In vitro Radiopharmaceutical Evidence for MCHR1 Binding Sites in Murine Brown Adipocytes

[11C]SNAP-7941 and its radiofluorinated, fluoro-ethyl derivative [18F]FE@SNAP have been developed as the first positron emission tomography tracers for melanin-concentrating hormone receptor 1 (MCHR1) imaging. Accumulation of these MCHR1 PET-tracers in rat brown adipose tissue (BAT) in vivo provided first indication of MCHR1 expression in rodent BAT. To rule out off-target binding, affinity of both MCHR1 ligands toward adrenergic beta-3 receptors (ADRB3) was examined. Further, specific binding of [11C]SNAP-7941 to brown adipocytes and effects of MCHR1 ligands on brown adipocyte activation were investigated. SNAP-7941 and FE@SNAP evinced to be highly selective toward MCHR1. [11C]SNAP-7941 binding to brown adipocytes was shown to be mainly MCHR1-specific. This data strongly indicates MCHR1 expression in rodent BAT and moreover, a peripheral, anti-obesity effect of MCHR1 antagonists directly exerted in BAT is proposed. Moreover, MCHR1 expression in murine brown adipocytes was confirmed by protein and mRNA analysis. We conclude that MCHR1 PET imaging contributes to basic research in endocrinology by elucidating the involvement of the MCH system in peripheral tissues, such as BAT.


INTRODUCTION
Although the physiological relevance of brown adipose tissue (BAT) was underestimated for a long time, it nowadays attracts scientific interest as a therapeutic target for obesity and related metabolic diseases such as type 2 diabetes (1)(2)(3). The groundbreaking discovery of functional BAT depots in adult humans contributed significantly to the re-introduction of BAT in metabolic research and was enabled by positron emission tomography (PET) using the glucose analog 2-[ 18 F]fluoro-2-deoxy-D-glucose ([ 18 F]FDG) combined with computer tomography (CT). The fact that BAT activity, quantified as [ 18 F]FDG uptake, was negatively correlated to body mass index and percent body fat indicated a potential role of BAT in energy expenditure in humans (4)(5)(6). PET is a sensitive and non-invasive in vivo imaging technique that directly visualizes different molecular interactions or metabolic processes depending on the used radiolabeled tracer in a target tissue. Thus, besides [ 18 F]FDG, also other PET-tracers were investigated to study BAT function, namely [ 11 C]meta-hydroxyephedrine, [ 11 C]acetate, 14-[ 18 F]fluoro-6-thiaheptadecanoic acid, and (S,S)-O-[methyl- 11 C] methylreboxetine thereby enhancing the understanding of BAT activation, control, and metabolism (7)(8)(9).
Evolutionary, BAT functions as an energy dissipating organ by exerting non-shivering thermogenesis and thus maintaining body temperature during cold exposure. This heat-producing process is controlled by the sympathetic nervous system leading to noradrenaline release followed by activation of adrenergic beta-3 receptors (ADRB3) in brown adipocytes. Subsequent lipolysis or increased uptake of glucose and its conversion into fatty acids finally lead to the uncoupling of ATP synthesis via uncoupling protein 1 (UCP-1) resulting in increased fatty acid oxidation and dissipation of excess energy as heat. Thereby, BAT regulates fuel metabolism by increasing glucose and triglyceride uptake, decreasing blood glucose and lipids, and by serving as glycogen storage (1,(10)(11)(12)(13)(14).
Obesity results from a chronic energy imbalance, when food intake exceeds total body energy expenditure (15). The melanin-concentrating hormone, a neuropeptide predominantly expressed in the lateral hypothalamus and zona incerta, is involved in the control of appetite and food intake (16,17). In fact, upregulated MCH expression was found in the hypothalamus of obese and leptin-deficient mice and is moreover induced by fasting in wild-type mice (16). Accordingly, MCHdeficient mice are lean due to hypophagia and have an increased metabolic rate (18). Besides, MCH was shown to stimulate leptin secretion in rat white adipocytes and MCH was detected in rat plasma (19). In rodents, MCH exerts its effects solely by stimulation of the melanin-concentrating hormone receptor 1 (MCHR1), as rodents do not express melanin-concentrating hormone receptor 2 (MCHR2). Several centrally active MCHR1 antagonists have been developed for the treatment of obesity (20). In a diet-induced obesity mouse model it was shown that the anti-obesity effects of the tested MCHR1 antagonist are not only due to suppression of feeding, but also to a stimulation of energy expenditure. A significantly increased body temperature in MCHR1 antagonist-treated mice suggested a potential involvement of the MCH system in the regulation of energy expenditure via BAT (21). It was reported that a large proportion of neurons in the lateral hypothalamus projecting to BAT contain MCH (22). Thus, a central effect of the MCHR1 antagonist and subsequent transmission to BAT was presumed, as a direct effect on BAT could not be shown (21).
[ 11 (30). The discrepancy in anesthetized and conscious rats suggests a potential influence of the applied anesthesia on µPET acquisition.
Based on these in vivo findings, MCHR1-selectivity of FE@SNAP and SNAP-7941 has to be proven to avoid misleading interpretation of PET imaging data. To evade molecular alterations caused by anesthesia, we decided in favor of an in vitro approach. Therefore, in this preclinical in vitro study, affinity of both ligands toward the ADRB3, which is the receptor predominantly involved in BAT activation, was determined. Moreover, the potential involvement of the MCHR1 in BAT was investigated using brown adipocytes and the respective PET-tracer [ 11 C]SNAP-7941 and additionally, [ 18 F]FDG as a surrogate marker for brown adipocyte activity. Within the scope of this in vitro study, we aimed at applying conventional binding assays and kinetic radioligand binding measurements to cover binding interactions with single receptor sites as well as whole cell binding studies.

