The animals used in this study were obese 3- to 5-months male homozygous loxTB Mc4r^−/− mice (stock number: 006414, Jackson Laboratories, Bar Harbor, Maine, United States; n = 22) and age-matched C57BL/6J (n = 6) mice for the first “preventative” study, and severely obese 8-month-old male homozygous loxTB Mc4r^−/− mice (n = 15) and C57BL/6J (n = 10) for the second “treatment study.” Homozygous mice exhibit severe obesity due to a loxP-flanked transcriptional blocking (loxTB) sequence that prevents normal endogenous gene transcription and translation from the endogenous locus. As such, homozygous mice are devoid of functional mRNA in all tested regions of the brain that endogenously express Mc4r. Animals used in this study were reared in an in-house breeder colony at Georgetown University and were bred from a homozygous breeder pair purchased from Jackson Laboratories. The control C57BL/6J mice were also purchased from Jackson Laboratories, and together with Mc4r^−/− mice were housed in a 12-h light/dark cycle under controlled temperature and humidity. All animals had ad libitum access to food and water except the night prior to surgery and during the recovery phase where the animals were pellet-deprived and fed a liquid diet of berry-flavored Ensure Enlive. All procedures were performed in accordance with the National Institutes of Health guidelines for use of animals in research and with the approval of the Georgetown University Animal Care and Use Committee.

Mice were acclimated to individual housing for at least 3 days prior to baseline measurement of their food intake (FI) and body weight (BW) over a 15 day period. All mice were provided access ad libitum to regular mouse chow that was pre-weighed (Low fat Purina Diet 5001; composition: 23.0% protein, 4.5% fat, 5.3% crude fiber, 49% carbohydrate, total digestible nutrient 76%, 3.04 kcal/g metabolizable energy). To assess FI, daily measurements were made between 3 and 5 p.m. by weighing each pellet manually and taking into account food spillage. FI was recorded for 4 weeks post-surgery.

Prior to surgery, all animals were solid food-deprived overnight (18–24 h), while access to water and a berry flavored nutrition liquid (Ensure Enlive) was provided ad libitum. This was done to ensure as little solid food in the stomach as possible. All animals were anesthetized with isoflurane [induction 4%; maintenance 1.5–2% of minimum alveolar concentration (MAC)] and their body temperature was maintained at 37°C with an infrared heating lamp. Following adequate depth of anesthesia (as monitored via the toe pinch reflex), a midline abdominal incision (~1 cm) was made along the linea alba running ~1 cm caudal from the xiphisternum. The liver was retracted with a saline dampened cotton swab, and gentle traction was applied to the esophagus by lifting the stomach out of the peritoneal cavity using an umbilical tape wrapped around the gastric antrum. A curved 22-gauge glass rod was placed carefully under the esophagus to gently lift it and hold it in an exposed position. All identifiable vagal afferent and efferent fibers running along the esophagus orad to the esophagogastric junction (~1 cm) were visualized with the aid of a surgical microscope (Bausch & Lomb, Inc.) and excised with a microsurgical hook having an interior cutting edge (Circon, Corp.). In each case, a vagal segment (as long as possible) was exposed and isolated, which was then excised. This consisted of the removal of all four major components of the abdominal vagus (celiac, hepatic, gastric, and both ventral and dorsal vagal trunks). In addition, all neural tissue surrounding the esophagus immediately beneath the diaphragm was excised with a microsurgical hook to capture any residual vagal branches. In all animals that received a vagotomy, extreme care was taken to avoid any damage to the subdiaphragmatic esophagus. As a result, both food intake and defecation were similar in these animals to those of sham-vagotomized and pyloroplasty alone controls [Note: Signs of dysphagia-associated damage to the upper esophageal sphincter or pharynx (abnormal licking, mastication patterns, or abnormal breathing) were not observed in any animals that underwent vagotomy or sham-vagotomy]. After this procedure, the umbilical tape and glass rod were removed. For sham-operated animals, the vagus nerves were similarly exposed but not cut, and the stomach received gentle traction with the umbilical tape.

