Overexpression of melanocortin 2 receptor accessory protein 2 (MRAP2) in adult paraventricular MC4R neurons regulates energy intake and expenditure

Objective Melanocortin 2 receptor accessory protein 2 (MRAP2) has a critical role in energy homeostasis. Although MRAP2 has been shown to regulates a number of GPCRs involved in metabolism, the key neurons responsible for the phenotype of gross obesity in MRAP2 deficient animals are unclear. Furthermore, to date, all the murine MRAP2 models involve the prenatal deletion of MRAP2. Methods To target Melanocortin 4 receptor (MC4R)-expressing neurons in the hypothalamic paraventricular nucleus (PVN), we performed stereotaxic surgery using AAV to selectively overexpress MRAP2 postnatally in adult Mc4r-cre mice. We assessed energy homeostasis, glucose metabolism, core body temperature, and response to MC3R/MC4R agonist MTII. Results Mc4r-crePVN-MRAP2 female mice on a standard chow diet had less age-related weight gain and improved glucose/insulin profile compared to control Mc4r-crePVN-GFP mice. These changes were associated with a reduction in food intake and increased energy expenditure. In contrast, Mc4r-crePVN-MRAP2 male mice showed no improvement on a chow diet, but improvement of energy and glucose metabolism was observed following high fat diet (HFD) feeding. In addition, an increase in core body temperature was found in both females fed on standard chow diet and males fed on HFD. Mc4r-crePVN-MRAP2 female and male mice showed increased neuronal activation in the PVN compared to controls, with further increase in neuronal activation post MTII treatment in females. Conclusions Our data indicate a site-specific role for MRAP2 in PVN MC4R-expressing neurons in potentiating MC4R neuronal activation at baseline conditions in the regulation of food intake and energy expenditure.


INTRODUCTION
MRAP2 is predominantly expressed in the hypothalamus, in particular within the paraventricular nucleus (PVN), a region known to express MC4R and with a critical role in energy homeostasis [1e3]. Mice with global MRAP2 deletion and conditional MRAP2 deletion in SIM1 expressing neurons developed marked obesity, while rare loss-offunction or missense heterozygous variants in MRAP2 were identified in humans with severe early-onset obesity [4,5]. MRAP2's critical role in the control of energy homeostasis has been linked to action on MC4R signaling [4,6]. Further evidence that MRAP2 acts via MC4R signaling came from a study on the role of Mrap2 in zebrafish feeding and growth [7].
Although data point to an MC4R dependent function, Mrap2 À/À mice do not fully phenocopy the Mc4r À/À mice. In particular, there is the paradox that while Mrap2 À/À mice become obese without detectable changes in food intake or energy balance, Mc4r À/À mice have hyperphagia and reduced energy expenditure (EE) [8,9]. Mrap2 À/À mice remain responsive to treatment with MTII, a MC3R/MC4R agonist, while the anorexic response to MTII is abolished in Mc4r À/À mice, suggesting at least some preservation of MCR function centrally [4,10]. The phenotype of gross, early onset obesity without detectable change of food intake and energy expenditure, replicated in an independent Mrap2 deficient model, is particularly intriguing [4,6]. However, the mechanism by which MRAP2 knockout animals become obese is still unclear. Plasma corticosterone, thyroid function, faecal energy measurements, body temperature, and brown fat function in response to 1 cold challenge were all indistinguishable between Mrap2 À/À and Mrap2 þ/þ mice [4,6]. Some of the complexity may arise from the fact that MRAP2 is a promiscuous accessory protein [1]. In addition to the melanocortin receptor family (MCR) [11], MRAP2 interacts and regulates other G Protein-Coupled Receptors (GPCRs) beyond the MCR family [1,12e14]. Interaction with Prokineticin 1 receptor, orexin 1 receptor, and Ghrelin receptor have been reported [12e14]. However, these interactions would result in a lean phenotype in the absence of MRAP2. Thus, while interaction with these additional GPCRs forms part of the growing understanding of MRAP2 action in the neuronal control of energy homeostasis, it does not explain the obese phenotype of MRAP2 deficient animals. With data pointing to the PVN and MC4R as the key to unraveling the obesity phenotype in MRAP2 deficient mice and because all existing mouse models to date involve developmental deletion of MRAP2, we undertook this study to assess the effect of postnatal overexpression of MRAP2 in MC4R neurons of the PVN. Furthermore, using this methodology we are able to exclude the effects of MRAP2 interaction with non-PVN MC4R GPCRs, in particular those GPCRs that have been described to interact with MRAP2 in the arcuate nucleus [12e14].

