INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

SODIUM DEPRIVATION activates compensatory mechanisms that preserve dwindling sodium stores and promote acquisition of salt through creation of a sodium appetite (increased intake of NaCl). This is accompanied by changes in gustatory sensitivity in the chorda tympani (7, 8, 10), nucleus of the solitary tract (NTS) (23, 30), and parabrachial nucleus (PBN) (39). The common finding is that responsiveness to the taste of hypertonic saline declines, especially among the most salt-sensitive neurons. This has led to the hypothesis that rats would perceive hypertonic saline as iso- or hypotonic, shifting it perceptually from an unacceptable into an acceptable concentration range (8). However, this intensity-based interpretation fails to capture the primary feature of sodium appetite: consumption of hypertonic NaCl at levels that untreated replete rats would not show for any concentration of salt. There is more to sodium appetite than muted aversiveness.

A clue to the gustatory character of salt during deprivation may have been revealed by Jacobs et al. (23). They reported not only that salt-sensitive neurons in the NTS responded less to NaCl in deprived rats, but that sugar-oriented cells responded more briskly and that the across-neuron profile evoked by NaCl became more similar to those of the sugars. Shimura et al. (39) subsequently extended the electrophysiological findings to the PBN, reporting that the neural responses to NaCl and sucrose became more similar in sodium-deprived rats.

The behavioral response to NaCl also becomes more like that to sugars during sodium deprivation. The taste of hypertonic saline normally elicits an aversive orofacial response in rats, but in those that are sodium deprived, the response is fully appetitive (3), as is that normally shown to sugars. Moreover, the opportunity to gain access to NaCl acquires the capacity to compete with the reinforcement of hypothalamic stimulation, a trait normally reserved for sugars (6). Finally, Frankmann and colleagues (17) have offered preliminary data that sodium-deprived rats had some measure of their sodium appetite satisfied by consuming sucrose, as if these two stimuli, sugar and salt, now shared an appetitive channel. If NaCl increased its capacity to drive the cells that typically respond primarily to sugars, there would be a plausible basis for the appetitive nature of salt to these animals.

It is still unknown whether the observed changes in gustatory neural responding actually cause increased NaCl ingestion, or whether they are independent consequences of the treatments used to create sodium appetite. For instance, sodium deprivation causes taste receptor cells to be created in a salivary environment poor in sodium (44), a condition that reduces the number of amiloride-sensitive channels through which sodium is reportedly transduced (22, 48). Such an effect at the receptor level would be expected to cause decreased responding to NaCl in salt-sensitive neurons found at higher-order gustatory relays (38). Also, sodium-deficient diets tend to be high in sugar content to enhance their palatability, so deprived rats bring a unique gustatory history to the recording table. It is also possible that the state of sodium depletion may place constraints on the ability of neurons to fire, because action potentials depend on the sodium ion as a carrier of electrical charge.

It has been difficult to identify the mechanisms responsible for the observed neural changes because of the lengthy period required to activate a sodium appetite: 1 day with furosemide injections and 8-10 days without it (9, 32, 41). This period exceeds the duration over which isolation can be maintained on a single neuron with the use of standard recording techniques, and thus experimenters have been forced to compare mean responses across neurons from independent groups of rats, sodium-replete and deprived. This does not permit an assessment of which individual neurons are affected or the manner or time course of the effects. To gain this level of resolution, either the creation of a sodium appetite must be brought into the 2-h time frame of neural recording or the ability to maintain isolation on one neuron's responses must be extended to days.

We have chosen the former course. We were able to create a sodium appetite rapidly by combined administration of DOCA, the precursor of aldosterone, followed by renin, which induces the formation of ANG II. This treatment is similar to those that have been used in prior experiments (15, 35): low doses of mineralocorticoids and ANG II act synergistically to cause increased NaCl intake in as little as 4 min. The resulting appetite does not depend on a need for sodium, but it is thought to arise because it mimics the physiological changes that are normally responsible for depletion-induced sodium appetite (11). We used this approach to monitor the activity of single taste neurons in the NTS as rats passed from normal to enhanced motivation to consume NaCl.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Rats were ultimately assigned to one of two groups, Neural and Behavioral. The Neural rats were used to address the central question of this experiment: how is taste-evoked responding in NTS altered by induction of sodium appetite? Because the Neural group's salt preference could not be tested during electrophysiological recording, the Behavioral group was necessary to confirm that a sodium appetite could be elicited on a consistent basis in animals receiving an identical pharmacological treatment.

