Male Wistar rats (~290g) were obtained from the Animal Facility of the Department of Physiological Sciences, Institute of Biological and Health Sciences, Federal Rural University of Rio de Janeiro (UFRRJ), or from the Center for the Development of Experimental Models for Biology and Medicine (CEDEME) Federal University of São Paulo (UNIFESP). The rats were housed under controlled conditions of temperature (22 ± 2 °C) and 12/12 hours light-dark cycle (lights on at 6 a.m.). All procedures performed were in accordance with current Brazilian legislation and the “Guide for the Care and Use of Laboratory Animals” (17) and were evaluated and approved by the ethical committees for animal use of Federal Rural University of Rio de Janeiro (CEUA-ICBS, protocol number 001/2017) and Federal University of São Paulo (CEUA-UNIFESP, protocol number 7236281119, ID 009443, 2019).

Figure 1 illustrates the experimental procedures performed in this work. A first set of rats was randomly separated into the following groups: control (CT) which had free access to filtered water and standard chow (1% w/v NaCl, Rhoster, São Paulo, Brazil); WD group, in which water was removed for 24 or 48 hours with free access to standard chow; and SL group, in which the only fluid available was 1.8% NaCl solution for 24 or 48 hours, also with free access to standard chow. The exploratory and anxiety-like behaviors were evaluated via 5 minutes of exposure to the elevated plus maze (EPM) test. A different set of animals of control, 48h WD and 48 SL groups were tested for 10 min in the open field test. The behavioral tests were performed at night (7:00 to 11:00 p.m.) because the dark period is the normal activity period of rats (18) and also because Martelli et al., 2012 (9) observed that changes in the locomotor behaviors occur during the dark but not during the light period in water deprived rats. Immediately after the EPM test, the animals were euthanized and blood was collected to determine the hematocrit and plasma osmolality.

A third set of rats was submitted to the control procedure, 48h WD or 48h SL and euthanized at night (7:00 to 11:00 p.m.) for gene expression evaluation by qPCR. The brains were rapidly removed from the skull, frozen with dry ice, and stored at −80°C.

Additionally, to investigate the contribution of OXTR signaling to the WD induced hypoactivity, a fourth set of rats was used to test whether the microinjection of OXTR antagonist in the CeA would be able to alter the dehydration-induced hypoactivity observed in the 48-h WD rats when tested in the EPM. So, 7 days after CeA cannulation, the rats were submitted to a control procedure or 48 h of WD and on day 9 received the OXTR antagonist or vehicle microinjection 30 min before EPM testing. At the end of the test, the rats were returned to their home cages and were allowed to drink water freely. The water consumption, measured in grams, was recorded at 30 and 120 minutes.

The EPM was employed to analyze the exploratory and anxiety-like behavior in rats (19). This apparatus consists of two opposed open arms (50 × 10 cm each), and two opposed closed arms of the same size with 40 cm high sidewalls, connected by a central area (10 × 10 cm) and raised 50 cm from the ground. Rats were placed into the center area of the apparatus facing a closed arm. Each animal was tested in the EPM for 5 minutes and the apparatus was cleaned between one rat and another with a 10% ethanol solution. The EPM is used to evaluate anxiety-like behavior in rodents. The number of arm entries reflects the locomotion of the animal and the open arm exploration reflects the anxiety-like behaviors. Thus, increased exploration of open arms is associated with decreases in anxiety levels. In addition, the ethological parameter head dipping was evaluated to support the assessment of emotional reactivity. The entries in closed arms and rearing indicate horizontal and vertical exploratory activity, respectively. The time in the center area and the episodes of stretch-attend posture are associated with decision making (19, 20).

The open field test was carried out as previously described (21) to confirm that 48 h of WD induced changes in exploratory activity. Each rat was placed individually in the periphery area at the beginning of the test and allowed to explore it freely for 10 min. The arena was cleaned between one rat and another with a 10% ethanol solution. Total, peripheric and central distance travelled and percentage of time spent in the central area were subsequently analyzed using the video-tracking software EthoVision 8.5 (Noldus Information Technology, Leesburg, VA, USA).

Trunk blood was collected in heparinized tubes and centrifuged (20 min, 3000 rpm at 4 °C), after which plasma osmolality was measured in 10 μl aliquots by using a benchtop osmometer (model 5005, Precision Systems, Natick, MA, USA), based on the freezing-point method. For hematocrit determination, the blood was placed into heparinized capillaries and centrifuged at 2000 rpm for five minutes. Next, a microhematocrit scale was used to determine the percentage of blood composed of erythrocytes.

