Hormonal and biochemical changes in female Proechimys guyannensis, an animal model of resistance to pilocarpine-induced status epilepticus

The Amazon rodent Proechimys guyannensis is widely studied for hosting various pathogens, though rarely getting sick. Previous studies on male Proechimys have revealed an endogenous resistance to epilepsy. Here, we assess in female Proechimys, whether sex hormones and biochemical aspects can interfere with the induction of status epilepticus (SE). The lithium-pilocarpine ramp-up protocol was used to induce SE, and blood sera were collected at 30 and 90 min after SE, alongside brains, for biochemical, western blot and immunohistochemical analyses. Results from non-ovariectomised (NOVX) Proechimys were compared to ovariectomised (OVX) animals. Data from female Wistars were used as a positive control of SE inductions. SE latency was similar in NOVX, OVX, and female Wistars groups. However, the pilocarpine dose required to induce SE in Proechimys was higher (25- to 50-folds more). Despite a higher dose, Proechimys did not show strong SE like Wistars; they only reached stage 2 of the Racine scale. These data suggest that female Proechimys are resistant to SE induction. Glucose and progesterone levels increased at 30 min and returned to normal at 90 min after SE. A relevant fact because in humans and rodents, SE leads to hypoglycaemia after 30 min of SE and does not return to normal levels in a short time, a typical adverse effect of SE. In OVX animals, a decrease in GABAergic receptors within 90 min of SE may suggest that ovariectomy produces changes in the hippocampus, including a certain vulnerability to seizures. We speculate that progesterone and glucose increases form part of the compensatory mechanisms that provide resistance in Proechimys against SE induction.

Status epilepticus (SE) a condition characterised by is a prolonged seizure activity resulting from the failure of mechanisms responsible for ending seizures or from mechanisms that lead to seizures with a prolonged duration 1 . In humans and experimental animal models, sustained SE lead to neuronal loss in selective regions of the brain that can cause epilepsy or death.
The transition to SE can be preceded by biochemical changes and the gradual development of pathological brain electrical activity characterised by the involvement of an epileptogenic focus usually located in the hippocampus. Studies have shown that susceptibility to SE can be analysed by measuring biochemical and brain dynamics (i.e., neurochemical and electrical activities) 2,3 .
Susceptibility to SE in women is related to hormonal fluctuations during the menstrual cycle phases 4 . Traditionally, in the follicular phase, oestrogen is produced, a hormone classified as a proconvulsant as it reduces the convulsive threshold. In contrast, in the luteal phase, progesterone, a hormone considered an anticonvulsant agent due to its neuroprotective role, reaches high levels 4,5 . In rodent brain, oestradiol decreases gamma-aminobutyric acid (GABA)ergic inhibition 6 , whereas progesterone, via allopregnanolone, binds to GABA receptors

Results
High doses of pilocarpine were required for female Proechimys to reach SE. In female Proechimys, an initial dose of 10 mg/kg of pilocarpine was administered, followed by 10 mg/kg every 30 min for up to five doses of pilocarpine, as described in the lithium-pilocarpine ramp-up protocol 17 . As the initial dose was not effective, we followed certain steps until reaching the ideal dose of pilocarpine as described in Supplementary Information A and schematic Fig. 1.
The amount of pilocarpine, latency to reach SE and the intensity of SE in female Proechimys were compared with female Wistar rats used as positive control. The mean SE latency in female Proechimys (29.83 ± 15.05 min) was no longer than that in female Wistar rats (28.30 ± 11.84 min). However, while Wistar rats easily reached SE with a 10 mg dose of pilocarpine, Proechimys only reached SE after high doses of pilocarpine (250-500 mg) (Fig. 1).
