Glucocorticoid Receptor Genetic Variants and Response to Fluoxetine in Major Depressive Disorder

Major depressive disorder (MDD), a prevalent psychological disorder associated with considerable morbidity,¹ is the third leading cause of disability worldwide and the first in middle- to high-income countries.² Almost 350 million people worldwide are suffering from MDD, among whom only 25% receive effective cure.³ Based on a World Health Organization report, the prevalence of depression in Asia and Africa is less compared to America and Europe.² Research has shown that along with environmental parameters, endogenous factors contribute substantially to the development of MDD.^4–6 Moreover, the role of genetic and epigenetic factors as the main contributors to this illness is further pronounced.⁷

Dysregulation of hypothalamic-pituitary-adrenocortical (HPA) system activity is one of the major neuroendocrine abnormalities in MDD, resulting in elevated plasma levels of corticotropin and cortisol.^8,9 Function of the HPA axis interrelates with glucocorticoid (GC) signaling, which is disrupted in MDD,^10–12 as well as perfectionism¹³ and other pathologic conditions, such as metabolic syndrome¹⁴ and asthma.¹⁵ GCs regulate the HPA axis through binding to GC receptors (GCRs). These receptors are expressed in almost every part of the body, such as several regions of the forebrain.¹⁶ There are reports on the impact of forebrain GCRs in HPA axis function in depression-like behavior, which highlight the role of GCRs in the development of MDD¹⁰ and at the same time propose therapeutic implications. Alongside the reduction in GCR mRNA levels in brain areas like the hippocampus and the frontal cortex in MDD patients,¹⁰ some studies have reported an association between polymorphisms in the GCR and MDD.^17–19 Further evidence is provided by Won et al.,²⁰ who argue the possible influence of GCR genetic variants on hippocampal shape and integrity of parahippocampal subdivision of the cingulum in MDD. The structural changes secondary to the GCR altered activity can also explain the association between different variants of GCR and childhood depression in Castellini’s report.²¹ Other psychiatric conditions with close ties with GCR activity include posttraumatic stress disorder (PTSD)²² and suicidal thoughts and behavior,²³ which are reportedly associated with GCR genetic variations.

Antidepressants are believed to stimulate GCR function and expression in human and animal models of depression.^24–27 They induce GCR nuclear translocation and transactivation mediated by the cyclic adenosine monophosphate (cAMP)/protein kinase A (PKA) cascade. The negative feedback of GCR on HPA axis in turn reduces HPA-axis activity.^26,28,29 Altogether, these observations make the GCR gene a prime candidate for associations with altered clinical response to antidepressant drugs. Among the antidepressants, selective serotonin reuptake inhibitors are common first-line treatments,³⁰ and fluoxetine is frequently prescribed because of its efficacy and tolerability.³¹ Fluoxetine is efficacious for MDD in all ages, although more in adults and youths compared with geriatric individuals.³² It can inhibit the steroid transporters present in neurons and the blood-brain barrier as well, resulting in an increased amount of cortisol in the brain, which negatively balances the HPA axis.³³

There is no doubt that pharmacotherapy has alleviated morbidity of individuals with depression, but only 30%−40% of patients seem to respond completely to treatment.³⁴ The hypothesis of this study is based on these observations and the interplay of fluoxetine with GCRs, which suggest clues for differential therapeutic response to fluoxetine upon the inheritance of different genotypes of GCRs. In the present study, we assessed the clinical response to fluoxetine to investigate whether the three polymorphisms in GCR, including rs41423247, rs6195, and rs6189/rs6190, might have any influence on response to treatment in a sample of depressed patients.

Methods

This work was carried out in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki) and Uniform Requirements for manuscripts submitted to biomedical journals. The study was approved by the local committee for ethics of medical experiments on human subjects of Shiraz University of Medical Sciences. The written consent was obtained from the participants prior to the interview. All patients were of Caucasian origin from the same geographical area.

