We previously reported that noradrenergic (NA) neurons in the nucleus of the solitary tract (NST) are necessary for exogenous CCK octapeptide to inhibit food intake in rats. To determine whether NST NA neurons also are necessary for lithium chloride (LiCl) to inhibit food intake and/or to support conditioned avoidance behavior, saporin toxin conjugated to an antibody against dopamine beta hydroxylase (DSAP) was microinjected bilaterally into the NST to ablate resident NA neurons. DSAP and sham control rats subsequently were tested for the ability of LiCl (0.15M, 2% body wt) to inhibit food intake and to support conditioned flavor avoidance (CFA). LiCl-induced hypophagia was significantly blunted in DSAP rats, and those with the most extensive loss of NST NA neurons demonstrated the most attenuated LiCl-induced hypophagia. Conversely, LiCl supported a robust CFA that was of similar magnitude in sham control and DSAP rats, including rats with the most extensive NA lesions. A terminal c-Fos study revealed intact LiCl-induced c-Fos expression in the lateral parabrachial nucleus and central amygdala in DSAP rats, despite significant loss of NST NA neurons and attenuated c-Fos activation of corticotropin-releasing hormone-positive neurons in the paraventricular nucleus of the hypothalamus (PVN). Thus, NST NA neurons contribute significantly to LiCl-induced hypophagia and recruitment of stress-responsive PVN neurons but appear to be unnecessary for CFA learning and expression. These findings support the view that distinct central nervous system circuits underlie LiCl-induced inhibition of food intake and conditioned avoidance behavior in rats.
- nucleus of the solitary tract
- parabrachial nucleus
- saporin toxin
- sickness behavior
- hypothalamic-pituitary-adrenal axis
lithium chloride (licl) is commonly used in experiments to study central neural pathways and mechanisms involved in “sickness behavior.” Systemic administration of LiCl in rats dose-dependently inhibits food intake, activates the hypothalamic-pituitary-adrenal (HPA) stress axis, and promotes conditioned flavor avoidance (CFA) (2, 3, 6, 9, 10, 34, 41, 50, 51, 54). CFA represents a learned association between a novel flavor and an aversive postingestive event (48). Because toxic, rancid, and other potentially dangerous foods coexist with most animals, CFA is believed to have evolved to protect animals (especially nonemetic species such as rats) from poisoning themselves out of the gene pool (47). The ability of LiCl to inhibit food intake and to support robust conditioned avoidance of associated flavors is evidence that LiCl has aversive qualities and that it recruits central neural circuits that underlie both hypophagia and adaptive associative learning.
It is widely held that LiCl-induced hypophagia is a direct consequence of LiCl-induced malaise. However, relatively low doses of LiCl that do not inhibit ad libitum or deprivation-induced food intake can nevertheless support conditioned taste aversion and avoidance behavior in rats (2, 41), and an experimental dissociation between the ability of LiCl to produce CFA and to inhibit food intake has been reported (6). In the latter study, LiCl-induced CFA was abolished in rats with aspiration lesions of the chemosensitive area postrema (AP), although LiCl still promoted hypophagia and increased plasma levels of oxytocin (6). Thus, LiCl inhibits food intake and recruits hypothalamic endocrine responses in AP-lesioned rats that are either incapable of experiencing LiCl-induced malaise or are incapable of associating malaise with a novel flavor.
First-order central nervous system (CNS) regions that are activated to express the immediate-early gene product c-Fos after LiCl treatment include the AP, which lacks a blood-brain barrier, and the subjacent nucleus of the solitary tract (NST). Activated AP and NST neurons include noradrenergic (NA) neurons that are immunoreactive for dopamine beta hydroxylase (DbH), and non-NA neurons (30, 36). NA and non-NA neurons in the AP and NST relay LiCl-initiated signals to other brain regions involved in sickness behavior, CFA learning, and stress responses, including the lateral parabrachial nucleus (laPBN), the central nucleus of the amygdala (CeA), and the paraventricular nucleus of the hypothalamus (PVN). Conditioned taste aversion and CFA responses to LiCl in rats require an intact AP, PBN, and amygdala (3, 6, 13, 17, 31, 32, 40–42, 58). However, the requisite neural substrates for LiCl-induced hypophagia are less well defined.
