Melanin-Concentrating Hormone Receptor 1 Deficiency Increases Insulin Sensitivity in Obese Leptin-Deficient Mice Without Affecting Body Weight

MCHr1 knockout mice were generated at AstraZeneca Transgenics and Comparative Genomics as reported previously (11) and backcrossed seven generations to C57BL/6. The ob/+ mice were supplied by Harlan Olac (Blackthorn, Bicester, U.K.) fully backcrossed to C57BL/6. The mice were group housed and fed standard diet containing 4% total lipids, 18.5% protein, 55.7% carbohydrates, and 3.5% fibers (R36; Lactamin, Stockholm, Sweden) in accordance with good animal practice. After 4 weeks of age the mice were single housed. Genotyping was done by PCR. For the MCHr gene, one primer was located upstream in the short arm (5′-GAGTCCCCAGCATTGAGAAC-3′), a second primer located in the exon part (5′-AGCTCCCACTGACATCACCT-3′), and the third located in the PGK-NEO promoter (5′-AGCGCATGCTCCAGACTGCCTT-3′). The ob mutation was detected using a combined PCR and restriction digest approach utilizing the fact that the ob mutation creates a new restriction enzyme site. PCR fragments were amplified using the primers (5′-GACTTCATTCCTGGGCTTCA-3′) and (5′-ATCCAGGCTCTCTGGCTTCT-3′), and the resultant products were digested with DdeI (In Vitro Sweden, Stockholm, Sweden). The study was performed in accordance with the ethical certificate approved by the local ethical committee for animal experimentation.

In the first step of breedings, MCHr1 knockout (−/−) mice were crossed to ob/+ mice to generate offspring that were heterozygous for both MCHr1 (+/−) and the ob mutation (ob/+). Next, these compound heterozygous mice were intercrossed and MCHr1 knockout (−/−) ob/+ as well as MCHr1 wild-type (+/+) ob/+ were identified. Finally, MCHr1 knockout (−/−) ob/+ were intercrossed to produce male MCHr1 knockout ob/ob and MCHr1 knockout ob wild-type (+/+) for the study. Also, MCHr1 wild type ob/+ were intercrossed to give male MCHr1 wild-type ob/ob and MCHr1 wild-type ob wild-type (+/+) mice that were studied.

Body weights were measured weekly between 4 and 23 weeks of age. To measure food intake, cages (23 × 16 cm) were prepared with normal diet and dried at 80°C for 1 h to correct for any differences in humidity. After 12 h at room temperature the cages were accurately weighed. Mice at 20 weeks of age were fasted for 12 h, over the dark period, before being put in the preweighed cages with free access to food and water. The mice were left in the cage for 48 h and were then returned to their original cages. All excrements were collected for further analysis. The cages were reincubated 1 h at 80°C to dry out water spill and urine and reweighed after 12 h. Feces from the animals were dried at 55°C overnight and stored in airtight containers at −20°C until assayed. The gross energy content of the fecal pellets was determined using a bomb calorimeter (C 5000; IKA Werke, Staufen, Germany).

Indirect calorimetry was measured as previously described (18) in 24-week-old mice at 22°C. Measures were taken for 5 s every 9 min. Data from the first 2 h was excluded from analysis to allow for acclimatization to the novel environment. Locomotor activity was monitored both in the calorimetric chambers over 48 h and in activity boxes (Kungsbacka mät-och reglerteknik, Kungsbacka, Sweden) over 1 h, 10:00–11.00 a.m. Rectal temperatures were recorded in 26-week-old conscious mice using a rectal probe (components supplied by ELFA, Järfälla, Sweden). Temperature readings were taken 1 min before and at the following time points after the mice had been placed in a 6°C-cold room: 15, 30, 45, and 60 min.

Body fat (%) and lean body mass (g) were determined by densitometry in 25-week-old mice as previously described (19).

