Divergent effects of ERα and ERβ on fluid intake by female rats are not dependent on concomitant changes in AT₁R expression or body weight

Abstract

Estradiol (E₂) decreases both water and saline intakes by female rats. The ERα and ERβ subtypes are expressed in areas of the brain that control fluid intake; however, the role that these receptors play in E₂'s antidipsogenic and antinatriorexigenic effects have not been examined. Accordingly, we tested the hypothesis that activation of ERα and ERβ decreases water and saline intakes by female rats. We found a divergence in E₂'s inhibitory effect on intake: activation of ERα decreased water intake, whereas activation of ERβ decreased saline intake. E₂ decreases expression of the angiotensin II type 1 receptor (AT₁R), a receptor with known relevance to water and salt intakes, in multiple areas of the brain where ERα and ERβ are differentially expressed. Therefore, we tested for agonist-induced changes in AT₁R mRNA expression by RT-PCR and protein expression by analyzing receptor binding to test the hypothesis that the divergent effects of these ER subtypes are mediated by region-specific changes in AT₁R expression. Although we found no changes in AT₁R mRNA or binding in areas of the brain known to control fluid intake associated with agonist treatment, the experimental results replicate and extend previous findings that body weight changes mediate alterations in AT₁R expression in distinct brain regions. Together, the results reveal selective effects of ER subtypes on ingestive behaviors, advancing our understanding of E₂'s inhibitory role in the controls of fluid intake by female rats.

the ovarian hormone estradiol (E₂) influences many behaviors in adulthood and has developmental effects that organize the nervous system to produce key sex differences, including susceptibility to many psychiatric conditions, such as depression, eating disorders, and anxiety disorders (29). An obvious starting point in understanding how E₂ contributes to sex differences is identifying how E₂ signals to influence behavior. Our understanding of the multiple estrogen receptor (ER) subtypes, their neuronal distribution, and their function has expanded greatly from the early acceptance of ER as a single nuclear receptor (34). Unfortunately, it is still unclear how signaling through multiple ER subtypes mediates E₂'s behavioral effects.

Measurement of fluid intake by laboratory rats has emerged as a useful approach to study the receptor-subtype requirements for E₂'s behavioral effects. E₂ decreases water and saline intake in female rodents (3, 34). Nocturnal fluid intake and intake stimulated by extracellular dehydration (or treatments with endocrine components of extracellular dehydration, such as administration of ANG II) are both lowest on the day of estrus (1, 5, 12, 14, 44). When rats are ovariectomized (OVX), these cyclic changes in fluid intake are lost, and daily baseline intake increases (14, 44), but E₂ treatment reduces intake to levels comparable to those seen on the day of estrus (15, 16, 21, 23). Although both E₂ and progesterone fluctuate across the estrous cycle, progesterone treatment alone or in combination with E₂ does not influence fluid intake (23, 40), suggesting that E₂ is sufficient for the estrus-related changes in fluid intake.

Despite clear evidence that E₂ decreases fluid intake and the strong likelihood that this effect is mediated by binding to an ER, the diversity in ER subtypes provide a number of ways by which E₂ can act at the cellular level to influence fluid intake. Our previous studies suggest that both nuclear and membrane-associated ERs contribute to E₂'s inhibitory effect on fluid intake (35); however, little research has focused on identifying which specific ER subtypes mediate E₂'s antidipsogenic and antinatriorexigenic effects. Despite the apparent simplicity in determining which receptor is involved, the situation is markedly complicated because of the diverse ER subtypes and their varied cellular localization. For example, ERα and ERβ proteins function as classic nuclear transcription factors that regulate gene expression over hours to days, but also localize to the plasma membrane, where they may engage signaling pathways more traditionally used by receptors on the plasma surface (2, 9). Moreover, the recent identification of novel membrane-associated estrogen receptors, GPER-1, ER-X, and Gq-mER, provide additional mechanisms by which E₂ can rapidly change neuronal activity and more gradually influence gene expression by activating second messenger signaling pathways (47). Although recent studies in our laboratory suggest that GPER-1 selectively reduces saline intake (33), the roles of other ER subtypes, particularly ERα and ERβ, in these behaviors related to body fluid balance remain an open question.

ERα and ERβ are expressed in areas of the brain that are critical for the control of fluid balance. For example, the subfornical organ (SFO) plays a major role in the antidipsogenic effect of E₂ through projections to and inputs from other areas of the brain (17, 27, 41, 42), and ERα is the only known ER subtype expressed in this structure (38). The paraventricular nucleus of the hypothalamus (PVN) acts downstream of the SFO in mediating E₂'s antidipsogenic effect (41, 43), and this structure has strong ERβ expression (38, 39). As such, it seems reasonable to hypothesize that these receptor subtypes play different, but perhaps coordinated, roles in fluid intake, with separable contributions to salt and water intake. Accordingly, we tested the roles of ERα and ERβ in both overnight water intake and in ANG II-stimulated water and saline intakes by female rats. We also examined changes in ANG II type 1 receptor (AT₁R) expression after selective ERα and ERβ activation to test for a downstream effector mechanism that underlies changes in fluid intake.

METHODS

Animals and housing.

Female Long-Evans rats (Harlan Laboratories, Indianapolis, IN) weighing 175–200 g upon arrival were used in all of the studies described. For behavioral experiments (experiments 1, 2, and 5), rats were singly housed in hanging wire-mesh stainless-steel cages with ad libitum access to food (Teklad 2018; Harlan Laboratories) and tap water unless otherwise noted. Rats in two-bottle intake tests (experiments 2B) had access to an additional bottle containing a 1.5% saline solution. All testing occurred in the rats' home cages. Rats in experiments 3 and 4 were singly housed in standard shoebox cages with ad libitum access to food and tap water unless otherwise noted. Body weight was monitored daily and at specific times during experimental protocols. The temperature- and humidity-controlled colony room was maintained on a 12:12-h light-dark cycle (lights on at 0700). All experimental protocols were approved by the Animal Care and Use Committee at the University of Buffalo, and the handling and care of the animals were in accordance with the National Institutes of Health's (NIH's) “Guide for the Care and Use of Laboratory Animals.”

