Hypothalamic signaling in anorexia induced by indispensable amino acid deficiency

Experimental Animals and Diets

Male Sprague Dawley rats (7–8 wk old) were purchased from Charles River Laboratories (Wilmington, MA). Male and female melanocortin receptor 4 (MC4R) knockout (MC4R-KO) and proopiomelanocortin (POMC)-enhanced green fluorescent protein (eGFP) mice (8–10 wk old) were generated as previously described (10, 23). Age-matched C57BL/6J wild-type (WT) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Rats and mice were individually housed in rooms with controlled temperature (22 ± 2°C) and illumination (12:12-h light-dark cycle). All animals were allowed ad libitum access to food and water unless otherwise stated and were allowed to acclimate for at least 3 days before being used in procedures. All studies were conducted according to the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee of Oregon Health & Science University. No mortality was observed in any of these studies.

All animals were initially fed standard laboratory rodent chow (Rodent Diet 5001; Purina Mills, St. Louis, MO). Experimental diets were obtained from Harlan Teklad Laboratories (Indianapolis, IN). The control diet (CD, catalog no. TD. 99366) was balanced in amino acids and contained 15.4% protein, 65.0% carbohydrate, and 8.0% fat by weight. The valine-deficient diet (VDD, catalog no. TD. 08578), an isonitrogenous modification of the CD, was deficient in valine and enriched in alanine, aspartic acid, glutamic acid, and glycine but was otherwise identical in nutrient composition to the CD.

One day before each experiment commenced, the animals were weighed and divided into treatment groups such that the mean body weights of each group were similar. During the experiments, food intake and body weight were measured at the same time each day, unless otherwise noted.

Short- and Long-Term Feeding Studies in Rats

All rats were switched from standard chow to the CD for at least 3 days. Rats that continued to gain weight on the CD were divided into the following dietary groups: CD, VDD, or PF (n = 5–6/group). PF rats were given the same amount of the CD as the mean amount of VDD consumed on the previous day by the VDD group. Body weight and food intake were measured daily for 3 (short-term study) or 21 (long-term study) consecutive days. In the long-term study, the VDD rats began to shred their food after 3 days on the VDD. Thereafter, the shredded food was carefully collected, weighed, and then replaced by the same weight of fresh food pellets each day.

Long-Term Feeding Study in Mice

Male WT and MC4R-KO mice were fed the CD for at least 3 days and then were divided into the following treatment groups: WT/CD, WT/VDD, WT/PF, MC4R-KO/CD, MC4R-KO/VDD, and MC4R-KO/PF (n = 7–8/group). Mice in the PF groups were given the same amount of the CD as the mean amount of VDD consumed on the previous day by the VDD mice of the same genotype. Body weight and food intake were measured for 14 consecutive days. Damaged food pellets were first seen in the VDD groups' cages after 6 days of consuming the VDD. For the remainder of the study, the food pellets were replaced each day as described above.

Acute Nocturnal Mouse Feeding Studies

Male WT and MC4R-KO mice were fed the CD for 10 days before initiating study diets. The mice were then divided into the following four groups: WT/CD (n = 15), WT/VDD (n = 24), MC4R-KO/CD (n = 14), and MC4R-KO/VDD (n = 21). Two hours before the onset of the dark phase on the day of the experiment, all animals were placed in clean individual cages without food. Just before lights off, preweighed CD or VDD pellets were placed in each cage. For the next 12 h, food weights were measured hourly under red light illumination. Care was taken to minimize stress to the animals during the frequent food measurements.

In a separate study, female WT (n = 8) and MC4R-KO (n = 7) mice were fed the CD for 10 days. On the day of the experiment, both groups of mice were switched to the VDD. Food intake was recorded hourly throughout the 12-h dark phase using the same procedure described above. We observed identical results in female mice as we observed in male mice (data not shown).

Melanocortin Receptor Antagonism Study

Body Composition

Body composition (fat and lean body mass) was measured using magnetic resonance relaxometry (EchoMRI 4-in-1 Live Animal Composition Analyzer; Echo Medical System, Houston, TX).

Tissue Collection

At the end of each long-term experiment, the animals were killed by decapitation under isoflurane anesthesia. Blood was collected for hormone analysis via cardiac puncture or decapitation into EDTA blood collection tubes, and protease inhibitor (Complete, Roche, Germany) was added according to the manufacturer's instructions. Plasma was isolated and stored at −80°C until used. Brains were immediately removed, and hypothalamic blocks were dissected as described previously (37). Interscapular BAT and epididymal white adipose tissue (WAT) were also isolated. Hypothalami, WAT, and BAT were stabilized in RNAlater at 4°C for subsequent RNA extraction.

