How to Cite this Article: Ishitobi Y, S Nakayama, K Yamaguchi, M Kanehisa, H Higuma, Y Maruyama, T Ninomiya, S Okamoto, Y Tanaka, J Tsuru, H Hanada, K Isogawa, J Akiyoshi. 2012. Association of CRHR1 and CRHR2 with Major Depressive Disorder and Panic Disorder in a Japanese Population. Am J Med Genet Part B 162B:78–85.
Dysfunction of the stress-responsive hypothalamic–pituitary–adrenal (HPA) axis is a common feature of anxiety and mood disorders [Bale and Vale, 2004; Nemeroff and Vale, 2005; Swaab et al., 2005; Hauger et al., 2006]. Activation of the HPA axis is controlled and regulated by hypothalamic corticotrophin-releasing hormone (CRH), which activates CRH receptor 1 (CRHR1) in the anterior pituitary to mediate the production of adrenocorticotrophic hormone (ACTH) [Smith et al., 1998; Timpl et al., 1998]. ACTH in turn promotes the synthesis and release of cortisol from the adrenal gland. Blood cortisol levels are normally regulated by a multi-point feedback loop at HPA, limbic, brainstem, and prefrontal brain areas. Extrahypothalamic effects are primarily mediated directly through CRHR1s in the central amygdala, which causes downstream effects on the serotonin neurotransmitter system, among others.
HPA system abnormality is often noticed in major depressive disorder (MDD) patients presenting anxiety as a symptom. The responsiveness of ACTH in MDD patients is lower than that observed in healthy subjects as measured by the CRF test [Pintor et al., 2007]. When a dexamethasone suppression test is given to panic disorder (PD) patients, the rate of the nonsuppression is the same or higher compared to healthy control subjects. In the CRF test in PD patients, the cortisol response to CRF is blunted [Charney and Drevets, 2002]. Such phenomena are hypothesized to depend on chronic over-secretion of CRF. Excessive levels of CRF desensitize CRF receptor function and induce a state of diminished pituitary sensitivity to CRF. A combined dexamethasone/CRF test is widely used investigate MDD patients [Künzel et al., 2003]. ACTH and cortisol responsiveness to CRF after dexamethasone are markedly increased in MDD patients compared to healthy control subjects [Isogawa et al., 2005]. This phenomenon is generally explained as a decreasing negative feedback of glucocorticoid and vasopressin signaling on the HPA axis [Watson et al., 2006].
Regarding morphological HPA changes, both pituitary gland hypertrophy and adrenal gland hypertrophy are reported in MDD patients [Nemeroff et al., 1992]. ACTH and cortisol blood concentrations are also increased in MDD patients [Carroll et al., 2007]. These reports demonstrate an HPA-axis hyper-activation in MDD. Interestingly, these abnormalities are known to normalize during remission.
The association between PD and HPA axis function have not solved yet. Increasing evidence suggests that CRF may play a critical role in mediating some types of anxiety responses, and that the CRF system may be a possible therapeutic target for the treatment of anxiety disorders [Bailey et al., 2011]; Central injections of CRF activate “panic/defense” responses [Ku et al., 1998] and injections of the panic-inducing agent doxapram activate CRF neurons [Choi et al., 2005]. Consistent with these preclinical findings, clinical evidence suggests that polymorphisms in the CRF1 receptor gene may be associated with panic [Keck et al., 2008] and that the CRF system may also be an important therapeutic target for PD [Risbrough and Stein, 2006]. The fear response without HPA axis activation has been activated during naturally occurring, spontaneous panic attacks [Kellner and Wiedemann, 1998; Abelson et al., 2007; Preter and Klein, 2008].
The biological effects of CRH and the urocortins (UCNs) are mediated by two distinct receptors, CRH receptor type 1 (CRHR1) and 2 (CRHR2), which both belong to the G protein-coupled receptor superfamily of proteins [Perrin et al., 2006]. Two separate genes encode the CRH receptors. CRHR1, a 415-amino acid protein, exhibits high affinity towards CRH and UCN, but low affinity towards UCN2 and no affinity towards UCN3. CRHR1 is primarily expressed in the CNS and the anterior pituitary. The CRHR2 receptor shares 70% sequence identity with CRHR1 and is expressed primarily in extra-CNS sites. The CRHR2 receptors exhibit high affinity towards UCNs and no affinity towards CRH [Bale, 2005].