In our initial studies of mice that underwent bilateral SDV, none of the mice survived more than a few days due to marked gastric distention and pyloric stenosis. To circumvent this problem, we performed a modified Heineke-Mikulicz pyloroplasty to facilitate gastric drainage and relieve pyloric stenosis.

After bilateral subdiaphragmatic vagotomy, a pyloroplasty procedure was subsequently performed in the anesthetized mice. To accomplish this, the gastric antro-pyloro-duodenal region was gently exposed and lifted with a 22-gauge bent gavage needle and a fine forceps were placed under the pylorus to stretch it. A longitudinal incision (~2–4 mm) was made through the pylorus, extending from the distal antrum to the proximal duodenum. Four (5-0 Vicryl) sutures were placed in the pyloric tissue such that by closing the incision transversely, the outlet diameter of the pylorus was increased. A small piece of absorbable gelatin sponge was placed over the pyloroplasty and secured in place with the remaining thread of the last pyloric tissue suture to further ensure against possible leakage of intestinal contents into the peritoneal cavity. The duodenum/stomach was then placed back in the peritoneal cavity and the abdominal incision closed in a two-step procedure using a 5-0-vicryl suture. Postsurgical medication consisted of Baytril (5 mg/kg SC) for infection prevention, 1cc SC saline for prophylactic hydration, and Rimadyl (5 mg/kg SC) for analgesia. No gastric distention was observed in animals undergoing this procedure, indicating that the antro-pyloro-duodenal junction remained patent. For pyloroplasty only animals, the subdiaphragmatic vagal trunks were similarly isolated and exposed but not excised; only a pyloroplasty was performed.

For 1 week, all animals were monitored for piloerection, ptosis, porphyrin, hunching, mobility, and ruffled coat. Any animal displaying any of these adverse effects was removed from the study. To minimize post-surgery disturbances of intestinal motility, during the post-recovery phase, animals were maintained on Ensure Enlive berry-flavored nutrition drink.

After the completion of both studies, each animal received a single intraperitoneal injection of FluoroGold (0.8 mg/0.4 ml saline; Fluorochrome, Denver, CO). One week after injection, all animals were euthanized with an intraperitoneal administration of pentobarbital (200 mg/kg) and perfused via the ascending aorta with a phosphate buffered saline wash (0.1 M, pH 7.4) followed by a buffered 4% paraformaldehyde (0.1 M, pH 7.4) fixative delivered over a 15 min period. The brains were removed and placed in fixative (4% buffered paraformaldehyde-20% sucrose) for at least 24 h. After cryoprotection, the brainstem was separated and cut on a cryostat (Leica, 1500) into 50 μm coronal sections, which were mounted on slides for later imaging using episcopic fluorescence microscopy. Each brainstem section (~9 sections/slide; total slides = 5/brainstem) was examined for FluoroGold label in the DMV. The presence of fluorescent label in DMV neurons was accepted as a marker of incomplete vagotomy. Only animals with complete or no vagotomy were included in the study.

After the completion of both studies, animals were euthanized (see above) and the fat pads dissected from the following compartments: mesenteric, peri-renal, sub-scapular, and epididymal. The wet weights of the different fat pads were then weighed.

Adult male Mc4r^−/− mice (n = 13) were acclimated in metabolic chambers (PhenoMaster: TSE Systems, Bad Homburg, Germany) for 3 days before the start of recordings. After acclimation for 3 days, oxygen (VO2) consumption and carbon dioxide (VCO2) production were acquired every 10 min for 72 h via the PhenoMaster software. Metabolic parameters that include energy expenditure (EE) and respiratory exchange ratio (RER = VCO2/VO2) were calculated by algorithms native to the supplied acquisition software. A detailed description and protocol of these can be obtained directly from the manufacturer or found on the website (www.tse-systems.com). Briefly, the software generates 3 types of calorimetric datasets of which the first takes into account the animal's body weight and the second takes into account the lean body mass (LBM), calculated from a user-defined power of the animal's weight in a range of 0.001–0.999. In the third set, neither the animal's weight or LBM is taken into consideration.