Mice aged 28 weeks were fasted for 12 h (12:00 a.m.to 12:00 p.m.) before oral administration of glucose solution (2 g/kg). Tail blood was collected 1 min before and 5, 15, 30, and 60 min after glucose administration. Glucose levels were measured by using an Accu-chek device and plasma calibrated test strips (Roche Diagnostics, Mannheim, Germany). Insulin levels were determined by using an ultrasensitive insulin enzyme-linked immunosorbent assay kit (Crystal Chem, Downers Grove, IL). To further investigate insulin sensitivity, a QUICKI calculation was done as follows: 1/[log(I₀)+log(G₀)], where I₀ is the fasting insulin (ng/μl) and G₀ is the fasting glucose (mmol/l). QUICKI has previously been shown to correlate well to insulin sensitivity in humans (20).

Tissues and serum were collected as previously described (21). Livers and BAT were weighed before freezing. Triglyceride content of liver biopsies were determined using a Triglyceride CP kit (ABX Diagnostics, Montepellier, France).

Corticosterone was measured using a radioimmunoassay kit (Code RPA 548; Amersham Biosciences, Uppsala, Sweden). For triglycerides and total cholesterol, a CHOD-PAP (TG/GB, no. 12146029216, Cholesterol no. 2016630; Roche Diagnostics, Mannheim, Germany) was used and nonesterified fatty acids levels were analyzed using nonesterified fatty acids C assay kit (Cat. no. 999-75406; Wako Chemicals, Neuss, Germany). The cholesterol distribution profiles were measured as previously described (19).

Total RNA extraction, cDNA synthesis and quantification was performed as previously described (21). Sequences for the primers and probes used in the Taqman PCR are presented in Table 1.

Values are presented as group means ± SE. Statistical analysis of multiple comparisons was done by using one-way ANOVA followed by Bonferroni post hoc test, and pairwise comparisons were done by using Student′s t test. In some experiments, more than two groups were investigated over time. In these cases, a type of two-way ANOVA (a standard repeated-measures model) was used with time and group as main effects and with a time and group interaction term included in the model. The within-animal correlation structure was modeled through autoregressive1 covariance structures. Group comparisons were made using simultaneous confidence intervals of Tukey type, and in some cases by post hoc Holm-Bonferroni adjustments. The analysis was conducted using the statistical software SAS PROC MIXED. The level of significance was taken as P < 0.05.

To study the importance of a functional MCHr1 for the phenotype of the obese leptin-deficient ob/ob mice, MCHr1 knockout ob/ob double-null mice were generated and compared with ob/ob littermates.

Both MCHr1 knockout ob/ob and the ob/ob mice had significantly higher body weight after 6 weeks of age compared with the wild-type mice (37 and 36% increase, respectively). The body weight of the MCHr1 knockout ob/ob and ob/ob groups were not distinguishable and did not differ statistically at any time point during the study. At 23 weeks of age, both MCHr1 knockout ob/ob and ob/ob mice had >40% higher body weight compared with the wild-type mice (Fig. 1A). At 30 weeks of age the MCHr1 knockout ob/ob and ob/ob mice were still similarly overweight (62.8 ± 2.4 g vs. 63.7 ± 2.0 g, respectively). There was no difference in food intake measured over 48 h between the MCHr1 knockout ob/ob and ob/ob mice. However, both groups consumed more diet than the wild-type mice (Fig. 1B). Energy in feces were not different between the MCHr1 knockout ob/ob, ob/ob, or wild type (data not shown).