To comprehensively map the polymorphic genetic variation in CRHR1 and CRHR2 in relation to MDD and PD, we analyzed 12 single nucleotide polymorphisms (SNPs) in MDD patients, PD patients, and healthy controls.
MATERIALS AND METHODS
All cases and control subjects were ethnically Japanese and were recruited in the vicinity of Oita prefecture, Japan. Subjects comprised 173 unrelated Japanese with MDD (91 males and 82 females; age = 41.6 ± 16.2 years) and 180 unrelated Japanese with PD diagnosed according to DSM-IV criteria (79 males and 102 females; age = 40.9 ± 10.5 years (mean ± SD), while 285 unrelated healthy volunteers (173 males and 112 females; age = 32.8 ± 8.7 years) served as controls. Patients in the MDD group had either pure MDD or MDD comorbid with anxiety disorders (31.8%). Among the patients with PD, 33.9% had comorbid MDD and 5.6% had other anxiety disorders. Cases for MDD and PD were recruited from Oita University Hospital in Oita. All cases were outpatients or stable in-patients. Each diagnosis was confirmed using the Mini-International Neuropsychiatric Interview (MINI) [Sheehan et al., 1998] and clinical records were also reviewed. The controls received a short interview (MINI) conducted by one of the authors (J.A.) to exclude a history of major psychiatric illness. Exclusion criteria were: (1) central nervous system disorders and medical disorders clearly affecting cerebral function; (2) alcoholism or drug dependence; (3) personality disorder (SCID-II); (4) presence of brain injury or disease; (5) pregnancy; (6) mental retardation; (7) age <18 years; (8) immediate danger of suicide; (9) treatment with lithium and/or carbamazepine; and (10) endocrine disorder of the pituitary or the adrenal gland. The objective of the present study was clearly explained and written informed consent was obtained from all subjects. The study was approved by the Ethical Committee of the Faculty of Medicine, the University of Oita.
Genomic DNA was extracted from leukocytes using the standard phenol–chloroform method. The CRHR SNP selection was basically performed according to the MAF standard (minor allele frequency, HapMap Project, 2003) with estimates higher than 0.1. We selected six SNPs (rs4076452, rs7209436, rs110402, rs242924, rs242940, and rs173365) in CRHR1 and six SNPs (rs4722999, rs3779250, rs2267710, rs1076292, rs2284217, and rs226771) in CRHR2. Individual genomic DNAs were genotyped for every SNP on the Taqman®PCR SNP genotyping assay, using a Roche LightCycler480 (Basel, Switzerland). The standard 10 ml PCR reaction containing 2 ng genomic DNA was carried out using the Taqman®PCR including Universal PCR Master Mix under the protocol guidelines.
All of the parameter calculations, allele and genotype frequencies, and pair-wise linkage disequilibrium analyses were conducted using the software Haploview 4.1 version [Barrett et al., 2005]. Hardy–Weinberg equilibrium calculations, haplotype analyses, and P-value permutations were carried out using the program SHEsis (http://analysis.bio-x.cn) [Shi and He, 2005; Li et al., 2009], a robust and user-friendly platform with integrated analysis tools particularly suited to association studies. P values were corrected using 10,000-permutation corrections. The permutations considering both of the two test types (allelic tests and genotype tests) were performed for each of the two disease analyses (e.g., MDD vs. control, PD vs. control). The significance level was set at α = 0.05.
In the 638 Japanese samples, genotype distributions were in Hardy–Weinberg equilibrium for all of the SNPs. The allele and genotype frequencies of 12 SNPs in the two-patient sample groups and the healthy controls are listed in Tables I and III. The haplotype analysis of the two-patient sample groups and the healthy controls are shown in Tables II and IV. The linkage disequilibrium among the 11 SNPs is shown in Figures 1 and 2. The selection criteria for haplotypes used in the haplotype analyses were adjacent SNPs with pairwise D′ > 0.85. In our analyses, haplotypes with frequencies above 0.03 were tested.
Table I. Genotype and Allele Frequency of the Six SNPs in the CRHR2 Gene
Corrected P-value after 10,000 permutations and the global P was the P-value for analysis of all haplotypes in one block. The significance level was set at α < 0.05.
