Ginkgo biloba extract (GbE) attenuates obesity and anxious/depressive-like behaviours induced by ovariectomy

While several pieces of evidence link obesity and mood disorders in menopause, the mechanisms involved are not yet fully understood. We have previously demonstrated that Ginkgo biloba extract (GbE) both attenuated diet-induced obesity of male rats and restored serotonin-induced hypophagia in ovariectomized female rats. The present study aimed at exploring whether GbE treatment ameliorates ovariectomy-related obesity and anxious/depressive-like behaviours. Wistar female rats were either ovariectomized (OVX) or sham-operated (Sham). After 2 months, either 500 mg/kg of GbE or vehicle were administered daily by gavage for 14 days. Anxious/depressive-like behaviours were assessed by the Elevated Plus Maze and the Forced Swim Tests, respectively. Ovariectomy caused high visceral adiposity, hyperleptinemia, and hypercholesterolemia, and increased the anxiety index (p = 0.048 vs. Sham + GbE) while it decreased the latency to immobility (p = 0.004 vs. Sham). GbE treatment in OVX rats improved body composition, adiponectin levels and blood lipid profile. It also reduced the anxiety index (p = 0.004) and increased the latency to immobility (p = 0.003) of OVX rats. Linear regression analysis demonstrated that leptin (p = 0.047) and total cholesterol levels (p = 0.022) were associated with anxious-like behaviours while body adiposity (p = 0.00005) was strongly associated with depressive-like behaviours. The results showed that GbE therapy was effective in attenuating the deleterious effects of ovariectomy on body composition, lipid profile, and anxious/depressive-like behaviours. Further studies are warranted to better understand the therapeutic potential of GbE in menopause.

In relation to anxiety related-behaviours evaluated in elevated plus maze, the anxiety index was negatively correlated with HDL-cholesterol levels (p = 0.025). The number of entries (p = 0.046), the distance travelled   Since the univariate analysis detected the existence of significant associations of body composition/metabolic/ hormonal parameters and behavioural assessments denotative of anxious-like and depressive-like behaviours, multivariate linear regression models were constructed to identify the significant predictors for the depressive/ anxious-like behaviours (Table 3).
For the climbing behaviour, the initial model tested body weight and sum of adipose tissues. The final model (F (2,36) = 3.4284, p < 0.0433) indicated that climbing was positively associated with body weight while the sum of the fat depots showed no significant effect (p = 0.06520). Considering the immobility frequency, the model included only the retroperitoneal fat depot and it showed a significant positive association (F (1,36) = 38.548, p < 0.00000). The latency to immobility was negatively associated with the sum of fat depots (F (2,36) = 16.777, p < 0.00001). Body weight was also tested but it showed no significant association (p = 0.0860). In

Discussion
The present findings indicated that ovariectomy caused anxious like-behaviours, as indicated by increased permanence in the closed arms and short permanence in the open arms, during the EPM test. The preference for the closed over the open arms has been attributed to spontaneous fear, which competes with the exploratory nature of rodents 27 .
The grooming behaviour, commonly observed when the animal is placed in an unfamiliar stressful situation, has been associated to obsessive-compulsive symptoms and considered as a measure of anxiety 28 . However, we failed to observe significant differences in this parameter among the groups.
In relation to depressive-like behaviours, the ovariectomized rats showed the highest frequency of immobility and the shortest latency to immobility, features highly indicative of a depressive-like state 29 . The present results agree with previous demonstrations of induction of anxious-and depressive-related behaviours by ovariectomy in rodents 4,30 as well as of increased depression and anxiety rates in postmenopausal women 2,12 . Similarly, it has been reported, in both rats and mice, that estrogen replacement attenuated the increased immobility and decrease of active behaviours induced by ovariectomy 31 .
Anxiety and depression disorders are commonly treated with antidepressants that also have anxiolytic properties, such as sertraline, as well as with benzodiazepines. However, the long-term therapy with these drugs induce undesirable side effects, including sleep disturbances, body weight gain, sexual dysfunctions and dependence. Thus, the finding of more effective, safer and shorter-lasting therapies for the adequate control of these diseases is of great relevance 26 .