RNA Preparation and Quantitative PCR

Total RNA preparation and real-time PCR were performed as described previously (37). RNA was extracted using RNeasy kits (Qiagen, Valencia, CA). Reverse transcription and quantitative PCR reagents were obtained from Life Technologies (Carlsbad, CA). 18S or GAPDH RNA were used as endogenous controls. Gene expression is reported as fold change relative to the CD group using the 2^−ΔΔC_t method. Statistical analyses were performed on the ΔC_t values.

Plasma Leptin, Insulin, Insulin-Like Growth Factor-I, Corticosterone, and Ghrelin Quantification

Concentrations of plasma leptin and insulin-like growth factor-I (IGF-I) in mice and rats were determined using enzyme-linked immunosorbent assay (ELISA) kits [mouse leptin and rodent IGF-I ELISA kits (R&D Systems, Minneapolis, MN) and rat leptin ELISA kit (Millipore, Billerica, MA)]. Concentrations of plasma insulin in mice and rats (Millipore) and plasma corticosterone (MP Biomedicals, Santa Ana, CA) and acyl-ghrelin (Millipore) in rats were quantified with appropriate RIA kits. All assays were performed according to the manufacturers' instructions.

Anterograde Tract-Tracing Study

Male rats (280–320 g; n = 10) were used in this study. Animals received a single iontophoretic injection of phaseolus vulgaris leucoagglutinin (PHA-L; Vector Laboratories, Burlingame, CA) within the APC. Each animal was anesthetized with 4% isoflurane, and a glass micropipette (tip diameter = 10–15 μm) filled with 2.5% PHA-L solution was stereotaxically positioned into the APC. The coordinates used were as follows: x = ± 3.7, y = 2.4, z = −6.3; x = ± 3.2, y = 2.7, z = −6.8; or x = ± 3.6, y = 3.0, z = 5.8 (mm, x = lateral to midline, y = rostral to bregma, z = below surface of skull). A +4.5-μA current pulsed at 7-s intervals was applied for 20 min using an iontophoresis pump (BAB-501; Kation Scientific, Minneapolis, MN). After a survival time of 14–16 days, animals were transcardially perfused with 4% paraformaldehyde in 0.01 M PBS, and brains were processed for immunohistochemistry (IHC). Six series of 30-μm coronal sections were cut and collected from the APC through the hypothalamus. Sections were processed for IHC with anti-PHA-L primary antibody (1:1,000; Vector Laboratories) and secondary antibody conjugated with Alexa Fluor 594 or 488 (1:500; Invitrogen, Carlsbad, CA) for visualization. One complete series of sections from the APC through the hypothalamus was stained for confirmation of injection sites and transporting fibers. Only animals with injection sites centered within the morphological APC borders as well as visible PHA-L-labeled fibers in sections were included in the analysis. To observe connections between PHA-L-labeled fibers and CRH-, somatostatin (SS)-, melanin-concentrating hormone (MCH)-, or orexin-immunoreactive neurons in the hypothalamus, a series of hypothalamic sections was processed for double- or triple-label IHC. Floating sections were incubated for 48 h with primary antibodies against PHA-L (1:1,000), CRH (1:4,000; Bachem, San Carlos, CA), SS (1:100; GeneTex, Irvine, CA), MCH (1:1,500; Santa Cruz Biotechnology), and orexin (1:1,500; Santa Cruz Biotechnology), followed by incubation in secondary antibodies conjugated with Alexa Fluor 488, 594, and 647 for visualization. Sections were viewed under a fluorescence microscope, and a rat brain atlas (32) was used to verify cytoarchitecture. Confocal photomicrographs were taken using a Zeiss LSM710 microscope (Carl Zeiss, Oberkochen, Germany) under identical microscope settings. Images were then processed using Imaris software (Imaris x64 7.3.1; Bitplane).