With regards to the SE development according to Racine's scale 18 , female Proechimys showed similar patterns as those described for male Proechimys with abnormal behaviours typical of Racine stage 2, i.e., all four limbs on the floor, extended digits, mastication and head nodding, but no evolution to further stages 12,13 . The mean latency to reach SE was similar for NOVX and OVX groups. Of the 12 non-ovariectomised (NOVX) rodents, 4 (33%) developed SE after the first pilocarpine dose (250 mg/kg, ip), whereas the remaining 8 (67%) rodents converted to SE after the second dose. The mean latency of NOVX was 29.83 ± 15.05 min (mean ± SD). Of the 12 ovariectomised (OVX) rodents subjected to pilocarpine, 2 (17%) and 10 (83%) animals developed SE after the first and second dose, respectively, with a mean SE latency of 34.00 ± 9.83 min. The seizure latencies between OVX and NOVX animals were not significantly different. Based on the Racine scale, Proechimys OVX and NOVX reached a score of 2.
Proechimys submitted to SE showed high blood levels of glucose and progesterone. Glucose levels in NOVX and OVX animals were significantly increased (p < 0.001) at 30 min after SE (311.00 ± 72.85 and 433.00 ± 87.00 mg/dL, respectively) and returned to control levels within 90 min after SE (117.00 ± 62.93 and 233.00 ± 136.32 mg/dL, respectively). The ovariectomy did not interfere with blood glucose levels when compared to controls groups (OVX and NOVX) (Fig. 2a, Table 1).
Progesterone levels in NOVX animals significantly increased (p < 0.001) at 30 min after SE (22.71 ± 3.12 ng/ mL) and returned to control levels at 90 min after SE (10.48 ± 1.53 ng/mL). In OVX animals, 30 min after SE onset, the progesterone level significantly increased (11.91 ± 2.71, p < 0.05) compared to CTR-OVX (5.84 ± 0.45 ng/mL), and remained increased in 90 min of SE. There was no considerable difference between controls (OVX and NOVX) animals, i.e. the ovariectomy did not lead to a significant decrease in the level of progesterone. When comparing the NOVX with OVX groups, 30 min after SE onset, ovariectomised animals (PILO 30-OVX) showed a significant decrease in progesterone levels (Fig. 2b, Table 1).
In terms of triglyceride levels, there was a significant increase (p < 0.001) in the CTR-OVX group (214.50 ± 43.00 mg/dL) compared to the CTR-NOVX group (48.50 ± 10.31 mg/dL). The ovariectomy itself resulted in an increase in triglyceride levels. The higher triglyceride levels could be associated with the lack of sex hormones such as in menopause 19 . After treating OVX animals with pilocarpine, there was a significant decrease in triglycerides in the 30 min (90. 40  Ovariectomy produced changes in the hippocampus of Proechimys with characteristics of vulnerability to seizures. The expression of the mGluR2/3 excitatory receptor in the hippocampus showed no change in NOVX animal after pilocarpine treatment. In the OVX animals, the receptor significant decrease (p < 0.001) in the PILO 90-OVX group as compared to the CTR-OVX group. When comparing CTR-NOVX and CTR-OVX groups, although there was no difference between groups, there was a tendency for glutamate receptors to be increased in OVX animals (Fig. 3a). The analysis of GABA B inhibitory receptor expression in the hippocampus showed no change in NOVX groups (CTR, PILO 30, and PILO 90) after pilocarpine treatment. However, the CTR-OVX group (0.77 ± 0.02) showed a significant decrease in GABA B levels in the PILO 90-OVX group (the expression of GABA B receptors was reduced below the detection level) (Fig. 3b). When comparing CTR-NOVX and CTR-OVX groups, the CTR-OVX group showed a significant decrease in GABA B levels (p < 0.05).
The expression of parvalbumin receptors was also analysed in the hippocampus due to its critical role in generating network oscillations and its involvement in epileptic seizures. There was no difference among the groups, except for the PILO 90-OVX group, compared to PILO 90-NOVX, which showed a decreased (p < 0.01) expression of these receptors. This may demonstrate a weakness in this type of GABAergic neuron at 90 min after SE onset, in ovariectomised Proechimys. This emphasises that no differences were found between CTR-NOVX and CTR-OVX groups (Fig. 3c).