Fasting glucose levels were not significantly different between the groups (Fig. 2A). Interestingly, fasting insulin was 49% lower in the MCHr1 knockout ob/ob mice compared with the ob/ob mice (31 ± 8 and 61 ± 9 ng/ml, respectively, Fig. 2B). QUICKI calculations revealed that the MCHr1 knockout ob/ob mice had significantly higher index compared with the ob/ob (0.46 ± 0.04 and 0.38 ± 0.01, respectively, P < 0.05). QUICKI in wild-type mice was 0.96 ± 0.05. The MCHr1 knockout ob/ob mice had significantly lower glucose levels 15 min after glucose administration when compared with the ob/ob mice (23%, P < 0.05). At time points 15 and 60 min after glucose administration, the MCHr1 knockout ob/ob had 62 and 69% lower insulin levels compared with the ob/ob mice and were significantly lower over the entire time course (P < 0.001, Fig. 2B). At 60 min after the glucose load, both groups had similar glucose levels and neither group differed from wild-type mice (Fig. 2A). Thus, the decline in glucose levels 15 min after the glucose load in ob/ob mice was significantly faster compared with the MCHr1 knockout ob/ob mice (P < 0.05). For comparison, MCHr1 knockout mice were also analyzed regarding glucose tolerance. The MCHr1 knockout mice had 43% lower fasting insulin levels compared with wild-type mice (0.7 ± 0.2 and 1.2 ± 0.1 ng/ml, respectively). The glucose administration to the MCHr1 knockout mice resulted in significantly lower serum insulin levels at all time points compared with wild-type mice (P < 0.01, Fig. 2B). The serum corticosterone levels were >40% lower in the MCHr1 knockout ob/ob mice compared with the ob/ob mice (Table 2). The expression level of corticotrophin-releasing hormone (CRH) in hypothalamus was not different between the groups (data not shown). Serum levels of triglycerides and cholesterol as well as the cholesterol distribution in the different lipoprotein classes were not different between the MCHr1 knockout ob/ob mice and ob/ob mice (Fig. 2C and Table 2).

The 48-h indirect calorimetric measurement revealed no difference in respiratory exchange ratio (RER) or energy expenditure between the MCHr1 knockout ob/ob and the ob/ob mice even though both the groups differed from wild-type mice (Fig. 3A). The spontaneous locomotor activity was measured both over 48 h in the indirect calorimetry cages (Fig. 3C) as well as over 1 h during daytime in activity boxes (Fig. 3D). The MCHr1 knockout ob/ob mice had higher locomotor activity compared with the ob/ob mice both over the 48 h time course (P < 0.001) and over 1 h during light time (P < 0.05). The body composition measurements showed that the MCHr1 knockout ob/ob mice had lower percentage of fat compared with the ob/ob groups (45 ± 1 and 51 ± 1%, respectively; Fig. 4A). Also, the MCHr1 knockout ob/ob mice had significantly more lean mass normalized for body length compared with the ob/ob mice (2.8 ± 0.05 and 2.6 ± 0.04 g/cm, respectively, Fig. 4B).

At room temperature, there were no statistically significant differences in body temperature between the groups of mice (MCHr1 knockout ob/ob 36.8 ± 0.2°C, ob/ob 36.6 ± 0.2°C, and wild type 37.2 ± 0.2°C; Fig. 5A). After 60 min at 6°C, the body temperatures of the MCHr1 knockout ob/ob mice were significantly higher compared with the ob/ob mice (36.9 ± 0.3 and 36.2 ± 0.2°C, respectively, P < 0.05, Fig. 5A). However, neither BAT weights normalized for individual lean masses nor mRNA expression of UCP-1 in BAT were significantly different between the MCHr1 knockout ob/ob and the ob/ob groups (Fig. 5B).

Liver weights normalized to lean mass and liver triglycerides content were not different between the MCHr1 knockout ob/ob and ob/ob mice (Fig. 6A). A high SCD-1 enzyme expression has been shown to be important for the phenotype of the ob/ob mice (22). Both the MCHr1 knockout ob/ob and the ob/ob mice had similar levels of SCD-1 expression, showing that the MCHr1 is not involved in the regulation of SCD-1 (Fig. 6B).

In this study, leptin-deficient (ob/ob) mice also lacking the MCHr1, were examined to investigate the importance of a functional MCHr1 for the obese phenotype of the ob/ob mice. We found that MCHr1 knockout ob/ob mice had a body weight gain, food intake, RER, energy expenditure, and lipoprotein profile comparable to the ob/ob mice. At 30 weeks of age, the MCHr1 knockout ob/ob and the ob/ob mice were both severely overweight with average body weights of >60 g. Interestingly, an oral glucose tolerance test revealed that the MCHr1 knockout ob/ob had lower glucose levels 15 min after glucose administration, lower insulin response, and lower fasting insulin levels when compared with the ob/ob mice. QUICKI, which has been previously validated as an index that correlates well with insulin sensitivity (20), was higher in the MCHr1 knockout ob/ob mice compared with the ob/ob mice. These findings clearly indicate enhanced insulin sensitivity and suggest that the MCHr1 contributes to the poor insulin sensitivity of the ob/ob mice. In addition, the MCHr1 knockout ob/ob mice had higher lean mass, lower body fat, higher locomotor activity, and a better cold tolerance than the ob/ob mice.