Initially, a rat's status was undetermined. Subjects were treated in pairs, and assignment to a particular group was delayed until the final day. This strategy ensured that a behavioral evaluation of the pharmacological regimen occurred regularly throughout the entire experiment, and it minimized the chances that the Neural and Behavioral rats could have been treated differently.

Behavioral Testing

After at least 6 days of recovery, rats were acclimated to a testing chamber equipped with a lick-counter circuit in which they were given access to distilled water for 30 min. In all further tests, only 0.5 M NaCl was available, and the number of licks to this stimulus was recorded every 15 min for 1 h. Testing took place daily at approximately the same time, in the middle of the animal's light cycle.

Subjects were given two 1-h sessions to determine their baseline licking of 0.5 M NaCl. It was later determined that there was no significant difference in the number of licks between these two sessions, so these values were combined to create an average baseline score for each rat.

Subjects were then treated chronically with the hormone DOCA, the precursor of aldosterone, to prime them to show a sodium appetite. This involved 6 or 7 days of twice daily subcutaneous injections of 0.125 mg each, delivered in a volume of 0.2 ml of propylene glycol. The concentration was selected to be below the threshold for causing a sodium appetite by itself, although still causing an upregulation of ANG II receptors, which in turn enhances the behavioral effects of components of the renin-angiotensin system (16, 24). All subjects were tested for NaCl intake on the penultimate day of treatment, and lick rates were compared with the average baseline rate to confirm that rats did not have an elevated salt preference due to DOCA alone.

Members of a pair were assigned randomly to their individual groups on the fifth day of DOCA injections. Neural recording required a full day of focused commitment, so testing of the Behavioral rat after renin infusion could not be conducted on the same day. Therefore, in half the cases, recording was performed on the Neural rats on day 6 and the paired Behavioral rats were tested on day 7, whereas the opposite was true in the other half.

On the final day of DOCA treatment, Behavioral rats received intracerebroventricular infusions of 25 ng of renin dissolved in BSA, delivered in a volume of 1 µl, over a period of 5-10 s. They were immediately offered 0.5 M NaCl, and the number of licks was compared with that on the prior day (after DOCA alone) to confirm that they had a salt appetite. There are reports that renin alone stimulates NaCl intake only after water drinking (32), which would have been problematic in the present experiment because Neural rats were not capable of drinking after infusion. Thus the complete regimen used here (DOCA plus renin) was chosen to stimulate a sodium appetite that would not require previous water intake.

A separate group of six rats underwent the same procedure as Behavioral rats, except that they received an infusion of the BSA vehicle alone to confirm that the infusion procedure itself was not responsible for the sodium appetite. The mean (± SE) intake of these rats after DOCA alone was 78.7 ± 45.3 licks, which did not differ significantly from the intake after BSA infusion (192.5 ± 187.1 licks).

Recording

Single units in the NTS were isolated with the use of glass microelectrodes with a tip diameter of ~1 µm (Z = 5-10 M at 1 kHz); the electrodes were filled with 1.6 M potassium citrate. Typical recording coordinates were 2.7 mm anterior to obex, 1.7 mm lateral to the midline, and 1 mm ventral to the surface of the medulla. The signal was amplified, filtered, and displayed with the use of conventional methods, and it was stored on audio tape for off-line analysis. The typical signal-to-noise ratio for cells that were recorded was 5:1.

It is critical in any single neuron study that isolation of a cell's response be uncompromised throughout the recording period. That requirement was put to a stern test in this experiment, for the lengthy procedure demanded the maintenance of isolation on 10- to 15-µm somas for at least 1 h in the notably unstable caudal hindbrain. Therefore, we did not rely on automated discrimination systems, but rather we monitored the discharges continuously on the oscilloscope, with the requirement that the isolated spike be clearly discriminable from background activity throughout the recording period. The irrevocable act, in a rat that had taken 2 wk to bring to this point, was the renin infusion. Therefore, this commitment was made only if a neuron had shown a high degree of stability during the 30- to 45-min preinfusion recording period. In ~25% of the recordings, such stability was never achieved; in another 25%, the cell was lost after renin but before the full stimulus array could be reapplied. Data were derived from the remaining 50% of the rats.