Surgery.

Rats were anesthetized with an injection of a mixture of ketamine (80 mg/kg im; Fort Dodge Animal Health, Fort Dodge, IA) and xylazine (4.6 mg/kg im; Akron, Decatur, IL) and underwent bilateral ovariectomy surgery, as previously described (35). OVX rats in experiments 2 and 5 were also implanted, in the same surgery, with an intracranial cannula (coordinates: 0.9 mm posterior and 1.4 mm lateral to bregma and 2.8 mm ventral to the surface of the skull) to allow delivery of drug to the right lateral ventricle, as described previously (35). After surgery, all rats received a single injection of carprofen (5 mg/kg sc; Pfizer Animal Health, New York, NY), as a perioperative analgesic. One week later, accurate cannula placement was verified by measuring the drinking response to an injection of 10 ng ANG II (Sigma-Aldrich, St. Louis, MO). Only rats that drank at least 5 ml in 20 min after ANG II treatment were included in the study. Rats were allowed to recover for 2 wk before the experiments started.

Drugs.

The ERα agonist 4,4,4-(4-propyl-[1H]- pyrazole-1,3,5-triyl)trisphenol (PPT) and the ERβ agonist 2,3-bis(4-hydroxyphenyl)-propionitrile (DPN) (Tocris, Minneapolis, MN) were dissolved in 50:50 mixture of DMSO: TBS (50% DMSO) and were used in experiments 1–3. ANG II was dissolved in sterile TBS and used in experiments 2 and 5. Estradiol benzoate (EB; Sigma-Aldrich) was dissolved in peanut oil and used in experiment 4. Water-soluble β-estradiol (wE₂; Sigma-Aldrich) and its control 2-hydroxypropyl-β-cyclodextrin (Sigma-Aldrich) were dissolved in sterile TBS and used in experiment 5. To make E₂ water-soluble, this latter form is synthesized with the addition of a carrier molecule, 2-hydroxypropyl-β-cyclodextrin, which accounts for 95.31% of the drug dry weight. Therefore, the amount of drug was adjusted so that 20 μg of E₂ was present in the injected solution diluted in sterile TBS. This dose was chosen on the basis of pilot work in our laboratory.

Experiment 1: does activation of ERα or ERβ decrease overnight water intake?

To determine whether ERα, ERβ, or a combination of both receptors decreases water intake, we measured water intake after selective ER subtype activation, using a counter-balanced, repeated-measures design. Three hours before lights off, OVX female rats (n = 9) received subcutaneous injections of either 200 μg PPT, 250 μg DPN, a combination of 200 μg PPT, and 250 μg DPN, or 50% DMSO vehicle. Rats received injections once a day for 2 days to mimic the typical regimen of estradiol replacement that reduces water intake (23–25). This model of hormone replacement typically reduces intake on day 4 when EB is administered (23–25). PPT, however, reduces food intake within the first 24 h (37, 45). Water intake and drinking microstructure, therefore, were measured for 4 days to ensure that we captured the reduction in intake. Fluid intake was measured by bottle weight and using a contact lickometer, as described previously (35, 36). These procedures were repeated over the course of 4 wk to generate a complete within-subjects design. Doses of PPT and DPN were based on previous studies (32, 37) and pilot experiments in our laboratory.

Experiment 2: does activation of ERα or ERβ decrease ANG II-stimulated fluid intake?

Next, we tested whether activation of ERα, ERβ, or both receptors decreases ANG II-stimulated fluid intake. In addition to a specific test of the effect of ER activation on the dipsogenic actions of ANG II, this condition allowed us to measure water intake in the absence of feeding, thereby removing any confounding effects that a reduction in food intake may have on fluid intake. Using a counter-balanced, repeated-measures design, OVX female rats received subcutaneous injections of either 200 μg PPT, 250 μg DPN, a combination of 200 μg PPT, and 250 μg DPN, or 50% DMSO vehicle during the early light phase. Three and a half hours later, rats received a 1-μl icv injection of 100 ng ANG II, and fluid intake was measured for the next 30 min. Rats had access to a single bottle of tap water (experiment 2A; n = 12) or to tap water and 1.5% saline (experiment 2B; n = 9). The time delay after agonist treatment was based on results from experiment 1. This dose of ANG II was chosen to be consistent with previous studies examining estrogen effects on fluid intake (5, 20, 21). These procedures were repeated over 4 wk to generate a complete within-subjects design.

Experiment 3A: does activation of ERα or ERβ influence AT₁R mRNA expression in areas of the brain that control fluid intake?

To determine whether ER subtypes differentially influence AT₁R mRNA expression in different brain regions, OVX female rats received an injection of either 200 μg sc PPT (n = 7), 250 μg sc DPN (n = 6), or 50% DMSO vehicle (n = 7). Food and water were removed from the cage. Three-and-a-half hours later, rats were anesthetized by a 90-s exposure to isoflurane and then decapitated. Brains were immediately removed from the skull, flash frozen with 2-methyl-butane (Sigma-Aldrich), and stored at −80°C. The anteroventral third ventricle (AV3V), SFO, and PVN regions of the brain were obtained by sectioning 300-μm coronal sections on a cryostat and then taking four 2-mm punches from each brain region. Tissue punches were stored at −80°C until processing for AT₁R mRNA content by RT-PCR, as described previously (36). Three samples from the SFO and PVN were not processed for AT₁R mRNA due to sample contamination.

Experiment 3B: does activation of ERα or ERβ influence AT₁R binding in areas of the brain that control fluid intake?

