In many previous studies, one or the other forebrain circumventricular organ, the subfornical organ (SFO) or organum vasculosum laminae terminalis (OVLT), was lesioned to test whether it was critical for the behavioral or physiological responses to sodium depletion and hypernatremia. These studies conflict in their conclusions. The present study was designed to create discrete lesions of both the SFO and OVLT in the same animals and to compare these with rats having a lesion of only the SFO or OVLT. Both the OVLT-lesioned group and the combined SFO + OVLT-lesioned group drank significantly more water and saline on a daily basis than Controls or SFO-lesioned rats. In both sodium depletion and hypertonic saline testing, rats with SFO lesions displayed transient deficits in salt appetite or thirst responses, whereas the rats with single OVLT lesions did not. In the sodium depletion test, but not in the hypernatremia test, rats with lesions of both the SFO and OVLT exhibited the largest deficit. The data support the hypothesis that a combined lesion eliminates redundancy and is more effective than a single lesion in sodium depletion tests. The interpretation of the OVLT lesion-only data may have been complicated by a tendency to drink more fluid on a daily basis, because some of those animals drank copious water in addition to saline even very early during the salt appetite test.
- subfornical organ
- organum vasculosum laminae terminalis
three circumventricular organs (CVOs) form a sensory interface between the blood and brain, the subfornical organ (SFO), organum vasculosum laminae terminalis (OVLT), and area postrema (reviewed in Refs. 3, 19, 26). All lack a blood-brain barrier, contain receptors for many substances that circulate in the blood, and contain both afferent and efferent connections with other neural structures in the brain. The SFO and OVLT lie on the rostral wall of the third ventricle in the forebrain, and the area postrema lies near the obex of the fourth ventricle in the brain stem. In the rat, all three CVOs contain receptors for ANG II and putative osmoreceptors that might facilitate thirst or salt appetite after a rise of plasma ANG II concentration or osmolality. Recent attention has focused on the two forebrain CVOs, the SFO and OVLT, as the most probable sites for the transduction of these events in the plasma into the behaviors of thirst or salt appetite because 1) thirst or salt appetite after physiological stimuli, such as sodium depletion or water deprivation, are completely abolished either by intercollicular decerebration (14) or by a lesion of the anteroventral third ventricle (AV3V), a lesion that destroys the OVLT, most of the afferent and efferent connectivity of the SFO, the ventral median preoptic nucleus (MnPO), and much of the midline hypothalamic periventricular tissues of the preoptic area (2, 19); and 2) infusions of ANG II or antagonists of ANG II into the forebrain ventricles are effective in stimulating or inhibiting ANG II-induced thirst or salt appetite, whereas infusions of these agents at even greater doses into the hindbrain are completely ineffective (6).
Many investigators have studied the effects of discrete CVO lesions on thirst or salt appetite in rats in an effort to determine the extent to which these behaviors rely on the integrity of the SFO or OVLT during dehydration or other challenges. For example, if the only mechanism the brain possessed for detecting a depletion of body sodium was to detect ANG II at the SFO, then a lesion of the SFO should abolish both thirst and salt appetite during sodium depletion. Single lesions of either the SFO or OVLT have indeed been reported to reduce or abolish thirst or salt appetite in response to a variety of stimuli related to sodium depletion or cell dehydration (1, 8, 18, 20–23, 27, 29–36, 39, 40, 42, 45, 47). In some cases, one lesion has proved more effective than the other (4, 5), and in other cases the same lesion has produced different results in different laboratories or different species (8, 29, 39, 40, 45).
Because the SFO and OVLT appear to share many of the functions related to the detection of circulating ANG II and hypernatremia, it is reasonable to consider that the remaining organ may at least partially compensate for the missing one in such lesion studies. If so, this could account for conflicting results of lesion studies between and even within laboratories. This would also predict that simultaneous lesions of both the SFO and OVLT would be more effective than a lesion of either organ alone in abolishing thirst or salt appetite after appropriate challenges. It was the goal of this study to compare the effects of combined lesions of SFO and OVLT with the effects of individual lesions of SFO or OVLT after sodium depletion or hypernatremia.
It is critical to the interpretation of this study that the CVO lesions be small and confined as much as possible to the CVOs alone, particularly in the region of the OVLT. It is already known that large lesions of the AV3V or slightly smaller lesions of the entire ventral MnPO abolish behavioral responses after many different challenges to hydration (2, 11, 12, 19). Both of those larger lesions produce chronic adipsia, hypernatremia, and other derangements of body fluid compartmentation and homeostasis (12, 19). By comparison, a lesion of the SFO produces only brief postsurgical adipsia, no chronic hypernatremia, and specific rather than general deficits in drinking behaviors (31, 32, 40, 42, 45, 47). A lesion confined to the OVLT and only a small piece of the ventral MnPO produces variable adipsia but no hypernatremia and specific rather than general deficits in drinking behaviors (1, 4, 8, 27). Thus we attempted to produce lesions that were confined as much as possible to these CVOs rather than lesions that damaged much of the surrounding tissue of the lamina terminalis and preoptic area.