Presentation Of Stimuli

Five milliliters of each stimulus at room temperature were sprayed over the entire tongue and oral cavity at a rate of 1 ml/s. The procedure followed that of Chang and Scott (5), who determined from inspection of the mouth after dye injections in four animals that the stimulus regularly contacted the fungiform, foliate, and circumvallate papillae, plus taste buds on the nasoincisor ducts, the Geschmackstreifen, and most of the posterior palatine field. Only buds on the epiglottis, estimated to represent 4% of the rat's total (29), were not consistently contacted.

Stimulus presentations were followed by a distilled water rinse of 25-30 ml and separated by at least 1 min. The order of presentation was varied, and to avoid adaptation effects, chemicals with similar taste qualities were not run consecutively. Typically, the four accepted prototypical stimuli (0.1 M NaCl, 0.01 M HCl, 0.01 M QHCl, and 0.5 M sucrose) were presented twice before renin infusion to ensure that the cell's responding was stable. In some cases, other chemicals were also reapplied.

If a neuron remained well isolated after all stimuli had been given, renin was infused with the use of the same procedure as described for the Behavioral rats. Reapplication of the entire stimulus array did not begin until 15 min after infusion to allow time for renin to exert its effects. However, three chemicals (0.1 M NaCl, 0.01 M HCl, and 0.5 M sucrose) were presented beginning at 5 min to examine neural effects in this initial period. Stimuli were then presented for as long as unequivocal isolation of the cell could be maintained.

No more than one renin infusion was performed in each animal, and no new cells were sought after an infusion had been given. This limitation was necessary because experience with sodium appetite, or the hormones that cause it, can have permanent effects on an animal's subsequent behavior toward NaCl (13, 36) and on NTS responding (43).

Verification Of Cannula Placement

Cannula placement was confirmed in all rats except for one in the Behavioral group and three in the Neural, all of whose postrenin behavioral data were discarded. Other behavioral data from these animals (baseline intake of 0.5 M NaCl and intake after DOCA alone) were kept as these tests should not have been influenced by the presumed misplacement of the cannulas. For the three Neural rats, all pre- and postinfusion neural data were discarded.

Analysis

Comparisons were made between the two conditions (before and after renin infusion) on the following measures: spontaneous rate, evoked rates in response to each stimulus, and breadth of tuning. T-tests, repeated-measures ANOVAs, and post hoc comparisons were performed with the use of the SYSTAT package, with P  0.05 considered significant.

Further analyses allowed neurons to be classified into groups on the basis of their response properties. First, correlation coefficients were calculated between every pair of cells (n = 39 × 38/2 = 741 coefficients) on the basis of how similarly they responded to the four basic stimuli. The resulting matrix of correlations was then subjected to a cluster analysis using Clustan (47), with the result being a dendrogram indicating the relative similarity among response profiles. Clusters of neurons were identified visually, and ANOVAs and post hoc comparisons were performed to confirm that groups differed significantly on responses to their defining (best) stimuli. Each statistically identifiable group was then investigated independently for an effect of sodium appetite.

Comparisons among stimuli were also made within each condition. Correlation coefficients were calculated between each pair of stimuli (n = 14 × 13/2 = 91 coefficients) on the basis of the similarity of the profiles they generated across all neurons, and the resulting correlation matrix was used to perform multidimensional scaling (SYSTAT).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Summary

Behavior

Behavioral animals increased their acceptance of 0.5 M NaCl to 921.4 ± 63.5 licks after renin was infused on the following day [F(1,77) = 167.6; P < 0.001]. Figure 1 shows the mean intake after DOCA alone or DOCA plus renin for each 15-min period. Although the most rapid rate of licking occurred in the first 15 min after infusion, rats continued to consume an exaggerated volume of NaCl throughout the hour.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Mean (± SE) cumulative licks to 0.5 M NaCl in Behavioral subjects after renin () or on the day before, after 5 or 6 days of DOCA treatment (). ** P < 0.01 relative to DOCA alone condition.