Ethics and animal husbandry
All animal studies on male and female mice were approved by Yale University Institutional Animal Care and Use Committee. Animals had free access to water and food was provided ad libitum, unless otherwise stated. Mice were fed on a standard chow diet (SD) (diet no. 2018; 18% calories from fat; Teklad Diets, Harlan Laboratories) for up to 6 months of age. For high fat diet (HFD) experiments, male mice were exposed to HFD (category no. 93075; 45% fat; Teklad Diets, Harlan Laboratories) starting at 2 weeks after PVN injection. All animals were kept in temperature and humidity-controlled rooms, in a 12 h dark and 12 h light cycle, with lights on from 7:00 AM to 7:00 PM.

Generation of Mc4r-t2a-cre mice and genotyping
Mc4r-t2a-cre mice were generated as previously described [15]. Genomic DNA was isolated from tails or yolk sacs by standard methods. Mc4r-t2a-cre mice were genotyped by polymerase chain reaction (PCR parameters: 42 cycles, 93 C for 30 s, 56 C for 1 min, and 72 C for 5 min). Amplification of a wild-type (WT) allele generated a 3.6-kb product, and a 4.1-kb product in the case of a mutant allele using the following primers: cre 350 FRT5:

Stereotaxic viral injection of Adeno-associated virus (AAV) into the PVN
The AAV2-CMV-DIO-GFP [AAV-DIO-GFP] and AAV2-CMV-DIO-MRAP2-FLAG [AAV-DIO-MRAP2] virus (VECTOR BIOLABS) were injected bilaterally into the PVN of Mc4r-cre animals. Moreover AAV2-CMV-GFP (VECTOR BIOLABS cat No. 7004) virus was co-injected to determine successful PVN targeting. Eight weeks old Mc4r-cre mice were anesthetized with 100 mg/kg ketamine and 10 mg/kg xylazine (IP) and placed in stereotaxic apparatus. A guide cannula with dummy cannula (Plastics one, Roanoke, VA) was inserted into the PVN according to the atlas of Franklin and Paxinos (Franklin KBJ) (co-ordinates, bregma: anterior-posterior À0.7 mm; lateral AE0.2 mm; and dorsale ventral À5.2 mm), virus was infused at a rate of 40 nl/min (w2 Â 10 12 viral particles/mL) for 15 min and the injector (Plastics one) remained in place for an additional 5 min before. The injector was connected with a Hamilton syringe and infusion was administered at a rate of 33.3 nL/min.

GFP immunostaining
Fed mice were perfused and brains were processed for immunofluorescence staining to confirm the injection site in the PVN, using anti-GFP antibody (ab13970; Abcam). Mice in which viral injections were located outside the PVN were studied separately. The sections were incubated overnight in anti-GFP antibody (diluted 1:5000 in 0.1 mol/L sodium phosphate buffer) and then incubated in secondary antibody (category no. A11039, Alexa Fluor 488ecoupled goat anti-chicken, 1:500 dilution; Life Technologies) for 2 h. Sections were then analyzed with fluorescent microscope.