To determine whether ER subtypes differentially influence AT₁R binding in different brain regions, OVX female rats (n = 6/group) received an injection of either 200 μg sc PPT, 250 μg sc DPN, or 50% DMSO vehicle. Food and water were removed from the cage. Three-and-a-half hours later, they were anesthetized by a 90-s exposure to isoflurane and then decapitated. Brains were immediately removed from the skull, flash frozen with 2-methyl-butane (Sigma-Aldrich), and stored at −80°C until processed for AT₁R binding by autoradiography, as described previously (36). Briefly, coronal sections (20 μm) were preincubated in buffer before being incubated for 2 h in either 500 pM ¹²⁵I-labeled sarcosine¹ ANG II (¹²⁵I-Sar¹AngII) or 500 pM ¹²⁵I-Sar¹ANG II with 3 μM nonlabeled ANG II. Sections were washed in DIW before exposure to X-ray film for 24 h. Film was digitized, and densitometry was used to evaluate AT₁R binding in the AV3V, dorsal median preoptic nucleus (dMnPOA), SFO, and PVN with ImageJ software (NIH, Bethesda, MD) using background-corrected values. After 2.5 half-lives (to allow the decay of the ¹²⁵I), slides were Nissl stained to confirm anatomical localization of AT₁R binding.

Experiment 4: what is the influence of estradiol treatment and body weight on AT₁R mRNA expression?

Studies using in situ hybridization suggest that EB decreases AT₁R mRNA in the SFO (25). As a positive control, OVX female rats were treated once per day for 2 days with 20 μg EB (n = 8) or oil vehicle (n = 9). Forty-eight hours after the second injection, rats were anesthetized by a 90-s exposure to isoflurane and then decapitated. Brains were immediately removed from the skull, flash frozen with 2-methyl-butane (Sigma-Aldrich), and stored at −80°C until processed for SFO AT₁R mRNA by RT-PCR, as described previously (36).

Experiment 5: is E₂'s inhibitory effect on water intake secondary to changes in body weight?

Systemic E₂ treatment, which decreases ANG II-stimulated water intake, also decreases body weight and AT₁R mRNA (25). Thus, it is unclear whether an E₂ replacement protocol that does not influence body weight can still influence ANG II-stimulated water intake and AT₁R expression. Therefore, we employed an E₂-treatment protocol that does not influence body weight to assess the effect on water intake and AT₁R expression. Specifically, ∼2 h into the light phase, rats (n = 7) were given 1-μl icv injections of either 20 μg wE₂ or its vehicle (cyclodextrin dissolved in TBS). Eight hours later, rats received a second 1-μl icv injection of 100 ng ANG II, and water intake was measured for the next 30 min. The dose and timing of these injections were based on pilot studies in our laboratory. One week later, the experiment was repeated, and rats received the counterbalanced treatment to achieve a within-subjects design. Body weight was measured before each intracerebroventricular injection. A second group of OVX rats (n = 6/group) received an injection of either 20 μg icv wE₂ or its vehicle. Food and water were removed from the cage, and 8 h later, they were anesthetized by a 90-s exposure to isoflurane and then decapitated. Brains were immediately removed from the skull, flash frozen with 2-methyl-butane (Sigma-Aldrich), and stored at −80°C until processing for SFO AT₁R mRNA by RT-PCR, as described previously (36).

Statistical analysis.

Data are presented as means ± SE. All statistical analyses were performed using Statistica Software (StatSoft, Tulsa, OK). The number of licks, burst size, and burst number were analyzed in 10-min (experiment 2) or 3-h (experiment 1) bins. Fluid intake, burst size, and burst number in experiments 1 and 2 were analyzed with a two-factor repeated-measures ANOVA (PPT × DPN). The number of licks during the experiment were analyzed with a three-factor repeated-measures ANOVA (time × PPT × DPN). RT-PCR values were calculated using the ΔΔCT quantification method with 18s as the housekeeping gene to which values were normalized. One-way ANOVAs were used to analyze changes in AT₁R mRNA for each brain region, AT₁R binding for each brain region, and changes in body weight after agonist treatment. Fisher's least significant difference tests were used to probe significant main effects or interactions. Student's t-tests were used to analyze changes in water intake, AT₁R mRNA expression, and body weight after EB or wE₂ treatment. Correlations were run to analyze the relationship between AT₁R mRNA expression and absolute body weight for each brain region. To normalize AT₁R mRNA expression to body weight, the delta value of the ΔΔCT formula for each animal was expressed as its body weight (g)/100. The variability in AT₁R mRNA expression in each brain region was calculated by comparing the absolute value of the deviation from the mean for each data point. A t-test then was used to compare the deviations from the mean in the original data set and in the data set corrected for body weight.

RESULTS

Experiment 1: does activation of ERα or ERβ decrease overnight water intake?

Overall, water intake was reduced during the first 24 h after agonist treatment. Initial analysis revealed no difference between water intake during the first and second days of treatment; therefore, the data during these time points were averaged. After PPT treatment, rats drank less water in a 24-h period than did rats given vehicle treatment (F_1,8 = 17.71, P < 0.01; Fig. 1A). There was no difference in water intake when rats were treated with DPN [F_1,8 = 0.28, P = not significant (n.s.); Fig. 1A]. Furthermore, DPN treatment did not affect PPT's inhibitory effect on intake (F_1,8 = 0.36, P = n.s.; Fig. 1A). There was no change in intake on days 3 and 4 (data not shown).

Fig. 1.