Changes in progesterone and oestrogen receptor patterns in the hippocampus. The analysis of progesterone receptor patterns revealed differences between the PILO 30 and CTR among NOVX groups, with an increased at 30 min (p < 0.05). In contrast, there was a reduction in the progesterone level expression in the PILO 30-OVX group (p < 0.01) and an increase expression in the PILO 90-OVX group (p < 0.01), as compared to the CTR-OVX group (Fig. 4a). www.nature.com/scientificreports/ Regarding oestrogen receptor patterns during SE, in the NOVX groups, the ER-alpha expression was significantly decreased (p < 0.001) at 90 min after SE onset, compared to CTR-NOVX. In OVX animals, no statistical differences were found among CTR-OVX, PILO 30-OVX, and PILO 90-OVX groups (Fig. 4b). Ovariectomy by itself decreases ER-alpha expression in the hippocampus, according to comparative data between CTR-NOVX and CTR-OVX groups. An ovariectomy by itself did not interfere with the expression of ER-alpha in the hippocampus.
With regards to the ER-beta, the expression in the NOVX group significantly increased (p < 0.01) at 90 min after SE onset, compared to the CTR-NOVX group. In OVX animals, no statistical differences were found among CTR-OVX, PILO 30-OVX, and PILO 90-OVX groups (Fig. 4c). When analysing the proportion of ER-alpha receptors in relation to ER-beta expression in the hippocampus, the expression of ER-alpha was greater than that of ER-beta, except for PILO 90-NOVX group, which had a higher expression of ER beta than ER-alpha (Fig. 4d).

Rearrangement of receptors in the Proechimys hippocampus after pilocarpine.
We also costained Proechimys hippocampi for the hormonal ER-alpha, ER-beta, and progesterone receptors (stained in red) with GABA B and mGluR2/3 receptors (stained in green; Fig. 5a-i). The co-expression of ER-alpha, ER-beta, and progesterone receptors with GABA B was detected in all analysed hippocampal regions. Likewise, the co-expression of ER-alpha, ER-beta, and progesterone with mGluR2/3 was analysed; progesterone receptor co-expression with mGluR2/3 was only observed in the cell membrane, whereas ER-alpha and ER-beta co-expression was detected in the nucleus and the cell membrane. The co-localisations patterns were quantified in the hippocampal areas: CA1, CA3, DG, and hilus ( Table 2). Figure 5 outlines the statistical differences found at 30 min and 90 min after SE onset. www.nature.com/scientificreports/ When comparing the co-localisations patterns of receptors in PILO 30 (NOVX or OVX) and PILO 90 (NOVX or OVX) with the respective control groups NOVX or OVX, we observed several receptor reorganization trends.
In OVX rodents, all changes can be seen in Table 2 and Fig. 6.

Discussion
To the best of our knowledge, this study is a pioneer in using females P. guyannensis for SE induction analyses. Comparative analyses were performed with NOVX and OVX Proechimys undergoing on SE, and respective controls (CTR-NOVX, CTR-OVX) were analysed. Female Wistars were used as a positive control. There is a strong incentive from the National Institutes of Health to include females because results will be more consistent with the general population 15 . We conducted studies in female Proechimys (NOVX group) selecting the oestrus stage to minimise the bias related to hormonal changes. In our previous work analysing the oestrous cycle of Proechimys, low levels of progesterone were observed in the oestrus stage 14 , which can correspond to the progesterone withdrawal phenomenon leading to the increased excitability and susceptibility to seizures, as proposed to explain the increase in seizures that usually occur during the perimenstrual period in women with epilepsy 22,23 . In this context, the oestrus stage chosen to carry out this work would be more closely associated with a decrease than an increase in the threshold for SE induction.
