Preliminary evidence of acylated ghrelin association with depression severity in postmenopausal women

We have previously shown increased depression and anxiety scores in postmenopausal overweight women, when compared to overweight premenopausal women. The mechanisms responsible for these alterations are not understood. Although ghrelin involvement in mood modulation has been suggested, its role is still ambiguous and has not been evaluated in postmenopause. Here we investigated the association of ghrelin with depression and anxiety symptoms in postmenopausal women. Fifty-five postmenopausal women with depression symptoms, who were not in use of hormones or antidepressants, were included in the study. Depression symptoms were evaluated by Beck’s Depression Inventory (BDI) and Patient Health Questionnaire-9 (PHQ-9) and anxiety symptoms were evaluated by Beck’s Anxiety Inventory (BAI). Women were allocated into three groups, according to BDI classification of mild, moderate, or severe depression symptoms. Anthropometric, biochemical and hormonal parameters were analyzed. Total and acylated ghrelin levels were higher in the severe depression than in the mild depression group. Multivariate regression analyses showed positive associations of BDI scores with acylated ghrelin and BMI, and of PHQ-9 scores with acylated ghrelin and homeostasis model assessment of insulin resistance (HOMA-IR). BAI scores associated positively with waist-to-hip ratio. To the best of our knowledge, this is the first demonstration of an association between acylated ghrelin and the severity of depression symptoms in postmenopausal women. This association may reflect either a physiological response aimed at fighting against depression symptoms or a causal factor of this mental disorder.

Ghrelin involvement in modulation of mood disorders has also been investigated but its role in depression and anxiety is still ambiguous. Although some studies in rodents have shown antidepressant and anxiolytic effects of total and acylated ghrelin [20][21][22] , other reports have evidenced depressogenic and anxiogenic roles [23][24][25] . Clinical data about ghrelin levels in depressive patients are also controversial. While some studies have reported no changes in acylated ghrelin levels in depressive patients 26 , others have reported decreased 27 or increased levels 28 . The clinical heterogeneity observed in major depressive disorder might contribute, at least in part, to these inconclusive results. An association between total ghrelin levels and eating disorders in depressive individuals has been shown 29 . Intriguingly, among depressive patients, the ones who were experiencing increased appetite presented lower acylated ghrelin levels than both the ones reporting decreased appetite and the healthy subjects 30 .
Reduced serum total and acylated ghrelin has been observed in postmenopausal women [31][32][33] and estrogen replacement therapy was able to increase acylated ghrelin levels 34 and to improve hypothalamo-pituitary sensitivity to acylated ghrelin 35 . Weight loss has also been shown to raise total ghrelin levels in this population 36 . In rodents, the involvement of estrogens in the regulation of total ghrelin expression has been indicated 37 and an antidepressant-like effect induced by ghrelin in ovariectomized mice has been shown 38 .
We were not able to find any studies investigating ghrelin's association with mood disorders in postmenopausal women. A better comprehension of the postmenopause-associated changes and the factors influencing them will certainly improve the management of the symptoms affecting the quality of life of middle-aged women. In the present study, we hypothesized that ghrelin influences the depression and anxiety symptoms in postmenopausal women.
There were no significant differences in anthropometric parameters among the 3 groups ( Table 2). As shown in Table 3, the severe depression group had higher levels of total and acylated ghrelin than those of the mild depression group (H (2, n=50) = 6.17622, p = 0.045 and H (2, n=49) = 8.63559, p = 0.013, respectively). All other clinical parameters failed to show significant differences among the 3 groups.
The Pearson's correlations were calculated including all women and all the variables assessed in the study, which included psychological scores, metabolic hormones and anthropometric measurements. Table 4 shows only the variables having at least one significant correlation with the psychological scores. A complete table containing all the correlations across all parameters is included as supplementary material (Supplementary Table 1).
For BAI scores, SMM, BMR, and insulin showed non-significant influence. Due to multicollinearity, we tested SMM and not FFM. The final model equation predicted 18% of BAI scores variations (1 predictor, statistical power = 0.99, effect size = 0. 19) and showed that only WHR was positively associated with BAI scores ( Table 6). The model implies that every 0.1 unit of WHR increase (95% CI 27.8-119.6) is associated with 8.1 increases in BAI scores.