ERα activation significantly reduced water intake. After treatment with 4,4,4-(4-propyl-[1H]- pyrazole-1,3,5-triyl)trisphenol (PPT), but not 2,3-bis(4-hydroxyphenyl)-propionitrile (DPN), 24-h water intake was significantly reduced compared with vehicle treatment (A). The number of licks for water during the first two 3-h bins were significantly reduced in PPT-treated rats, compared with vehicle-treated rats (B). Analysis of drinking microstructure revealed that PPT treatment significantly reduced the number of bursts (C) and burst size (D) during the first 3-h bin. During the second 3-h bin, there was no change in burst number (E), but PPT treatment significantly reduced burst size compared with vehicle treatment (F). *Less than vehicle, P < 0.05.

To further investigate the nature of the inhibitory effect on water intake after PPT treatment, analysis of licks in 3-h bins was conducted. This analysis revealed a main effect of time (F_7,56 = 24.90, P < 0.001; Fig. 1B), PPT (F_1,8 = 8.69, P < 0.05; Fig. 1B), and an interaction between time and PPT (F_7,56 = 3.10, P < 0.001; Fig. 1B). Post hoc analysis revealed that after PPT treatment, rats had fewer licks for water than did rats given vehicle treatment during the first two 3-h bins (P < 0.05).

We also conducted an analysis of drinking microstructure during the first two 3-h bins. During bin 1, burst number (F_1,8 = 6.18, P < 0.05; Fig. 1C) and burst size (F_1,8 = 41.95, P < 0.001; Fig. 1D) were significantly reduced after PPT treatment. DPN had no effect on burst number (F_1,8 = 1.54, P = n.s.; Fig. 1C) or burst size (F_1,8 = 2.03, P = n.s.; Fig. 1D). During bin 2, there was no effect of PPT treatment on burst number (F_1,8 = 1.41, P = n.s.; Fig. 1E), but a significant reduction in burst size (F_1,8 = 53.52, P < 0.001; Fig. 1F). Again, DPN had no effect on burst number (F_1,8 = 0.09, P = n.s.; Fig. 1E) or burst size (F_1,8 = 3.32, P = n.s.; Fig. 1F).

Experiment 2: does activation of ERα or ERβ decrease ANG II-stimulated fluid intake?

After PPT treatment, rats drank less water in response to ANG II than did rats given vehicle treatment (F_1,11 = 10.03, P < 0.01; Fig. 2A). There was no difference in water intake in response to ANG II when rats were treated with DPN alone (F_1,11 = 0.43, P = n.s.; Fig. 2A) nor did DPN affect PPT's inhibitory effect (F_1,11 = 0.95, P = n.s.; Fig. 2A).

Fig. 2.

Estrogen receptor-α (ERα) activation significantly reduced ANG II-stimulated water intake. After treatment with PPT, but not DPN, ANG II-stimulated water intake was significantly reduced compared with vehicle treatment (A). B: number of licks for water during the first 10 min bin was significantly reduced in PPT-treated rats, compared with vehicle-treated rats. Analysis of drinking microstructure during the first 10-min bin revealed that PPT treatment significantly reduced the number of bursts (C) but had no effect on burst size (D). *Less than vehicle, P < 0.05.

To further investigate the nature of the inhibitory effect on water intake after PPT treatment, an analysis of licks in 10-min bins was conducted. This analysis revealed a main effect of time (F_2,22 = 109.51, P < 0.001; Fig. 2B), and an interaction between time and PPT treatment (F_2,22 = 4.63, P < 0.05; Fig. 2B). Rats treated with PPT licked less for water in response to ANG II during the first 10 min compared with vehicle treatment (P < 0.05). We also analyzed drinking microstructure during the first 10-min bin. After PPT treatment, the number of bursts was significantly less than what was observed after vehicle treatment (F_1,11 = 6.63, P < 0.05; Fig. 2C). Burst size, however, was not influenced by PPT treatment (F_1,8 = 2.51, P < n.s.; Fig. 2D).

During the two-bottle test, there was no effect on water intake in response to ANG II after PPT (F_1,8 = 0.92, P = n.s.; Fig. 3A), DPN (F_1,8 = 0.99, P = n.s.; Fig. 3A) or a combination of PPT and DPN (F_1,8 = 2.20, P = n.s.; Fig. 3A). This lack of effect was not surprising given the low volume of water consumed during these conditions. Analysis of saline intake (the preferred solution in this test), however, revealed an interesting contrast with the data generated from the single-bottle test, in which PPT reduced water intake. Specifically, in the two-bottle test, we found no difference in ANG II-induced saline intake when rats were treated with PPT (F_1,8 = 0.35, P = n.s.; Fig. 3B). ANOVA revealed an interaction between PPT and DPN in saline intake in response to ANG II (F_1,8 = 5.45, P < 0.05; Fig. 3B), and post hoc analysis showed that, after DPN alone, rats drank less saline in response to ANG II than did rats given vehicle treatment (P < 0.05). Analysis of licks for water in 10-min bins revealed only a main effect of time in response to ANG II (F_2,16 = 5.39, P < 0.05; Fig. 3C). The number of licks during the first 10-min bin was greater than licks during the 2nd and 3rd bin (P < 0.05). Analysis of licks for saline in response to ANG II revealed a main effect of time (F_2,16 = 35.26, P < 0.001; Fig. 3C) and an interaction between time and DPN treatment (F_2,16 = 6.09, P < 0.05; Fig. 3C). As time increased, the number of licks decreased (P < 0.05). Rats treated with only DPN licked less for saline in response to ANG II during the first 10 min compared with vehicle treatment (P < 0.05).

Fig. 3.

ERβ activation significantly reduced ANG II-stimulated saline intake. A: there was no effect of either PPT or DPN on water intake when rats had access to both water and saline. After treatment with DPN, but not PPT, ANG II-stimulated saline intake was significantly reduced compared with vehicle treatment (B). There was no effect of either PPT or DPN on the number of licks for water (C). The number of licks for saline during the first 10 min bin was significantly reduced in DPN-treated rats, compared with vehicle-treated rats (D). Analysis of drinking microstructure during the first 10-min bin revealed no effect of PPT or DPN on the number of bursts (E) or burst size (F). *Less than vehicle.