In analyzing the data, we took into consideration which subset of the data generated by the PhenoMaster software would most accurately reflect the calorimetric parameters associated with our experimental conditions. It is well-documented that indirect calorimetry to compare EE among animals that differ in body weight has inherent inaccuracies (Arch et al., 2006; Butler and Kozak, 2010; Tschöp et al., 2012). First, differences in body weight are usually associated with differences in tissue distribution, making it difficult to calculate their specific contribution. For example, adipose tissue may have low EE in obese animals and although the metabolic rate of fat tissue may be low, it is not zero and may represent a relatively high proportion of the animal's total mass. Second, adipose tissue releases adipokines, notably leptin that affect EE in other tissues. Third, bigger animals with their larger body mass are widely accepted to expend more energy (Felig et al., 1983; Arch et al., 2006). Hence, to avoid these potentially confounding variables, many investigators choose to normalize VO₂ using body weight or, preferably, to LBM, which many investigators feel accurately reflects total EE (Tschöp et al., 2012; Speakman, 2013). It is worth noting, however, that normalization to LBM has its own inherent flaws since the lean tissue is not distributed evenly in all compartments and, as a result, LBM does not compare isometrically with total body mass (Arch et al., 2006; Speakman, 2013). In the present study, since a body composition analyzer (EchoMRI) was not available to us, we did not measure LBM. Additionally, a ratio-based method for normalizing EE to LBM is not presented. We do present EE normalized to body weight [kcal/h/kg], as well as non-normalized EE data [kcal/h], because of substantial differences in the body size of the study animals.

Data are presented as mean ± SEM. In all cases, p < 0.05 was the criterion used to determine statistical significance. Statistical analysis of the fat pads was performed using a one-way analysis of variance (ANOVA) and unpaired-t- tests, whereas body weight and cumulative food intake were analyzed employing a two-way repeated measure ANOVA and Bonferroni's multiple comparison post-hoc tests.

To analyze the metabolic data, a linear regression model was employed to determine the relationship of EE (kcal/hour) to body mass within and across groups. EE was averaged with respect to time and the resulting averaged EE values were fitted using a linear regression model with body weight as an independent variable. Locomotor activity (LA) was measured across the long axis of the cage by an array of infrared sensors. Activity data is reported here as the combination of ambulatory activity (mouse crosses two adjacent beams) and fine activity (i.e., grooming where the mouse crosses the same beam twice). Normalized EE (kcal/h/kg), RER, and LA were analyzed for both dark and light cycle phases using repeated measures one-way analysis of variance (ANOVA) with post-hoc analysis of the data using the Tukey multiple comparison test. [Note: for the dark cycle LA measurements, light cycle phase data was filtered to only include data points where the animals showed no LA (i.e., animals are presumed to be at rest)]. The 24 h EE, RER, and LA data was computed from the first 24 h of indirect calorimetry data.

The effectiveness and completeness of the SDV surgery in these studies was verified by the retrograde tracer Fluro-Gold (IP). As expected, in sham operated mice (sham SDV and sham pyloroplasty; SS group; n = 7), Fluoro-gold labeled neurons were abundantly present in the dorsal motor nucleus of the vagus (DMV; Figure 1A) compared to those undergoing SDV and pyloroplasty where labeled neurons were absent in the DMV (PV group; n = 7; Figure 1B). The presence and absence of fluorescence label in the DMV in the two groups attest to the intactness of the vagus nerve in the SS group and the completeness of SDV in the PV group, respectively. Furthermore, these observations were corroborated by differences in body weight (BW) between the two groups of mice. In the SS group, following an initial reduction in post-surgery, the weight exceeded that of pre-surgical values (Figure 1C) which in the case of the PV group did not recover to baseline levels (Figure 1D).