FLAG immunostaining
Immunofluorescence staining was performed using anti-FLAG antibody (F1804; Sigma). Brains were sectioned with a vibratome, and sections were incubated for 24 h at room temperature in anti-FLAG antibody (diluted 1:2000). After several washes with phosphate buffer (PB), sections were incubated in secondary antibody (category no. BA 2000 biotinylated horse anti-mouse IgG; 1:200 in PB; Vector Laboratories) for 2 h at room temperature, then rinsed in PB three times 10 min each time, and incubated for 2 h at room temperature with Alexa Fluor 594 streptavidin (Life Technologies, 1:2000 in PB). Sections were mounted on glass slide with vectashield (Vector lab) and analyzed with fluorescent microscope.
2.6. Indirect calorimetry system and body composition Body weight was measured weekly after stereotaxic surgery. Body composition was measured in vivo by MRI (EchoMRI; Echo Medical Systems, Houston, TX). Twelve weeks after PVN injection male and female mice were acclimated in metabolic chambers (TSE Systems, Germany) for 4 days before the start of the recordings. Animals were continuously recorded for 3 days with the following measurements being taken every 30 min. Measurements include food intake, locomotor activity (in X and Y axes), and gas exchange (O 2 and CO 2 ) every 30 min using the TSE LabMaster System. Respiratory exchange ratio (RER) was calculated as a ratio of CO 2 production and O 2 consumption. Energy expenditure (EE) was calculated according to the manufacturer's guidelines (PhenoMaster Software, TSE Systems) and analyzed relative to body weight using ANCOVA analysis [16]. Food intake was determined continuously using weighing sensors integrated within the sealed cage.

Glucose and insulin tolerance tests
Glucose tolerance test (GTT) was performed using 2 mg/kg glucose in saline (DeltaSelect) given intraperitoneally (IP) to 16 h fasted animals as previously described [17]. Blood glucose levels were then monitored at 15, 30, 60, and 120 min from the injection. Insulin tolerance test (ITT) was performed using an insulin dose of 0.75 U/kg (Actrapid; Novo Nordisk A/S Denmark) delivered by IP in mice fed ad libitum. Blood glucose was measured before IP injection and at 15, 30, 60, and 120 min after insulin injection.

cfos immunostaining
Ad libitum fed mice were anesthetized and transcardially perfused with 0.9% saline with heparin followed by 4% paraformaldehyde. In another set of experiments, mice were injected with either MTII (200 nM, IP) or equal volume saline. Animals were perfused 1 h post injection. Brains were collected and post-fixed overnight before several sections of the entire hypothalamus were taken at every 50 mm. Sections were washed and incubated with the goat anti-cfos antibody (Santacruz, 1:2000), and rabbit anti-POMC antibody (Phoenix Pharmaceuticals, 1:2000) in PB containing 4% normal donkey serum, and 1% Triton X-100 for 24 h at room temperature. After several washes with phosphate buffer (PB), sections were incubated in the secondary antibodies (biotinylated donkey anti-goat immunoglobulin G [IgG]; 1:200 in PB; Vector Laboratories and donkey anti-rabbit Alexa-fluor 488; 1:500 in PB; Life Technologies) for 2 h at room temperature, then rinsed in PB five times, 10 min each time. Sections were then mounted on glass slide with Vecta-Shield antifade (Vector Laboratories). Fluorescent images of five to seven brain sections were captured with fluorescent microscope and analyzed by imaging Software (Image J). 2.10. Statistical analysis All statistical analysis was performed using GraphPad Prism. Data is plotted as mean AE S.E.M. Student's t test was used to compare two groups; for more than two groups one-way ANOVA was performed followed by Bonferroni multiple comparison test. In all analyses, a twotailed probability of <5% (i.e., P < 0.05) was considered statistically significant.  [15]. Moreover, to assess the overexpression of MRAP2 in the PVN, immunostaining for FLAG epitope was performed in female ( Figure S1C) and in male Mc4r-cre PVN-MRAP2 mice ( Figure S1F). To determine whether selective MRAP2 overexpression in PVN MC4R affects metabolism, metabolic analyses were performed in both male and female mice. Mc4r-cre PVN-MRAP2 female mice fed on a standard chow diet showed lower body weight (n ¼ 11 per group; Figure 1A) starting at 3 weeks from the viral injections compared with Mc4rcre PVN-GFP controls. The lower body weight was due to a significant reduction of fat mass (4.49 AE 0.61 g n ¼ 11; Figure 1B) in Mc4rcre PVN-MRAP2 female mice compared to female controls (7.80 AE 0.51 g Mc4r-cre PVN-GFP female mice; n ¼ 11) evidenced after 4 weeks post viral injections. No differences in lean mass were observed between the 2 experimental groups (n ¼ 11; Figure 1C). This was associated with decreased food intake (2.79 AE 0.  Figure 3A, E). Furthermore, when MTII, a MC3R/MC4R agonist, was peripherally injected, no difference in food intake ( Figure 3F) and cfos activation in the PVN (Figure 3CeE Figure 5A) that was accompanied by greater BAT UCP1 ( Figure 5B) and Dio2 ( Figure 5C) mRNA levels. Finally, to test that the observed improved metabolic phenotype was not due to a differential activation of POMC neurons, we quantified  Figure S6D and E), and found no difference between Mc4rcre PVN-MRAP2 and Mc4r-cre PVN-GFP male mice fed on HFD. Altogether these data indicate that overexpression of MRAP2 in MC4R neurons affects energy metabolism in male mice when challenged on HFD feeding.   Figure 6F). Interestingly, unlike in females, no statistical difference in cfos positive cells was observed between Mc4rcre PVN-MRAP2 male mice treated with saline compared to Mc4r-cre PVN-MRAP2 mice treated with MTII. Together these data suggest a role for MRAP2 in PVN MC4R-expressing neurons in potentiating neuronal activation within the PVN at baseline, and in males without further increase after MTII administration.