To further investigate the nature of the inhibitory effect of DPN on saline intake, analysis of drinking microstructure to saline during the first 10-min bin was conducted. DPN treatment did not affect either burst number (F_1,8 = 3.06, P = n.s.; Fig. 3D) or burst size (F_1,8 = 2.90, P = n.s.; Fig. 3D). We then analyzed licking patterns during the first two 5-min drinking bins (Table 1). During the first 5 min, there was no effect of DPN treatment on burst number (F_1,8 = 0.35, P = n.s.), but there was a decrease in burst size (F_1,6 = 13.53, P < 0.05). During the second 5-min bin, there was no effect of DPN on burst number (F_1,8 = 3.72, P = n.s) but after PPT treatment, there was an increase in burst number (F_1,8= 8.57, P < 0.05). ANOVA also revealed an interaction between DPN and PPT on burst size during the second 5-min bin (F_1,8= 6.02, P < 0.05), and post hoc analysis showed that after DPN treatment, burst size was significantly reduced compared with vehicle treatment (P < 0.05).

Table 1.

Analysis of drinking microstructure during the first 10 min of the intake test shown in Figure 3

	V	DPN	PPT	P + D
Burst Number
Bin (5 min)
1	11.89 ± 2.51	10.56 ± 2.25	9.89 ± 2.16	8.67 ± 1.72
2	11.33 ± 2.00	11.11 ± 2.65	19.56 ± 2.66†	12.22 ± 2.19†
Burst Size
Bin (5 min)
1	90.78 ± 19.70	52.34 ± 11.34*	85.36 ± 15.87	54.98 ± 9.21*
2	83.52 ± 14.59	52.34 ± 12.57*	53.99 ± 7.50	69.81 ± 8.03

Values are expressed as means ± SE.

^*Less than vehicle, P < 0.05.

^†Greater than vehicle, P < 0.05.

Experiment 3A: does activation of ERα or ERβ influence AT₁R mRNA expression in areas of the brain that control fluid intake?

Next, we tested the hypothesis that activation of ERα and ERβ affect levels of AT₁R mRNA, in an attempt to elucidate a mechanism responsible for the behavioral effects observed in experiment 2. Treatment with PPT or DPN did not, however, alter AT₁R mRNA expression in the AV3V (F_2,17 = 1.82, P = n.s.; Fig. 4A), SFO (F_2,14 = 0.26, P = n.s.; Fig. 4B), or PVN (F_2,14= 1.67, P = n.s.; Fig. 4C).

Fig. 4.

Activation of specific ER subtypes did not influence AT₁R mRNA expression. Treatment with either PPT or DPN did not influence AT₁R mRNA in the AV3V (A), SFO (B), or PVN (C).

Experiment 3B: does activation of ERα or ERβ influence AT₁R binding in areas of the brain that control fluid intake?

Changes in AT₁R mRNA do not always parallel changes in protein expression (26, 48). Therefore, we tested the hypothesis that activation of ERα and ERβ alter AT₁R binding. Analysis of brain sections using autoradiography found that AT₁R binding differed as a function of treatment in distinct brain regions (Fig. 5). There were no treatment effects on AT₁R binding in the AV3V (F_2,15 = 0.07, P = n.s.; Fig. 5, A and B), SFO (F_2,15 = 0.44, P = n.s.; Fig. 5, E and F), or PVN (F_2,15 = 1.02, P = n.s.; Fig. 5, G and H). In the dorsal portion of the MnPO (dMnPO), there was more AT₁R binding in the group treated with DPN than in the other groups (F_2,15 = 4.71, P < 0.05; Fig. 5, C and D).

Fig. 5.

Activation of ER subtypes influenced AT₁R binding in distinct brain regions. Neither DPN nor PPT treatment influenced AT₁R binding in the AV3V (A and B). There was an increase in AT₁R binding in the dMNPO after rats were treated with DPN, but this increase was not observed after PPT treatment (C and D). There were no treatment effects of PPT or DPN on AT₁R binding in the SFO (E and F) or PVN (G and H). *Greater than all other groups, P < 0.05. Representative photomicrographs are from vehicle-treated rats.

Experiment 4: what is the influence of estradiol treatment and body weight on AT₁R mRNA expression?

The lack of ER-induced changes in AT₁R mRNA and binding seemed to conflict with previous reports showing E₂ effects on AT₁R that were presumed to underlie the effect of E₂ on fluid intake (8, 20, 22, 25). In an attempt to reconcile these differences, we first attempted to replicate previous findings by measuring AT₁R mRNA in the SFO from OVX rats treated with vehicle and EB. Consistent with previous reports, rats treated with EB had lower levels of AT₁R mRNA in the SFO than was found in vehicle-treated rats (t₁₅ = 2.27, P < 0.05; Fig. 6A). Furthermore, the regimen of hormone replacement influenced body weight. Rats treated with EB lost weight during the 4-day protocol compared with vehicle-treated rats (t₁₅ = 8.34, P < 0.001; Fig. 6B). Because a change in AT₁R mRNA expression was found in animals that showed a change in body weight, we analyzed the changes in body weight during the testing regimen used in experiment 3. There was no influence of treatment on body weight during the 3.5 h-testing protocol (F_2,15 = 0.49, P = n.s.; Fig. 6C).

Fig. 6.

Peripheral EB treatment influences AT₁R mRNA expression and body weight. A: As expected, rats that were peripherally treated with EB had less AT₁R expression in the SFO compared with oil-treated rats. EB-treated rats lost body weight during the hormone replacement paradigm, while oil-treated rats gained weight (B). In experiment 3, body weight change did not differ between ER agonist and vehicle treatment (C). *Less than oil, P < 0.05.