In the present study, GbE treatment was able to reduce the anxious-like behaviours of OVX rats, as indicated by a lower anxiety index and a higher number of non-protected head-dipping events. An anxiolytic role of GbE has been described in both animals and humans. In elderly patients diagnosed with mild cognitive impairment, daily treatment with 240 mg of EGb761 for 24 weeks reduced anxious symptoms in comparison to placebo 32 . In aged female rats, GbE (EGb761) oral supplementation (100 mg/kg, for 30 days) attenuated anxious-like behaviours while in young and middle-aged male mice, the same treatment caused a significant improvement of anxiety symptoms related to acute stress 33,34 .
The present results also indicated that the oral administration 500 mg/kg of GbE for 14 days promoted an antidepressant effect in ovariectomized rats. A similar effect has been observed in mice receiving an intraperitoneal dose of 10 mg/kg of EGb761 for 17 days. The authors associated the antidepressant action of EGb761 with an antioxidant effect in midbrain, prefrontal cortex and hippocampus 35 . The central mechanisms through which GbE affects mood and cognition have not been fully elucidated. Bilobalides present in GbE extracts have been shown to block long-term depression in rats, through increased synaptic plasticity in the medial perforant path-dentate gyrus (MPP-DG) 36 .
The present results data showed that ovariectomy induced increments of the retroperitoneal fat pad and carcass fat contents, in comparison to both the Sham and the Sham + GbE groups. In agreement, it has been reported that both OVX mice and rats developed overweight, with excess visceral and subcutaneous adipose tissues, which were related to increased adipocyte size and pre-adipocyte differentiation, as result of hypoestrogenism 37 . Similarly, postmenopausal women have been shown to present increased total and abdominal fat masses and reduction of lean body mass, resulting from the estrogen fall allied to a raise of testosterone levels 9 .
In the present study, ovariectomy induced hyperleptinemia, probably as a consequence of increased adiposity. It is well established that obesity-related hyperleptinemia is associated to leptin resistance, which leads to hyperphagia and decreased energy expenditure 38 . In premenopausal women, estrogens levels have been closely related to leptin levels and regulation of leptin receptors expression and sensitivity. Additionally, both post-menopausal women and ovariectomized animals presented an elevation of serum leptin and a reduction of adiponectin levels, what may promote insulin resistance 38,39 .
Although the ovariectomized rats of the present study exhibited higher blood levels of leptin, no signs of insulin resistance were observed. In ovariectomized mice, glucose intolerance developed only when the animals www.nature.com/scientificreports/ were exposed to a high-fat diet 40 . In another study, the authors noted that both diet-induced obesity and ovariectomy were able to lead to a state of insulin resistance in rats. However, those rats were submitted to ovariectomy at the age of 13 weeks, i.e., 5 weeks later than in our study 41 . The present data also showed that GbE improved body composition, as it decreased retroperitoneal fat pad mass and protected against adipose tissue accumulation in ovariectomized rats. In addition, GbE raised adiponectin levels of OVX animals, what may have been a consequence of the lower adiposity, since this adipokine secretion is inversely proportional to body fat 11 . Corroborating the present findings, it has been reported that supplementation with Morus alba extract, an anti-inflammatory and anti-oxidant herbal medicine, to female rats fed a high-cholesterol diet, reduced body adiposity, up-regulated adiponectin gene expression, down-regulated leptin and resistin gene expressions, and improved insulin sensitivity in the visceral adipose tissue 42 .
In agreement with the present data, we have previously demonstrated, in obese rats, that GbE reduced both body weight and adiposity, epididymal adipocytes volume and incorporation of fatty acids into triglycerides, what allowed us to suggest a potential use of GbE in the control of obesity [22][23][24] . Ginkgo biloba biflavones have been found to exert a lipolytic action in the epididymal adipose tissue of rats, through inhibition of cAMP-phosphodiesterase 43 .
Decline in estrogens production has been associated with changes in lipid profile. The dyslipidaemia after menopause is characterized by high levels of TC and LDL-cholesterol, followed by decreased HDL-cholesterol levels, what may accelerate atherosclerotic processes 39 . In the present study, high levels of TC and LDL-cholesterol were observed in the ovariectomized rats, in agreement with a previous report 44 . However, TAG levels were decreased in both OVX and OVX + GbE rats, probably due to estrogen deficiency. In fact, ovariectomy has been demonstrated to impair the lipolytic response to norepinephrine in fat cells, as well as to increase the activity of adipose tissue lipoprotein lipase (LPL), mechanisms potentially linked to the increased fat deposition associated with ovariectomy 45 .