As result of pilocarpine treatment, the mean SE latency in NOVX was no longer than that in female Wistar rats, but the dose of pilocarpine required to reach SE was higher for Proechimys (25-to 50-folds more) (Fig. 1). During SE, Proechimys evolved only up to stage 2 on the Racine scale, unlike Wistar rats which usually reach stage 5 of this scale-the lower the scale, the less strong SE. The pilocarpine dose to induce SE, latency to SE and the Racine scale were similar to NOVX and OVX Proechimys. Ovariectomised animals remain resistant to induced seizures, i.e. ovarian hormones do not appear to have interfered with seizure threshold and intensity in Proechimys. This result is in contrast to our previous study in Wistar rats, in which ovariectomy decreases the latency to SE, indicating that these rodents without ovaries became more sensitised to pilocarpine 9 . The systemic injection of pilocarpine in rodents is one of the most frequently used models for evoking experimental SE 24,25 . Pilocarpine-induced SE results from acute peripheral proinflammatory actions of pilocarpine leading to blood-brain barrier leakage, as well as a direct cholinergic system activation by pilocarpine in the brain 26 . The dose of pilocarpine influences substantially the induced effects. The pilocarpine dose schedule is critical for obtaining animals experiencing SE with minimal mortality. We perform the lithium-pilocarpine ramp-up protocol due to its high proportion of rodents developing epilepsy and its low mortality (Supplementary Information A). In female Wistar, pretreatment with lithium results in a conspicuous reduction in the pilocarpine dose required to induce SE: 320 mg/kg without lithium 9 and 10 mg/kg with lithium (present study).
In female Proechimys, pretreatment with lithium results in a higher dose required to induce SE (250 to 500 mg/ kg) compared to those used in female Wistar rats. Surprisingly, lithium may not potentiate the convulsive effects of pilocarpine in female Proechimys, similar to observations obtained in mice 27,28 . Although an explanation for the lack of lithium potentiating the susceptibility of mice to pilocarpine has not been provided, it may be due to species differences in the modulation of inflammatory activity, glutamate receptors 29 , or the lack of inositol depletion by lithium 30,31 .
Previous studies in male Proechimys, epileptogenic treatment with pilocarpine without lithium, hippocampal inflammatory cytokines (IL-1β, IL-6, IL-10, and TNF-α) and vascular endothelial growth factor remain unchanged after SE 32 , suggesting that Proechimys exhibit attenuated inflammatory responses to injury. Furthermore, although male Proechimys develop SE, they do not become epileptic in contrast to Wistar rats. In these animals, the SE rarely exceeds 2 h, in contrast to the 8-12 h in Wistar rats 13 .
The SE undergoes temporal stages based on pathophysiological changes and the relationship between SE duration and epilepsy development. Hippocampal damage with subsequent complex synaptic reorganization  Figures S1-S3). Data are presented as mean ± SD. Levels marked with asterisk (*) indicate significant differences: *p < 0.05, **p < 0.01, ***p < 0.001.

Scientific Reports
| (2020) 10:20982 | https://doi.org/10.1038/s41598-020-77879-1 www.nature.com/scientificreports/ is only produced by more than 30 min of SE 33 . Temporal changes occur in the brain during the progression of tonic-clonic SE; around 30 min after SE onset, physiological changes are compensatory. With prolonged SE, these compensatory mechanisms break down. As the SE continues for more than 60 min, the risk of brain damage progressively increases along with the depletion of glucose and glycogen concentrations in the brain 34 .
Here, we assess serum blood and hippocampus at 30 and 90 min after SE onset, to identify changes that may contribute to SE resistance. Anti-epileptogenesis in Proechimys. Proechimys (NOVX and OVX) presented high glucose levels at 30 min of seizures, and at 90 min of seizures glucose levels remained equal to controls (CTR-NOVX and CTR-OVX, respectively) (Figs. 2a, 6a). This differs from other rodents and humans presenting a decrease in blood glucose after 30 min of SE 35,36 .
It has already been suggested that glucose regulation may be a critical factor linked to the susceptibility to seizure-induced cell death 36 . Glucose metabolism is impaired in humans under SE 37 , and some seizures usually improve with glycaemic control 38,39 . Cholinergic agents, as pilocarpine, are known to activate insulin production in rodents 40,41 which leads to hypoglycaemia. In Proechimys, treatment with pilocarpine generated hyperglycaemia rather than hypoglycaemia. In hormonal analyses, the equipment was unable to measure the insulin concentration in the blood serum of Proechimys. The reason underlying this phenomenon remains unknown, and may be linked to the lack of immunospecificity of the antibody to detect Proechimys insulin.