Sodium appetite was created with a high, but not perfect, level of consistency. Ninety-five percent of Behavioral rats showed greater intake after renin infusion than they had on the previous day, and in most cases, the differences were pronounced. The threshold for increase in licking that we established to indicate a significant sodium appetite was the mean change in lick rate between baseline days 1 and 2 plus 2.33 standard deviations (P < 0.01, 1 tailed). This yielded a criterion of 289 licks, which was exceeded fully by 85% of the Behavioral rats.

Neural Activity

Spontaneous Firing Rate

Response Reliability

Breadth of Tuning

Evoked Responses Across All Neurons

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Mean (± SE) net responses to each chemical across all neurons before (solid bars) or after (open bars) renin infusion. * P < 0.05 and ** P < 0.01 relative to preinfusion. N1, 0.01 M NaCl; N2, 0.03 M NaCl; N3, 0.1 M NaCl; N4, 0.3 M NaCl; N5, 0.5 M NaCl; M, 0.1 M monosodium glutamate; H, 0.01 M hydrochloric acid; C, 0.01 M citric acid; Q, 0.01 M quinine hydrochloride; K, 0.1 M KCl; S, 0.5 M sucrose; F, 1.0 M fructose; G, 1.0 M glucose; and SA, 0.03 M Na saccharin.

Responses In Neural Groups

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Dendrogram showing classification of neurons on the basis of their preinfusion profiles of responding across the 4 basic chemicals. Numbers along the y-axis refer to the level at which groups of cells are correlated. Numbers along the x-axis identify neurons on the basis of their order of recording in the experiment and are followed by the basic stimulus that evoked the largest response in the neuron. N, 0.1 M NaCl; other abbreviations as in Fig. 2.

View larger version (36K): [in this window] [in a new window]   Fig. 4.   Mean net response rates to each chemical in the 3 groups of neurons before (solid bars) or after (open bars) infusion. A: N-cells; B: S-cells; and C: H-cells. * P < 0.05. Stimulus abbreviations as in Fig. 2.

Figure 5 depicts, for each of the 39 neurons identified by group, the mean percent change after renin in the average response across all five NaCl concentrations. Ninety percent (18/20) of N-cells and 67% (6/9) of H-cells gave lower responses to NaCl after renin, whereas 75% (6/8) of S-cells gave greater responses. This distribution was significantly different from chance (² = 11.63; P < 0.005) and arose from a difference in how S- and N-cells were affected (² = 11.83; P < 0.01). H-cells did not differ from N- or S-cells. Therefore, responses to NaCl shown by S- and N-cells were affected in opposite directions by the procedure.

View larger version (12K): [in this window] [in a new window]   Fig. 5.   Percent change in the average NaCl response as a result of renin infusion in each individual neuron. Neurons are ordered along the abscissa according to group. , N-cells; , S-cells; , H-cells; and , quinine-best neurons.

Salt-sensitive neurons (N-cells). The responses of N-cells to each chemical before and after infusion are shown in Fig. 4A. There was a general decline in responding after renin [effect of condition; F(1,19) = 4.37; P = 0.05] that nevertheless was larger for certain chemicals [condition × chemical interaction; F(13, 247) = 4.02; P < 0.001]. Most salient is that the mean response across all five concentrations of NaCl was reduced from 75 to 61 spikes/s after renin (t₁₉ = 2.33; P < 0.05). Individual stimuli that evoked significantly lower responses were 0.5 M NaCl, HCl, and Na saccharin [F(1,19)  4.66; P < 0.05 in all cases].

View larger version (29K): [in this window] [in a new window]   Fig. 6.   Mean (± SE) net responding to each stimulus in the 50% of N-cells that evoked the smallest (A) or largest (B) responses to 0.1 M NaCl before infusion. Closed bars, preinfusion; open bars, post-infusion. * P < 0.05 relative to preinfusion. Chemical abbreviations as in Fig. 2.