MATERIALS AND METHODS
The subjects were 101 male Long-Evans rats weighing 300–600 g. They were maintained individually in hanging wire mesh cages with Harlan Teklad rat chow and tap water ad libitum. The rats also had access to both water and 0.3 M NaCl solution for drinking for at least 1 wk before the surgery. The temperature was controlled at 23°C, and the rats were maintained on a 12:12-h light-dark cycle. All procedures were approved by the Institutional Animal Care and Use Committee at the University of Washington.
After adaptation to the laboratory, all rats received either a sham lesion or a pair of lesions targeted at or near the SFO and the OVLT. The intention was to generate groups of lesioned rats having 1) a good SFO lesion but a missed OVLT lesion; 2) a good OVLT lesion but a missed SFO lesion; 3) a good lesion of both SFO and OVLT; and 4) both lesions missed. Because the OVLT lesion was done with a single penetration of the electrode, it was more likely to be missed than the SFO lesion. Therefore, we included some intentionally mistargeted SFO lesions to ensure sufficient sample size in the OVLT lesion group.
Rats were anesthetized with Equi-Thesin (Drug Services of the University of Washington Hospital, Seattle, WA; 0.35 ml/100g ip) and secured in a Kopf stereotaxic instrument. The skull was leveled between bregma and lambda, and two overlapping 2.5-mm holes were drilled on the midline just rostral and caudal to bregma. Lesions were made using a 1.0-mA current with a 31-gauge tungsten wire electrode insulated with Teflon, except for the cross-sectional area at the tip.
SFO lesions used four penetrations of the electrode for 7 s each on the midline at anteroposterior −0.1, −0.3, −0.5, and −0.7 mm and dorsoventral (DV) −5.0, −4.9, −4.7, and −4.5 mm relative to the midsaggital sinus. Intentionally missed SFO lesions had DV coordinates 1.0 mm shallower than the intentional lesions. OVLT lesions were made with one penetration for 12–15 s at anteroposterior +1.1–1.2 and DV −7.3. For all lesions, the electrode passed through the midsagittal sinus. Sham lesions were made by advancing the electrode through the midsaggital sinus without passing current. Bleeding resulting from the penetration of the sinus was controlled with direct pressure. Bone wax was applied to seal the hole in the skull before closing the wound. Betadine was administered topically, and the rats were allowed to recover from surgery on a warm heating pad before being returned to their home cages.
After experimentation, the lesioned rats were deeply anesthetized with pentobarbital sodium and perfused through the heart with saline for exsanguination followed by 10% formalin saline for fixation. Brains were removed and stored in 10% formalin saline until cutting on a freezing microtome in 50-μm sagittal sections. Sections were mounted on glass slides, stained with thionine, coverslipped, and examined by using a microprojector or light microscope. Rats were considered to have a good lesion of the SFO if the rostroventral pole containing the efferent fibers from the SFO was destroyed (4, 5, 21, 27, 29, 32–34, 36, 47). OVLT lesions were considered to be complete if they destroyed no more than the most ventral one-third of the ventral MnPO and the entire dorsal cap of OVLT (4, 5, 8, 27). Complete OVLT lesions extended in a rostrocaudal dimension from the ventral surface of the brain to the third ventricle, thus destroying all dorsal connections of the OVLT. Lesions that destroyed greater than one-third of the ventral MnPO were not counted in the OVLT lesion group.
Because our lesion groups included some animals with fair amounts of undamaged SFO or OVLT, our conclusions apply most strongly to the functions of the connectivity of the organs rather than to the functions of the organs themselves. Lind et al. (21) were the first to apply this criterion to an electrolytic SFO lesion, and it rests on the assumption that the only possible influence of the nucleus is through its connectivity. The coordinates that we use bias the lesion slightly in a rostral-ventral direction so that an incomplete SFO lesion is still likely to damage the connectivity (we have very few caudal misses). Our laboratory's original name for this type of OVLT lesion was a “ventral ventral MnPO lesion” (4, 8) to emphasize that the lesion always damages some part of MnPO and that it may not completely lesion the OVLT. We always compare the data of complete CVO lesions with those of the disconnection-type lesions, and we have never observed a significant difference. Nevertheless, the possibility remains that a fragment of undamaged CVO could act by secreting into the third ventricle or by some other unknown influence.