Patients with MDD were recruited from Hafiz Hospital in Shiraz, Iran, between 2011 and 2012. All patients were at least 18 years old, with the diagnosis of MDD according to the DSM-V criteria.³⁵ A total of100 newly diagnosed MDD patients, defined as a negative previous diagnosis of depression and no history of antidepressant medications use, were included in the study (male: 33, female: 67, mean age±SD=32.87±10.65). The 21-item Hamilton Depression Rating Scale (HAM-D) was used to evaluate the severity of depression.²⁶ Exclusion criteria were as follows: a family history of schizophrenia, a personal history of bipolar disorder, a family history of bipolar disorder in first-degree relatives, a personal history of schizophrenia, manic or hypomanic episode, mood incongruent psychotic symptoms, active substance dependence, current treatment with antipsychotics or mood stabilizers, and significant medical conditions.

Drug Administration

Fluoxetine (FLUOXETINE-ABIDI) was administered at a fixed-dose regimen of 20 mg for one week initially, followed by dosages of 20–80 mg/day according to clinical response over the 6-week treatment duration. Only hypnotics such as zolpidem or anxiolytics such as chlordiazepoxide were allowed for severe anxiety. A positive clinical response to fluoxetine treatment was considered if at least 50% reduction in the baseline HAM-D score was observed through the sixth week. Evaluations were done at baseline and weeks 1 and 6 of treatment.

Prior to interviewing, 5 ml of venous blood samples were collected for further genotyping.

DNA Extraction and Genotyping

Genomic DNAs were extracted from whole blood using a standard method.³⁶

PCR amplification of rs41423247, rs6195, and rs6189/rs6190 was carried out using primers mentioned in Table 1. The PCR protocol included 1 cycle of denaturation at 95°C for 5 minutes, followed by 30 cycles of denaturation at 94°C for 1 minute, annealing at 55°C for 1.5 minutes (except for rs6195 polymorphism [exon 2–5] at 51°C), extension at 72°C for 1.5 minutes, and a final cycle at 72°C for 5 minutes.

Restriction fragment-length PCR (RFLP) analysis was performed to determine genotype frequencies.

Forrs41423247 polymorphism, 5μL of relevant PCR product (335 base pair [bp]) was digested by 10 units of rs41423247 restriction endonuclease (Roche Applied Science, Mannheim, Germany) for 3–16 hours at 37°C.

The CC genotype produced two bands (117 and 222 bp), the CG variant gave three fragments (117, 222, and 335 bp), and the GG genotype remained undigested (335 bp).

The rs6189/rs6190 polymorphism genotypes were determined by digesting 10μL of relevant PCR product (448 bp) with 1.25 μL of MnlI restriction enzyme (New England Biolabs, Beverly, MA) at 37°C for 3–16 hours. Endonuclease digestion yielded fragments of 142 and 163 bp for the wild-type allele (ER22/23ER genotype); heterozygous allele (ER22/23EK genotype) appeared as three bands of 142, 163, and 177 bp; and homozygous allele (EK22/23EK genotype) gave 163 and 177 bp fragments.

To determine rs6195 polymorphism genotypes, 10μL of relevant PCR product (243 bp) were digested with 5 μL of Tsp509I restriction endonuclease (New England Biolabs, Beverly, MA) for 3–16 hours at 65°C. Digestion produced 3 fragments of 19, 95, and 134 bp (wild type allele, N363N) or two bands of 95 and 153 bp (heterozygous mutant allele, N363S).

Digested fragments were separated by electrophoresis on agarose (InvitrogensUltraPure) gel 2% after 3–16 hours of incubation (Table 1). They were then stained by ethidium bromide and visualized in an ultraviolet transilluminator. It is important to mention that all of the samples were genotyped at least twice and reconfirmed.