Sugar-sensitive neurons (S-cells). An ANOVA performed across all eight neurons in this cluster failed to demonstrate a significant effect of renin on evoked responding. It can be seen in Figs. 4 and 5, however, that there was a tendency for S-cells to give greater responses to NaCl after infusion. Two factors prevented these differences from being significant. First, one neuron showed a 41% decrease in salt responsiveness (see Fig. 5). Second, the other seven S-cells did not increase responding to NaCl over the entire 5-s evoked period. In one neuron, the NaCl response was dampened in the first second, but it increased by 76% during the tonic portion (last 4 s). Among the other six neurons, however, there was a consistent pattern of change: salt sensitivity was increased during the first second of the response, whereas the tonic portion was unaffected. When these six neurons were tested for an effect of renin with the use of their net responses for the first second, significant increases were found for 0.03 M, 0.1 M, and 0.5 M NaCl, as well as for MSG (t₅  2.51; P  0.05 in all cases). Figure 7 shows the time course of responding to these chemicals before and after infusion. These cells showed no differences in their 1-s responses to other stimuli.

View larger version (19K): [in this window] [in a new window]   Fig. 7.   Poststimulus time histograms displaying time course of responding in a group of 6 S-cells before (thin line) and after (thick line) infusion. Abbreviations as in Fig. 2.

Acid-sensitive neurons (H-cells). After renin, these neurons showed one significant change, a reduction in response to 0.5 M NaCl [F(1,8) = 7.67; P < 0.05]. The mean decline to this stimulus was only 11%, but it occurred consistently across the nine H-cells.

Stability Of Neural Groups

The majority of neurons (79%) remained in the same neural group after renin. There was also no change in the proportion of neurons contained in each group compared with preinfusion data. Among the eight neurons that did become reaffiliated, the most common change was found in the N-cells, where four neurons moved to the S-cell cluster. The other 16 N-cells continued to be grouped with each other along with three neurons that had been H-cells before infusion. There was also one S-cell that switched to the H-cell cluster.

Across-Neuron Profiles Of Activity

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Multidimensional spaces in which stimuli are compared with each other on the basis of their across-neuron profiles of activity before (A) or after (B) renin infusion. Abbreviations as in Fig. 2.

Onset Of Neural Effects

Comparison Of Preinfusion Data With Prior Control Data

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Electrophysiological Basis For Sodium Appetite

These effects are similar in nature to those observed most commonly following other means of generating a sodium appetite (7, 10, 23, 39). In the present experiment, however, they cannot be explained by factors that may have been present in prior studies, such as maintenance on a diet high in sugar or generation of taste receptor cells with altered characteristics. Our results also support preliminary work by Tamura and Norgren (42) indicating that the changes appear even when the animal is not in need of sodium. Presumably, then, these neural changes specifically serve to increase salt ingestion rather than being an independent consequence of the various procedures used to generate the appetite.

Is The Nature Of The Neural Effect Consistent With The Behavior?

We hypothesize that the basis for reversed acceptance-rejection reflexes and access to the dopamine system during sodium appetite may have been revealed in this study and prior work (23). In both cases, the creation of a sodium appetite resulted in a shift in the distribution of NaCl-induced activity: responding was dampened in N-cells but enhanced in S-cells. Although these changes do not mean that salt becomes sweet tasting, NaCl may acquire an enhanced ability to activate neurons in the NTS and its targets that are normally driven most effectively by sugars. Two sets of those targets mediate the behavioral components of a sodium appetite: 1) salivatory and pharyngeal efferents in the reticular formation, the hypoglossal nucleus, and the facial and ambiguus nuclei through which acceptance-rejection reflexes are orchestrated and 2) ventral forebrain regions associated with hedonics and reinforcement. Thus the nature of the electrophysiological effects we find here may offer a neural basis for sodium appetite.

Is The Magnitude Of The Neural Effect Consistent With The Behavior?