A novel immunotoxin comprising an antibody to the NA synthetic enzyme DbH conjugated to saporin toxin (DSAP) destroys NA neurons selectively and site-specifically (1, 11, 12, 25, 35, 39, 56). A study from our laboratory used DSAP lesions to reveal that NA neurons in the caudal NST are necessary for exogenous CCK to inhibit food intake in rats, likely because of disruption of caudal brain stem circuits that mediate CCK-induced hypophagia (35). Caudal medullary NA neurons also are necessary for CCK to activate c-Fos expression in the PVN but are unnecessary to activate the laPBN and CeA, in which CCK-induced c-Fos expression appeared to be increased normally in rats with complete bilateral DSAP lesions of the caudal NST (35). The present study used a similar DSAP-lesioning approach to test the hypothesis that the neural substrates of LiCl-induced hypophagia require NA neurons within the caudal dorsomedial medulla, whereas the same NA neurons are unnecessary for the ability of LiCl to recruit AP and laPBN neural circuits that underlie CFA learning.
MATERIALS AND METHODS
All procedures conformed to the National Institutes of Health guidelines and were approved by the University of Pittsburgh Animal Care and Use Committee. Data from 43 adult male Sprague-Dawley rats (Harlan Laboratories, Indianapolis, IN) are included in this report. Singly housed rats had ad libitum access to water and pelleted rat chow (Purina, St. Louis, MO), except during food intake and CFA experiments, as detailed below. Colony rooms were maintained at 22–23°C and kept on a 12:12-h light-dark cycle, with lights on from 0700 to 1900.
Experimental Groups and Body Weights
Rats received bilateral injections of toxin (DSAP; Advanced Targeting Systems, San Diego, CA; n = 27) or vehicle (0.15 M NaCl; sham control; n = 16) into the caudal NST, as described below. Each rat was subsequently used in a CFA experiment, in a food intake experiment, and in a terminal c-Fos study. Thus, each rat received three LiCl treatments over the course of the study.
Rats weighed 187–212 g at the time of DSAP or sham surgery, 240–276 g at the time of CFA and food intake experiments, and 265–297 g at the time of the terminal c-Fos study. Body weights did not differ significantly at any experimental time point in DSAP vs. sham control rats. A previous report from our laboratory described a transient anorexia and BW loss that emerged in some rats ∼1 wk after bilateral DSAP injections into the caudal NST (35). The initial anorexia observed in that study resolved, and all rats gradually recovered normal body weight growth curves (35). With that result in mind, all rats in the present study were offered a palatable liquid diet (described in CFA experiment) in addition to their regular chow for 2 days before brain stem injection surgery and for 5–7 days after surgery. The lack of lesion-related body weight loss in the present study (see results) could be due to the dietary intervention and/or to the smaller number of DSAP injection sites (i.e., four per rat as described below, vs. six per rat in our previous study). A smaller number of DSAP injection sites were used in the present study in an effort to reduce toxin spread to NA neurons within the ventrolateral medulla.
DSAP or vehicle microinjections were made bilaterally into the NST at four sites in each rat, with two injection sites in the left NST and two in the right. DSAP was freshly prepared from a frozen stock solution within 2 h of injection. Injection sites encompassed the caudal medial NST region that contains the highest incidence of c-Fos expression by DbH-positive NA neurons after LiCl treatment, corresponding to the location of the A2 NA cell group (30, 37). Rats were anesthetized (2–5% halothane in 100% oxygen) and secured in a stereotaxic frame using blunt ear bars, with the nose ventroflexed. The skin over the dorsal neck surface was shaved, sterilized, and incised, and the neck muscles were retracted and bluntly dissected to expose the meninges overlying the dorsal surface of the caudal medulla. With the aid of a surgical microscope, the meninges were cut with a sterile needle to reveal the AP. A glass micropipette tip (outer diameter 50–75 μm) filled with DSAP or vehicle and affixed to a 1.0-μl Hamilton syringe was positioned on the midline at the caudal limit of the AP, then moved 0.25 mm lateral and 0.5 mm below the medullary surface for the first injection site. A second more rostral injection site was located 0.25 mm below the lateral border of the AP at its widest rostrocaudal extent (35). The two NST injection sites were then duplicated contralaterally. At each injection site, 50 nl of sterile 0.15 M NaCl vehicle containing 0 or 5 ng DSAP was delivered by manual pressure injection over ∼30 s. The pipette tip was left in place for 2 min after each microinjection, then withdrawn. The skin incision was closed with sutures after the final injection. Rats were returned to their home cages after recovery from anesthesia. In addition to their regular chow diet, a palatable liquid diet (Vanilla flavored Ensure, 25 ml/day; Abbott Laboratories, Abbott Park, IL) was offered for 2 days before and for 5–7 days after DSAP or sham injection surgery. All rats consumed both liquid and solid diets, and DSAP and sham control rats gained equivalent amounts of body weight during the postsurgical period.
Two to three weeks after NST microinjection surgery, DSAP and sham control rats were tested for the ability of LiCl to support CFA. A two-bottle choice paradigm (8) was used to determine whether rats avoid consuming water that contains flavors previously paired with LiCl treatment. Flavor exposure during CFA training and testing was conducted near the end of the light cycle of the photoperiod, between 1600 and 1800. Rats were acclimated for 3–4 days to gentle handling before the start of the CFA experiment.
