Assessing Interactions Between Ghsr and Mc3r Reveals a Role for AgRP in the Expression of Food Anticipatory Activity in Male Mice

Animal husbandry

All studies were approved by the Scripps Research Institute Institutional Animal Care and Use Committee. Mc3rKO [B6(Cg)-Mc3r^tm1Butl/J] generated by our laboratory using C57BL/6J (B6) embryonic stem cells (25) were on a mixed C57BL/6J and C57BL/6N background. GhsrKO had been backcrossed onto the C56BL/6J background for more than 12 generations (13). Double heterozygous (+/−) mice generated by mating male Mc3rKO with female GhsrKO were intercrossed to generate mice used in subsequent breeding steps: Mc3r−/−;Ghsr+/− males bred with Mc3r+/−;Ghsr−/− females to produce Mc3r−/−;Ghsr−/− mice (DKO). GhsrKO mice were generated using homozygous GhsrKO males and females. Mc3rKO were generated using Mc3rKO males crossed with Mc3r+/− females. Controls for the study were obtained from wild type (WT) mated with Ghsr+/− mice. The studies presented here used males housed at a constant temperature of 23°C with food and water provided ad libitum under a 12-hour light, 12-hour dark (LD) cycle, unless otherwise specified.

Assessment of FAA using running wheels

WT (n = 12), Mc3rKO (n = 6), GhsrKO (n = 10), and DKO (n = 10) male mice aged 8–12 weeks were individually housed in cages equipped with running wheels. After a 2-week acclimation period in a LD cycle, the mice were transferred to constant darkness (DD) for 3 weeks. FAA was then assessed using a hypocaloric RF protocol previously used by our laboratory (24–26). Briefly, mice received 3 g of purified low-fat high-carbohydrate diet (D12450B; Research Diets, Inc) every 24 hours in the middle of their inactive phase. After 4 weeks of RF, the mice were returned to ad libitum feeding conditions for 12 days, before being transferred back to LD. After 2 weeks in an ad libitum feeding condition in LD, the mice were submitted to a second RF protocol for 2 weeks. Note that the use of the same animals for multiple RF protocols separated by a recovery period has already been reported (38, 39). Mice were then euthanized 1 hour before the scheduled meal presentation and brains collected and snap frozen using an isopentane/dry ice mush. Wheel-running data were analyzed using Clocklab software (Actimetrics). Cage changes were performed every 2–3 weeks, with night-vision goggles used when mice were housed in DD.

Assessment of body composition and metabolic parameters

Body composition was determined using nuclear magnetic resonance (Bruker Minispec). Metabolic parameters were assessed using a 16-chamber comprehensive laboratory animal monitoring system (Columbus Instrument) using 10-week-old male mice fed regular chow. After 24 hours of acclimation, the respiratory exchange ratio (RER), oxygen consumption (VO₂) and activity (breaks in infrared light beams in the X axis) were measured for 2 days. The mice were then fasted overnight (16 h), followed by 7 hours of food access. Diet-induced obesity (DIO) was investigated using male mice fed a high-fat diet (HFD) (60% kcal from fat; Research Diets, Inc) for 8 weeks starting at 20 weeks of age. At the end of the HFD experiment, trunk blood was collected after an overnight fast (16 h).

Blood chemistries

Trunk blood was collected after decapitation and centrifuged at 1000g at 4°C for 10 minutes. Plasma glucose and triglycerides (TGs) were assessed using the Cobas c311 (Roche). Plasma insulin was measured with a mouse insulin ELISA kit (Crystal Chem). The intra-assay coefficient of variation for the insulin assay was less than 10%, whereas the inter-assay coefficient of variation was 25%. Homeostasis model assessment of insulin resistance (HOMA-IR) was calculated as previously described (40). Intraperitoneal glucose tolerance tests were performed before the introduction of the HFD and after 6 weeks on this diet using a One Touch Ultra 2 glucometer (Lifesccan, Inc). Briefly, the mice were fasted for 16 hours, and blood glucose was measured before and at 15, 30, 60, and 90 minutes after an injection of D-glucose (1 g/kg; Sigma). Blood glucose was also monitored every 2 hours from lights on (zeitgeber time 0) until lights off (zeitgeber time 12) on the 15th day of the second RF protocol.

Gene expression analysis

Hypothalami were dissected from the frozen brains on dry ice, and total RNA was extracted using the TRIzol reagent (Invitrogen, Life Technologies). Total RNA was treated with a DNA-free kit (Ambion, Life Technologies) to remove genomic DNA contamination, and 1 μg was used for cDNA synthesis using a SuperScript III Reverse Transcriptase kit (Invitrogen). Quantitative PCR was then performed in 96- or 384-well plates using TaqMan gene expression and a One Step Plus or ABI PRISM 7900 HT Detection System (Applied Biosystems, Life Technologies). Glyceraldehyde 3-phosphate dehydrogenase or cyclophilin A was used as reference gene.