Daily water and 0.3 M NaCl intakes were measured in all rats from the time of the lesion until the day before the sodium depletion test 3–4 wk later. SFO lesions cause transient adipsia (usually only the first night) (42), but OVLT lesions can cause more prolonged and even fatal adipsia (8). For this reason, any rat, regardless of lesion condition, that exhibited adipsia (<10 ml) for a day received 15 ml of water by gavage the next morning. If the adipsia persisted >2 days, the rat was given a 5% sucrose solution in addition to the water and saline solution, and the daily gavages with water were continued. As the previously adipsic animals began drinking copious amounts of the sucrose solution, the gavages were halted, and the sucrose was withdrawn (sometimes by fading the concentration a bit each day) until the rats were drinking at least 20 ml of water per day and eating normally. With this procedure, mortality was negligible, and the rats quickly began to recover lost body weight. Experiments were conducted when all rats had been in healthy condition and gaining weight for at least 1 wk.
All rats were then given two challenges 1 wk apart: the first was a depletion of body sodium with furosemide, and the second was a depletion of intracellular volume with a subcutaneous (SC) injection of hypertonic NaCl.
For the furosemide test, each rat was weighed, given a single SC injection of 10 mg/kg furosemide, and placed into a wire-bottomed metabolism cage designed for separating urine and feces. No food or fluids were available at this time. After 1 h, the volume of urine that had appeared beneath the cage was recorded, and the rats were given water to drink in an inverted graduated cylinder fitted with a drinking spout. Water was allowed for drinking overnight without any food or sodium present. At ∼23 h after the furosemide, water was removed, the rat was weighed, the total urine volume from the time of the injection was recorded, and a sample of urine was taken for later analysis of sodium and potassium concentrations by flame photometry. About 24 h after furosemide, fresh water and 0.3 M NaCl solution were provided in burettes calibrated to 0.1 ml, and the latencies from this provision until the rats drank water and saline were each recorded. Intakes of the solutions were measured at 15, 30, 60, and 90 min.
For the hypertonic saline test, the rats were given a burette of fresh water in the home cage at least 30 min before the challenge. Each rat was weighed, given a single SC injection of 2 ml/kg of 2 M NaCl, and replaced into the home cage. The latency to drink water was recorded, and intake was measured at 15, 30, 60, and 90 min.
Dependent variables that were not complicated by large differences in variance or normality were analyzed by ANOVA or analysis of covariance (ANCOVA). ANCOVA was used when it was necessary to reduce the effect of differences in the body weights of the animals on the intake or excretion data. All regressions from the ANCOVA that are reported as significant are positive correlations. This method adjusts the group means for differences in body weight to the extent that intake and body weight are actually observed to be positively correlated within the groups. Post hoc comparisons used Tukey's honestly significant difference test. Some intake data were analyzed by using nonparametric statistics because of large differences in the variances or normality between the groups. First, a Kruskal-Wallis test was conducted across the groups at each time interval. If that test was significant, we proceeded to examine all pairwise comparisons with the use of the Mann-Whitney test. Tests were conducted on the uncumulated data at 15, 30, 60, and 90 min. Nonparametric tests were also used for drinking latencies. Data are presented as means ± SE. A probability of <0.05 was required for significance.
Histological examination of the animals revealed that 20 rats had sham lesions and 12 rats had lesions that completely missed both the SFO and OVLT (many of these were intentionally missed SFO lesions and targeted OVLT lesions). The data for these sham- and control-lesioned animals were statistically very similar and were combined into one Control group for purposes of further analysis (n = 32).
This Control group and four other groups were included in the main analysis (total n = 69). The group with complete lesions of both the SFO and OVLT included nine rats (SxOx), complete lesions of SFO with missed OVLT lesions included 11 rats (Sx), and complete lesions of OVLT with missed SFO lesions included 6 rats (Ox). Several other classifications were possible, but we included in the main analysis one additional group of 11 rats that had a complete SFO lesion and some partial damage to the OVLT that was not sufficient to call complete (SxOxp).
Of the 35 rats with good SFO lesions, 14 had 100% complete lesions, and 21 had disconnection-type lesions with variable amounts of damage to the body of SFO. Of the 15 rats with good OVLT lesions, 10 had complete lesions, and 5 had disconnection-type lesions.
Other groups with barely sufficient numbers for comparison were a group of four rats having variable amounts of damage to the ventral MnPO (instead of or in addition to the OVLT) without damage to the SFO and another group of four rats with similar damage to ventral MnPO and complete destruction of the SFO (SxVmnx). Because of the small sample sizes, these were analyzed separately. Twenty-four rats were omitted because the damage to the CVOs was only partial, the lesion was too large, or for a few other isolated technical reasons.
Photomicrographs of the various types of lesions are presented in Fig. 1.
The body weights of the five groups in the analysis immediately before surgery, furosemide injection, or hypertonic NaCl injection are shown in Table 1. The groups did not differ significantly before surgery, but the combined SxOx lesion group began as the lightest group and lagged significantly in body weight throughout the experiment. Both the SxOx group and the Ox group experienced some adipsia postsurgically, but they recovered after a few days. The furosemide experiment was begun by the time that all groups had regained their initial body weights. By 1 wk later when the hypertonic challenge was given, the groups were gaining weight similarly.