Rats were water deprived for 22 h. Approximately half of the rats within each NST injection group (i.e., DSAP or sham) were then presented with a single bottle of almond-flavored tap water to drink from a graduated tube, and the others were presented with a single bottle of vanilla-flavored water (0.5% McCormick brand almond or vanilla extract). The left-right position of the drinking tube on each cage was switched after 15 min, with cumulative intake recorded at the 30-min time point. Thirty minutes after the end of this initial single-flavor exposure session, all rats were injected intraperitoneally with 0.15 M NaCl (2% body weight). Plain tap water was returned 30 min later, and rats had ad libitum water access for the next 24 h. Rats were then water deprived for 22 h, followed by presentation of a single bottle containing the alternate flavor to drink for 30 min, with the bottle position on each cage switched after 15 min. Thirty minutes after the end of this second single-bottle flavor exposure session, rats were injected intraperitoneally with 0.15 M LiCl (2% body weight). Plain water was returned 30 min later, and rats had ad libitum water access for 24–48 h.
For the final two-bottle choice test, rats were water deprived for 22 h, and then given 30-min simultaneous access to two bottles of water, one containing the saline-paired flavor (either almond or vanilla) and the other containing the alternate LiCl-paired flavor. The volume consumed by each rat from each bottle was recorded after 30 min of access, with bottle positions switched at the 15-min time point. Rats then were returned to ad libitum water access.
Flavor preference ratios displayed by each rat during the two-bottle choice test were averaged within each treatment group to obtain group preference ratios (means ± SE) for intake of saline-paired flavors relative to LiCl-paired flavors. Outcomes indicating significantly shifted preference ratios (e.g., 75%:25%) were interpreted as evidence for conditioned avoidance of the flavor represented by the lower value in the ratio. Paired t-tests were used to determine whether differences in preference for saline-paired vs. LiCl-paired flavors were statistically significant, with significance set at P < 0.05.
Food intake experiment.
One week after completing the CFA experiment, DSAP and sham control rats were tested for the ability of LiCl treatment to inhibit food intake. For this purpose, rats were moved from their original colony room to a new environment with the same 12:12-h light-dark cycle. Rats were housed individually in clear Plexiglas boxes (25 cm × 30 cm floor, 22 cm height) with stainless-steel rod floors, each equipped with a computer-driven pellet delivery and monitoring system (Med Associates, St. Albans, NY). Drinking water was available ad libitum from sipper tubes within each box. Food access was restricted to a daily 3-h period beginning at lights out (1900), when a single 45-mg chow pellet (Precision Dustless Pellets; Bio-Serv, Frenchtown, NJ) was delivered automatically to a shallow feeding trough in each cage, thereby breaking a laser photobeam crossing the base of the trough. A new pellet was delivered automatically each time the preceding pellet was removed. The cumulative number of pellet deliveries (i.e., photobeam breaks) was recorded automatically at 30-min intervals during the 3-h feeding period. Cage trays beneath the open-rod floors were routinely inspected at the end of the feeding period to confirm that delivered pellets were actually consumed. Data were collected and stored using Med PC software (Med Associates). Rats were acclimated to the new environment and feeding schedule for 1 wk before testing, by which time stable daily 3-h food intakes of ∼20–22 g per rat (i.e., ∼7.5–8.5% body weight) were achieved.
On testing day 1, rats were injected intraperitoneally with 0.15 M NaCl (2% body wt) at 1830. Food access was initiated 30 min later, at lights out (1900). Cumulative pellet delivery data were collected for 3 h. On testing day 2 (24 h later), rats were injected intraperitoneally with 0.15 M LiCl (2% body wt) at 1830. Food access was initiated 30 min later, at lights out (1900). Cumulative pellet delivery data were collected for 3 h. Thus, each DSAP or sham control rat served as its own control for the effects of intraperitoneal saline vs. intraperitoneal LiCl on subsequent food intake. Rats were returned to their original home cages with ad libitum chow access after testing day 2.
Food intake data (i.e., the number of 45-mg pellets consumed over 3 h) were combined a priori according to NST injection group (DSAP vs. sham control) and expressed as group means ± SE at each 30-min time point. Group- and treatment-related differences in food intake were tested for statistical significance by using ANOVA, with NST injection (DSAP vs. sham), intraperitoneal treatment (NaCl or LiCl), and time (30-min intervals) as independent variables. When F values indicated significant main effects and interactions among these variables on food intake, ANOVAs were followed up with post hoc t-tests using Dunn's (Bonferroni) correction for multiple comparisons. Differences were considered significant when P < 0.05. Feeding data also were analyzed by correlating the magnitude of LiCl-induced feeding suppression at the final 3-h time point with DSAP lesion extent, defined by the number of DbH-positive NST neurons (see results).