In Situ Hybridization

After receiving either CD, VDD, or being PF for 5 days, rats were killed, and brains were removed, snap-frozen, stored at −80°C, and processed for in situ hybridization as previously described (20), with the following modifications. Hypothalamic coronal sections (20 μm) were cut on a cryostat, collected in a 1:6 series, and stored at −80°C. Antisense ³³P-labeled rat CRH riboprobe (corresponding to bases 344–751 of rat CRH; GenBank accession no. NM_031019.1; 0.15 pmol/ml), SS riboprobe (corresponding to bases 14–447 of rat SS; GenBank accession no. NM_012659; 0.01 pmol/ml), or MCH riboprobe (corresponding to bases 122–449 of rat promelanin-concentrating hormone; GenBank accession no. NM_012625.1; 0.005 pmol/ml) were denatured, dissolved in hybridization buffer containing tRNA (1.7 mg/ml), and applied to slides. Slides were covered with glass cover slips, placed in a humid chamber, and incubated overnight at 55°C. The following day, slides were treated with RNase A and washed under conditions of increasing stringency. Slides were dipped in 100% ethanol, air-dried, and then dipped in NTB-2 liquid emulsion (Eastman Kodak, Rochester, NY). Slides were developed and coverslipped 1 day (SS and MCH) or 7 days (CRH) later. Blinded counts of the number of silver grain clusters (corresponding to radiolabeled CRH, SS, or MCH mRNA) as well as the number of silver grains/cell were made using Grains 2.0.b software (University of Washington, Seattle, WA). After counts were analyzed, all slides were decoded, and representative slides were chosen for Nissl counterstaining to verify important anatomical landmarks and cytoarchitecture.

Nuclear c-Fos Analysis in the PSTN

Male rats were individually housed and fed the CD for 7 days before the experimental day. Two hours before the onset of the dark phase on the experimental day, the rats were placed in clean cages without food. Just before lights off, preweighed CD or VDD pellets were placed in each cage (n = 3 rats/group). Ninety minutes after feeding commenced, the remaining CD and VDD pellets were weighed for confirmation of reduced food intake in the VDD rats. The rats were then deeply anesthetized with a ketamine-xylazine-acepromazine cocktail and transcardially perfused with 0.01 M PBS followed by ice-cold 4% paraformaldehyde in 0.01 M PBS. Brains were removed, postfixed for 2–4 h, cryopreserved overnight in 20% sucrose, frozen on dry ice, and stored at −80°C.

Coronal hypothalamic sections were cut at 30 μm on a freezing microtome. Six sets of sections were collected from the diagonal band of Broca caudally through the mammillary bodies. Free-floating sections were stored in 0.01 M PBS at 4°C. The sections were incubated for 1 h at room temperature in blocking reagent (5% normal donkey serum in 0.01 M PBS and 0.1% Triton X-100; PBST). After the initial blocking step, the sections were incubated in rabbit anti-c-Fos antibody (PC38; EMD Biosciences, San Diego, CA) diluted 1:50,000 and rabbit anti-c-Fos antibody (SC-52; Santa Cruz Biotechnology, Santa Cruz, CA) diluted 1:25,000 in blocking reagent for 72 h at 4°C. The sections were then incubated in donkey anti-rabbit Alexa Fluor 594 (1:500; Invitrogen) in PBST for 2 h at room temperature. Between each step, the sections were washed thoroughly with 0.01 M PBS. Some sections were incubated in the absence of primary antibodies to confirm specificity of the secondary antibody. Sections were mounted on gelatin-coated slides and coverslipped using Vectashield mounting media (Vector Laboratories). The number of c-Fos-immunoreactive cells was manually counted in sections containing the PSTN using a fluorescence microscope (Leica 4000 DM; Leica Microsystems, Bannockburn, IL) by an investigator blinded to the treatment groups.

Nuclear c-Fos Analysis in POMC Neurons

To ensure robust simultaneous activation of POMC neurons in the experimental groups, we utilized a validated food-entrained mouse model (21). Male POMC-eGFP mice (n = 40) were allowed access to the CD from 0900–1100 each day for 8 days to establish strong and predictable food anticipatory behaviors. Mice were divided into the following nine treatment groups (n = 3–5/group): 3 diet treatments (CD, VDD, and fasted) × 3 time points (1.5, 2, and 3 h of feeding time). On day 9 at 0900, the CD and VDD mice were fed their respective diets for 2 h. Food was withheld from the fasted groups. At 1.5, 2, or 3 h after feeding commenced, mice were deeply anesthetized with a ketamine-xylazine-acepromazine cocktail and transcardially perfused with 0.01 M PBS followed by ice-cold 4% paraformaldehyde in 0.01 M PBS. Brains were processed, sectioned, and used for c-Fos IHC as described above. The number of c-Fos-immunoreactive green fluorescent protein-expressing cells was manually counted in sections containing the arcuate nucleus (ARC) by an investigator blinded to the treatment groups.