This discrepancy in body weight gain after surgery complicated the interpretation of the intake data in the subsequent tests. As will be detailed below, the SxOx group drank significantly more water and saline than Controls during the last 7 days of daily intakes before the furosemide injection, but they tended to drink less water or saline during the two challenge periods. Thus water intake by the rats was not strictly in proportion to body weight, and a correction of all intake by body weight (e.g., ml/100 g body wt) would be misleading.
For that reason, we will present uncorrected data and analyses; however, we will also provide the adjusted means resulting from an ANCOVA with the pretest body weight as the covariate where appropriate. This analysis corrects the data not simply by the body weight of the animal but by the observed relationship between body weight and the intake or excretion variables within the groups. This ANCOVA adjustment sometimes did and sometimes did not change the interpretation of the analysis.
Rats with CVO lesions were occasionally adipsic for ≥1 day after the lesion. This phenomenon was more severe in the groups with a complete OVLT lesion than in the groups with only a lesion of the SFO. During this transition period, some OVLT-lesioned rats would drain a 100-ml graduated cylinder overnight if it contained 5% sucrose but would nevertheless return to the adipsic pattern the next night if the sucrose was withdrawn. All rats included in the study were successfully weaned from sucrose and were then given ∼2–3 wk to ingest only water, 0.3 M NaCl, and food before the furosemide challenge. The water and saline intakes were averaged for the last 7 days of this period, and the data are presented in Fig. 2. Daily water intake differed significantly by group [F(4,63) = 6.33, P = 0.0002], and follow-up tests demonstrated that the SxOx group drank a greater amount of water than the Control group or than either of the other SFO-lesioned groups, Sx or SxOxp. Daily saline intake also differed significantly by group [F(4,63) = 4.47, P = 0.003], and follow-up tests determined that the SxOx and SxOxp groups both drank more saline than the Control group. Although the OVLT-lesioned group, Ox, had an absolutely larger mean saline intake than either the SxOx or SxOxp groups, the difference was not significant because of the smaller sample size in the Ox group. Obviously, the total daily fluid intake (water plus saline) was also significant [F(4,63) = 7.21, P = 0.0001], and follow-up tests determined that the SxOx group drank more total fluid than either the Control group or the Sx group. In addition, the Ox group also drank more total fluid than the Control group. In summary, the Control group and Sx group drank the least, the group with partial OVLT damage drank an intermediate amount, and the SxOx and Ox groups drank the most.
Excretion data from the furosemide experiment are given in Table 2, where the adjusted mean values from the ANCOVA are given in parentheses. Despite the significant difference in body weight, the urinary excretion of water at 1 h [F(4,62) = 1.89, P = 0.12] and 24 h [F(4,62) = 2.05, P = 0.10] and the excretion of sodium at 24 h [F(4,62) = 1.16, P = 0.34] did not differ significantly among the groups in the raw analysis. Potassium excretion was significantly reduced in the SxOx group [F(4,62) = 2.98, P = 0.03]. However, the SxOx group did have the smallest mean for 1-h urine volume and 24-h sodium excretion as well as for potassium excretion. When analyzed by ANCOVA, we found that the regression of the 1-h urine volume and prefurosemide body weight was significant [F(1,61) = 11.11, P = 0.001], and the effect of groups was still not significant (P = 0.34). For 24-h urine volume, neither the regression (P = 0.15) nor the effect of groups (P = 0.18) was significant after ANCOVA. For 24-h sodium excretion, the regression was significant [F(1,61) = 8.39, P = 0.005], but the effect of groups was not significant after ANCOVA (P = 0.66). For 24-h potassium excretion, the regression was significant [F(1,61) = 8.64, P = 0.005], but the effect of groups was not significant after ANCOVA (P = 0.24). Thus the groups did not differ greatly in water and sodium excretion after furosemide, and the correction for body weight in the ANCOVA adequately controlled for the group difference in body weight when there were differences. This was particularly evident in the potassium excretion, where a significant difference disappeared after adjustment for the body weight covariate.
The water and saline intakes consumed by these groups during the 90-min salt appetite test are presented in Fig. 3. The left-hand panels present the mean amount of fluid ingested by each group during the period represented, i.e., 0–15, 15–30, 30–60, or 60–90 min. The right-hand panels present the cumulative 90-min intake of each group as both the observed mean intake and the adjusted mean intake from the ANCOVA.
Intakes during the four time periods had quite different variances because some groups drank highly variable amounts of fluid and some groups drank virtually none at all. For that reason, each time period was analyzed by nonparametric statistics as described in Statistics.
For water intake during the salt appetite test 24 h after furosemide, the Kruskal-Wallis ANOVA for ranked data was significant during the 15–30 min (P = 0.03) and 30–60 min (P = 0.02) intervals. Follow-up Mann-Whitney tests at 15–30 min demonstrated that the SxOx and Sx groups consumed significantly less than the Control group and that the Sx group was also significantly less than the Ox group. Between 30 and 60 min, the SxOx group drank significantly less water than the Control, Sx, and Ox groups. Thus the combined lesion group had a larger deficit in drinking than either of the groups with a lesion of only one CVO. The Ox group did not drink less water than the Control group at any time point. Variability was wide in the SxOxp group at 30–60 min: seven rats in the group drank no water, but, of the other four responding rats, two drank ∼9 ml each.