LiCl-induced c-Fos experiment.
A terminal c-Fos study was performed 7–10 days after the end of the food intake experiment. Previous work demonstrated that little or no c-Fos labeling is present within the AP, NST, PBN, CeA, or PVN in either DSAP or sham control rats perfused after control intraperitoneal injections of 0.15 M NaCl (35). Thus, all rats in the present study (n = 27 DSAP, n = 16 sham control) were injected intraperitoneally with 0.15 M LiCl (2% body wt) before perfusion. LiCl injections were made between 0900 and 1000. Rats were returned to their home cages immediately after intraperitoneal injection and left undisturbed for 90–120 min. Rats then were anesthetized with pentobarbital sodium (Nembutal; 50 mg/kg ip) and perfused through the heart with 50–100 ml of 0.15 M NaCl followed by 500 ml of fixative solution (26) containing 4% paraformaldehyde, 1.4% lysine, and 0.2% sodium metaperiodate in 0.1 M sodium phosphate buffer (hereafter, buffer). Fixed brains were removed from the skull, postfixed overnight, then cryoprotected for 24–48 h in 20% sucrose. A freezing-stage microtome was used to cut coronal sections with a thickness of 35 μm from the caudal extent of the NST through the rostral extent of the corpus callosum. Sections were collected in six serially adjacent sets and stored at 20°C in cryopreservant solution (55).
Primary and secondary antisera were diluted in buffer containing 0.3% Triton-X100 and 1% normal donkey serum. Dual immunoperoxidase localization of cytoplasmic DbH and nuclear c-Fos labeling was performed to simultaneously assess the extent of NST lesions (i.e., the loss of DbH labeling) and to determine LiCl-induced neuronal activation (i.e., c-Fos labeling). For this purpose, one set of tissue sections (1:6 frequency) from each DSAP and sham control rat was incubated for 48 h at 4°C in rabbit anti-c-Fos antiserum (provided by P. J. Larsen and J. D. Mikkelsen, Denmark; 1:50,000). The specificity of this antibody for c-Fos protein has been reported (38). Sections were rinsed and processed for blue/black immunoperoxidase localization of nuclear c-Fos labeling using biotinylated donkey anti-rabbit IgG (1:500; Jackson ImmunoResearch Laboratories, West Grove, PA), Elite Vectastain avidin-biotin reagents (Vector Laboratories, Burlingame, CA) and a nickel-enhanced diaminobenzidene (DAB) reaction. After c-Fos immunostaining, sections were incubated for 48 h at 4°C in a monoclonal mouse anti-DbH antibody (1:50,000; Chemicon, Temecula, CA) and processed for brown immunoperoxidase localization of cytoplasmic DbH using biotinylated donkey anti-mouse IgG (1:500), Elite Vectastain reagents, and a nonenhanced DAB reaction. Additional adjacent sets of tissue sections from each rat were processed similarly for dual immunoperoxidase localization of c-Fos and CGRP (1:50,000; Peninsula Laboratories, San Carlos, CA), or for c-Fos and corticotropin-releasing hormone (CRH; 1:20,000; Peninsula Laboratories). CGRP provides a distinct neurochemical marker for the viscerosensory laPBN-to-CeA projection pathway (22, 59), and in our previous study, the CGRP-positive pathway was activated normally in DSAP- lesioned rats after CCK treatment (35). CRH immunostaining identifies parvocellular PVN neurons that comprise the apex of the HPA stress axis, and in our previous study, CCK-induced activation of CRH neurons was markedly attenuated in DSAP-lesioned rats after CCK treatment (35). Immunoreacted sections were mounted out of buffer onto charged glass microscope slides (SuperFrost Plus; Fisher Scientific, San Jose, CA), dehydrated in a graded ethanol series, defatted in xylene, and coverslipped with Histomount (VWR, Pittsburgh, PA).
Quantitative analysis of immunolabeling.