Statistical Analysis

All data are presented as means ± SE for each group. Statistical analyses were performed using GraphPad Prism 5 (La Jolla, CA). Comparisons between two groups at a single time point were performed using two-tailed Student's t-tests. Comparisons between three or more groups were performed by one-way ANOVA followed by post hoc analysis using Bonferroni-corrected t-tests. All comparisons involving two variables or multiple time points were performed using two-way ANOVA followed by post hoc analysis using Bonferroni-corrected t-tests. P < 0.05 was considered statistically significant.

Dietary Valine Deficiency Reduces Food Intake and Body Weight in Rats

We first evaluated the consequences of chronic dietary valine deficiency on food intake and body weight in rats. In rats fed the experimental diets for 21 days, daily food intake was significantly decreased in the VDD group compared with the CD group during the entire experimental period (P < 0.001; Fig. 1A). Whereas the CD group continued to gain weight over the course of the experiment, the VDD group weighed significantly less than the CD group as early as 4 days after introduction of the VDD and continued to lose weight for the duration of the experiment (P < 0.01; Fig. 1B). In contrast to the VDD group, the initial loss of body weight in the PF group was followed by an extended period of stable weight, typical of calorically restricted animals. The VDD rats weighed significantly less than the PF rats during the last 8 days of the experiment (P < 0.05; Fig. 1B).

Dietary Valine Deficiency Reduces Fat and Lean Body Mass in Rats

Whereas the CD group gradually gained fat and lean body mass over the course of the 21-day experiment, the VDD rats lost fat and lean body mass after only 7 days on the VDD (Fig. 1, C and D). The PF group also lost a significant amount of fat and lean body mass after 7 days but retained more fat and lean mass than the VDD group on days 14 and 21. Indeed, after 7 days, the PF group did not lose any further lean mass, whereas the VDD group exhibited a steady loss of lean body mass throughout the study. At the end of the experiment, gastrocnemius weights were significantly reduced in the VDD rats compared with the CD or PF animals (Fig. 1E). Furthermore, the expression of the muscle catabolic genes, MAFbx and MuRF1, was significantly elevated in rats that were fed the VDD for 3 days (compared with CD animals; Fig. 1F) or 21 days (compared with PF rats; Fig. 1G), indicating that VDD consumption activates an atrophy transcriptional program in skeletal muscle. Finally, the percentage of lean mass that is comprised of water was the same between all groups after 21 days of VDD exposure, suggesting that dehydration was not present in these animals (CD: 79.7 ± 0.3%, VDD: 82.2 ± 1.0%, PF: 82.7 ± 0.9%; P > 0.05).

Pharmacological Melanocortin Receptor Blockade Does Not Reverse Chronic Anorexia Induced by Dietary Valine Deficiency

We next investigated whether pharmacological blockade of central MC3/4R signaling reverses the effects of VDD consumption on food intake, body weight, and body composition. After 2 days on the CD (days −2–0), rats were switched to the VDD and were chronically infused with aCSF or SHU-9119 (1 nmol/day) for 21 days. For the first 3 days of treatment (days 1–3), the rats that were infused with SHU-9119 exhibited significantly greater daily food intake than the aCSF-treated animals (Fig. 2A). After the 3rd day of treatment, food intake declined in the SHU-9119-treated rats and no longer differed between the two groups for the remainder of the experiment. Similarly, the SHU-9119-treated animals lost less body weight than the aCSF-treated rats during the first 3 days of treatment but not on days 3–21 (Fig. 2, B and C). At the end of the 21-day treatment period, both groups had lost fat and lean body mass compared with pretreatment levels (Fig. 2D). Whereas the aCSF group lost 73% of their baseline fat mass, the SHU-9119 group only lost 49% of their baseline fat mass, indicating that body fat was preserved by SHU-9119 treatment. There was no difference in lean body mass loss between the two treatment groups.

Body temperature gradually decreased in both treatment groups during sustained VDD consumption (Fig. 2E). Notably, the daily nadir was reduced by more than 2°C compared with when the rats consumed the CD (days −2–0). As a result, the difference between daily peak and nadir temperatures gradually increased with prolonged VDD consumption, regardless of drug treatment.