The ability of Proechimys to maintain glucose homeostasis along with the apparent recovery from seizures after SE induction is peculiar (pilot study findings). A different scenario involves the homeostasis of glucose in Proechimys. Atypically, the glucose regulation mechanism in these animals is not based on the action of insulin, as it normally occurs in mammals. Proechimys have an insulin molecule with amino acid substitutions that confers a lower potency (< 10%) compared to bovine insulin [42][43][44] . Consequently, this leads to an impaired binding  Figures S4-S6). Data are presented as mean ± SD. Levels marked with asterisk (*) indicate significant differences: *p < 0.05, **p < 0.01, ***p < 0.001.

Scientific Reports
| (2020) 10:20982 | https://doi.org/10.1038/s41598-020-77879-1 www.nature.com/scientificreports/ affinity for insulin receptors in the hepatic plasma membrane and an inability to stimulate glucose uptake and, thus, its function in lipid conversion 43 . However, Proechimys regulate glucose levels like other mammals 45 , suggesting compensatory traits that may permit maintenance of standard values of plasma glucose 45 . May Proechimys resistance to epilepsy be caused by an unknown mechanism that regulates their glucose? Glucose homeostasis is performed by a highly sophisticated network of hormones and neuropeptides released mainly from the pancreas, brain, liver and intestine 46,47 . Molecules, enzymes and interactions may triggers signalling pathways underlying insulin secretion 47 .
Analysing the biochemical and hormonal blood concentrations, NOVX and OVX Proechimys behaved similarly. The only difference found in the blood serum between CTR-OVX and CTR-NOVX was the increase in the level of triglycerides in the OVX animals. Interestingly, even without the ovaries, ovarian hormones are present in OVX animals. Moreover, OVX increases progesterone levels during SE, as NOVX. Progesterone is an anticonvulsant hormone 48 that can be synthesised by the adrenal glands and the nervous system, besides being synthesised in the ovary 46 . Usually, in Proechimys, the highest level of progesterone is found in the proestrus stage, lasting 2 to 3 days 14 . The increase in progesterone in the blood serum over a short time period, over the course of 30 min of SE, was an interesting finding, seen mainly in the OVX animals (Fig. 2b).
Regarding the hippocampal analysis, an ovariectomy by itself leads to a decrease in GABA B receptors. In the hippocampus, neuronal network is composed of excitatory (glutamatergic) and inhibitory (GABAergic) neurons, the imbalance in this network leads to seizures. In Proechimys NOXV treated with pilocarpine, expressions of glutamatergic (mGluR2/3) or GABAergic (GABA B ) receptor did not change. When ovariectomized animals received pilocarpine, a decrease in mGluR2/3 and GABA B occurred at 90 min of SE. So far, it is possible that the ovariectomy in these animals leads to the vulnerability of GABAergic and glutamatergic neurons if the SE lasts 90 min (Fig. 3a,b).  50 . We analysed parvalbumin receptor expression in the OVX Proechimys hippocampus to assess whether any remodelling in the interneuronal network could be occurring during SE. In this study, SE did not involve the remodelling of parvalbumin interneurons.
When analysing hormonal receptors in the hippocampus, changes in the expression patterns of progesterone and oestrogen receptors were found in NOVX and OVX Proechimys, according to the treatment. After SE induction, expression of progesterone receptor increased in 30 min for NOVX and 90 min for OVX (Fig. 4a). This increase in 90 min in OVX may be related to the local synthesis of progesterone by glial and neuronal cells 51,52 . Which may indicate optimised use of progesterone by the hippocampus. Progesterone is an anticonvulsant agent with neuroprotective role, Proechimys may also use this mechanism in an attempt to protect themselves from seizures.