CRHR1 and Major Depressive Disorder
According to the selection criteria, five SNPs (rs7209436, rs110402, rs242924, rs242940 and rs173365) of CRHR1 with strong D′ > 0.85 were in one block for MDD (Fig. 1). Two SNPs (rs7209436 and rs110402) with strong pairwise D′ > 0.85 were in one block and another two SNPs (rs242940 and rs173365) of CRHR1 in another block for PD (Fig. 2). Regarding CRHR2, two SNPs (rs4722999 and rs3779250) with strong pairwise D′ > 0.85 were in one block and another two SNPs (rs1076292 and rs2284217) were in another block for MDD (Fig. 3) while three SNPs (rs1076292, rs2284217, and rs2267716) with strong pairwise D′ > 0.85 were in one block for PD (Fig. 4).
We found that rs110402 and rs242924 in CRHR1 were positively associated with MDD in terms of both allele and genotype distributions (rs110402, allele: P = 0.001, genotype: P = 0.008, odds ratio (OR) = 1.81 [95% CI = 1.25–2.60]; rs242924, allele: P = 0.013, genotype: P = 0.013, OR = 1.02 [95% CI = 0.07–1.54]) (Table I). After a 10,000-permutations correction, rs110402 was still significant in the genotype distribution (P = 0.008). Haplotype analysis revealed that the T–A–T–G–G haplotype consisting of rs7209436 and rs173365 in CRHR1 was positively associated with MDD (P = 0.0003). The corrected global P-value was 1.14E-5. Haplotype analysis revealed that the T–A haplotype consisting of rs7209436 and rs110402 in CRHR1 was positively associated with MDD (P = 0.0013). The corrected global P-value was 3.75E-5 (Table II).
CRHR2 and Major Depressive Disorder
Furthermore, we found that rs3779250 in CRHR2 was positively associated with MDD in both allele and genotype distributions (allele: P = 1.75E-11, genotype: P = 2.08E-11, odds ratio (OR) = 2.83 [95% CI = 2.08–3.85]) (Table III). After 10,000-permutation correction, rs3779250 was still significant in the genotype distribution (P = 0.000). Haplotype analysis revealed that the C–C haplotype consisting of rs4722999 and rs3779250 in CRHR2 was associated with MDD. The corrected global P-value was 5.55E-16 (Table IV).
Table III. Genotype and Allele Frequency of the Six SNPs in the CRHR2 Gene
Corrected P-value after 10,000 permutations and the global P was the P-value for analysis of all haplotypes in one block. The significance level was set at α < 0.05.
CRHR1 and Panic Disorder
We found that rs242924 in CRHR1 was positively associated with PD in both allele and genotype distributions (rs242924, allele: P = 0.022, genotype: P = 0.033, odds ratio (OR) = 1.55 [95% CI = 1.06–2.27]) (Table I). Haplotype analysis revealed that the T–A–T–G–G haplotype consisting of rs7209436 and rs173365 in CRHR1 was not associated with PD (P = 0.1354). The corrected global P-value was 0.0017. Haplotype analysis revealed that the T–A haplotype consisting of rs7209436 and rs110402 in CRHR1 was not associated with PD (P = 0.4612).
CRHR2 and Panic Disorder
Finally, we found that rs3779250 in CRHR2 was not associated with PD in both allele and genotype distributions (Table III). Haplotype analysis revealed that the C–C haplotype consisting of rs4722999 and rs3779250 in CRHR2 was associated with PD (P = 0.0115). The corrected global P-value was 2.00E-15 (Table IV).
The main finding of the present study was a significant association of the CRHR1 and CRHR2 genes with MDD and PD in a Japanese population. The data suggest that CRHR1 and CRHR2 might be involved in MDD and PD susceptibilities, and support previous reports that MDD and PD may have a substantial overlap in terms of pathogenesis [Simon and Fischmann, 2005]. MDD and PD are severe mental disorders underlying genetic components as well as contributory experiential and environmental factors [Anisman et al., 2008]. In our study, following 10,000-permutation correction, we found that the CRHR1and CRHR2 genes were associated with MDD and PD.