To confirm the bioactivity of SHU-9119 at the end of this experiment, a separate group of animals was injected intracerebroventricularly with the residual drug remaining in the infusion pumps at the end of 21 days. This treatment significantly increased food intake and body weight in this cohort, confirming ongoing bioactivity of the SHU-9119 (data not shown).

Genetic MC4R Deletion Does Not Reverse Anorexia Induced by Chronic Dietary Valine Deficiency

We used MC4R-KO mice to further examine the role of melanocortin signaling in VDD-induced anorexia. WT and MC4R-KO mice that consumed the VDD for 14 days ate significantly less food than their respective CD groups (P < 0.001; Fig. 3A). Unexpectedly, the MC4R-KO/VDD mice reduced their food intake to a greater extent than the WT/VDD mice. When normalized to body weight, cumulative food intake in the MC4R-KO/VDD mice was significantly less than in the WT/VDD animals (WT: 1.12 ± 0.05 g/g and MC4R-KO: 0.85 ± 0.06 g/g; P < 0.01).

The WT/VDD and MC4R-KO/VDD groups both exhibited an aggressive, sustained loss of body weight compared with the CD and PF mice of the same genotype (P <0.001; Fig. 3B). Initially, weight loss did not differ between the two VDD groups, but there was a gradual trend over time for exaggerated weight loss in the MC4R-KO mice in both the PF and VDD groups, achieving statistical significance (P < 0.05) after 10 days of diet manipulation.

Both the WT and MC4R-KO mice fed the VDD lost a significant amount of fat and lean body mass compared with their respective CD and PF groups (Fig. 3, C and D). The MC4R-KO/VDD mice demonstrated a relative sparing of fat mass but lost more lean body mass than the WT/VDD mice.

Genetic MC4R Deletion Modestly Attenuates Anorexia Induced by Acute Dietary Valine Deficiency

To investigate acute changes in nocturnal food intake, we divided the 12-h dark phase into three periods (1900–2300, 2300–0300, and 0300–0700) and compared food intake in WT and MC4R-KO mice fed the CD or VDD. Although cumulative food intake did not differ between the WT/CD and MC4R-KO/CD mice during the first 4 h (1900–2300) of diet consumption, cumulative food intake was significantly elevated in the MC4R-KO/VDD mice compared with the WT/VDD mice during this same time period (Fig. 4). Food intake did not differ between the WT/VDD and MC4R-KO/VDD groups for the remainder of the dark phase. The MC4R-KO/CD mice consumed more food during the final 4 h (0300–0700) of the dark phase than the other three groups of mice because of the occurrence of two early morning feeding peaks in the MC4R-KO/CD mice that were absent in the MC4R-KO/VDD mice.

VDD Activates BAT Uncoupling Protein-1 Expression Via a Melanocortin-Independent Mechanism

Compared with animals in the CD or PF groups, uncoupling protein-1 (UCP-1) mRNA was significantly increased in the BAT of rats and mice after chronic VDD exposure (Fig. 5, A and B). There was a trend toward increased UCP-1 gene expression in rats that were fed the VDD for 3 days, but the difference did not reach statistical significance (Fig. 5A). Central infusion of SHU-9119 for 21 days did not alter UCP-1 gene expression in VDD-fed rats (data not shown), and the induction of UCP-1 mRNA was enhanced in MC4R-KO/VDD mice compared with WT/VDD mice (Fig. 5B). Leptin is a key regulator of central melanocortin tone. We therefore collected WAT to determine how VDD affects leptin gene expression in this tissue. Leptin mRNA expression was dramatically decreased in the WAT of rats fed the VDD for 3 days compared with rats that consumed the CD (Fig. 5C). Rats that consumed the VDD for 21 days had lower WAT leptin mRNA expression than either CD or PF rats.