We also analysed the expression patterns of oestrogen receptor in the hippocampus, although there was no increase in the serological level during SE, the increase in expression of oestrogen receptor can make the hormone more bioavailable. A higher ER-alpha than ER-beta expression was found (CTR-NOVX; Fig. 4d), as in mice 53 . On the other hand, ER-beta is more prevalent in humans and rats hippocampi [54][55][56] . Through these receptors, oestrogen activates rapid signalling in neurons 57 and both oestrogen receptors (ER-alpha and -beta) seem to be involved in neuroprotection [58][59][60] . ER-alpha is known to plays a critical role in regulating reproductive neuroendocrine function 61 . ER-beta agonist significantly reduce damage by 55% in the CA1 region of hippocampus after injury 60 , Table 2. Colocalization of oestrogen and progesterone receptors with GABA B and mGluR2/3 receptors in the hippocampi of females Proechimys guyannensis. ER-alpha: Oestrogen receptor alpha; ER-beta: Oestrogen receptor beta; GABA B : Gamma-aminobutyric acid; mGluR2/3: Metabotropic glutamate receptor; CTR-NOVX: Non-ovariectomised control group, animals no submitted to SE; CTR-OVX: Ovariectomised control group, animals no submitted to SE; PILO 30-NOVX, Non-ovariectomised animals analysed 30 min after SE onset; PILO 30-OVX: Ovariectomised animals analysed 30 min after SE onset; PILO 90-NOVX, Non-ovariectomised animals analysed 90 min after SE onset; PILO 90-OVX: Ovariectomised animals analysed 90 min after SE onset. Data are presented as mean ± SD. Kruskal-Wallis test followed by Dunn's post hoc test. *p < 0.05, **p < 0.01 versus CTR-NOVX or CTR-OVX. # p < 0.05, ## p < 0.01 versus PILO 30-NOVX or PILO 30-OVX. www.nature.com/scientificreports/ this receptor is crucially involved in hormonal and behavioural responses associated with stress 62,63 . A decrease in ER-alpha and an increase in ER-beta occurred within 90 min of SE onset; in NOVX animals (Fig. 4b,c, respectively), the pilocarpine treatment and the following seizures may be the causes of this receptor reorganization. Although ovariectomy in Proechimys does not influence the SE, changes in protein receptor expression patterns are observed in the brain. Analysing different areas of the hippocampal formation (CA1, CA3, DG ,and hilus) using immunohistochemistry, differences were found between the OVX and NOVX groups showing that ovarian hormones may play an important role in hippocampal rearrangements at SE induction (Fig. 5, Table 2). We made schematic resume of all the changes found in this study relation to serum, western blot and immunohistochemistry (Fig. 6a,b).
Finally, although the latency to SE observed in female Proechimys is similar to that in other rodents, the dose of pilocarpine required to induce SE was higher in Proechimys than in Wistar rats. Moreover, the intensity of seizures in Proechimys is lower when compared to other rodents, measured on the Racine scale. Of all biochemical analyses, glucose and progesterone appear to have an important influence. We can speculate whether the mechanism behind the glucose compensation during SE may be a part of the compensatory mechanisms that make the induction of seizures in these rodents difficult. Rearrangements that occur in the hippocampus (in mGluR2/3, GABA B , progesterone, ER-alpha and ER-beta receptors) may slightly interfere with SE. A decrease in GABAergic receptors within 90 min of SE in OVX animals may suggest ovariectomy produces changes in the hippocampus with characteristics of vulnerability to seizures. SE episodes commonly lead to recurrent seizures and epilepsy. This rodent, with an unknown mechanism underlying the innate resistance to SE, thereby represents a special opportunity to better understand SE.

Methods
Proechimys rodents were bred in a colony established at Universidade Federal de São Paulo (UNIFESP). All animals handling and experimental procedures complied with the guidelines for animal care and use of laboratory animal and were approved by the Ministry of the Environment (IBAMA-Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renováveis), protocol nº 1561643, and by the Board for Ethics in the Use of Animals (CEUA, Comissão de Ética no Uso de Animais), an institutional ethics committee of the UNIFESP, protocol nº 5594280316.