For saline intake during the salt appetite test 24 h after furosemide, the Kruskal-Wallis ANOVA for ranked data was significant during the 0- to 15-min (P = 0.0004) and 15- to 30- min (P = 0.02) intervals. Follow-up Mann-Whitney tests at 0–15 min demonstrated that all three SFO-lesioned groups, Sx, SxOx, and SxOxp, drank significantly less than the Control or Ox groups. Between 15 and 30 min, the SxOx and SxOxp groups drank significantly less saline than the Control groups. The Ox group did not drink significantly less saline than the Control group at any time point.
For total fluid intake during the salt appetite test, the Kruskal-Wallis ANOVA for ranked data was significant at the 0- to 15- (P = 0.0001), 15- to 30- (P = 0.0003), and 30- to 60-min (P = 0.036) intervals. Follow-up Mann-Whitney tests demonstrated that all three SFO-lesioned groups, SxOx, Sx, and SxOxp, drank less total fluid than the Control group during the 0- to 15- and 15- to 30-min intervals; however, the Sx and SxOxp groups rebounded in total fluid intake at 30–60 min and were not significantly different from the Controls. The SxOx group at 30–60 min drank significantly less total fluid than the Control group and the Sx group.
For cumulative water and saline intakes at the end of the 90-min test, there was less problem with heterogeneous variances, so the data were analyzed by ANOVA and again by ANCOVA with body weight as the covariate. In Fig. 3, the adjusted means representing the intakes after an adjustment for differences in body weight are displayed to the right of each of the actual means. Unadjusted water intake was significantly affected by the lesion treatments [F(4,64) = 4.17, P = 0.005], and follow-up Tukey tests revealed that the SxOx group drank significantly less than both the Control and the Ox groups. The regression of water intake and body weight was not significant [F(1,63) = 0.09, P = 0.76], and the difference in intake was not abolished by an adjustment for body weight using ANCOVA [F(4,63) = 3.46, P = 0.01]. Similarly, saline intake was significantly reduced in the SxOx group compared with the Control and Ox groups [F(4,64) = 5.73, P = 0.0005]. Although the regression of saline intake and body weight was significant [F(1,63) = 7.35, P = 0.009], the group difference in saline intake was preserved after ANCOVA [F(4,63) = 3.56, P = 0.011]. Total fluid intake was significantly affected by the different lesions [F(4,64) = 6.12, P = 0.0003], and follow-up Tukey tests revealed that the SxOx group drank significantly less total fluid than the Control group or the Ox group. The regression of total fluid and body weight was not significant [F(1,63) = 2.85, P = 0.096], and the effect of lesions remained significant after the adjustment for differences in body weight [F(4,63) = 4.29, P = 0.004].
Overall, the rats in the Ox group drank highly variable amounts of saline (5.8–22.8 ml), water (0.0–15.7 ml), and total fluid (5.8–35.8 ml), including one rat that drank 11.1 ml of water and 13.8 ml of saline during the first 15 min of the test and one rat that drank no water and only 5.8 ml of saline during the entire test.
Latency data for water and saline intakes were analyzed nonparametrically by a Kruskal-Wallis test. If a rat did not drink a fluid at all during the 90 min, it was given a score of 90 min. The lesion groups did not differ significantly for either the latency to drink water (P = 0.87) or the latency to drink saline (P = 0.80).
The water intakes of these groups are presented in Fig. 4. Intakes during the four time periods had quite different variances because some groups drank highly variable amounts of fluid and some groups drank virtually none at all. For that reason, each time period was analyzed by nonparametric statistics, as described in Statistics.
For raw water intake immediately after hypertonic NaCl, the Kruskal-Wallis ANOVA for ranked data was significant during the 0- to 15-min (P < 0.0001) and 30- to 60-min (P = 0.01) intervals. Follow-up Mann-Whitney tests at 0–15 min demonstrated that all SFO-lesioned groups (Sx, SxOx, and SxOxp) drank significantly less than the Control and Ox groups. However, by 30–60 min, both the SxOxp and Sx groups drank significantly more than the Ox group, and the SxOxp group drank more than the Control group. The combined lesion (SxOx) group was the only SFO-lesioned group that did not differ from either the Control or Ox groups during this period. The Ox group did not drink significantly less water than the Control group at any time.
For cumulative water intake during the entire 90 min, the total effect of groups was not significant [F(4,64) = 1.84, P = 0.12]. The regression of water intake and body weight was significant [F(1,63) = 10.01, P = 0.002], and the effect of groups became even more remote after the ANCOVA [F(4,63) = 0.64, P = 0.64].