The electrophysiological effects, however, were small and not fully consistent. One basis for reduced consistency could be the variable effectiveness of the manipulation. The hormonal regimen was only 85% successful in generating a significant (P < 0.01) appetite in Behavioral rats, so a 15% failure among Neural rats is to be expected as well. This leads to complexities in data interpretation. The decreased response to NaCl among N-cells has been widely reported, and its appearance here in the most sodium-responsive neurons was anticipated. The fact that the reduced response characterized 90% (18/20) of the N-cells is in accord with the proportion of Behavioral rats in which the protocol was successful. More uncertain was the finding of an increased response to NaCl among S-cells, to which we were alerted by the data of Jacobs et al. (23). The fact that such an increase occurred, and was specific to NaCl, is remarkable, for it requires that N- and S-neurons have their sensitivities modified in opposite directions with a single manipulation. However, that it was transient (only during the first second of the response) and inconsistent (6 of 8 neurons) among S-cells reduces its power to serve as an explanatory mechanism in two ways: 1) it makes the demonstration statistically fragile and 2) it raises the question of whether so robust a behavioral phenomenon could be mediated by so subtle an alteration in afferent activity.

How might we relate the magnitude of changes in gustatory neural activity to alterations in taste-induced behavior? One insight is provided by comparing the neural responses to 0.3 M and 0.1 M NaCl, stimuli that elicit aversive and appetitive responses, respectively, in naive rats. During the past 11 years, we have collected three data sets from cells in the NTS that include responses to both concentrations of NaCl (20, 23, 27). Mean responses to 0.3 M NaCl among N-cells were 53, 52, and 52 spikes/s for a grand mean of 52 spikes/s; mean responses to 0.1 M NaCl were 44, 48, and 40 spikes/s for a grand mean of 44 spikes/s. Therefore, a difference of 15% (8 spikes/s) was associated with a reversal in behavioral response to NaCl. By comparison, the mean response to 0.3 M NaCl among N-cells in the present study was 17% higher than the mean response to 0.1 M NaCl, and it declined 13% after the renin infusion. Thus renin brought the response to aversive 0.3 M NaCl nearly to the level evoked by appetitive 0.1 M NaCl in a naive (prerenin) rat.

The broad use of the term "appetitive," however, does not capture the difference between the casual acceptance a replete rat shows toward 0.1 M NaCl and the avid consumption of 0.3 M NaCl by one with a sodium appetite. If the reduction in response to 0.3 M NaCl among N-cells can help explain the tempering of its aversive character, we suggest that the increase to NaCl among S-cells may serve as one factor generating increased NaCl intake. Here, where the neural effect is weakest and most in need of interpretation, quantification of the neural-behavioral relationship is unavailable. Across all eight S-cells and over the full 5-s recording period, the response to 0.3 M NaCl increased 12%, from 48 to 54 spikes/s. However, we know of no NTS data that would provide a context in which to evaluate the impact on behavior toward stimuli that elicit such a difference in response among S-cells. Thus we are left with an effect on S-cells that is inconsistent and, lacking further analysis, appears disproportionately small to the appetitive behavior it may be helping to mediate, yet which offers a logical basis for such behavior.

It is likely that other gustatory areas alter their responding during sodium appetite, and the changes in NTS activity serve as one contributing component. Another factor that may have limited the size of the neural effects we observed is that we created our sodium appetite rapidly. Additional elements may come into play when the appetite is generated more gradually. For example, a decrease in the number of amiloride-sensitive channels on taste receptor cells, one clear basis for a reduced response to sodium in the NTS (38), has been reported after several weeks of restricted sodium intake in rats (22, 48). This mechanism would have had no impact in the present study where sodium appetite was created rapidly. This may also explain why decreased responding to hypotonic concentrations of NaCl has never been observed when the appetite was created in 48 h or less (2, 39, 43), whereas responses to hypo- and hypertonic concentrations have been reduced in studies where it was generated over 9 days or more (8, 23, 30, 46).

There was also a longer onset for significant neural effects than for the behavioral effects in most cases. However, this can be explained in part by the smaller sample size of neural groups, which would result in less statistical power. This discrepancy may also reflect the fact that the rat is sensitive to alterations of its own perceptions long before the neural changes that underlie those alterations achieve statistical significance. After all, a test of significance is a demand to show that an event is almost certainly out of the ordinary (<5% of the distribution). Changes in behavior do not require such unlikelihoods.