Based on our published study on characterization of the oestrous cycle in female Proechimys 14 , oestrus stage was chosen in the current work to carry out analyses on NOVX Proechimys due to low levels of progesterone observed. www.nature.com/scientificreports/ Ovariectomy. Bilateral flank incisions were performed in 18 rodents anaesthetised with 10% ketamine hydrochloride (150 mg/kg, ip) and 2% xylazine hydrochloride (11 mg/kg, ip). The ovaries were removed, and the uterine horns were replaced. The muscular wall and skin were closed using absorbable and non-absorbable sutures, respectively. Rodents were treated with tramadol hydrochloride (5 mg/kg, ip) as an analgesic and enrofloxacin (5 mg/kg, im) as an antibiotic for 3 days. The ovariectomy success was verified 1 week after surgery through vaginal smears. Ovariectomies were performed 1 month before SE induction.

SE induction.
Proechimys were treated with lithium chloride (127 mg/kg, ip) and methylscopolamine-bromide (1 mg/kg, ip, both Sigma-Aldrich) 24 h and 30 min, respectively, before pilocarpine injections. The rampup pilocarpine protocol 17 was adapted for Proechimys rodents; pilocarpine (250 mg/kg, ip) was administrated every 30 min until SE developed. This model was chosen as it efficiently induces SE with low mortality rates 17 . Seizures were scored using Racine's scale 18 . The beginning of the SE was defined as the onset of continuous generalised seizure activity without regaining normal behaviour, and pilocarpine injections were limited to five injections per animal. The SE was terminated by euthanasia of the animals after 30 or 90 min, and the brains were quickly removed.
Serum analysis. The animals were decapitated without anaesthesia for the blood collection. After 30 min, the blood samples (n = 36) were centrifuged (239×g, 15 min, 24 °C), and the serum was stored at -80 °C until analysed. The measurement of glucose, triglycerides, total cholesterol, and fractions (HDL, not HDL), GGT, T3, T4, sodium, creatinine, urea, progesterone, and oestradiol was outsourced to a clinical laboratory. The biochemistry parameters were analysed with ADVIA-2400 (Siemens, NY) equipment to determine glucose, triglycerides, total cholesterol, HDL, not HDL, GGT, and creatinine. The urease enzyme method was used to quantify urea, and the ion-electron selective method was used for sodium determination (Aptec-Biosys Ltda, Belgium). The chemiluminescence method (Beckman Coulter, Unicel DXI-800, USA) was performed for T3, T4, oestradiol, and progesterone.
Western blots. The hippocampi (n = 3 per group) were dissected and stored at -80 °C until assayed. Samples were homogenised in lysis buffer (5 M NaCl, 0.5 M HEPES, 0.5 M EDTA, pH 8.0, 1% Nonidet P-40) with protease inhibitor cocktails (Sigma-Aldrich). The protein content was determined using the Bradford method 64 . Samples were diluted in Laemmli buffer and boiled for 5 min. Six standard curves were generated using several protein concentrations. Protein samples of 40 μg for ER-alpha, ER-beta, mGluR2/3, 15 μg for progesterone and GABA B receptor, and 60 μg for parvalbumin were within the linear range of this assay. Thus, an equivalent amount of protein was electrophoresed on 10% polyacrylamide mini-gels (8-16% gel for parvalbumin) and transferred to 0.45 µm nitrocellulose membrane sheets (Millipore) through electroblotting. Samples (NOVX and OVX) were load in two gels, it was unable to load all sample in one single gel. The gels were run in parallels. The procedure and loading of all the samples were carried out in a similar fashion. To avoid any mistakes, we load one sample of control NOVX on each gel and compare all band intensities to the intensity of the standard. Finally, all bands were normalised with housekeeping protein (β-Actin) (uncut gels can be found in Supplementary Information B). Membranes were blocked with 5% non-fat milk (15 min, 24 °C) and incubated with mouse monoclonal antibodies against ER-alpha (1:500; Novus, NR3A1), ER-beta (1:2500; Novus, NR3A2), progesterone (1:1000; Abcam, ab2765), or parvalbumin (1:1000; Sigma, P3088) or rabbit polyclonal against mGluR2/3 (1:1,000; Novus, NB100-1760) or GABA (1:500; Sigma, A2052) in TBS-T (50 mM Tris-HCl, 154 mM NaCl, pH7.5, 0.1% Tween 20) plus 2% non-fat milk at 4 °C overnight. Afterwards, blots were washed with TBS-T and incubated for 1 h with peroxidase-conjugated anti-mouse or anti-rabbit immunoglobulin (Calbiochem), diluted in TBS-T plus 2% non-fat milk.