Immediately after the EPM test, rats were decapitated and their brains were frozen on dry ice and stored at −80°C. The brains were cut into 60 μm coronal sections using a cryostat (Leica Biosystems, CM 1860, Wetzlar, Germany), and the brain nuclei were identified and delimited according to the rat brain atlas (22). A 1 mm diameter micropunch needle (Fine Science Tools) was used to bilaterally collect LHb (coordinates from Bregma −2.30 to −4.70 mm), BLA (coordinated from Bregma −1.80 to −3.80 mm) and CeA (coordinated from Bregma -1.80 to -3.30 mm). The micropunch material was immediately placed in microtubes containing TRIzol^® reagent (Life Technologies, Waltham, MA, USA) and stored at −80°C. Sections were mounted on glass slides and stained with 0.1% toluidine blue to confirm the location of the punches using a light microscope. Micropunches from incorrect dissection were not used. Total RNA was extracted from punch samples using TRIzol^® reagent as recommended by the manufacturer, and was quantified in a nanospectrophotometer (DS-11^® Denovix Inc. Wilmington, DE, USA). Samples presenting a ratio of optical density (OD) 260/280 in the 1.8-2.1 range were used for reverse transcription. The synthesis of cDNA was performed using the QuantiTect Reverse Transcription Kit (Qiagen) with 500 ng of total RNA. The cDNA obtained was diluted in a proportion of 1:3 and stored at –20°C.

The qPCR was performed in duplicate or triplicate using SYBR green (Applied Biosystems) or Taqman Universal PCR Master Mix kit (Carlsbad, CA, USA). The qPCR was performed with the QuantStudio 3^® system (ThermoFisher Scientific). With the SYBR green, the following primers (Applied Biosystems) were used: Avpr1a (5′-ATCTGCTACCACATCTGGCG-3′ and 5′-TTATGAAAGGGACCCACGGC-3′), Oxtr (5′-CTTCATCCAACCCTGGGGAC-3′ and 5′-CTTGAAGCTGATGAGGCCG–3′), and Rpl19 (5′-GCGTCTGCAGCCATGAGTA-3′ and 5′-TGGCATTGGCGATTTCGTTG-3′) as the endogenous control gene. With the Taqman Gene Expression Assays (Applied Biosystems), the following probes were used: Gad1 (Rn00690300_m1), Slc17a6 (Rn00584780_m1), and Actb (Rn00667869_m1) as endogenous control. The reactions were performed using an ABI 7500 Sequence Detection System (ABI, Warrington, UK), with universal cycling conditions carried out according to the manufacturer’s instructions. The genes RPL19 and β-Actin give highly reproducible and stable expression measurements between different physiological challenges within the CeA, BLA and LHb. For relative quantification of gene expression, the 2^−ΔΔCT method was employed (23).

Rats were anesthetized with ketamine (100 mg/Kg; i.p.) and xylazine (10 mg/Kg; i.p.). After the onset of anesthesia, animals received ketoprofen (3 mg/kg; s.c.) as analgesic, pentobiotic (30000 I.U./Kg; i.m.) as antibiotic, polyacrylic acid as eye lubricant and lidocaine 0.5% as local anesthetic in ears and at the incision site. A 23 G and 14 mm length cannula was implanted bilaterally using the following stereotaxic coordinates relative to the bregma: posterior −2.2 mm, lateral ± 4.3 mm and ventral −7.6 mm. A stylet was inserted into the guide cannula for obturation. The cannula guide and the stylet were fixed with dental cement. Twenty-four hours after surgery, the ketoprofen injection was repeated. Rats were allowed to recover for one week after the surgical procedure. Two hours before the EPM test, rats were individually placed in standard polyethylene cages and the stylet was removed from the guide cannula. Thirty minutes before the EPM test, a 32 G needle connected to a polyethylene tube (PE 10), which in turn was connected to a 1 µL syringe (Hamilton) was inserted and extended 0.2 mm from the end of the cannula guide. Then, 0.4 µl of the OXTR antagonist L-371,257 (Santa Cruz Biotechnology, Inc., Dallas, TX, 1; sc-204038; 4 µM in 0.9% saline-10% DMSO) or vehicle (0.9% saline-10% DMSO) was manually infused during 1 min. The needle was left for 1 additional minute to allow the complete diffusion of the drug and then it was removed. The antagonist dose was chosen based on the literature (4, 24).

In order to verify the right infusion site at the CeA ( Supplementary Figures 1 and 2 ), at the end of the experiment the rats were administrated identically as described above with 0.4 µl of 0.2% Evans blue, anesthetized with isoflurane and decapitated. Then the brains were removed and fixed by placing them into a solution of 4% paraformaldehyde in 0.1 M phosphate buffer for two days. Fixed brains were sliced into 60 µm coronal sections with a cryostat (Leica Microsystems CM1850 Cryostat; Wetzlar, Germany) and light microscopy was used to confirm the infusion site according to the Paxinos atlas (22). Data from rats with one or two misplaced cannulas were not included in the analyses.

All values are presented as means ± SD. Data from hematocrit, plasma osmolarity, EPM, OF and qPCR with normal distribution (depending on the Shapiro–Wilk test) were subjected to 1-way analysis of variance (ANOVA; the value of the F-statistic is reported) with 3 levels: control, WD and SL. Significant differences between groups were further analyzed through the Tukey post hoc test. When these variables were not normally distributed, data were subjected to the Kruskal-Wallis test (the value of the H-statistic is reported) with three levels: control, WD and SL. Significant differences between groups were further analyzed through Dunn’s multiple comparisons test. Data from OXTR antagonist were submitted to 2-way ANOVA with the hydration status factor having two levels (control and WD) and the drug administration factor also with two levels (vehicle or OXTR antagonist). The values of stretch-attend posture and rearing in EPM did not have normal distribution and were rank transformed before statistical analysis according to Hora and Conover, 1984 (25). All statistical analyses were conducted with the GraphPad Prism software (version 8, San Diego, USA). In all cases, p-values smaller than 0.05 were considered to indicate a significant effect.