Statistical analysis

All data are presented as mean ± SEM. Statistical analyses were performed using Prism 6 statistical software (GraphPad Software, Inc). Differences between genotypes were assessed by either ANOVA followed by Tukey's post hoc test or unpaired t test according to the experimental design. Effect of genotype and diet were analyzed using two-way ANOVA followed by Tukey's post hoc test. Pearson correlation was used to evaluate the relationship between FAA and the level of each hypothalamic transcript studied. P < .05 were considered significant.

Effect of genotype on body composition and energy homeostasis

The inheritance of the null alleles exhibited a normal Mendelian distribution, indicating no lethality associated with loss of both Mc3r and Ghsr genes. Although Mc3rKO mice weaned onto rodent chow exhibited normal body weight at 8 and 20 weeks of age (Figure 1, A and D), they exhibited the increased fat mass (FM) and reduced fat-free mass (FFM) that is a defining feature of this strain (Figure 1, B and E). Body weight and composition of GhsrKO mice was normal (Figure 1, A, B, D, and E). DKO mice exhibited decreased body weight compared with all other genotypes at 8 weeks (Figure 1A) with reduced FFM (Figure 1B) and adiposity (FM % as a body weight) when compared with Mc3rKO (Figure 1C). However, by 20 weeks, adiposity of DKO and Mc3rKO was not significantly different, suggesting a delayed onset of obesity (Figure 1E). We also observed delayed DIO in GhsrKO mice relative to DKO and Mc3rKO (two-way ANOVA, effects of genotype F_(15,252) = 31.96, P < .0001; effects of time F_(5,252) = 96.16, P < .0001; genotype X time F_(3,27) = 31.93, P > .05) (data not shown). GhsrKO also exhibited improved insulin sensitivity, with lower fasting insulin levels and HOMA-IR compared with controls and Mc3rKO after 8 weeks on a HFD (Table 1). Intraperitoneal glucose tolerance test data were generally consistent with these findings (Supplemental Figure 1), although the effect of genotype on glucose tolerance after high-fat feeding did not achieve statistical significance.

Because the feeding protocol used to induce FAA involves periods of intermittent fasting, we first examined energy balance in DKO in ad libitum-fed and fasting conditions using a separate group of 10-week-old mice housed in a temperature-controlled comprehensive laboratory animal monitoring system. In this group, genotype had a similar effect on body composition to that observed at 8 weeks (Figure 2A). The delayed increase in adiposity of DKO relative to Mc3rKO may be due to increased energy expenditure (Figure 2). Food intake averaged over 24 hours was not significantly affected by genotype (F_(3,27) = 1.972, P > .05). Although the amount of food removed from the container was significantly different during the dark phase (F_(3,27) = 3.174, P < .05) (Figure 2B), this was due to excessive spillage by 2 GhsrKOs and 1 DKO. Strong trends for increased VO₂ during the light phase and fasting were observed in DKO mice when data were normalized to lean mass (genotype effect F_(3,25) = 2.875, P = .06 and F_(3,24) = 2.883, P = .06, respectively) (Figure 2, E and F). The RER, an indicator of whole-body substrate preference, and activity were normal in all of the mutant strains in fed conditions (Figure 2, G and H). We also assessed the response of the mutant mice to fasting and refeeding in the CLAMS, observing no significant differences (Supplemental Figure 2, A–J).

Analysis of FAA in mice housed in DD

When transferred to the cages equipped with wheels, the mice were 9–10 weeks old and exhibited similar differences in body weight to mice aged 8 weeks. DKO weighed significantly less compared with the other genotypes (WT, 24.6 ± 0.4 g; GhsrKO, 25.1 ± 0.9 g; Mc3rKO, 24.6 ± 1.1 g; and DKO, 21.5 ± 0.5g; F_(3,34) = 5.819, P < .05). During acclimation, DKO exhibited increased wheel-running activity during the dark phase that translated to increased 24 hour activity compared with GhsrKO (Supplemental Figure 3, A and B). However, this difference was transient and was no longer observed after transfer to DD (data not shown). All strains exhibited a robust circadian rhythm with no effect of genotype on period length (τ) or amplitude (data not shown).

After 3 weeks of free running, food was removed and the mice then were subjected to RF for 4 weeks. Food (3 g) was presented during the middle of their inactive phase. Before the start of the RF protocol, no activity was detected in the 3 hours before the projected food presentation (Figures 3 and 4, A and B), except for 1 Mc3rKO and 1 GhsrKO mice that exhibited shorter τ and were thus removed from the analysis. WT mice exhibited a robust FAA after 4 days of RF (Figures 3 and 4, C and D). All KOs exhibited marked reductions in FAA after 3–4 days of RF (Figures 3 and 4). After 7 days of RF, the robust FAA observed in WT mice had also not developed in Mc3rKO and DKO (Figures 3 and 4, E and F). However, by that time, GhsrKO had begun to exhibit FAA.