For PHQ-9 as the dependent variable, WHR showed non-significant influence. Due to multicollinearity between total and acylated ghrelin and among glucose, insulin, and HOMA-IR variables, we selected acylated ghrelin and HOMA-IR for the regression analyses. The final model explained 26% of PHQ-9 variations (2 predictors, statistical power = 0.94, effect size = 0.35), showing positive association of acylated ghrelin and HOMA-IR with PHQ-9 scores ( Table 7). The final model implies that for every one-unit increase in HOMA-IR the PHQ-9 score increases 2.37 (95% CI 0.73-4.01), and that for every one-unit increase in acylated ghrelin the PHQ-9 score increases 0.015 (95% CI 0.008-0.032).

Discussion
The present results indicated that, after menopause, depressive and anxious symptoms are associated with both acylated ghrelin levels and parameters indicative of obesity and insulin resistance.
Among the postmenopausal women evaluated in the present study, all having some degree of depressive symptoms, we observed that the anxiety symptoms were higher in the groups with moderate and severe depression, when compared to the mild depression group. Moreover, the BAI scores correlated positively with those of the BDI and PHQ-9 inventories. These data are in agreement with previous studies reporting a bidirectional association between depression and anxiety symptoms in the general population [39][40][41] . Suggested mechanisms for this association emphasized the existence of common neurobiological basis for anxiety and depression symptoms, including abnormal activity of the hypothalamic-pituitary-adrenal axis (HPA axis), dysfunction of the γ-amino butyric acid (GABA) system, increased oxidative stress, and disturbance of neural plasticity [42][43][44] .
The three groups of depressive postmenopausal women presented median values of body parameters categorized as overweight, abdominal obesity, and high percentage of body fat 45 (WHO, 2008), in agreement with previous reports 46,47 . These alterations have been associated with metabolic, sleep, cardiovascular, and psychological disorders in postmenopausal women 1,46,48 . These considerations emphasize the importance of monitoring the anthropometric parameters and encouraging healthy lifestyle habits in order to avoid weight gain in this population.
Moreover, an important influence of total and abdominal obesity on mental disorders has been described in middle-aged women 1 as well as in the general adult population 49 . Here, we demonstrate such a connection in postmenopausal women, as WHR associated positively with BAI scores and BMI associated positively with www.nature.com/scientificreports/ BDI scores. Also, PHQ-9 scores associated with HOMA-IR. These findings indicate that obesity has a relevant influence on psychological symptoms in postmenopausal women.
In the general population, the deleterious influence of obesity exerts on psychological status has been associated with production of inflammatory mediators, stimulation of the HPA axis, and insulin resistance. Additionally, a role has been attributed to the dissatisfaction with body image and low self-esteem [49][50][51] . These mechanisms are likely involved in the pathogenesis of anxiety and depression in postmenopause with overweight. Additionally, impairment of monoaminergic function due to loss of estrogens could be relevant 52,53 . Stressor agents typical of this period of women`s life may also play a part, such as changes in social roles, financial insecurity, and empty nest syndrome 48,54,55 .
Increased incidence of depression has also been demonstrated in type I and type II adult diabetic patients 56 . A relevance to this association has been attributed to the central actions of insulin, since the hormone mediates neuromodulatory, neuroprotective and neurotrophic effects in the brain 57 . Impairment of central insulin action has been associated with cognitive impairment and depression, through mechanisms involving alterations in mitochondrial function, monoamine oxidase expression, and dopamine turnover 58 .
In the present study, we found that total and acylated ghrelin levels were significantly higher in the group with severe depression, in comparison to the group with mild depression, and that both total and acylated ghrelin levels correlated positively with both BDI and PHQ-9 scores. Moreover, in the linear regression analyses, acylated ghrelin appeared as an independent factor associated with depression symptoms in both questionnaires. This is the first study showing that the higher the serum acylated ghrelin levels, the more severe the depression symptoms in depressed postmenopausal women.
Although an association of depression symptoms and ghrelin levels has been studied in distinct populations, whether there is a positive or a negative relationship between them has not been established. In eutrophic young adults, elevated levels of acylated and des-acyl ghrelin were described in patients with severe and moderate depression 28 and in individuals evaluated after suicide attempts 59 . In both eutrophic and overweight depressive adult patients, total ghrelin levels were higher than those in the respective non-depressive controls and showed an important reduction after depression treatment 60,61 . Reduction of the hormone levels after citalopram has also been reported in young eutrophic adults who, however, had pre-treatment levels of both acylated and des-acylated ghrelin even lower than those of the control group 27 . In contrast, there are reports of normal total and acylated ghrelin levels in overweight adults diagnosed with depression 62 or eutrophic adults showing depressive symptoms 26,63 . Studies focusing on the ghrelin/depression association in women only are scarce. In eutrophic adults, total ghrelin levels were positively correlated with depression incidence in female but not in male individuals 64 .