Importantly, GbE promoted a healthier lipid profile in ovariectomized rats, increasing HDL-cholesterol levels while protecting against the elevation of TC and LDL-cholesterol levels. These findings agree with our previous study showing GbE to decrease serum TAG and tending to decrease TC of obese male rats 22 . In addition, we observed a decrement of NEFA serum levels in OVX + GbE rats. It is well established that the elevation of blood NEFA may stimulate inflammatory pathways by activation of toll-like receptor 4, leading to an impairment of insulin sensitivity 46 . A protective role of GbE has also been demonstrated on the lipid profile alterations induced by ethanol intake and diabetes in rats 47,48 .
Obesity, increased visceral fat, insulin resistance and dyslipidaemias are components of the metabolic syndrome, whose prevalence becomes potentially high in post-menopausal women 9 . In addition, due to the gonadal hormones fluctuations from menarche until menopause establishment, women exhibit a higher vulnerability to develop anxiety and depression than men 1,2 . Although hormone replacement therapy (HRT) shows good results for the treatment of these menopause-associated problems, HRT is known to increase the risk of developing cardiovascular diseases, breast cancer and other tumors 9 .
In human breast cells in culture, GbE blocked the cell proliferation induced by high estrogen levels, presenting a chemoprotective role 49 . Considering the beneficial actions of GbE, it might be suitable to prevent menopauserelated disturbances, without inducing breast cancer, the main unwanted effect of HRT.
A link between obesity and psychological diseases has been described. Excess visceral adiposity and unbalanced feeding habits have been implicated in the development of depressed mood during obesity establishment. In addition, depressive individuals are more prone to weight gain due to poor food choices and physical inactivity 50 . In American men and women, obesity has been highly associated with depression and anxiety 51 .
Here, we performed linear regression analyses to determine whether body composition and serum parameters might be predictors for the anxious/depressive-like behaviours associated with menopause. The data showed that leptin and total cholesterol were positive predictors for anxious-like behaviours. In agreement with the present findings, the hypercholesterolemia and hyperleptinemia of obese female rats have been shown to be positively associated with mood and anxiety symptoms. Interestingly, this study also reported a reduction of depressiveand anxious-like behaviours by the oral supplementation with Morus alba extract 42 .
The sum of fat depots was strongly associated with depressive-like behaviours, in agreement with the proposition of an influence of visceral adiposity on the development of depression 52 . In fact, even in healthy individuals, inadequate eating patterns have been linked to psychological disturbances 53 . Considering these observations, it seems reasonable that a relation of body adiposity gain and depression exists in both eutrophic and obese animals and humans.
It has been reported that obese individuals had a 55% higher risk of developing depression while depressed individuals presented a 58% higher chance of becoming obese. Excessive release of pro-inflammatory cytokines from increased visceral fat may induce depressive symptoms through the activation of hypothalamus-pituitary-adrenal axis, generating hypercortisolaemia and more fat deposition 54 . Maintenance of healthy body weight reduced depressive disorders in post-menopausal women 55 .
Finally, it has been reported that central inflammation may trigger behavioural disturbances, such as anxiety and depression, in menopausal women [56][57][58][59] . Despite the fact that TNF-alpha was similar among the groups, we cannot rule out the possibility of ovariectomy-induced inflammation, as observed by other authors 5 . The present lack of a statistically significant difference in TNF-alpha levels may have been a consequence of the short period after ovariectomy or of the age of the rats at the moment of euthanasia. In addition, other pro-inflammatory cytokines may be altered, as increased levels of IL-6 and of its soluble receptor have been reported in menopausal women with depressive mood 56 .
The present data showed that GbE reduced body adiposity and increased carcass protein and serum adiponectin levels in the ovariectomized rats, while it ameliorated anxious-and depressive-like behaviours. Thus, the present data suggest a potential for GbE to ameliorate menopause-related obesity and mood disorders.