Latency data for water intakes were analyzed nonparametrically by a Kruskal-Wallis test followed by Mann-Whitney tests when the global test was significant. If a rat did not drink a fluid at all during the 90 min, it was given a score of 90 min. The groups did differ significantly with respect to latency (P = 0.0001), and follow-up tests revealed that all three SFO-lesioned groups (Sx, SxOxp, and SxOx) required longer to begin drinking than the Control group (each P < 0.003). In addition, the SxOx group's median latency was significantly longer than the Ox group's (P = 0.013).
Twenty-three rats were removed from the overall analysis because their ventral lesions were too large to be considered discrete OVLT lesions or because the OVLT was not lesioned at all. Of these, the brains of eight rats contained a large amount of damage to the ventral part of the MnPO, and one-half of these also had SFO lesions. The behavioral data from these eight animals are summarized informally in Table 3 alongside the pertinent data for the Control group from the overall experiment, and a 95% confidence interval is provided for the eight rats. It is interesting to note for both of these MnPO-lesioned groups 1) the large daily water and saline intakes during the baseline condition, 2) the low intakes of water and saline in the salt appetite test of the furosemide-treated condition, and 3) the complete lack of response to the hypertonic saline injection through the first 15 min.
The principal hypothesis of this study was that simultaneous lesions of both the SFO and OVLT would be more effective than a lesion of either organ alone in reducing thirst or salt appetite after sodium depletion or hypernatremia. The results support this hypothesis for the sodium depletion condition but not for the hypernatremia condition. Water and saline intakes were significantly reduced over the entire 90-min salt appetite test only in the combined lesion group. This resulted from a consistently low intake of both fluids throughout the test. Rats in the Sx and SxOxp groups also had reduced intakes during the early part of the test, but they tended to have a spurt of intake later in the test that nullified the significance of the difference in cumulative intake relative to the Control group. Intakes by rats in the Ox group were so variable that it is impossible to generalize about the group, and the mean values never differed significantly from those of the Control group. By contrast, with the sodium depletion test, water intake in response to hypernatremia was greatly reduced in all SFO-lesioned groups, Sx, SxOxp, and SxOx, early in the test, and they all tended to catch up with the Controls in cumulative intake by the end of the test. The combined lesion was not exceptional in this regard.
Our laboratory (8) previously reported that daily water and saline intakes were elevated in rats with a lesion of the OVLT (then called a “ventral-ventral MnPO lesion”). Various brain lesions have been reported to induce polydipsia for one reason or other, but one well-studied example is the ventral MnPO lesion (9, 10). Some rats with ventral MnPO lesions drank excessive amounts of water or saline on a daily basis, and others did not, and the investigators were unable to detect any specific damage in the brain that could explain the results (9, 10). The animals in those studies did not have elevated renin or aldosterone levels, and they could concentrate urine or reduce sodium output appropriately when challenged with water or sodium deprivation. Thus the elevated daily water and saline intakes appeared to result directly from the lesion rather than from physiological consequence of the lesion. We observed a similar elevation in the few MnPO-lesioned rats identified in this study (Table 3) as well as in OVLT-lesioned rats. Although we have not presented the data that way, we also attempted and failed to determine whether there were any commonalities among the lesions of the rats that did and did not increase their intakes. Specifically, a lesion of the OVLT did not guarantee that the hyperdipsia would occur, and conversely the OVLT lesion was not required to generate the effect.
An interesting corollary to this phenomenon is the finding that an OVLT lesion greatly sensitizes rats to salt appetite induced by deoxycorticosterone (DOCA) (4, 8). OVLT-lesioned rats given large doses of DOCA increase saline intake two- to threefold compared with sham-lesioned rats or OVLT-lesioned rats with smaller doses of DOCA (4). Furthermore, the threshold for producing a significant increase in saline intake with low doses of DOCA is much closer to the physiological level in OVLT-lesioned rats (4). Lumped together, the phenomena seem to suggest that the OVLT exerts inhibitory as well as excitatory control over water and saline intakes, and that the inhibitory effect is released once the OVLT is lesioned. One potential substrate for this effect that has recently been described is a GABAergic input from the OVLT to the MnPO that modulates noradrenaline release in the MnPO (43).
Several investigators have previously tested the effects of SFO lesions on sodium depletion-induced salt appetite, and, whereas some of these found a significant reduction or abolition of intake (40, 45), we have had difficulty reproducing that effect (29, 33, 36). In our laboratory, the SFO-lesioned group usually has the smallest mean cumulative intake, but this results from a few individual animals with very small intakes, and it is not sufficient to achieve significance. In one previous study, our laboratory examined whether the effectiveness of the lesions resulted from differences in the completeness of the lesion (29). We found no apparent or significant difference, and one of the three of the rats that did have a consistent individual deficit in salt appetite throughout multiple depletions had a rostral stalk lesion instead of a complete SFO lesion (29).