The presence of a positive association of ghrelin and depression, as evidenced in the present study, could reflect either a physiological response aimed at fighting against depression symptoms or a causal factor of this mental disorder. Unfortunately, the available data are not conclusive. In healthy male rats, either juvenile or adult, the intracerebroventricular administration of total and acylated ghrelin increased depressive-like behaviors 23,25 , implicating the hormone as a possible pathophysiological mechanism for depression. A study in humans reported that the administration of acylated ghrelin to 7 eutrophic adults with unmedicated major depression failed to significantly modify depressive symptoms in both men and women 65 . At the best of our knowledge, no other interventional studies evaluated ghrelin effects on depression in humans. Contrastingly, there are demonstrations of an anti-depressant effect of ghrelin in murine models of depression induced by chronic stress 20,22,66 , what suggests that increased endogenous ghrelin levels could be a mechanism aimed at fighting the disease.
The relevance of ghrelin for depression in postmenopause also remains to be determined. The triad ghrelin/ depression/estrogen relationships are probably complex, as indicated by demonstrations that blockade of estrogen receptors hindered the anti-depressant effect of ghrelin in ovariectomized mice 38 and that, conversely, estradiol administration increased the expression of ghrelin receptors in the arcuate nucleus of female mice 67 . There is also evidence that estradiol replacement augmented ghrelin-induced stimulation of GH levels in postmenopausal women 35 .
The above data indicate that, although the majority of animal studies have indicated an anti-depressant effect of ghrelin, more studies in humans are necessary to ascertain this aspect.
It is important to clarify that the current findings may not be generalized due to the small sample size of the severe depression group. Therefore, the findings need to be interpreted cautiously pending replication in larger samples. Future researches direction should note to include a control comparison.
In summary, the present findings confirmed the existence of an association between obesity and anxiety/ depression in postmenopausal women. Moreover, we demonstrated for the first time that acylated ghrelin levels associate positively with depression in this population. Further investigations are warranted to assess whether the elevated acylated ghrelin levels reflected a risk factor or a protection mechanism.

Methods
This cross-sectional study was designed according to STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines and was approved by the Ethics Committee of the Universidade Federal de São Paulo (CEP #921.394/2014). The investigation was conducted in accordance with the Declaration of Helsinki. All participants have signed the informed consent.
Study design and study population. Two hundred and thirty-three volunteers that were recruited by the Universidade Federal de São Paulo media got in contact with our group. All volunteers were patients of the Gynecology Division of the Hospital São Paulo, the main affiliated hospital of the Universidade Federal de São Paulo. Of these, 24 refused to participate, 48 were classified with no depression symptoms, and 106 did not meet www.nature.com/scientificreports/ the eligibility criteria due to: premenopausal stage, age, absence of depressive symptoms, surgical menopause, hormone therapy, use of antidepressants, menopausal transition, and use of insulin. The inclusion criteria were: women aged 50-65 years, amenorrhea for at least 12 months, follicle-stimulating hormone (FSH) levels greater than 30 mUI/mL, depression symptoms, no use of hormonal replacement therapy, antidepressants or insulin. The volunteers were medically evaluated for the confirmation of the menopause status.
At the day of the study, the participants arrived between 7 and 8 a.m., after 12 h of fasting. Saliva and blood samples were collected and, then a trained dietitian performed the anthropometric assessments. The participants were then instructed about the depression and anxiety inventories and answered the questionnaires.
Statistical analysis. The SPSS software version 18.0 (IBM, Armonk, NY, USA) was used to perform all statistical analysis. The outliers were excluded according to the Chauvenet's criterion. Data distribution was analyzed by the Shapiro-Wilk normality test, while the homogeneity of variance was assessed by Levene test. Parametric variables were analyzed by ANOVA and Tukey post hoc test. Nonparametric variables were analyzed by Kruskall-Wallis and multiple comparisons test. Relationships between hormonal/anthropometric parameters and depression/anxiety scores were determined by the Pearson's correlation coefficient. All the variables showing significant correlations with the psychological scores were tested in multivariate linear regression analyses to identify predictors for depression and anxiety symptoms. However, some of our data were missing at random (MAR) due to the fact that some samples were out of range for the assays used. Thus, the model for BDI included 49 women, the model for BAI included 53 women, and the model for PHQ-9 included 44 women.
The sample size of a minimum of 44 participants was calculated considering a statistical power of 0.85, an effect size of 0.31, alpha level of 0.05, and 3 predictors. www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.