To further investigate the relationship between body weight and AT₁R mRNA expression, correlation analyses were performed between AT₁R expression and body weight from the rats in experiment 3. We found a significant negative relationship between body weight and AT₁R mRNA expression in the AV3V (r = −0.47, P < 0.05) and SFO (r = −0.52, P < 0.05), but not the PVN (r = 0.22, P < n.s.).

Because of the relationship between body weight and AT₁R mRNA expression, we reanalyzed the data from experiment 3A by normalizing AT₁R mRNA expression to body weight. Again, neither PPT nor DPN had any influence on AT₁R mRNA expression in the AV3V (F_2,17 = 0.52, P = n.s.), SFO (F_2,14 = 0.42, P = n.s.), or PVN (F_2,14 = 0.06, P = n.s.) (Table 2); however, normalizing AT₁R mRNA expression to body weight did significantly reduce overall variability in each brain region examined. This difference was detected by calculating the absolute value of the differences from the mean for each animal with and without normalizing to body weight and using t-tests to compare these data sets (AV3V, t₁₉ = 3.11, P < 0.05; SFO, t₁₆ = 3.17, P < 0.05; and PVN, t₁₆ = 3.02, P < 0.05) (Table 2).

Table 2.

AT₁R mRNA shown in Figure 4, normalized by body weight

	Treatment Group			Variability
Brain Region	V	DPN	PPT	Pre	Post
AV3V	1.00 ± 0.11	0.93 ± 0.06	1.08 ± 0.10	0.44 ± 0.06	0.18 ± 0.04*
SFO	1.00 ± 0.11	1.11 ± 0.06	1.17 ± 0.12	0.41 ± 0.07	0.17 ± 0.03*
PVN	1.00 ± 0.20	1.16 ± 0.13	0.93 ± 0.09	0.61 ± 0.13	0.28 ± 0.06*

Values are expressed as means ± SE. Variability in AT₁R mRNA in each brain region before (Pre refers to data shown in Fig. 4) and after (Post refers to data shown in Table 2) normalizing to body weight. The measure of variability was calculated as the absolute value of the deviation from the mean and shown as the means ± SE

^*Less than Pre, P < 0.05.

Experiment 5: is E₂'s inhibitory effect on water intake secondary to changes in body weight?

Systemic E₂ treatment, which decreases ANG II-stimulated water intake, also decreases body weight and AT₁R mRNA (25). In an attempt to separate E₂ effects on intake from changes in body weight, we sought a treatment condition that would affect intake without concomitant effects on body weight. To this end, we used central injections of a water-soluble form of estradiol (wE₂) or its vehicle, and tested for effects on intake and body weight 8 h later. When treated with wE₂, rats drank less water in response to ANG II than when control-treated (t₆ = 2.6, P < 0.05; Fig. 7A). In contrast to the effect on fluid intake, we found no changes in body weight caused by wE₂ (t₆ = 0.08, P = n.s.; Fig. 7B). We also tested for changes in SFO AT₁R mRNA 8 h after wE₂ and found no effect of wE₂ (t₁₀ = 0.22, P = n.s.; Fig. 7C).

Fig. 7.

Central water-soluble E₂ (wE₂) treatment had distinct effects on water intake, body weight, and AT₁R mRNA expression. After central wE₂ treatment rats drank less water in response to ANG II compared with control treatment (A). There was no difference in 8-h body weight change between wE₂- and control-treated rats (B). There was no difference in AT₁R mRNA expression in the SFO between wE₂ and control-treated rats (C). *Less than vehicle, P < 0.05.

DISCUSSION

Overnight and stimulated water and saline intake, which are critical for body fluid balance, are inhibited by estradiol in female rodents (19, 21, 23, 24), yet very little is known about the mechanism(s) underlying these effects. In this series of experiments, we demonstrate a divergence in E₂'s inhibitory effect on fluid intake. Specifically, activation of ERα decreased overnight- and ANG II-stimulated water intake, whereas activation of ERβ decreased only ANG II-stimulated saline intake. We hypothesized that these selective effects of ERα and ERβ are attributable to differential effects on AT₁R expression in distinct areas of the brain. These treatments, however, did not decrease AT₁R mRNA or binding in areas of the brain that control fluid intake. This finding was surprising because past studies report changes in AT₁R mRNA and binding after E₂ treatment (8, 20, 22, 25). Our data, however, are consistent with a previous report (25), showing that a change in AT₁R mRNA expression in distinct brain regions may be secondary to the change in body weight after E₂ treatment. Importantly, we demonstrated that the inhibitory effect of E₂ on fluid intake, unlike the change in AT₁R mRNA expression, is not dependent on changes in body weight. Together, these data further our understanding of E₂'s inhibitory role in the control of fluid intake in females. Nonetheless, the underlying mechanism remains to be elucidated.

Interestingly, overnight fluid intake was decreased after ERα, but not ERβ, activation during the first 24-h period. In this initial experiment, agonists were administered for two consecutive days to be consistent with the regimen of E₂ replacement typically used in fluid balance/ingestion studies (23–25). It would not have been surprising if 2 days of agonist treatment were needed to decrease fluid intake because higher doses of E₂ are needed to reduce water intake, compared with the doses needed to reduce food intake (35). The more rapid effect on fluid intake observed here, however, is similar to the time course of PPT's inhibitory effect on food intake (37, 45). Consequently, we cannot rule out the possibility that the 24-h antidipsogenic effects were secondary to an effect on feeding. Future studies are, therefore, needed to more conclusively demonstrate that the overnight drinking effect occurred independently from any effect on feeding. Although it is possible that higher doses of each agonist would reveal additional changes in fluid intake, higher doses would increase risk of nonselective binding of the drugs, confounding the interpretation of the results. Future studies using other approaches (e.g., examining fluid intake in knockout animal models) could be used to address this issue without the concern about drug specificity.

