence on serum T3 as a source of T3 in the brain (Dentice et al., 2005).Thus, the contribution of DIO2 polymorphisms to the etiology of depressive disorders could be related to the D2 increased activity and subsequently to TH levels in the brain.
The findings that the CC genotype may act as a protective factor appears to be contradictory to previous observations that found low TH levels in patients with depression (Rao et al., 1996).The results can also enlarge discussion on the hyperthyroidism involvement in the depressive disorder.Hyperthyroidism is accompanied by some psychiatric symptoms including depressive symptoms (Tylor, 1975).The possible protective role of CC genotype may be related to the lower levels of TH as C allele is associated with decreased thyroid stimulating hormone-stimulated (TSH) release of T3 and the lower levels of TSH (Butler et al., 2010;Zevenbergen et al., 2014).On the other hand, it is possible that the Thr92Ala variant may have no specific function or may be in linkage equilibrium with other variants.Given the involvement of the inflammatory process in the etiology of depression (Anisman, 2011;Maes et al., 2011) and taking into consideration more recent data that suggests the role of iodothyronine deiodinase in the inflammation process (Boelen et al., 2011), another potential explanation for the obtained results could be proposed.A significant increase in both DIO2 expression and D2 levels was observed after lipopolysaccharide-induction (LPS) in increased lung injury, while the Ala (C) allele was considered a protective factor in sepsis and severe lung injury (Ma et al., 2011).Further, a markedly higher amount of D2 protein positive cells are characteristically found in osteoarthritis (Bos et al., 2012).Thus, an increase in D2 activity would be characteristic for the inflammation process.Similar to the above mentioned diseases, depression is characterized by an increased number of leukocytes and pro-inflammatory cytokine levels in the peripheral blood (Anisman, 2011;Maes et al., 2011).In addition, LPS which is associated with an increased levels of both DIO2 and D2, was shown to be able to induce depression (Maes & Leunis, 2008).Thus, our results showing a significantly higher frequency of the CC genotype of Thr92Ala in healthy subjects with no inflammatory com-Polymorphisms of the DIO2 gene and depression ponent involved, which potentially could be linked to lower D2 activity, would be in agreement with the above hypothesis.
There are several limitations associated with our study.First, a modest study cohort used to determine genetic variants associated with rDD.It is acknowledged that a relatively larger study cohort would be required in order to confirm trends observed in the current study.Second, the current study focuses on two functionally known polymorphisms and other DIO2 polymorphisms should be investigated in order to fully characterize the potential involvement of the DIO2 gene in the etiology of depression disorder.Further studies involving larger cohort, characterization of TH levels, as well as markers of the inflammation process should provide a better understanding of the role of DIO2 polymorphisms in depression.

CONCLUSIONS
DIO2 polymorphisms may constitute potential risk factors in depression.
Depression is not simply a thyroid-related illness but should be rather considered as a complex process with an inflammatory component involved.

Table 1 . Descriptive statistics for demographic and clinical data in rDD patients and the CG for different genotypes of Thr92Ala and ORFa-Gly3Asp polymorphisms
p stands for p-value of appropriate statistical test: χ2 or Kruskal-Wallis Anova rDD, recurrent depressive disorder; CG, control group; p, level of statistical significance; F, female; M, male; ± standard deviation, HDRS, Hamilton Depression Rating Scale

Table 2 . Comparison of genotypes and allele frequencies of Thr92Ala and ORFa-Gly3Asp polymorphisms in rDD patients and the control group.
Odds ratios (OR) with 95% confidence intervals (95%CI) were calculated by the method of logistic regression with age and sex adjustment.

Table 3 . The DIO2 gene haplotype analysis for rDD patients and CG
rDD, recurrent depressive disorder; CG, control group; p, level of statistical significance; c 2 , Chi-square statistics; df, degrees of freedom; %, percentages; 95% CI, 95% confidence interval; OR, odds ratio E. Gałecka and others