Changes in Plasma Concentrations of Leptin, Insulin, IGF-I, Ghrelin, and Corticosterone in VDD Rats is Consistent With Severe Energy Deficit Relative to Caloric Restriction Alone

Consistent with leptin mRNA levels in WAT, the plasma concentration of leptin was dramatically decreased in both mice and rats fed long term with the VDD compared with CD or PF animals (Table 1). Similarly, plasma concentrations of insulin and IGF-I were significantly decreased in VDD animals compared with CD or PF animals (Table 1). At the 3-day time point, leptin, insulin, and IGF-I were equally suppressed in VDD and PF rats. In rats exposed to VDD for 21 days, leptin, insulin, and IGF-I continued to fall relative to the 3-day values, whereas PF animals demonstrated an increase in these hormones on day 21 relative to day 3. Ghrelin levels were increased in both the VDD and PF rats relative to CD rats on day 3 (Table 1), but there was no difference between the VDD and PF groups. Ghrelin levels remained elevated in VDD rats after 21 days, although the differences between groups on this experimental day did not reach statistical significance. Plasma corticosterone concentrations did not differ between rats after 3 days (CD: 263 ± 79 ng/ml, VDD: 179 ± 50 ng/ml, PF: 316 ± 106 ng/ml; P > 0.05) or 21 days (CD: 152 ± 25 ng/ml, VDD: 113 ± 37 ng/ml, PF: 231 ± 87 ng/mL; P > 0.05). We also measured a basic electrolyte panel in the rats after 3 and 21 days of VDD exposure. No differences between CD, VDD, or PF groups were found in glucose, creatinine, sodium, potassium, chloride, or bicarbonate at either time point (data not shown), again suggesting that these animals were not dehydrated. We did observe a small but statistically significant increase in plasma blood-urea-nitrogen in the VDD group at 21 days relative to CD and PF groups (CD: 11.2 ± 0.7 mg/dl, VDD: 17.6 ± 1.7 mg/dl, PF: 10.4 ± 1.1 mg/dl; ANOVA P < 0.01), likely reflecting the increased muscle catabolism (and subsequent nitrogen load) observed in this group.

VDD-Induced Anorexia is not due to Altered Expression of Appetite-Regulating Neuropeptides or Cytokines in the Hypothalamus

To further explore the role of hypothalamic neuropeptide systems in the anorexic response to the VDD, we measured the hypothalamic gene expression of neuropeptides that modulate feeding behavior. In rats fed the experimental diets for 3 days, agouti-related peptide (AgRP) gene expression was significantly upregulated in the VDD and PF groups (Fig. 6A). POMC and neuropeptide Y (NPY) gene expression did not differ between the groups. Rats that consumed the VDD for 21 days exhibited reduced POMC mRNA and increased AgRP and NPY mRNAs compared with either the CD or PF groups (Fig. 6B). There was no difference in MC4R mRNA expression between the three groups (data not shown). In mice fed the VDD for 14 days, WT animals showed similar changes in gene expression as observed in the 21-day rat study (data not shown). However, there was no difference in the expression of any of these genes between WT and MC4R-KO mice, regardless of diet (data not shown).

To rule out the possibility that VDD-induced anorexia occurs secondary to central inflammation, we measured hypothalamic interleukin (IL)-1β, IL-1 type I receptor, and tumor necrosis factor-α mRNA expression. In rats that were fed the experimental diets for 3 days, the expression of all three proinflammatory genes was similarly suppressed in the VDD and PF groups (Fig. 6A); however, long-term VDD consumption did not significantly alter the expression of these genes in rats or mice (data not shown).

Previous studies demonstrated an increase in hypothalamic SS gene expression in mice that consumed a VDD for 3 days (30). Our quantitative PCR data from mice on the VDD for 14 days are consistent with this observation, with a significant increase in SS expression in VDD animals relative to both the CD and PF groups (relative quantity for CD: 1.00 ± 0.07, VDD: 1.31 ± 0.06, PF: 0.95 ± 0.02; P < 0.01). However, we did not observe an increase in SS mRNA in rats that consumed the VDD for 21 days (Fig. 6B). Finally, we did not detect any global hypothalamic changes in CRH, MCH, or orexin gene expression in VDD rats or mice by quantitative PCR (Fig. 6B and data not shown).

In Situ Hybridization Analysis Reveals Novel Induction of Hypothalamic CRH, but not SS or MCH mRNAs, in Response to the VDD

Some neuropeptides implicated in the complex neuroendocrine response to IAA deficiency or dehydration have widespread expression in the brain, with only small neuroanatomically distinct subsets of cells demonstrating transcriptional responses to these challenges (41, 43, 44). We therefore used in situ hybridization analysis to investigate gene expression within specific hypothalamic nuclei. CRH is an anorexigenic neuropeptide that is expressed in neuroendocrine cells in the parvocellular paraventricular nucleus (PVN), with inducible expression in a distinct population of cells in the LHA in response to dehydration. We observed a small amount of CRH mRNA expression in the LHA of all rats, and there was a small but significant increase in grains per cell in the LHA of VDD animals as a whole (Fig. 7A). However, it was obvious that there was a marked induction of CRH mRNA in a distinct population of neurons located in the PSTN (Fig. 7B), found only in the VDD group (Fig. 7, C–E). Because there were no grain clusters in this nucleus in either the CD or PF groups, we did not perform grain-counting analysis on this distinct population of CRH-expressing cells.