Statistics

Data were analyzed using SPSS 21.0 for Windows (SPSS Inc., Chicago). Hardy–Weinberg equilibrium (HWE) for distribution of genotypes was calculated using chi-square test. Continuous variables are demonstrated as mean±SD. Genotype frequencies are shown in percentage (%). Distribution of all continuous variables was tested for normal distribution with the Kolmogorov–Smirnov test. Associations between categorical variables were determined by Pearson's chi-square or Fisher's exact test and for interval data by Student's t test. Univariate analysis of genotypes was performed using chi-square test. Odds ratio (OR) and 95% confidence intervals (CI) were obtained. We also analyzed the distribution of genotype frequencies under three different genetic models (additive [CC=0, CG=1, GG=2], recessive [CC and CG versus GG], and dominant [CG and GG versus CC]), using the SNPassoc package RV.3.0.1 (http://www.Rproject.org).³⁷ A p value <0.05 was considered as statistically significant.

Results

A total of 100 MDD patients were included in this study. In terms of response, 70 patients (70%) were responders to fluoxetine (baseline HAM-D: 23.8±6.9; HAM-D after week 6: 6.2±4.8). Table 2 demonstrates patients’ demographic information.

Table 3 shows genotype and allele frequencies of patients receiving fluoxetine based on 50% score reduction. The distribution of study genotypes was in agreement with Hardy-Weinberg equilibrium. Regarding rs41423247 polymorphism, carriers of CG and GG genotype responded significantly better to fluoxetine compared with CC carriers (p=0.008, OR=3.3, 95% CI=1.35–8.07; Fisher's exact test, two-sided). Moreover, the G allele of rs41423247 polymorphism was strongly associated with response to fluoxetine (p=0.032, OR=2.2, 95% CI=1.09–4.44; Fisher's exact test, two-sided). Further analysis under log-additive, recessive, and dominant models for rs41423247 revealed that log additive and dominant models are, respectively, 2.41 (95% CI=1.13–5.14) and 3.31 (95% CI=1.36–8.07) times more likely to respond to fluoxetine (Table 4).

There was no significant association between different genotypes and alleles of rs6195 and rs6189/rs6190 variants and response to fluoxetine (p=0.213 and 0.99, respectively).

Discussion

To our knowledge, this is the first study investigating the relationship between GCR genetic polymorphisms and response to fluoxetine in depressed Iranian patients. Here, we found that MDD patients who carry the G allele of rs41423247 respond approximately two times better than the carriers of the C allele (p=0.032, OR=2.2, 95% CI=1.09–4.44) and that carriers of CG and GG genotypes respond three times more to fluoxetine compared with CC carriers (p=0.008, OR=3.3, 95% CI=1.35–8.07).

According to a report by Van Rossum et al., the GG and CG genotypes were functionally equivalent because cortisol levels were significantly lower in the carriers of the two genotypes (G-allele carriers) than CC carriers after dexamethasone suppression tests.³⁸ Our results support this conclusion, showing the equivalence of the therapeutic response behavior of the G-allele carriers. Furthermore, the associated increased sensitivity to GCs in carriers of the G allele can explain the observed increased responsiveness to fluoxetine.³⁸ The high expression of GCRs in hippocampal formation underlies the atrophic response to stress hormones in these regions of the brain.³⁹ Being therefore at increased risk of depression secondary to these structural changes, carriers of the G allele of rs41423247 are more responsive to antidepressants like fluoxetine, because stress hormone levels are readily decreased in response to medication.^39,40

With the increasing rate of depression worldwide,² response and remission rates are quite unsatisfactory.⁴¹ While many studies highlight the contribution of genetics in susceptibility to this illness,^42,43 the role of genetics as a therapeutic determinant still remains elusive. Along with the association studies focusing on the monoaminergic hypothesis of MDD,^44–47 newer candidate genes affecting the HPA axis have gained attention as well.^17,48,49 There is some evidence of the involvement of HPA axis in MDD,^50,51 and previous findings suggest the interrelation between HPA axis and GC signaling as a pathophysiologic determinant of MDD development and progression.^25,50,52