In the present study, 11 of 22 rats in the Sx and SxOxp groups had cumulative 90-min saline intakes below the entire range of the sham-lesioned rats, but the cumulative intake of the Sx and SxOxp groups was not significantly different from that of Controls. As in the prior study (29), there was not a significant difference between rats with complete or disconnection-type lesions. However, the SFO lesion clearly did affect saline intake early in the test because all three SFO-lesioned groups drank significantly less than the Control group in the first 15 min. One big difference between the present study and all others that we have conducted is the additional damage to the ventral part of the brain in all SFO-lesioned animals of the present study. Such damage was not present in other SFO lesion studies, and nonspecific damage in the region of the OVLT may have augmented the effect of the SFO lesion.
Fewer studies have assessed the effects of specific OVLT lesions on drinking behaviors. The well-known AV3V lesion reduces depletion-induced salt appetite and drinking in response to hypernatremia (2, 19). This lesion does ablate the OVLT, but it also destroys most of the MnPO, the efferents and afferents of the SFO, and much of the periventricular hypothalamic tissues lining the ventricle in that region (19). As seen in the present results and elsewhere with MnPO-lesioned rats (11, 12), drinking and endocrine responses can be severely disturbed by lesions confined to the MnPO without damage to the OVLT, so the effects of the AV3V lesion cannot be ascribed to its effects on the OVLT alone. In our experiments, we have excluded rats with OVLT lesions that had damage to more than one-third of the ventral portion of the MnPO (4, 5, 8, 27). Rats with OVLT lesions, according to our criteria, have selective deficits in drinking of saline solutions in response to low-dose administration of captopril or sodium depletion (8), although the sodium depletion test in that experiment was short (6 h) and the difference in behavior did not appear until the sodium-rich chow had been returned for the night. The OVLT-lesioned rats did not exhibit deficits in drinking to yohimbine or isoproterenol (4, 5), and, as previously noted, they dramatically increased daily intake of saline in studies of DOCA administration (4, 8). In a different study, lesions termed “A3V lesions” that appear to have been quite similar to ours in scope were found to reduce saline intake in response to a peritoneal dialysis that caused sodium depletion without a change in plasma sodium concentration (1).
In the present study, the OVLT-lesioned group had highly variable behavior in both tests, and some individuals behaved near both the high and low extremes. Generalizing about the effect of this lesion is, therefore, difficult. It is interesting to note that 1) at least one rat per test in the Ox group drank far less water or saline than the entire range of the sham-lesioned group (different rats in each test), and 2) one of these rats drank ∼11 ml of water and 14 ml of saline, equivalent to 25 ml of 0.17 M NaCl, in the first 15 min of the salt appetite test. We were not able to discriminate these animals histologically. It is highly unusual for a rat to drink this much water this early in a salt appetite test because the animals have had access to water all night. Some have speculated that rats drink water during a salt appetite test only because of the large volume of hypertonic saline that they have consumed earlier in the test, but we demonstrated that at least some of this intake results from ongoing ANG II-mediated thirst (33). Nevertheless, rats almost always consume large volumes of saline before drinking any water. The large and rather nonselective intake of fluid by some rats in the OVLT-lesioned group during the salt appetite test is reminiscent of the large intakes of both water and saline on a daily basis in that group (see Fig. 2). It is probable that at least some of the intake during the test had little to do with the need conditions established during the overnight sodium depletion and instead resulted from the same cause, whatever that may be, as the high daily need-free intake. For this to be a good explanation, however, it would have to apply only to the Ox group and not to the SxOx group, which also had high daily intakes but drank the least during the tests. Unfortunately, we did not conduct mock depletions, so we are unable to answer this question directly.
The largest deficits in water and saline intakes during the salt appetite test were by the combined lesion group, SxOx. Five of nine rats in this group drank <5 ml of total fluid during the test, and this was the only condition that produced a significant deficit in water and saline intakes throughout the test period. Clearly, the effects of the combined lesion were greater than that of either lesion alone. This did not happen because the rats are poor responders and because this group had the highest daily total fluid intakes of any group, despite being lighter than the other groups. The reduced intakes by this group are not a result of their lower body weights, because an analysis using the pretreatment body weight of the animals as a covariate in the ANCOVA did not abolish the differences in intakes, even though ANCOVA did abolish all differences in water or electrolyte excretions after the furosemide injection.