Blots were washed in TBS-T and enhanced chemiluminescence reagents (Amersham Pharmacia Biotech) were applied to the blots. The membranes were then exposed to the chemiluminescence imaging system (UVITEC), and the bands were quantified by densitometry. Anti-β-actin immunoglobulins (1:30,000; Sigma) were used as controls. The ratios between the optical densities of the NOVX or OVX groups for each receptor and the internal standard of β-Actin bands are presented as mean ± SD. Staining preparations. Proechimys brains (n = 3 per group) were removed immediately and fixed in 0.1 M phosphate-saline buffer solution (PBS; pH 7.4) containing 4% formaldehyde for 24 h at 4 °C. The brains were soaked in a cryoprotectant solution containing 30% sucrose in PBS, then cut in 40-μm coronal sections using a freezing microtome. Sections were stored individually with an anti-freezing solution (30% sucrose, 1% polyvinylpyrrolidone 40, and 30% ethylene glycol in PBS, pH 7.2). Every sixth coronal slice was kept in sequential order in that anti-freezing solution until used.
For immunofluorescence, adjacent sections were selected. Three hippocampal sections per animal were mounted on slides for further dehydration and dyeing. Images were taken using an OLYMPUS BX60 microscope with a 20× objective and the SPOT software, version 4.6 (http://www.spoti magin g.com). The classical subdivisions of the hippocampus (CA1, CA3, dentate gyrus (DG), and hilus) were identified using stereotactic coordinates 65 . Immunofluorescence. A double-labelling protocol was used to examine the co-expression of ER-alpha, ER-beta, or progesterone receptors with GABA B or glutamate (mGluR2/3) receptors. Selected brain slices were washed with 0.1 M PBS (pH 7.4), then placed in a blocking solution (containing 10% foetal bovine serum, 2% albumin, and 0.1% Triton-X) for 120 min. Afterwards, the slices were incubated with primary mouse monoclo-Scientific Reports | (2020) 10:20982 | https://doi.org/10.1038/s41598-020-77879-1 www.nature.com/scientificreports/ nal antibodies against ER-alpha (1:500), ER-beta (1:500), or progesterone (1:50) for 48 h at 4 °C. The primary antibodies were removed by several washes in 0.1 M PBS. Then, the slices were incubated with Alexa Fluor 594-coupled secondary antibodies (red, Invitrogen) for 30 min. Next, the slices were washed again with 0.1 M PBS and incubated with the second primary rabbit polyclonal anti-GABA B antibody (1:5000) or mGluR2/3 (1:500) overnight at 4 °C. The next day, the slices were washed again with 0.1 M PBS and incubated with the Alexa Fluor 488-conjugated secondary antibody (green, Invitrogen). Sections were washed in PBS, coverslipped with FluoroShield mounting medium containing DAPI (Sigma, F6057), and stored at 4 °C to preserve the fluorescence. Samples were analysed using the confocal microscope TCS SP8 (Leica) with a 40× objective. Images were imported into Fiji (http://fiji.sc, a version of the ImageJ software, National Institutes of Health, http://image j.nih.gov/ij). The coloc2 plugin was used to analyse the colocalization of ER-alpha, ER-beta, or progesterone receptors (red) with GABA B or mGluR2/3 receptors (green) in distinct hippocampal regions (CA1, CA3, DG, and hilus). Three background-corrected representative images were used to quantify each region in each group, and Pearson's correlation coefficient was used as the co-localisation parameter 66 .
Statistical analysis. For SE latency, serum analyses, and western blots, data were first transformed into Z-scores, after which the two-way ANOVA and Tukey's post hoc tests were used. The statistical significance was set at p < 0.05. GraphPad Prism version 6.01 software (GraphPad Software, Inc., http://www.graph pad.com) was used in this study.