The decrease in overnight water intake after ERα activation was mediated by changes in both burst number and burst size. Previous studies demonstrate that changes in burst number reflect changes in satiety signals, whereas changes in burst size reflect changes in orosensory feedback (6, 7). The observed change in drinking microstructure is similar to a previous report from our laboratory in which E₂ treatment decreased overnight water intake through changes in both satiety and orosensory feedback (35). Therefore, our current data suggest that part, if not all, of these effects are mediated by ERα. Our previous report also demonstrated that activation of membrane-associated ERs decreases overnight intake, with some differences observed between E₂ treatment and activation of only the membrane ERs (35). It is, therefore, likely that both nuclear and membrane-associated ERα contributes to the decreased intake after ER activation. Future studies will be needed to further elucidate the contributions of membrane-associated ERα and nuclear ERα in the control of overnight water intake.

Our experiments show that ERα-activation decreases ANG II-induced water intake. Previous studies have shown effects of estrogens on ANG II-induced fluid intake (5, 20, 21), but until now, there was little information about the ER subtype responsible for the effect. In addition to providing a more complete understanding of the interactions between estrogens and the renin-angiotensin system, testing the effect of estrogens on ANG II offered the benefit of testing fluid intake that occurred without food intake. This is important because PPT decreases overnight food intake (37, 45), opening the possibility that the decreased water intake observed in experiment 1 was secondary to a hypophagic effect of the treatment. The finding that ERα activation decreased water intake in circumstances when no food occurred strongly supports the view that ERα activation has a direct effect on water intake that does not depend on concomitant decreases in food intake. This conclusion is consistent with previous studies using guinea pigs that showed separable effects of estrogens on fluid and food intakes (4), as well as studies showing the effects of E₂ on water intake stimulated by isoproterenol or water deprivation (19, 24, 25). Our data also suggest that ERβ does not play a role in decreasing ANG II-stimulated water intake. This, coupled with a previous report from our laboratory suggesting a minimal role of GPER-1 on ANG II-stimulated water intake (33) strongly suggests that ERα mediates most, if not all, of E₂'s inhibitory effect on ANG II-stimulated water intake. At present, however, we cannot rule out the possibility of involvement of G_q-mER in the inhibition of stimulated water intake. Additional studies will be needed to address this possibility.

It should be noted that in addition to stimulating fluid intake, the dose of ANG II used here increases blood pressure (46), which can counteract the behavioral effect of ANG II, thereby reducing its dipsogenic potency (31). On the surface, this observation might seem to suggest that differences in blood pressure could have mediated the observed differences in fluid intake. A recent study, however, showed that ERα knockdown enhances ANG II effects on blood pressure (50), suggesting that ERα activation would reduce the effect of ANG II on blood pressure. If so, this would be expected to increase fluid intake, rather than cause the decrease that we observed. This suggests that the magnitude of our reported ERα effect on water intake may be an underestimate of the potency of the treatment.

We further extended our studies by examining changes in saline intake attributable to ERα or ERβ because both water and saline intake are decreased by E₂ (3, 34). Surprisingly, activation of ERα had no influence on ANG II-stimulated saline intake, but activation of ERβ significantly reduced ANG II-stimulated saline intake. Interestingly, although ERβ activation influenced intake of a different fluid than did activation of ERα, the effects occurred with a similar time course. This observation could be viewed as evidence for similar mechanisms of action, although further experiments will be necessary to determine whether this is, in fact, the case. Nevertheless, it was surprising that saline intake after treatment with the combination of PPT and DPN was not different from vehicle treatment. This observation stands in contrast to the inhibitory effect (driven by PPT) that this combined treatment had on water intake. It has been reported that ERα and ERβ act in an antagonistic manner in certain systems (30), and this could account for the absence of the DPN effect when PPT was coadministered. The drinking microstructure analysis, however, revealed that burst size, which drove the decrease in saline intake, was reduced by a similar amount after DPN and PPT + DPN treatment. Reconciling the differences between total volume and drinking patterns and whether any antagonist action of the two receptors influences fluid intake are important areas for future research. In any event, the observed difference in the effects on water or saline intakes highlights a previously unappreciated divergence in E₂'s inhibitory influence on fluid intake. This divergence is further illustrated by differences in drinking microstructure. Specifically, activation of ERα decreased water intake through a change in burst number, suggesting changes in satiety signals, whereas activation of ERβ decreased saline intake through a change in burst size, suggesting changes in orosensory feedback. The effect of ERβ, but not ERα, is similar to that seen in a previous study in which we found that GPER-1 decreases ANG II-stimulated saline intake (33). That decrease, however, occurred through a change in burst number, suggesting changes in satiation, which is more like the present observed change in water intake after ERα stimulation (33). Nevertheless, these findings demonstrate separable, but coordinated, means by which E₂ controls fluid intake by female rats.

To test for downstream consequences of ERα and ERβ activation that could be involved in the control of ANG II-stimulated water and saline intakes, we tested the hypothesis that selective ER subtype activation differentially influences AT₁R expression in distinct regions of the brain that control fluid intake. This hypothesis is based on observations that E₂ decreases AT₁R mRNA and binding in regions of the brain that control fluid intake, such as the SFO, organum vasculosum of the lamina terminalis (OVLT) and PVN (8, 20, 22, 25). Given the distinct ER expression pattern within these nuclei [SFO expresses only ERα, OVLT expresses both ERα and ERβ, whereas the PVN expresses ERβ, but not ERα (38)], it seemed plausible that selective ER activation influences AT₁R expression in a nucleus-specific manner. Indeed, ER activation did not change AT₁R mRNA or binding in the AV3V, SFO, or PVN, but ERβ activation increased AT₁R binding in the dMnPO (an area of the brain too small to dissect in punches for RT-PCR analysis). Because an increase in AT₁R can increase responsivity to ANG II, leading to an increase in fluid intake, this change is in the opposite direction to explain decreased ANG II-stimulated saline intake. However, it does partially support findings from one study showing an increase in AT₁R mRNA in the MnPO after the combination of E₂ and progesterone (10) and suggests that the observed effect may underlie other physiological processes under the control of AT₁R.