Gustatory Neuron Types

As noted above, although the majority of sugar-sensitive neurons gave an enhanced phasic response to NaCl after renin, one showed clear suppression, and another showed a large increase in tonic rather than phasic firing. Among salt-sensitive cells, those that were most responsive to NaCl before infusion were also most suppressed, in relative as well as absolute terms. This relationship could not have been detected previously because comparisons had been made across responses in separate groups of subjects. In some reports, however, the responses of individual neurons were presented (7, 8, 30, 46). As was the case here, it can be seen that the highest sodium responders in the replete group are never matched by responses of cells from deprived rats, implying that it is these cells that were disproportionately suppressed as the appetite was created.

Despite this variability within putative groups, the major point is that partitioning cells according to their response profiles before renin served to identify those that perform different roles in sodium appetite.

Possible Mechanisms

It is also unlikely that the response decrements that we observed represent mere instability of the electrophysiological preparations, because we verified that there was high stability during the preinfusion condition, as described in Response Reliability. However, it could be argued that the high-responding N-cells, which showed the largest effect of the infusion, may be particularly vulnerable to decreases in responding to NaCl over time or with repeated stimulus application. For this reason, we examined instances in these neurons where 0.1, 0.3, or 0.5 M NaCl was presented twice before infusion. There were nine cases where 0.1 M NaCl was presented twice, and one and three cases (respectively) for 0.3 and 0.5 M NaCl. The mean (± SE) response to these stimuli after the first presentation (127.1 ± 10.3 spikes/s) did not differ significantly from the mean response after the second preinfusion presentation (127.6 ± 11.7 spikes/s). Thus in the absence of renin infusion, these neurons were capable of maintaining their extraordinarily high-response rates to NaCl. In contrast, after infusion, a significant decline in 0.1 M NaCl responding was observed within 5 min in these neurons.

The rapidity of the effects places further constraints on possible mechanisms. For example, it is unlikely that new neural connections or taste receptor cells were generated. Rather, the fact that changes occurred almost immediately, within 5 min in some cases, suggests that the NTS can be directly influenced by activation of the central renin-angiotensin system.

Although renin was infused, sodium appetite presumably was caused by rapid formation of ANG II in cerebrospinal fluid (37). ANG II receptors are distributed throughout the brain, including in rostral NTS (18). It is possible therefore that the effects were due to a direct influence of angiotensin on the NTS cells whose activity was recorded. There is evidence, however, that the brain stem alone is not competent to manage the expression of sodium appetite and that interaction with the forebrain is necessary (14, 21).

Areas around the anterior third ventricle, such as the median preoptic area and the paraventricular nucleus (PVN) of the hypothalamus, contain ANG II receptors (28) and are crucial for angiotensin-induced sodium appetite (15). The preoptic area sends fibers to the PVN (49), which in turn projects to the gustatory region of NTS (45), allowing for an influence of these areas on gustatory processing. Thus the effects observed in this experiment may have been caused by the binding of ANG II in regions near the anterior third ventricle, which in turn influenced NTS cells. This influence, as mentioned above, was specific in that neurons with different response profiles were affected in opposite directions, and only responses to particular chemicals were altered.

Effects Of DOCA On NTS Responding

This effect may have occurred as a result of DOCA's binding to the cells from which we recorded, as the rostral NTS contains mineralocorticoid receptors (1). Alternatively, it may have resulted from influences of the central nucleus of the amygdala or bed nucleus of the stria terminalis, areas that send projections directly to gustatory NTS and whose neurons are sensitive to the effects of chronic DOCA treatment (25).

In conclusion, sodium appetite was associated with a change in gustatory sensitivity to NaCl in the NTS. The nature of the effect depended on a cell's response characteristics. Neurons that were sodium oriented, especially those that responded most briskly before the appetite, showed decreased sensitivity to higher concentrations of NaCl. Most sugar-oriented neurons, in contrast, exhibited increased phasic activity to NaCl. These changes depended on central action of the renin-angiotensin system, but not on sodium deficiency, and occurred in as little as 5 min.