As indicated in the methods, bilateral SDV per-se is lethal in mice, therefore for it to be successful, it has to be performed in conjunction with a pyloroplasty. Hence, to assess the effect of SDV as a preventative measure for Mc4r^−/− deficient obesity, mice were assigned to three groups. These groups underwent the following treatments: Group 1, sham SDV and sham pyloroplasty (SS, n = 7); Group 2, pyloroplasty and sham SDV (PS, n = 8); and, Group 3, pyloroplasty and SDV (PV, n = 7). Prior to surgery, (see section Materials and Methods), initial baseline weights (obtained over a 2 week period) of the three groups of mice were 44.0 ± 1.1, 41 ± 1.2, and 40.0 ± 1.0 g, respectively (Figure 1E). After surgery, all mice were placed on a liquid diet for 7 days. Their normal diet was re-established on post-surgery Day 8. From thereon, body weight (BW) was monitored daily and plotted over an 8-week period (Figure 1E). As can be noted, the SS and PS groups after initially losing weight following surgery gained it at a fairly constant rate. At the end of the 8-week period, the SS and PS groups had attained weights of 54 ± 1.4 and 53 ± 1.3 g, respectively. Mice that underwent both pyloroplasty and SDV (PV) lost weight initially but then gained it, though at a much slower rate than those mice in the SS and PS groups. Indeed, at the end of the 8-week monitoring period, the PV group weighed only 38.7 ± 2.3 g, a value significantly less than the corresponding weights of the SS and PS groups (p < 0.05). Instead, the PV group of mice began to return to the weight that they had prior to surgery (Figure 1E). This represents a 2.5% decrease in BW over the 8-week period compared to an overall increase in BW that was seen for both the SS and the PS groups (22.7 and 29.2% respectively). More importantly, the >20% increase in BW over an 8-week period that would have been expected in Mc4r^−/− mice (as seen in SS and PS controls) was blunted in the PV group.

The change in weight of the three groups over the duration of the post-surgical monitoring period is depicted in Figure 1F. As can be noted, the SS and PS groups gained 10.2 g and 11.6 g respectively, while the PV group lost 0.8 g. Inspection of the data clearly indicates that pyloroplasty alone was not responsible for the ability of SDV to cause the Mc4r^−/− obese mice to gain weight at a much slower rate than the SS and PS groups. Instead, the slower rate of weight gain in the PV group appeared to be due entirely to the SDV procedure.

Examination of the data also raised the question as to whether the efficacy of SDV, in addition to preventing obesity as seen in Figure 1E, is robust enough to restore the weight of Mc4r^−/− mice to that of normal age-matched C57BL/6J mice (wild-type; WT). In other words, is SDV able to reverse the obesity phenotype associated with the Mc4r deficiency and bring the weight of the mice back to that of the normal weight of the control background strain (C57BL/6J). To address this question, we studied a fourth group of mice, namely a WT group (n = 6) that underwent both sham pyloroplasty and sham SDV. Their body weights at baseline, during recovery after surgery, and over the 8-week monitoring period are shown in Figure 1E (WT SS group). The starting average BW of this group was 27.0 ± 0.4 g. As can be noted from Figure 1E, except for an initial drop in weight during sham surgery, this group of mice steadily maintained their body weights over the 8-week period of observation (28.3 ± 0.6 g). This value was not significantly different from their baseline value, but clearly less than the corresponding value of the PV group (p < 0.05) at any time point. These data indicate that SDV while having a robust effect to counter the weight gain observed in the Mc4r^−/− mice, did not restore the BW of these animals to normal. With regards to the general well-being of the PV mice, each mouse appeared in good health and, based on general behavior, could not be distinguished from mice in the other groups; except by body size. Moreover, the fecal output of this group was of normal volume and consistency and could not be distinguished from those of other groups; neither could their eating, drinking, or urinating.

Solid food intake (FI) data recorded over a 4-week period in the three Mc4r^−/− groups are presented in Figure 1G. As is evident, cumulative weekly FI of the PV group was significantly less (p < 0.05) than that of the SS and PS groups in the first week after post-surgical recovery and significantly less (p < 0.05) than that of the PS group in the second week after recovery. However, due to the fact that FI of the PV group normalized to that of both controls in subsequent weeks, no overall decrease in FI was observed.