We performed a similar analysis of MCH expression and saw no differences between groups when observed across the entire LHA or when the dorsomedial nucleus and perifornical area were analyzed separately (data not shown).

Finally, analysis of SS expression in the ARC and PVN revealed no differences in the number of SS-expressing cells (ARC: CD 363 ± 14, VDD 336 ± 34, PF 374 ± 49, P > 0.05; PVN: CD 314 ± 20, VDD 272 ± 14, PF 282 ± 30, P > 0.05) or grains/cell (ARC: CD 101 ± 2, VDD 106 ± 2, PF 111 ± 5, P > 0.05; PVN: CD 191 ± 5, VDD 170 ± 5, PF 171 ± 11, P > 0.05) in either nucleus.

APC Efferent Fibers Project to the PSTN and Form Close Contacts with CRH Neurons in This Nucleus

To provide a neuroanatomical substrate for the regulation of hypothalamic feeding centers by the APC, we traced efferent projections from the APC using anterograde PHA-L tracing. Rats were iontophoretically injected in an area of the brain matching the coordinates of the APC. After sectioning and immunohistochemical processing, we found that 2 of 10 animals had injection sites outside of the APC, and 1 other animal had no visible axonal transport. In the remaining seven animals, the efferent projections from the APC presented a consistent pattern. In each animal, we observed a number of PHA-L-labeled cells and fibers within the APC with a large majority projecting ipsilaterally along and within the cortex to a widespread but anatomically distinct distribution throughout the thalamus, hypothalamus, and limbic area. Within the amygdala, extensive innervation was observed. Within the hypothalamus, we observed extensive PHA-L labeling within the LHA and PSTN, with the projections having a beaded appearance typical of en passant synaptic varicosities (Fig. 8A). In a few cases, fibers were seen projecting dorsally in the dorsal hypothalamic area, posterior hypothalamic area, and the mediodorsal thalamic area. In triple-label immunohistochemical analysis, we observed PHA-L-labeled axonal fibers in close contact with CRH-positive cell bodies within the LHA and PSTN (Fig. 8B). Although we did not seek direct evidence for synaptic contact between PHA-L fibers and CRH cell bodies, we did observe numerous instances of close overlap within the 1-μm confocal field. There was no obvious anatomical relationship between PHA-L-labeled fibers and SS-labeled (Fig. 8B), MCH-labeled (data not shown), or orexin-labeled (data not shown) cell bodies in the LHA or elsewhere. We did not observe any difference in PVN CRH labeling between the three groups (data not shown).

c-Fos Immunoreactivity Demonstrates Activation of PSTN Neurons by VDD Exposure

When we analyzed c-Fos expression in response to introduction of VDD vs. CD, we observed extensive induction of c-Fos in the PSTN in VDD rats (Fig. 8, C–F). The VDD rats had significantly greater c-Fos expression in the PSTN than the CD animals (Fig. 8G). We also observed c-Fos expression in the VDD rats in areas previously shown to express c-Fos after introduction of a threonine-deficient diet, including the infralimbic cortex, dorsomedial hypothalamic nucleus, and central amygdala (data not shown) (40).

ARC POMC Neurons Do Not Have Exaggerated c-Fos Responses After Acute VDD Exposure

In all of our studies, animals begin to reject the VDD within hours of introduction, and we did observe some attenuation of acute anorexia with melanocortin blockade. Therefore, we evaluated neuronal activation in ARC POMC neurons after acute (1.5, 2, or 3 h) exposure to the CD or VDD in food-entrained POMC-eGFP mice to explore the possibility that rejection of the VDD is secondary to enhanced activation of POMC neurons. Compared with fasted mice, introduction of either the VDD or CD evoked significant increases in the number of ARC POMC cells displaying c-Fos immunoreactivity at all three time points (Fig. 9). However, there were no differences between the CD and VDD groups except for a less robust increase in c-Fos immunoreactivity in the VDD mice at the 3-h time point.