Cell counting was performed to document DSAP lesion extent and LiCl-induced c-Fos expression in each animal. Counts of DbH-positive neurons in the NST and VLM were performed bilaterally in every sixth section (210-μm frequency) through the rostrocaudal extent of the AP (3 or 4 tissue sections per rat). Cells were counted using a 20× microscope objective on a Zeiss Axioplan 2 microscope. All DbH-immunopositive profiles that were clearly perikaryal (rather than dendritic or axonal) were counted, regardless of whether their nucleus was visible in the section. This cell-counting strategy was adopted to avoid undercounting of DbH-positive neurons that were c-Fos-negative, because c-Fos-positive nuclei are more readily visible than are unstained nuclei. Perikaryal DbH-positive profiles were ovoid shaped and at least as large in diameter as representative c-Fos-labeled nuclei within the section. LiCl-induced neural activation within the AP and NST was quantified by counting c-Fos-immunopositive profiles within each region bilaterally in the same tissue sections. The proportion of DbH-positive NST and VLM cells that were colabeled for nuclear c-Fos was determined. Cells were considered c-Fos-positive if their nucleus contained detectable blue/black immunolabeling, regardless of labeling intensity, and c-Fos-negative if they displayed no visible nucleus or a nucleus lacking c-Fos labeling. Thus, the reported proportion of DbH-positive neurons that were double-labeled for c-Fos is likely an underestimation of the actual proportion. However, the magnitude of this potential underestimation should not differ between experimental treatment groups.
c-Fos labeling in the visceral sensory region of the pontine PBN was quantified by determining the number of CGRP-positive laPBN neurons that were activated to express c-Fos, using counting criteria similar to those described above. Tissue sections were examined bilaterally through the rostrocaudal extent of cellular CGRP labeling in the laPBN (3 or 4 sections spaced by 210 μm per rat). c-Fos activation in the hypothalamus was quantified by determining the number of CRH-immunopositive neurons within the medial parvocellular PVN (mpPVN) that contained nuclear c-Fos labeling. For each rat, two tissue sections (spaced by 210 μm) through the mpPVN that contained the highest density of CRH immunolabeling were selected for analysis. c-Fos activation in the amygdala was determined using CGRP fiber immunolabeling to distinguish the cytoarchitectural boundaries of the CeA. All c-Fos-positive profiles within those CeA boundaries were counted in tissue sections through the rostrocaudal extent of fibrous CGRP labeling (5–7 sections spaced by 210 μm in each rat).
In each rat, regional cell count data were expressed as the number of cells counted within each specified brain region summed bilaterally, divided by the number of analyzed tissue sections through the specified region, to obtain “average per section” values. Group data were combined a priori by NST injection group (i.e., DSAP vs. sham control) and expressed as mean counts ± SE for each brain region. Between-group differences in cell count values within each region were tested for statistical significance by using separate t-tests, with NST injection condition (i.e., DSAP or vehicle) as the independent variable. Differences were considered significant when P < 0.05.
DSAP toxin injections eliminated a significant proportion of DbH-positive NST neurons. Quantitative analysis revealed an average loss of ∼49% of the NST NA neuronal population in DSAP rats relative to sham controls (Table 1). However, DSAP lesion extent was variable across animals. Lesion data are detailed in DSAP-Induced Loss of DbH-Immunopositive NA Neurons.
DSAP Lesions Do not Alter the Ability of LiCl to Support CFA
Rats in both surgical groups consumed similar volumes of novel almond- or vanilla-flavored water during initial presentation on single-bottle CFA training days (i.e., 13.5 ml ± 1.2 consumed on a saline-paired day, 13.3 ml ± 1.4 consumed on LiCl-paired day; range 11–18 ml; paired t-test P > 0.05). There were no significant effects of lesion group (DSAP vs. sham) or flavor pairing order (vanilla + intraperitoneal saline followed by almond + intraperitoneal LiCl, or vice versa) on cumulative 30-min fluid intake during training. Thus, vanilla- and almond-flavored waters were isopreferred by DSAP and sham control rats, and surgical group had no effect on the volumes consumed of either flavor after water deprivation.
In two-bottle choice tests, DSAP and sham control rats drank significantly lower volumes of flavors that previously were paired with intraperitoneal LiCl (DSAP 2.7 ml ± 0.3; sham 3.2 ml ± 0.4) vs. their intake of intraperitoneal saline-paired flavors (DSAP 15.3 ml ± 0.8; sham 14.1 ml ± 0.9) (P < 0.01 for each group when comparing intakes of LiCl- vs. saline-paired flavors in paired t-tests) (Fig. 1). The resulting flavor preference ratios for saline-paired vs. LiCl-paired flavors were ∼80%:20% in sham control rats and ∼85%:15% in DSAP rats (Fig. 1). Thus, DSAP and sham control rats displayed similarly robust CFA responses to flavors previously paired with LiCl treatment.