Further examination of the data raised the critical question as to whether or not the weight loss associated with SDV in the Mc4r^−/− mouse may be confounded to non-specific effects of the surgery itself that could occur in any mouse model independent of genotype. To address this question, we studied a separate group of WT mice (n = 10) that underwent both vagotomy and pyloroplasty (PV). Their body weights at baseline, during recovery after surgery, and 17 days post-surgery are shown in Figure 1E (insert graph, WT PV group). The starting average BW of the group at baseline was 25.1 ± 0.68 g, whereas, on post-surgery day 17, the average weight of this group of mice was 25.9 ± 0.67 g. As can be noted from Figure 1E (insert graph) there was no statistically significant difference in body weight of these mice at baseline compared to 17 days post-surgery (p > 0.05; Figure 1E insert graph). These data indicate that SDV while having a robust effect on BW in Mc4r^−/− mice, is not effective in reducing BW in age-matched normal weight WT mice. In fact, in both the prevention and reversal studies using the obese Mc4r^−/− mouse, robust weight loss was seen 17 days post-surgery in the PV group. In the “prevention” study, these mice lost 20.2% of their baseline weight (see Figure 1E), while in the “reversal” study, they lost 30.3% of their initial excess weight (Figure 3A). The data demonstrate that weight loss associated with SDV in the Mc4r^−/− mouse is not due to non-specific effects of the surgery itself, but rather due to the underlying influence of the Mc4r^−/−genotype.

We also determined whether SDV would prevent white adipose tissue (WAT) accumulation in Mc4r^−/−mice. Fat pads were dissected from the visceral compartments in the 3 groups of mice. In the PV group, the weight of WAT as a percentage of that in the SS controls (mesenteric = 2.18 ± 0.13 g; peri-renal = 1.34 ± 0.09 g; epididymal = 2.7 ± 0.21 g; brown adipose tissue = 0.72 ± 0.10 g) was significantly less in both the mesenteric (MES) and peri-renal (PR) sub-compartments (p < 0.05; Figure 1H). In contrast, the weights of the fat pads taken from the PS was not different from that of the SS group. Sub-scapular brown fat (BAT) and epididymal fat pad (EPI) in the PV group was not statistically significant from that of the other two groups (Figure 1H). In addition, we measured the snout-to-anus length (an indicator of body growth) as it known to increase in Mc4r^−/−obese mice (Huszar et al., 1997); SDV was found to significantly blunt this end-point, whereas pyloroplasty, per-se, had no effect on linear growth (p < 0.05; Figure 2).

Our purpose was to determine whether SDV could reverse the established obesity in the Mc4r^−/−mouse. Mice had already become obese before SDV treatment was performed. Prior to surgery, initial baseline weights (obtained over a 1 week period) were 51 ± 0.3 g (SS), 52 ± 0.3 g (PS), and 55 ± 0.3 g (PV), respectively (Figure 3A). After surgery, all mice were placed on a liquid diet for 7 days. Their normal diet was re-established on post-surgical day 8. From thereon, body weights were monitored daily and plotted over a 4-week period. As can be noted, the SS and PS groups initially lost weight after surgery, but gained it back over time (Figure 3A). At the end of the 4-week period, the SS and PS groups had attained weights of 51 ± 3.6, 55 ± 3.4 g, respectively. Obese mice that underwent both pyloroplasty and SDV (PV group) lost weight initially, and then continued to lose weight over the 4-week period of the study (Figure 3A). Indeed, at the end of the 4-week monitoring period, the PV group weighed only 38.4 ± 0.8 g, a value significantly less than the corresponding weights of the SS and PS groups (51.0 ± 3.58 and 54.8 ± 3.4 g, respectively; p < 0.05). The change in weight of the 3 groups over the duration of the post-surgical monitoring period is depicted in Figure 3B. As can be noted, at the end of the 4-week monitoring period, the SS and PS groups had returned to their baseline weights while the PV group did not; they had lost 17 g from their initial baseline weight.

Solid food intake (FI) recorded weekly for the three Mc4r^−/−groups are presented in Figure 3C. Similar to the prevention study (see Figure 1G), FI in the PV group of mice was significantly less (p < 0.05) than that of the SS and PS groups at post-surgical week 1 (week 3 of the overall study). In addition, FI of the PV group was reduced relative to the SS group during week 2. This is not reflective of an overall decrease in FI since in the PV group it normalized to that of the SS groups in the subsequent weeks (week 3 and 4). Interestingly, although FI for the PV group was not significantly reduced relative to the SS group during weeks 3 and 4, it was significantly less than that of the PS group (p < 0.05; Figure 3C). Furthermore, it should be noted that that the PS group increased their FI relative to baseline in weeks 3 and 4 (Figure 3C).