DISCUSSION
MRAP2 has been shown to have a critical role in mammalian metabolism [4]. Mice deficient in MRAP2 have severe early-onset obesity due to increased fat mass. The mechanism of how MRAP2 knockout animals become obese without detectable changes in food intake and energy expenditure remains unclear [1,3]. Deletion of MRAP2 in Sim-1 expressing neurons leads to obesity in mice to a similar extent as in global Mrap2 À/À animals pointing to these neurons, located also in the PVN, as key targets of MRAP2 action in metabolism regulation [4]. Although data suggest the involvement of MC4R signaling, the difference in phenotype between the Mrap2 À/À and Mc4r À/À mice  suggest possible non-MCR modes of action [4,6]. This has since been shown to be the case. In addition to MC4R positive neurons, MRAP2 has been shown to have a broader distribution in the Central Nervous System [4,12e14,18]. Within the hypothalamus, MRAP2 has been shown to interact and regulate other GCPRs, including ghrelin receptors expressed in the arcuate Neuropeptide Y/Agouti-related (NPY/ AgRP)-expressing neurons, where it positively regulates hunger signaling [13], adding to the complexity of the system. As Sim1-Cre mice express cre activity in sites outside the PVN [9,19] and because of the critical role of PVN MC4R in the regulation of metabolism and the changes in neuropeptide transcripts observed in the PVN of global Mrap2 À/À animals [6], we focused our study on investigating the role of MRAP2 in the MC4R-expressing neurons of the PVN. Our data provide evidence that MRAP2 in MC4R-expressing neurons of the PVN represents an important regulator of food intake and energy metabolism. Furthermore, the data indicate that postnatal manipulation of MRAP2 leads to changes in weight phenotype that are opposite of those observed in Mrap2 À/À mice in which MRAP2 is deleted during development. By selectively overexpressing MRAP2 in MC4R neurons of the PVN, we reveal a reduction in food intake and energy expenditure that supports the observed lean phenotype, which is unlike that of Mrap2 À/À in which no change in either measures were identified [4,6]. This difference could be due to action of MRAP2 on other GPCRs in the arcuate nucleus that leads to lean phenotype in the absence of MRAP2 [12e14]. In addition, unlike data on the global Mrap2-deficient mice that showed no change in BAT activation or body temperature phenotype [4,6], Mc4r-cre PVN-MRAP2 mice showed a significant increase in BAT activation with increased UCP1 and Dio2 mRNA levels that were associated with increased body temperature compared to control mice in both genders. We also found a sex-specific metabolic difference in the Mc4rcre PVN-MRAP2 mice compared with Mc4r-cre PVN-GFP matched control.
Whilst female mice overexpressing MRAP2 in PVN MC4R neurons demonstrate overt protection against obesity on a chow diet, this phenotype was only observed in male mice after challenge with high fat diet. Similarly, sex differences were also observed in locomotor activity and glucose handling. We have previously described a sexspecific increase in daytime locomotor activity and exploratory activity in global Mrap2-deficient mice [6]. However, when taken in isolation, focusing on MRAP2-overexpression in PVN MC4R neurons, we now demonstrate an increase in locomotor activity in both female and male Mc4r-cre PVN-MRAP2 mice compared to controls. A difference in glucose clearance and hyperinsulinemia was demonstrated in Mrap2 À/ À mice on a C57/BL6N background, while on a 129/Sv genetic background no changes in insulin and glucose handling were