Liu et al.  first reported that the CRHR1 gene might be a new susceptibility gene for depression. In their study, they found significant allele and genotype associations with rs242939, and the haplotype defined by alleles G–G–T for rs1876828, rs242939, and rs242941 was significantly over-represented in major depression patients compared to controls. However, in our study, we used rs4076452, rs7209436, rs110402, rs242924, rs242940, and rs173365 from Liu's study. In the present study, we found an association between two SNPs (rs110402 and rs242924) of the CRHR1 gene with MDD. Rs110402 was shown to be associated with an early onset of a first episode of depression [Papiol et al., 2007]. CRH is a crucial activator of the HPA axis, binding to receptors that initiate the stress response, concluding with the release of cortisol from the adrenal cortex. Recently, CRHR1 polymorphisms were shown to interact with childhood maltreatment to predict HPA axis reactivity, which has been linked to both MDD and early life stress [Tyrka et al., 2009]. Differences in the CRHR1 function have been linked to risk for depression, in the presence of childhood maltreatment [Bradley et al., 2008; Polanczyk et al., 2009]. The TAT haplotype formed by rs7209436, rs110402, and rs242924 was associated with a significant protective effect [Bradley et al., 2008; Polanczyk et al., 2009]. We could find five SNPs (rs7209436, rs110402, rs242924, rs242940, and rs173365) of CRHR1 with strong D′ > 0.85 were in one block for MDD (Fig. 1). The T–A–T–G–G haplotype consisting of rs7209436 and rs173365 in CRHR1 was positively associated with MDD. The T–A haplotype consisting of rs7209436 and rs110402 in CRHR1was positively associated with MDD (Table II). We did not find that the TAT haplotype was associated with MDD. We did not find that there was strong positive association between the rs7209436–rs173365 haplotype and MDD (Fig. 1). Another group examined SNPs in CRHR1 in a large population sample [Grabe et al., 2010]. They detected a significant moderator effect of the TAT haplotype on the risk of adult depression symptoms. However, the haplotype was associated with greater depressive symptoms only among individuals who experienced moderate-to-severe physical neglect. We could not examine the relationship between childhood maltreatment and MDD. Both rs110402 and rs242924 showed a significant interaction with maltreatment in the prediction of cortisol response to the DEX/CRH test [Tyrka et al., 2009]. That the GG genotype of both SNPs was linked to higher cortisol responsiveness to the test extends recent findings of an increase in depressive symptoms among subjects with the GG genotype of these SNPs who reported a history of childhood maltreatment [Bradley et al., 2008]. Maltreatment at this critical age might affect our results, and further studies are needed to determine the relationship between childhood maltreatment and MDD in the Japanese.
In our study, we found a significant association between rs3779250 in CRHR2 and MDD. The CRHR2 gene was reported unlikely to be involved in the genetic liability underlying HPA axis dysfunction and mood disorders [Villafuerte et al., 2002]. CRHR2 may function endogenously to dampen or modulate the stress response associated with CRHR1 activation, and therefore CRHR2 may play a counteracting role in the CRHR2 in the HPA axis stress response [Bale, 2005]. However, ethnicity and sample size may explain how the Villafuerte et al. results differ from the present study, as their 89 patients were of Belgian origin. Allele G of rs2270007 of CRHR2 gene was associated with a worse overall response to citalopram after treatment follow-up [Papiol et al., 2007]. Genetic variants like rs2270007, located on genes involved in HPA axis regulation, could make this axis more adaptable in presence of antidepressant treatment. Further exploration of the CRHR1 and CRHR2 findings in large case–control PD samples may provide more definitive evidence implicating these loci in the genetic etiology of PD.
We found a significant association between rs242924 in CRHR1 and PD. Some studies have reported an absence of associations of polymorphisms in CRHR1 in relation to PD [Keck et al., 2008; Hodges et al., 2009]. The SNP pair in the combined with rs878886 in CRHR1 and rs28632197 in arginine vasopressin (AVP) 1B was significant association in PD [Keck et al., 2008]. In these cases, ethnicity might again explain the differences between their data and the present study. We did not find the association between CRHR2 and PD. Some studies reported that the CRHR2 polymorphisms examined did not confer susceptibility to PD [Tharmalingam et al., 2006; Keck et al., 2008]. Their SNP is different from the SNPs of our present study. Further studies investigating additional polymorphisms in this gene and other components of the CRH signaling system may prove useful. In conclusion, these results provide support for an association of CRHR1 and CRHR2 with MDD and PD.
This study is supported by a Grant-in-Aid for Scientific Research(C) (No. 20591371) from the Japan Society for the Promotion of Science (JSPS). We are grateful to Ms. Kazumi Oda for excellent technical assistance.