We also determined whether SDV would reverse WAT accumulation in fully obese Mc4r^−/−mice. Fat pads were taken from the visceral components in the 3 groups of mice and weighed. The WAT content in PV group as a percentage of that present in the SS control tissue (MES = 1.8 ± 0.4 g; PR = 1.59 ±0.5 g; EPI = 2.02 ± 0.2 g; BAT = 0.3 ± 0.1 g) was significantly less in the MES sub-compartment (p < 0.05; Figure 3D). In the PR and EPI sub-compartments, the WAT content in the PV mice was not significantly different from that of the SS group, as was that of BAT, which was similar to those of the PS and SS groups (Figure 3D). Additionally, no significant differences were found in fat pads weights between the two control groups (PS and the SS group). The snout-to-anus length, reflective of body growth, was also unaffected by SDV in these fully obese animals (data not shown).

Data obtained from studying SDV as prevention and as a treatment strategy against Mc4r^−/− mouse obesity demonstrates a robust anti-obesity effect. The mechanisms for the anti-obesity effect initially include a transient reduction in FI (Figures 1G, 3C). However, the reduction in FI did not persist, which suggested to us that the SDV effect on BW maybe due to an increase in energy expenditure (EE). To determine this, we undertook a series of studies to measure the influence of SDV on metabolic parameters that are reflective of EE.

Metabolic parameters associated with EE were analyzed (see Materials and Methods) in relationship to BW in 3 groups of fully obese adult Mc4r^−/− mice as described above (SS, PS, and PV). We particularly focused on this relationship during the dark cycle, which is correlated with increased activity in rodents. While there was no significant correlation between BW and EE in the SS control group, animals undergoing SDV (PV group) showed an inverse linear relationship (p = 0.014, slope = −0.009, 95% confidence interval = [−0.014, −0.003], R² = 0.81; Figure 4A). The decrease in BW in the PV group was associated with an increase in EE. This contrasts with the relationship of EE to BW in animals subjected to pyloroplasty and sham vagotomy (PS group). In the PS group (see Figure 4A), there was an indication of a positive correlation between EE and BW indicating that as they became heavier, the likelihood of their energy consumption also increased (p = 0.07, slope = 0.01, 95% confidence interval = [0.002, 0.026], R² = 0.71). Despite the fact that there was no significant correlation in either control group, it is worth noting that both groups exhibited a positive relationship between BW and EE indicating that as they got heavier, they expended more energy. This was in contrast to the PV group that exhibited a negative relationship between the two variables. We also employed linear regression analysis to data obtained from the aforementioned three groups during the light cycle. No significant differences were observed in the relationship of EE to BW in all three groups (data not shown).

While body size and composition influences EE, there is no consensus on the best way to normalize VO₂ or VCO₂ data (Butler and Kozak, 2010; Kaiyala et al., 2010). We chose to present the data as normalization of EE to BW in the light and dark cycles to facilitate comparison to our the linear regression data (Figure 4A). As illustrated in Figures 4B–D, mean normalized EE [kcal/h/kg] was analyzed for the three groups in the dark and light cycles, as well as the first 24 h cycle of indirect calorimetry recording. As can be seen, PV mice displayed an increase in normalized EE relative to the SS control group in both the dark, light and 24 h cycle (Figures 4B–D). However, the mean normalized EE of the PS group (Figures 4B–D) was the highest of the three groups despite showing no weight loss (Figures 1E,F, 3A).

Mean respiratory exchange ratio (RER) was collapsed over time and analyzed for the three groups. In comparison to the SS and PS groups, the overall RER of the PV group was lower (p < 0.05; Figures 5A,B) during both the dark cycle (12 h) and light cycle (4 h). The RER for the 24 h cycle is shown in Figure 5C.

Locomotor activity (LA; a combination of axial/ambulatory and fine movements) during the light cycle (collapsed over ~18 h) and dark cycle (collapsed over ~ 28 h) revealed a significant overall increase in LA in the PV group relative to the SS and PS groups with no change seen between either of the control group (p < 0.05; Figures 6A,B). A representative graph of the initial 24 h LA cycle is shown in Figure 6C.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.