In order to validate our dehydration models (WD and SL), we evaluated the plasma osmolality and the hematocrit ( Figure 2 and Supplementary Table 1 ). As expected, osmolality was significantly affected by dehydration protocols after 24 [H=6.051; df=2; p=0.0485; Figure 2B ] and 48 hours [H=10.94; df=2; p=0.0042; Figure 2D ]. The hematocrit was affected by 48 h of WD [F_(2,25)=12.86; p=0.0001] since rats submitted to 48 h of WD showed higher hematocrit values than the control and the SL groups (p=0.0003 and p=0.0019; respectively; Figure 2C ). These results demonstrated that after 48 h, both WD and SL increased the plasma osmolality, thus validating our dehydration models.

We carried out the EPM test to assess whether dehydration triggered by WD and SL influences anxiety-like behavior and/or locomotion ( Figure 3 and Supplementary Table 2 ). Figure 3A–F shows that 24 h of WD or SL did not alter rats’ behavior in the EPM test. However, after 48 h of dehydration ( Figure 3G ), we found a significant effect on closed arm entries [F_(2,33)=6.192; p=0.0052], with WD rats entering less in the closed arms than the control (p=0.0085) and SL rats (p=0.0185). The other parameters assessed were not altered by 48 h of WD or SL ( Figures 3G–L ). These results suggest that 48 h of WD induces hypolocomotion in EPM without affecting anxiety-like behavior.

Additionally, the open field test ( Figure 4 and Supplementary Table 3 ) showed a decrease in total ( Figure 4A ) and peripheral ( Figure 4B ) exploration of 48 h WD group [F_(2,12)=5.05; p=0.0256 and H=8.42; p=0.0068; d.f. = 12; respectively]. On the other hand, the distance traveled and time spent in the central area were not affected by WD or SL, confirming that 48 h reduced nocturnal exploratory behavior without significantly affecting the anxiety-like behaviors.

Dehydration activates the MCNs, increasing AVP and OXT synthesis and secretion (1). In addition to their classical osmoregulation function, these neuropeptides have key roles in modulating anxiety-like and locomotory behaviors (6, 9). For this reason, we investigated the gene expression of AVP and OXT receptors in the CeA, BLA and LHb, limbic structures known to regulate anxiety-like and locomotory behaviors and to receive projections from MCNs (2, 5, 6). Additionally, we measured the mRNA expression of gene markers of GABAergic (Gad1) or glutamatergic (Slc17a6) neurons to infer whether the dehydration could modulate the level of excitation or inhibition of the neurons of the structures mentioned above ( Figure 5 and Supplementary Table 4 ).

While dehydration did not change the expression of these genes in the LHb ( Figures 5A–D ) and BLA ( Figures 5E–H ), it altered Oxtr gene expression in CeA [F_(2,15)=5.994; p=0.0122]. Our data showed that rats submitted to 48 h of WD had higher Oxtr gene expression levels in CeA compared to controls (p=0.0103; Figure 5L ). The gene expression values of Slc17a6, Gad1 and Avpr1a in CeA were not altered by the dehydration model ( Figures 5I–K ). Thus, we investigated whether OXTR in CeA can modulate the locomotor response of rats.

To further investigate whether OXTR signaling in the CeA mediates the behavioral alteration observed in WD rats, we injected the OXTR antagonist L-371,257 into the CeA of rats after 48 h of WD. Twenty minutes later, the animals were submitted to the EPM test ( Figure 6 and Supplementary Table 5 ). Figure 6A shows that rats submitted to 48 h of WD entered the closed arms less than the control rats (not water deprived) [F_(1,32)=8.154; p=0.0075]. However, the OXTR antagonist used failed to reverse the reduction of closed arm exploration in water-deprived rats. Open arm entries and time in the center area, head dipping and stretch-attend posture episodes ( Figures 6B–F ) were not affected by the OXTR antagonist or by WD. Note that these data match those presented in Figure 3 . There was a significant effect of antagonist administration on reduction of rearing episodes [F_(1,32)=5.670; p=0.0234], demonstrating the efficiency of the oxytocin receptor antagonist used ( Supplementary Figure 3 and Supplementary table 5 ).

At the end of the plus maze test, animals were allowed to drink water freely, and the respective intakes at 30 and 120 minutes ( Figures 7A, B and Supplementary Table 6 ) were recorded. As expected, water-deprived rats drank more water than the control animals at 30 and at 120 minutes [F_(1,33)=504.8; p<0.0001 and F_(1,33)=521.4; p<0.0001; respectively]. However, the oxytocin receptor antagonist had no effect on the water drank by the rats ( Figures 7A, B ).