After 3 weeks, WT mice exhibited very little activity after food presentation. In contrast, Mc3rKO and DKO exhibited significant levels of activity after food presentation at this stage (Figure 3). Further analysis of activity at selected time points confirmed these observations while also suggesting that deletion of both alleles had no additive effect (Figure 4, D and F). After 4 days of RF, WT mice exhibited a significant and progressive rise in activity preceding food presentation; this was attenuated in GhsrKO, and more severely in Mc3rKO and DKO (Figure 4, C and D). After 7 and 14 days of RF, the mutant mice exhibited a delayed onset of FAA (Figure 4, E–H). However, by the end of the 3rd week of RF, all mice were exhibiting high levels of activity before food presentation, perhaps related to the advance of free running activity into the feeding time (Figures 3 and 4, I–L). Indeed, we were surprised to observe that the mice did not appear to completely synchronize periods of peak anticipatory activity with food presentation, as we previously observed (24). Rather, all mice continued to free run, resulting in the feeding time coinciding with the period of peak activity.

After 4 weeks of RF, the mice were returned to ad libitum feeding for 12 days and then returned to a LD condition. After a period of adjustment, FAA was reassessed using the same protocol (Figures 3 and 5, A–H). The amplitude and onset of FAA were again markedly attenuated in Mc3rKO and DKO (Figures 3 and 5, C and E). However, GhsrKO mice exhibit a strong FAA with an advanced onset compared with the control mice after 4 days on a second RF protocol (Figure 5, C and D).

A correlation between FAA and the expression of melanocortin ligands

The response of hypothalamic melanocortin neurons expressing the endogenous melanocortin receptor ligands in Mc3rKO may be altered in situations of negative energy balance (28). We performed a pilot study analyzing the expression of orexigenic and anorexigenic neuropeptides in the hypothalamus of Mc3rKO and WT 1 hour before mealtime. After 7 days of RF in the home cage, Mc3rKO exhibited lower expression of the orexigenic neuropeptides AgRP and neuropeptide Y (Npy), and higher expression of proopiomelanocortin (Pomc) mRNA, which encodes the precursor for the anorexigenic melanocyte-stimulating hormones relative to WT controls (Figure 6A). No significant difference was observed in the expression of other neuropeptides involved in regulating feeding behavior (orexin, melanin-concentrating hormone).

This outcome suggested that the attenuated FAA observed in Mc3rKO might be due at least partially to an attenuated response of AgRP/Npy and Pomc neurons to signals of negative energy balance. We next measured expression of these transcripts in hypothalami collected 1 hour before food presentation from WT, Mc3rKO, GhsrKO, and DKO after 17–19 days of RF in LD. We were also able to verify Mc3r and Ghsr deficiency in the mutant mice (Supplemental Figure 4, A and B). A genotype effect was observed for AgRP (F_(3,30) = 3.42, P < .05) (Figure 6B) and Npy transcript levels (F_(3,30) = 2.957, P < .05) (Figure 6D) but not for Pomc transcript levels (F_(3,30) = 1.372, P > .05) (Figure 6F). Post hoc analysis revealed that DKO mice had significantly lower AgRP and NPY expression before meal presentation than control mice. No significant differences were noted in the expression of orexin and MCH (Supplemental Figure 4, C and D).

To further investigate the link between the expression of genes encoding neuropeptide that regulate feeding behaviors and the expression of FAA, we performed correlation analysis for AgRP, NPY, POMC, and FAA. A strong positive correlation was observed between the expression of AgRP and NPY transcripts and FAA (r = 0.5915, P < .001; r = 0.6418, P < .0001) (Figure 6, C and E). A weaker but still significant negative correlation (r = −0.3633, P < .05) was observed between Pomc expression and FAA (Figure 6G).

Blood chemistries during RF

We first assessed blood glucose levels at 2-hour intervals after 15 days of RF in LD. A predictable postprandial increase in blood glucose levels after food presentation was observed in WT, GhsrKO, and DKO (Supplemental Figure 5A). However, Mc3rKO mice exhibited high variability in their blood glucose levels both in terms of the levels between mice and in terms of showing high variability throughout the day, suggesting poor glycemic control (Supplemental Figure 5B). Table 1 shows the plasma levels of glucose and insulin 1 hour before scheduled meal presentation in the mice on the day of killing (after 17–19 d of RF). There was no significant effect of genotype on fasting plasma glucose (F_(3,34) = 1.164, P > .05). Insulin levels and HOMA-IR values were significantly affected by genotype (F_(3,34) = 3.049, P < .05; F_(3,34) = 3.532, P < .05). Post hoc analysis revealed that GhsrKO had lower insulin when compared with DKO (P < .05). Liver weight and plasma TG levels were not affected by genotype (F_(3,34) = 1.361, P > .05; F_(3,34) = 1.53, P > .05).