DSAP Lesions Attenuate LiCl-Induced Hypophagia
DSAP and sham control rats consumed similar numbers of food pellets over the 3-h monitoring period after intraperitoneal injection of 0.15 M NaCl (Fig. 2, upper dashed lines) [F(1,41) = 0.147, P > 0.05]. As expected, intraperitoneal injection of LiCl in sham control rats markedly suppressed their food intake vs. intake by the same rats after intraperitoneal NaCl (Fig. 2) [F(1,15) = 71.46, P < 0.05]. The hypophagic effect of LiCl treatment in sham control rats was maintained throughout the 3-h postinjection monitoring period, ranging from ∼70% inhibition of feeding at the 30-min time point to ∼45% inhibition at the final 3-h time point (Fig. 2). The rate of eating (as evidenced by the slope of the lines connecting time points; Fig. 2) was suppressed to the greatest extent during the first 30 min of feeding in both sham and DSAP rats after LiCl compared with feeding after saline injections. Between the 30- and 60-min time points, DSAP rats ate at a somewhat higher rate than sham rats, which produced the statistically significant group difference in food intake evident at the 60 min and later time points (Fig. 2). The ability of LiCl to inhibit food intake in DSAP rats was significantly attenuated at most time points (Fig. 2, asterisks), although DSAP rats as a group still exhibited LiCl-induced hypophagia across the 3-h monitoring period [F(1,26) = 53.39, P < 0.05]. However, within the 3-h monitoring period, neither group compensated for the reduced food intake that was most evident during the initial 30–60 min of the 3-h monitoring period.
The suppressive effect of LiCl on cumulative food intake at the final 3-h time point was more markedly attenuated in rats with more complete DSAP lesions (Fig. 3; lesion data described further, below). A significant correlation existed between lesion extent and the magnitude of LiCl-induced hypophagia at this time point in DSAP rats (Pearson r = −0.531; P < 0.05), such that larger lesions were associated with less feeding suppression after LiCl compared with intake after saline injections. Conversely, grouping animals by DSAP lesion extent did not reveal any differences among groups in results from the two-bottle choice test that demonstrated similarly robust CFA responses to LiCl (see Fig. 3).
DSAP-Induced Loss of DbH-Immunopositive NA Neurons
Bilateral toxin injections were associated with an average 49% reduction of DbH-positive NST neurons in DSAP rats compared with sham controls (Table 1, Fig. 4, A and B). The loss of DbH-positive NST neurons in DSAP rats was statistically significant (t-test, P < 0.01). Marked loss of DbH-positive neurons within the AP also was observed in DSAP rats (Fig. 4, A and B), although this was not quantified. Loss of DbH-positive NST neurons in individual DSAP rats ranged from 3% to 84% relative to average counts in sham controls. Thus, the overall effect of bilateral DSAP toxin injections to eliminate DbH-positive NST neurons was significant but was variable across animals within the DSAP group.
Compared with results from our previous study (35), in which six toxin injection sites were made in each animal, DSAP lesion extent in the present study was more variable (i.e., 3–84% loss of DbH-positive NST neurons), and we obtained a larger proportion of DSAP rats with NA lesions that were less than 50% complete. However, DSAP lesions in the present study were more anatomically restricted than in our previous report, with no significant loss of A1/C1 neurons in the caudal VLM (Table 1).
LiCl-Induced c-Fos Expression in DSAP and Sham Control Rats
The average numbers of c-Fos-positive profiles counted per tissue section within the AP and NST were similar in DSAP and sham control rats after LiCl treatment (Table 2, Fig. 4, C and D), evidence that DSAP lesions eliminated a relatively small subset of AP and NST neurons activated by LiCl. By extension, DbH-positive NA neurons represent a relatively small subset of the AP and NST neurons activated by LiCl. Despite the average 49% loss of DbH-positive NST neurons in DSAP rats (Table 1), the proportion of remaining DbH-positive NST neurons activated in DSAP rats after LiCl treatment (i.e., 70.5%) was statistically similar to the proportion activated in sham controls (i.e., 64.4%; Table 2). DbH-positive NA neurons within the caudal VLM also were similarly activated after LiCl treatment in DSAP and sham control rats (i.e., 34.8% and 38.0%, respectively; Table 2).
DSAP lesions were associated with reduced DbH immunolabeling throughout magnocellular and parvocellular subdivisions of the PVN (Fig. 5, A and B). DSAP lesions also were associated with reduced PVN c-Fos expression after LiCl treatment, including a significant attenuation of c-Fos expression by CRH-immunopositive neurons within the mpPVN (Table 2, Fig. 5, C and D). LiCl activated 44.1% fewer CRH-positive neurons in DSAP rats compared with CRH activation in sham controls (Table 2; P < 0.05). Conversely, LiCl-induced c-Fos expression within the lateral magnocellular PVN appeared similarly robust in DSAP and sham control rats (Fig. 5, C and D). A significant correlation existed between NA lesion extent in DSAP rats and LiCl-induced activation of CRH-positive neurons (Pearson r = −0.404; P < 0.01), such that rats with larger lesions exhibited less CRH neural activation. LiCl-induced CRH activation in DSAP rats also was positively correlated with the degree of LiCl-induced hypophagia assessed at the 3-h time point (Pearson r = 0.407; P < 0.05).