It is well known that obesity is the primary risk factor for insulin resistance and type 2 diabetes (rev. in 3). The MCHr1 knockout ob/ob mice had a significantly lower glucose and insulin response to an oral glucose load compared with ob/ob mice. Lower insulin levels indicate improved insulin sensitivity, suggesting a role for the MCHr1 in the regulation of insulin sensitivity. This is supported by a recent publication showing that central MCH injection to rats induced insulin resistance without affecting body weight (24). The higher lean body mass and the lower body fat may partly explain the improved insulin sensitivity. However, the marked decrease in insulin levels is probably also attributed to other physiological alterations than the small changes in body composition. It is interesting to note that, in humans, exercise improves glucose tolerance. Thus, the increased locomotor activity in MCHr1 knockout ob/ob mice may also play a part in the improved insulin sensitivity.

MCH and the MCHr1 also affect glucose metabolism in nonobese animals. MCHr1 knockout mice had lower fasting insulin and insulin response to an oral glucose challenge compared with wild-type mice in line with previous findings of low insulin levels in MCHr1 knockout mice (9,10). An additional mechanism by which MCH may affect glucose metabolism is insulin secretion. MCHr1 mRNA is present in insulin-producing B-cells, and MCH can stimulate insulin secretion (25). It has been suggested that MCH, as a neurotransmitter, plays an important role in the autonomic aspects of metabolism, including control of the pancreas (26). MCH is also present in rat plasma (14). Thus, since the MCHr1 knockout ob/ob mice have somewhat slower glucose clearance, we cannot rule out that MCHr1 deficiency in B-cells results in decreased insulin secretion in the MCHr1 knockout ob/ob mice.

Glucocorticoids are involved in glucoregulation by reduction of hepatic and peripheral insulin sensitivity (27). It is well known that leptin-deficient mice have increased glucocorticoid levels (28,29). In contrast, both MCH and neuropeptide EI can stimulate the hypothalamic-pituitary-adrenal axis and increase corticosterone levels (30). In our study, the MCHr1 knockout ob/ob mice had 40% lower corticosterone level compared with the ob/ob mice. However, the expression level of CRH was not statistically different between the groups, suggesting that the effect of MCHr1 was downstream of CRH. Lower corticosterone levels in the MCHr1 knockout ob/ob mice might contribute to the lower insulin levels and the improved glucose metabolism. The inhibitory effects of leptin on the hypothalamic-pituitary-adrenal axis may thus involve the MCHr1 and the MCH pathway, as also supported by findings in the MCH knockout ob/ob mice (31).

There is a large body of evidence stating that MCH and the MCHr1 affect body weight, food intake, and energy balance as well as leptin and insulin action (4–11,14,25). The orexigenic peptides neuropeptide Y and MCH are upregulated in the ob/ob mice, which may be an explanation for the observed hyperphagia (32,33). The MCHr1 knockout ob/ob mice used in this study were as hyperphagic as the ob/ob mice. Therefore, the contribution of these signals to the hyperphagia of ob/ob mice would appear not to require functional MCHr1. Moreover, the MCHr1 is not required as a mediator for another orexigenic peptide: ghrelin (21).

The MCHr1 knockout ob/ob mice had increased locomotor activity compared with their ob/ob littermates, which could contribute to the higher lean mass and the lower body fat. It is well established that dopamine and dopamine receptors alter locomotor activity (rev. in 34). Recently, it was shown that MCHr1 knockout mice have upregulated mesolimbic dopamine receptors and norepinephrine transporters, indicating that the MCHr1 might modulate mesolimbic monoamine functions (35). These findings might partially explain the increased locomotor activity.