Salt appetite after sodium depletion, as seen in this test, is largely dependent on circulating ANG II (41, 44). This circulating ANG II necessarily activates brain ANG II to generate salt appetite (46). Hypothetically, this occurs when the circulating peptide activates AT1 receptors in the forebrain CVOs, and then these neurons use ANG II as a neuromodulator to activate efferent pathways inside the blood-brain barrier. The salt appetite can be abolished by an intravenous infusion of the angiotensin-converting enzyme inhibitor captopril at a dose that does not block angiotensin-converting enzyme inside the blood-brain barrier (41, 44). The appetite can then be restored by an intravenous infusion of ANG II (7, 44), and a lesion of either the SFO or OVLT abolishes this salt ingestion in response to exogenous intravenous ANG II (27). However, neither single lesion reduces the appetite under more standard conditions of overnight sodium depletion in a way that is consistently repeatable among laboratories (1, 8, 29, 33, 36, 40, 45). This discrepancy led us to speculate that both CVOs were necessary in order for the rat to produce a rapid and robust response to a mild stimulus or one with a sudden onset (infusion), whereas either CVO alone could support the behavior if the stimulus was large enough and presented for hours instead of minutes (27). We suggested that a lesion of both forebrain CVOs should be more effective in eliminating this redundancy in conditions of overnight sodium depletion, and the data from the present study confirm this hypothesis.
Studies from different laboratories and species have concluded that the osmoreceptors (or sodium receptors) responsible for thirst and vasopressin secretion in response to an elevation of plasma osmolality must be in the CVOs (25, 38), because intravenous hyperosmotic treatments that dehydrated the CVOs (e.g., NaCl) potently generated these responses, whereas intravenous infusions that increased cerebrospinal fluid sodium concentration and dehydrated cells in the brain without affecting the CVOs (urea) were less effective. Certainly, hypernatremia elicits activation of neurons as inferred from c-Fos expression in the CVOs as it does in the hypothalamus, nucleus of the solitary tract, and elsewhere (13, 16, 17, 24, 28, 35). Nevertheless, studies of lesions of either the SFO or OVLT have produced varying results in drinking tests after hypertonic challenges. We have been able to demonstrate consistent suppression or delay of drinking in response to small loads of intragastric saline (32) or food intake (34), but studies using large loads of hypertonic saline administered SC found no lasting effect of a lesion of the SFO (31) or OVLT (8). For that reason, we used a large dose of SC hypertonic saline in the present study to determine whether complete ablation of both the SFO and OVLT produced a deficit in drinking that was not seen with either lesion alone. What we observed was that the SFO lesion, accompanied by partial to complete destruction of the OVLT, always decreased water intake during the first 15 min of the test. However, the OVLT lesion by itself did not affect osmotic thirst in a systematic way (Fig. 4) and did not interact with the SFO lesion to produce any significantly greater deficit.
These data must be interpreted with the same caution mentioned in Salt appetite with respect to nonspecific drinking by the groups of OVLT-lesioned animals that had greatly elevated daily water and saline intakes. Again, we did not do a mock cell dehydration experiment, so we do not know how much would have been consumed by these OVLT-lesioned rats in the absence of any physiological stimulus. Furthermore, lesions that were a fraction of a millimeter dorsal to the OVLT in the ventral MnPO potently suppressed drinking during the first 15 min of hypernatremia, similar to the results of the SFO lesion. Ventral MnPO lesions are known to inhibit osmotic thirst (11). This could be because of endogenous sodium receptors in the MnPO (15) or because the MnPO lesion destroys fibers of passage from other osmosensitive areas such as the SFO (20).
The delay in the onset of drinking by SFO-lesioned rats during hypernatremia is similar to the effect that we observed with food-related drinking in SFO-lesioned animals (34). A meal of dry chow produced drinking in ∼8 min in sham-lesioned rats, but the SFO lesion delayed the onset of drinking until ∼30 min. After that, the SFO-lesioned rats made up the difference quickly and were never severely dehydrated. Similarly, the SFO-lesioned rats in the present study, including the double-lesioned SxOx rats, began drinking between 15 and 30 min after hypertonic saline injection and rapidly made up the difference. Thus the lesioned animals in both circumstances are behaving as if they do not see the stimulus until after the sham-lesioned animals are able to detect it. This could be a natural consequence of the fact that a hypernatremic stimulus would be available earlier to osmoreceptors or sodium receptors in the CVOs than inside the blood-brain barrier. Nevertheless, a strong osmotic stimulus would eventually be detected by osmoreceptors inside the blood-brain barrier in areas such as the MnPO and hypothalamus as the brain gradually became osmotically dehydrated, and this could short-circuit the CVO pathway and trigger the drinking response. We observed a similar phenomenon with low and high intragastric saline loads in animals with abdominal vagotomy (37). To the extent that this is true, the SFO appears to be the more important of the two forebrain CVOs with respect to hypernatremia induced by SC hypertonic saline in rats. The role of the OVLT may be more evident with different routes or doses of hypertonic saline, or its influence may somehow be negated by extracellular fluid expansion after hypertonic saline loading. Certainly the OVLT was considered for a number of years to be “the” osmoreceptor for drinking because of pioneering work done in dogs (39). Species differences are probable.
This study was supported by National Institute of Neurological Disorders and Stroke Grant NS-22274 to D. A. Fitts.
We thank Dr. R. C. Speth for a critical reading of an early version of this manuscript.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society