The inconsistency between the mRNA and binding results in the present study and the results of previous studies (8, 20, 22, 25) led us to replicate previous studies that measured changes in AT₁R mRNA. To this end, we measured AT₁R mRNA in the SFO of oil- and EB-treated OVX rats. Consistent with a previous study (25), we observed less AT₁R mRNA in EB-treated OVX rats than was found in oil-treated rats. Furthermore, over the course of the 4-day hormone replacement paradigm, EB treatment prevented daily body weight gain, a finding that has been shown numerous times (11, 18). OVX rats that were pair-fed to mimic the decrease in body weight caused by E₂ treatment also had less AT₁R mRNA in the SFO compared with that in controls, suggesting that the change in AT₁R expression in the SFO after E₂ treatment is secondary to E₂'s effect on body weight (25). Our data support this finding because there was no body weight change during the 3.5-h time interval between agonist treatment and tissue collection in experiment 3. Together, these findings suggest that E₂ may not directly influence AT₁R expression in the brain and that the observed decreases are instead attributable to E₂-induced changes in body weight.

To further explore the role of body weight on AT₁R expression, we tested for a relationship between these two measures. In both the AV3V and SFO, body weight was negatively correlated with AT₁R mRNA expression, but we did not find a correlation between body weight and AT₁R mRNA expression in the PVN. The negative correlation in the AV3V and SFO was, however, surprising because an earlier report by Krause et al. (25) found less AT₁R mRNA expression in E₂-treated rats. Because E₂ decreases both fluid intake and body weight, and E₂ was associated with a reduction in AT₁R mRNA expression, it seemed reasonable to predict a decrease in body weight would also be associated with decreased AT₁R expression. Reconciling this apparent logical inconsistency is not possible without further study, but it highlights the need for additional research to understand the relationship between body weight and AT₁R expression in female rats. A number of factors that correlate with body weight appear to influence AT₁R expression [e.g., interleukin-6 and oxidative stress (13, 28, 49)], and are targets for our future studies. Regardless of the direction of the relationship, the correlation analysis and the relationship with body weight that it revealed, prompted us to normalize the AT₁R mRNA expression data by body weight. This normalization significantly reduced the variability between treatment groups in each brain region, allowing us to test for effects of the treatments, without the confounding influence of different body weights. Together, these findings point to the need to consider body weight in future examinations of AT₁R expression, although the functional implications of body weight regulation of AT₁R expression remain unclear.

The finding that body weight is related to AT₁R expression made it necessary to consider the possibility that E₂-mediated changes in fluid intake may be secondary to a change in body weight. Our data provide evidence against this conclusion. In experiment 2, there was no difference in body weight in the time frame between ER agonist and ANG II, but differences in fluid intake were detected. We cannot rule out the possibility that this finding is specific to the agonists used. Indeed, these agonists do not completely mimic E₂. For example, the time course of effect onset after PPT and DPN treatment is much shorter than it is after E₂ treatment (37, 45); however, we also controlled for body weight differences by using central infusion of a water-soluble form of estradiol, and measuring fluid intake during a time when no body weight changes were found. SFO AT₁R mRNA expression was not affected by central wE₂ treatment, which 1) suggests our null result in experiment 3 was not due to artefactual effects of the agonist, and 2) further supports the hypothesis that changes in AT₁R mRNA expression in the brain are mediated by changes in body weight after E₂ treatment. Thus, although altered AT₁R expression may not be directly mediated by E₂, reduced fluid intake after E₂ treatment clearly is not dependent on changes in body weight. Determining how this inhibition of fluid intake happens, independent of changes in AT₁R expression remains an important open question.

Perspectives and Significance

In conclusion, these findings provide strong support for divergent receptor mechanisms by which E₂ decreases water and salt intake by female rats, and solidifies the argument that drinking behavior in the rat can serve as a useful model for studying independent actions of specific ER subtypes. Use of this approach allows us to conclude that ERα plays a key role in the antidipsogenic effect of E₂, whereas ERβ plays a key role in E₂'s anti-natriorexigenic effect. Although selective ER-mediated changes in AT₁R expression do not appear to be the mechanism underlying either effect, our findings provide additional support to the report (25), suggesting that E₂-mediated change in AT₁R mRNA expression (8, 20, 22) is likely secondary to changes in body weight. Determining the mechanism of action for the effects observed here is an important next step, which will help further our understanding of the biology of fluid ingestion in females. Furthermore, understanding how E₂ signals to influence fluid intake will aid our understanding of the basic neurobiology of ovarian hormones and how they underlie sex differences in ingestive behaviors.

GRANTS

This work was supported by National Institutes of Health Grants HL-091911 to D. Daniels, DK-098841 to J. Santollo, HL-113905 and Pilot Award from the Translational Technologies Component of the Georgetown, Howard Universities Center for Clinical and Translational Science UL1TR000101 to R. C. Speth, and DA-024754 to S. D. Clark and a grant from OCAST (HR12-196) to K. S. Curtis.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.S., K.S.C., and D.D. conception and design of research; J.S., A.M., R.C.S., and S.D.C. performed experiments; J.S. and D.D. analyzed data; J.S., K.S.C., R.C.S., S.D.C., and D.D. interpreted results of experiments; J.S. and D.D. prepared figures; J.S. drafted manuscript; J.S., K.S.C., R.C.S., S.D.C., and D.D. edited and revised manuscript; J.S., A.M., K.S.C., R.C.S., S.D.C., and D.D. approved final version of manuscript.