Pons and amygdala.
In contrast to the reduced activation of c-Fos labeling in CRH-positive mpPVN neurons in DSAP rats, c-Fos expression within the laPBN (Fig. 6, A and B) and CeA (Fig. 6, C and D) was not attenuated in DSAP rats. Quantitative analysis confirmed that similar numbers of CGRP-positive laPBN neurons and similar numbers of cells within the CGRP-rich region of the CeA were activated to express c-Fos in DSAP and sham control rats after LiCl treatment (Table 2, Fig. 6).
Results from this study support the view that NA neurons in the caudal dorsomedial medulla contribute to LiCl-induced hypophagia and also are important for conveying LiCl-related interoceptive signals to CRH-positive neurons in the mpPVN. Conversely, DSAP lesions of this NA cell population did not affect LiCl-induced activation of neurons within the laPBN or CeA and did not reduce the ability of LiCl to support CFA. These results are striking, as they illustrate a unique experimental situation in which this “nauseogenic,” malaise-inducing agent maintains its ability to support robust conditioned avoidance in animals in which hypophagic responses to the same agent are significantly attenuated.
An earlier study reported that aspiration lesions of the AP block the ability of LiCl to support CFA without significantly affecting LiCl-induced hypophagia (6). That result is perhaps easier to reconcile with the knowledge that many natural conditions and experimental treatments inhibit food intake without also producing conditioned avoidance behavior, presumably because those situations and treatments are not aversive. On the other hand, conditioned avoidance also can occur in the absence of demonstrable hyophagia. Certain doses of LiCl that are subthreshold for inhibiting food intake in rats have been reported to support comparatively mild but still significant and dose-related conditioned avoidance (2, 9). Conversely, we are unaware of another experimental condition in which a robust CFA induced by a relatively high dose of LiCl is maintained without a similarly robust unconditioned hypophagic response, as occurred in rats with the most complete DSAP lesions.
DSAP Lesion Efficacy and Specificity
Bilateral NST microinjections of DSAP were designed to destroy NA neurons in the caudal medial NST, corresponding to the location of the A2 cell group and the caudal portion of the C2 cell group. Results from previous studies attest to the efficacy and specificity of the conjugated DSAP toxin (1, 11, 12, 25, 35, 39, 56). In the present study, toxin-induced loss of DbH cell and terminal immunostaining within the caudal dorsomedial medulla (both AP and NST) and loss of DbH terminal labeling in the PVN offer support for lesion efficacy. Despite the loss of DbH-positive neurons, similar numbers of c-Fos-positive NST and AP neurons were counted in DSAP and sham rats after LiCl treatment, supporting the view that NA neurons comprise a relatively small (albeit important) subset of the total population of NST and AP neurons activated by LiCl.
NA Cell Loss, Hypophagia, and CFA
As in our previous DSAP study examining the effect of NA lesions on CCK-induced hypophagia (35), a positive correlation was found in the present study between the number of surviving DbH-positive NA neurons in the caudal NST and the extent of LiCl-induced hypophagia. As shown in Fig. 3, the more effective the DSAP lesion, the more blunted the ability of LiCl to inhibit food intake. Rats with the most complete DSAP lesions (i.e., more than 75% loss of NA neurons) displayed significantly less LiCl-induced hypophagia than sham control rats or DSAP rats with less complete lesions. Thus, not only are NA neurons in the caudal dorsomedial medulla activated by LiCl and CCK, they underlie at least a significant portion of the ability of these agents to inhibit food intake. It should be noted that DSAP lesions in the present study were never 100% complete; thus, it remains unclear whether NA neurons are absolutely necessary for treatment-induced hypophagia. We cannot rule out the possibility that more extensive NA lesions would disrupt LiCl-induced CFA and/or c-Fos responses in the laPBN and CeA. Considered together, however, these findings support the view that NA neurons in the NST contribute importantly to brain stem circuits that suppress food intake. These circuits likely include known NA inputs to regions of the medullary and pontine reticular formation (7) that contain the premotor and motor neurons that ultimately control ingestive behavior (49).