found [4,6]. Others have confirmed a glucose metabolism phenotype in Mrap2-deficient mice on a C57/BL6N background [14]. In agreement with previous work, here we found differences in glucose handling when MRAP2 was manipulated. However, our data suggest that the improved glucose handling was rather a consequence of the leaner phenotype of the female Mc4r-cre PVN-MRAP2 mice. Indeed, no significant differences were observed in male mice on a chow diet. To determine a possible role for POMC neurons in the regulation of PVN MC4R neurons, we then assessed the activation levels of POMC neurons by cfos immunostaining. No difference in POMC neuronal activation was observed between Mc4r-cre PVN-MRAP2 and controls. Furthermore, no differences were found in POMC fiber staining in the PVN of controls and Mc4r-cre PVN-MRAP2 mice. This suggests that the metabolic phenotype observed might be independent from changes in POMC neuronal activation, thus pointing to the increased neuronal activation in the PVN as the principal cause of the lean phenotype in the Mc4r-cre PVN-MRAP2 model as we observe a significant increase in cfos staining in the PVN of Mc4r-cre PVN-MRAP2 mice compared to controls irrespective of sex. To assess the hypothesis that MRAP2 overexpression in MC4R neurons affect MC4R signaling, thus affecting food intake, we then administrated MTII, a MC3R/MC4R agonist, in Mc4r-cre PVN-MRAP2 and control mice. MTII induced a significant increase in PVN neuronal activation and a reduction in feeding in male and female Mc4r-cre PVN-GFP mice. In female Mc4r-cre PVN-MRAP2 mice MTII treatment results in further increase in PVN neuronal activation and decrease in food intake. At baseline, MRAP2 overexpression drives neuronal activation in the PVN in both males and females. However, response to MTII, differed between genders; whilst female Mc4r-cre PVN-MRAP2 mice maintained responsiveness to MTII this was lacking in male Mc4r-cre PVN-MRAP2 mice. Altogether these results indicate that MRAP2 in PVN MC4R neurons can be manipulated postnatally to result in a lean phenotype in which feeding is reduced and energy expenditure increased along with core body temperature. Importantly, this leads to an increase in neuronal activation with in the PVN. As the number of PVN cfos positive cells are in excess of the number of FLAG positive Mc4r-cre MRAP2 overexpressing cells, the increased cfos would suggest a more global effect on PVN neuronal activation beyond MC4R expressing neurons. The likelihood is that the action of MRAP2 on MC4R PVN neurons is due to the action on enhancing MC4R function as much of the phenotype in Mc4r-cre PVN-MRAP2 mice correlates to models of MC4R activation in which reduced food intake, increased energy expenditure and thermogenesis have been described [9,15]. Some other features such as sexual dimorphism of responses have not been seen in MC4R activation models. These differences could be due to estrogenic effects on MRAP2/MC4R interaction, which has been described with MRAP1 and MC2R [20,21], although action on other GPCRs expressed in MC4R expressing PVN neurons cannot be excluded.

CONCLUSION
In conclusion, our data provides the first evidence that MRAP2 acts postnatally, in a sex-specific manner, to play a role in the regulation of food intake and energy expenditure through the enhancement of MC4R neuronal signaling in the paraventricular nucleus of hypothalamus.

AUTHOR CONTRIBUTION
GB, JDK, and LFC conducted experiments, acquired and analyzed data. LFC and SD designed the research studies and analyzed data. GB, LFC, and SD wrote the manuscript.