The high RER indicates that the MCHr1 knockout ob/ob and the ob/ob mice mainly use carbohydrates as a source of energy while fat is stored. Hepatic fat content was high in both the obese groups. SCD-1 is involved in the biosynthesis of monounsaturated fats and is highly upregulated in ob/ob mice (22). The level of SCD-1 mRNA was not different between the MCHr1 knockout ob/ob and the ob/ob mice. This observation differs from findings in the MCH knockout ob/ob mice that had reduced SCD-1 expression compared with ob/ob mice, despite high hepatic triglycerides content (31). Thus, MCH and the MCHr1 may differ in the regulation of the SCD-1 gene.

The leptin-deficient MCH knockout and MCHr1 knockout mice also differ with respect to body growth, lean mass, insulin levels, and energy expenditure. Basal core body temperature was not different between MCHr1 knockout ob/ob and ob/ob, in contrast to findings in the MCH knockout ob/ob mice (31). As shown for MCH knockout ob/ob mice, the MCHr1 knockout ob/ob had improved thermogenesis in response to cold exposure, but the decrease in body temperature was less pronounced in the MCHr1 knockout ob/ob mice (this study and 31). The differences between the models indicate that MCH has other ways of signaling than via MCHr1. The MCH gene also encodes other neuropeptides, neuropeptide EI, neuropeptide GE, and MCH gene overprinted polypeptide (2,36). Another possible explanation is that the other neuropeptides derived after proteolytic processing from the MCH preprohormone influence these parameters via separate signaling pathways. Whether neuropeptide GE exists as a functional active peptide is not clear but neuropeptide EI has been shown to be coexpressed with MCH and alter locomotor activity (37,38), whereas MCH gene overprinted polypeptide influences somatostatin secretion (36).

Sympathetically mediated thermogenesis in BAT depends on activation of UCP-1, which generates heat by uncoupling of the proton gradient across the inner membrane in the mitochondria. Leptin administration increases sympathetic nerve activity in BAT (16), and ob/ob mice have lower body temperature (31). MCH on the other hand decreases synaptic activity in neurons from lateral hypothalamus, body temperature, and BAT UCP-1 levels whereas MCHr1 knockout mice have increased heart rate and body temperature (6,11,39). Interestingly, the MCHr1 knockout ob/ob mice had significantly higher body temperature than the ob/ob mice when exposed to a low-temperature environment. Thus, ablation of MCHr1 in the ob/ob mice ameliorates the effect of leptin deficiency on body temperature. This finding suggests that MCH and leptin have opposing effects on the sympathetic nervous system regulated thermogenesis. Both chronic intravenous and intracerebroventricular leptin infusion and overexpression of leptin result in increased arterial pressure (40–42), whereas ob/ob mice have lower arterial pressure than control mice (43). It is well known that leptin downregulates orexigenic peptides like neuropeptide Y and MCH and upregulates the anorexigenic peptide α-melanocyte-stimulating hormone (MSH). Intracerebroventricular infusion of α-MSH acutely increases mean arterial pressure and heart rate (44). Thus, since MCHr1 knockout mice have increased heart rate, leptin may affect mean arterial pressure and heart rate via MCH and α-MSH.

This study characterized the MCHr1 knockout on a leptin-deficient background, and our findings clearly show that the absence of MCHr1 affects insulin levels, locomotor activity, body temperature, body fat, and lean body mass. No statistically significant differences on food intake, body weight, energy expenditure, liver triglycerides, or serum lipids were found when compared with ob/ob mice. Our data suggests that the MCHr1 knockout ob/ob mice have improved insulin sensitivity despite being severely obese. The MCHr1 might also be important for the autonomic control of body temperature acting via sympathetic nervous system. When the MCHr1 knockout ob/ob was compared with the previously reported MCH knockout ob/ob, there were several differences including body weight, lean body mass, and energy expenditure. The reason for these differences might be attributed to other known peptides from the MCH gene or that MCH exerts its actions via multiple signaling systems in addition to the MCHr1.

We thank Harriet Andersén for her invaluable contribution in generating the MCHr1 knockout mice; Lena Amrot Fors, Anna Tuneld, Charlotte Lindgren, Anne-Cristine Carlsson, Martina Johansson, Gisela Häggblad, and Marie-Louise Berglund Zackrisson who kindly helped with the serum assays; and Johan Johansson who took care of the animals. Also, we thank Brian Middleton for invaluable help with statistical analysis.