Methods
Animals. All the procedures were approved by the Ethics Committee on Animal Research of the Universidade Federal de São Paulo and were performed in accordance with the principles of the Brazilian guideline for the use of animal models in research (Arouca Law-number 11794/08). Fifty nine 8-week-old female Wistar rats undergone bilateral ovariectomy (OVX) or false-ovariectomy (Sham) as previously described 20 . The survivor rate after surgical procedure was 100%. Eight weeks after surgery, the Sham and OVX rats were sub-divided according to the phytotherapy treatment described below. Figure 6 exhibits the experimental timeline.
Behavioural evaluation. During the phytotherapy treatment, the animals were submitted to the Elevated Plus Maze test (EPM) to evaluate anxious-like behaviours and to the Modified Forced Swim test (FST) in order to identify depressive-like behaviours. A two-day interval was allowed between the tests.
Analysis of anxious-like behaviour: EPM test. The plus-maze apparatus was made of wood and positioned 50 cm above the floor. It contained two closed arms (CA, 10-cm wide and 40 cm high walls) and two open arms (OA, 10-cm wide, without walls) connected by a central platform, as previously described 27,60 . All tests were conducted during the light phase of the light/dark cycle, between 9 am and 2 pm. All sessions were recorded for subsequent analysis.
At the 10th day of the gavage treatment, the animals were individually placed in the central platform, facing an open arm, and allowed 5 min to explore the apparatus.
The percentage of time spent in the OA [% time in OA = (time spent in the OA /time spent in all arms) × 100)] and the percentage of entries into the OA [% entries in OA = (OA entries/total entries in all arms) × 100] were measured to evaluate anxious-like behaviour. In addition, the total number of entries (into both enclosed and open arms) was used to evaluate spontaneous locomotor activity 61 .
The number of risk assessments, which consists of stretch-attend postures, head-dipping, grooming and rearing were used as complementary measures, since they have been used to determine when rodents evaluate and/or avoid the OA 60,62 . The anxiety index was calculated as: 1 − [(OA time (min)/5 min) + (number of OA entries/total entries)]/2. The results range from 0 (less anxious) to 1 (more anxious) 63 . The maze was thoroughly cleaned between subjects.
Analysis of depressive-like behaviour: FST. At the 12th and 13th days, each animal was placed in a transparent cylinder, measuring 30 cm in diameter and 50 cm in height, filed with water (24 ± 1 °C) up to 30 cm, to avoid both the touching of the bottom and the escaping from the tank 31 . All animals were subjected to two sessions. At the 12th day, a 15-min. training session was carried out, for acclimation to the testing environment. After 24 h, the 5-min test was performed. The latency to immobility was used as an indication of behavioural despair 29 . The www.nature.com/scientificreports/ animal was considered immobile when making only necessary movements to keep its head above the water or floating 3,29 . The frequency of immobility, and the number of swimming, climbing, and dip events were counted every five seconds 31,63 .

Carcass fat and protein contents.
After the removal of the tissues (gastrointestinal tract, liver, uterus, and retroperitoneal, gonadal and mesenteric fat pads), the carcasses were kept in freezer -20 °C until analyses. For the determination of carcass lipid and protein contents, carcasses were shaved, softened, homogenized and digested in 30% potassium hydroxide and 6 N sulfuric acid. Lipid was extracted from 5 g aliquots with petroleum ether and quantitated gravimetrically 64,65 . Protein from 2 g aliquots was dissolved in KOH 0.6 N and quantified according to Bradford method (Bio-Rad, Hercules, CA).

Statistical analyses.
Statistical analyses of behavioural evaluations, serum parameters, and body composition parameters were performed using PASW Statistics version 21 (SPSS Inc., USA), for p ≤ 0.05. Comparisons among the four groups were performed by two-way ANOVA followed by Tukey HSD test, adopting ovariectomy and GbE treatment as the fixed factors.
Linear regression analysis was used to identify body composition and serum parameters significantly influencing behavioural variables. The regression models were constructed based on the statistically significant correlations shown by the univariate analysis. Both Pearson's correlation coefficient and linear regression analysis were performed using STATISTICA 12.0 (StatSoft, Tulsa, OK, USA).