DSAP Lesions Reduce PVN Activation After LiCl
In addition to inhibiting food intake in rats, systemic LiCl activates the HPA stress axis and also increases plasma levels of oxytocin, and, to a lesser degree, vasopressin (6, 10, 50, 51, 54). Plasma hormone levels were not assayed in the present study, but the significantly blunted c-Fos expression by CRH-positive mpPVN neurons in DSAP rats predicts a similarly blunted HPA response to LiCl treatment. Indeed, a close association between treatment-induced c-Fos activation in CRH neurons and plasma corticosterone levels has been reported (1). Our finding of reduced c-Fos expression in mpPVN CRH-positive neurons in DSAP rats after LiCl treatment is consistent with a previous report of reduced mpPVN activation in DSAP rats after CCK treatment (35). Conversely, despite significant loss of DbH-immunoreactive fibers throughout the PVN in DSAP rats, LiCl-induced c-Fos activation in the lateral magnocellular PVN appeared to be intact, presumably due to remaining inputs from neurons other than dorsomedial medullary NA neurons. Likely candidates include NA neurons in the caudal VLM, which were not lesioned in DSAP rats and which project directly to the vasopressin-rich core of the lateral magnocellular PVN (4, 5, 44, 52).
DSAP Lesions Do not Attenuate laPBN or CeA Activation After LiCl
In contrast to the decremented mpPVN activation in DSAP rats after LiCl, activation of the CRGP-containing pathway from laPBN to CeA (59) was quantitatively unaffected. We interpret this dissociation as evidence that NA neurons in the caudal dorsomedial medulla are unnecessary for conveying LiCl-related viscerosensory signals to the CeA, probably because these signals are relayed through CGRP neurons in the laPBN that innervate the CeA. Direct NST projections to the amygdala include a prominent NA component that is activated by LiCl and other experimental treatments that inhibit food intake and promote conditioned avoidance behavior (28–30). However, the laPBN is a major relay for ascending viscerosensory inputs from the caudal medulla to the amygdala (15, 18, 33, 43, 53). NST and AP inputs to the laPBN are predominantly excitatory, and arise from both NA and non-NA neurons (15, 16, 19). Subpopulations of neurons within the NST and laPBN project differentially to brain stem and forebrain target regions, suggesting that different types of viscerosensory signals may be carried through different projection pathways. For example, systemic administration of endotoxin in rats activates 10 times as many CeA-projecting neurons in the laPBN than CeA-projecting neurons in the NST (53). Neurons within the laPBN, including CGRP-positive neurons, project rather weakly to the PVN and lateral hypothalamus but strongly to the CeA and bed nucleus of the stria terminalis (BNST) in rats and in humans (21, 22, 45, 59). As another example, NA inputs to the mpPVN arise from caudal medullary NST and VLM neurons that also provide collateralized axonal inputs to the BNST, whereas NA inputs to the lateral magnocellular PVN arise from a separate population of hindbrain NA neurons (1).
Although DbH-positive AP and NST neurons that project to the laPBN and amygdala undoubtedly were lesioned (to varying degrees) in DSAP rats, non-NA neurons that project to these regions presumably were spared. The latter include projections arising from serotoninergic AP neurons that target the laPBN (23) and from glucagon-like peptide-1 (GLP-1)-positive NST neurons that target the laPBN and CeA (24, 27) and are activated in intact rats after LiCl treatment (36). We did not examine LiCl-induced activation of these neurons in the present study. However, in our previous DSAP study, GLP-1 neurons remained intact and were activated to express c-Fos after CCK treatment in rats with more extensive dorsal medullary NA lesions (35). Thus, GLP-1 signaling pathways may contribute to the observed maintenance of LiCl-induced CFA in DSAP rats. Indeed, infusion of GLP-1 into the amygdala supports conditioned avoidance behavior in rats, and blockade of amygdalar GLP-1 receptors attenuates the ability of systemic LiCl to produce conditioned avoidance (20). Conversely, infusion of GLP-1 directly into the amygdala does not inhibit food intake (20). If GLP-1 pathways are sufficient to maintain CFA in rats with DSAP lesions, it is still unclear why they are not sufficient to maintain LiCl-induced hypophagia. Two laboratories have reported that global blockade of central GLP-1 receptors attenuates LiCl-induced hypophagia (34, 46), but in neither study was the hypophagia eliminated. It should be noted that the hypophagic effect of systemic endotoxin is significantly attenuated by pharmacological antagonism of hindbrain but not forebrain GLP-1 receptors (14), that hindbrain NA neurons express GLP-1 receptors, and that central infusion of GLP-1 activates these NA neurons (57). Thus, central GLP-1 signaling pathways may participate in hypophagic responses to LiCl because they recruit hindbrain NA neurons. It would be interesting to determine whether the ability of central GLP-1 to inhibit food intake is attenuated in rats with DSAP lesions of these neurons.
We conclude that NA neurons within the dorsomedial medulla (AP and/or NST) contribute importantly to LiCl-induced hypophagia and also to recruitment of stress-responsive mpPVN neurons but are unnecessary for the ability of LiCl to support robust CFA learning and expression. Our findings support the view that the behavioral effects of LiCl are mediated by CNS circuits that are at least partially dissociable.
This work was supported by National Institutes of Health Grant MH-59911.
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 © 2007 the American Physiological Society