Antidepressants stimulate GCR function and expression in depressed patients and animal models.³³ In vivo and in vitro studies suggest that GCR gene expression and, at the same time, sensitivity to glucocorticoid activation are increased by means of antidepressants.^25,33,52 They induce GCR nuclear translocation and transactivation mediated by the cAMP/PKA cascade^26,29,33 and control neurogenesis in hippocampus by GCR-dependent mechanisms requiring PKA signaling. This is accompanied by alterations in GCR phosphorylation and GCR-dependent gene transcription, which eventually result in enhanced neurogenesis. The effect that antidepressants exert on GCR signaling eventuates in a negative feedback on the HPA axis activity as well and, overall, suggests the diverse mechanisms of action for medications used in MDD.²⁶

Regarding the study polymorphism, rs41423247 has been shown to be associated with several psychiatric abnormalities. In a study conducted in Polish adolescent girls, rs41423247 was associated with higher incidence of anorexia nervosa,⁵³ a condition in which almost 80% of the patients suffer from MDD as well.⁵⁴ Role of stressful life events on PTSD by means of GCR gene polymorphisms has been reported by Lian et al.²² A study in overweight children also advocates for the role of this variant in developing emotional dysregulation.²¹ Contrary to these findings, lack of association between this variant and perinatal depression was observed in a group of Chinese descent.⁵⁵ Another report examined the impact of physical activity with MDD and suicidal ideation with respect to rs41423247 and found that this polymorphism does not influence these two outcomes.⁵⁶

Genetic components of GCRs as therapeutic determinants therefore emerged from both the pathophysiological involvement of GCRs in MDD and the effects of antidepressants on GCR-related pathways. This can even translate into the personalized response to certain medications and explain some of the variabilities associated with fluoxetine response rates among MDD patients. Since the proposition of the involvement of GCR-related pathways in the overall function of the antidepressants, there have only been a few studies reporting the association of rs41423247 polymorphism and response to antidepressants. Consistent with our findings regarding the association of rs41423247 polymorphism with response to fluoxetine, Takahashi et al. has reported the same result in a Japanese population.⁵⁷ On the other hand, van Rossum et al., despite describing a significant association between variants of rs41423247 and MDD, has reported lack of association between variants of this polymorphism and antidepressant treatment.¹⁸ In the mentioned study, however, different classes of antidepressants were used, and differential analysis regarding each class with respect to rs41423247 was not performed. Ventura-Junca et al. also reported lack of association between rs41423247 variant and fluoxetine treatment.⁵⁸ Brouwer et al., in contrast with our findings, reported that carriers of the G allele, especially male patients, seem to be more resistant to therapy.⁵⁹

Regarding rs6189/rs6190 and rs6195 polymorphisms, our results showed lack of association with response to treatment, which is consistent with a previous report in a Dutch population.⁵⁹ However, in a German population, rs6189/rs6190 polymorphism was associated with a faster response to antidepressant treatment.¹⁸ The rs41423247 and rs6195 polymorphisms have been shown to be related to hypersensitivity to glucocorticoids,³⁸ while the rs6189/rs6190variant is associated with glucocorticoid resistance.⁶⁰ It seems that both glucocorticoid resistance and increase in GCR effects in parts of the brain contribute to response to antidepressant treatment.

It is noteworthy that we only recruited patients labeled as newly diagnosed with MDD, which refers to the absence of any concomitant psychiatric disorders and parallel medication use in them. Therefore, the observed results, although not obtained from a clinical trial design, would reflect a more probable cause-and-effect relationship. This is especially important in weighing the evidence in observational studies like ours and can corroborate the relation of a better response rate to fluoxetine with carrying the G allele of GCR gene, in that the other confounders are kept at a minimal level. We believe a randomized allocation of fluoxetine versus placebo in G and C allele carriers is a mandatory next step to justify our results. Besides, regarding conflicting reports suggesting diverse responses to therapy in G allele carriers in different populations, more studies ought to be done to reproduce the results and make the underlying association clearer.

Conclusions

Our data reveal that rs41423247 polymorphism is associated with response to fluoxetine. This further supports the hypothesis of involvement of GCR in response to treatment in depression, in which it is an influential element on the HPA-axis regulation.