Cell Membrane Preparation
CHO-K1-ADRB3 cell membranes were prepared from 175 cm 2 cell culture flasks, when 80% confluence was reached. All procedures were performed at 4 • C. Cells were scraped off using ice-cold lysis buffer (10 mM Tris-HCl pH 7.4, 1 mM EDTA) and protease inhibitor cocktail was added according to the manufacturer's instructions (Sigma-Aldrich, St. Louis, USA). The cell suspension was homogenized using cannulas (29 G) and centrifuged for 10 min at 1,000 × g. The obtained supernatant was centrifuged for 30 min at 100,000 × g. Subsequently, the pellet was suspended using 50 mM Tris-HCl pH 7.4 and aliquoted. Aliquots were shock-frozen in liquid nitrogen and stored at −80 • C until usage. Protein concentration of the cell membrane suspension was determined using BCA Protein Assay Kit (Thermo Scientific, Waltham, USA).

Competitive Binding Studies: Affinity Determination
All stated concentrations refer to the final assay volume of 500 µL buffer (25 mM HEPES pH 7.4, 1 mM EDTA, 0.5% BSA). Competitive binding assays were performed using 13.6 ng/µL ADRB3 expressing CHO-K1 cell membrane suspension (cf. Total binding was determined in the presence of 1% respective solvent. Incubation was performed for 90 min at 27 • C. Filtration through GF/B filters (Whatman R , presoaked in 0.05% PEI) was performed using a cell harvester (Brandel R , Gaithersburg, MD, USA) and was followed by two washing steps using ice-cold wash buffer (10 mM HEPES, 500 mM NaCl, pH 7.4). Filter pieces containing receptor-bound radioactivity were shaken for 30 min in scintillation cocktail (Ultima Gold TM , PerkinElmer, Waltham, USA) before liquid scintillation counting (Hidex 300 SL, Turku, Finland). IC 50 fitted binding curves were generated using the GraphPad Kinetic Measurements: Displacement of [ 125 I]Iodocyanopindolol Binding 2.5 × 10 5 cells (CHO-K1, CHO-K1-ADRB3) per cell culture dish were seeded 2 days prior to the experiment and incubated in an inclined position to allow an adherent cell pole at one side of the cell dish. Cell dish preparation was performed as previously described (32,33). Real-time kinetic radioligand binding measurements were performed using a dedicated device for low energy gamma radiation (LigandTracer R Gray Technology, Ridgeview, Sweden). The assay protocol compromised consecutive radioactivity measurements of the target region (cell pole) and of the opposite pole of the petri dish, where no cells were seeded (background signal). Radioactivity was counted in each region for 3 s with a delay of 2 s over the time course of the experiment. Raw counts per second (cps) of the target region were corrected for background signal and for radioactive decay. Displacement of [ 125 I]Iodocyanopindolol binding was conducted after 1 h association phase of radioligand binding by consecutively applying rising concentrations of SNAP-7941 and FE@SNAP. The same volume of ethanol was added as vehicle control. Experiments were performed on at least three different days.

Kinetic Measurements: Glycolytic Activity
Kinetic measurements of [ 18 F]FDG uptake were performed in pre-adipocytes and differentiated brown adipocytes to investigate metabolic activity. Cell seeding and the assay protocol followed a standard procedure described above (cf. 2.5). One hour prior to the experiment, post-differentiation medium was discarded and unsupplemented, glucose-free DMEM was added, followed by incubation with [ 18 F]FDG (150 kBq per dish) for another 50 min. As glucose concentration is high in cell culture media thus interfering with [ 18 F]FDG uptake, cells were starved to achieve reasonable signal. Experiments were performed on at least three different days.

[ 18 F]FDG Uptake Experiments: Effects of MCHR1 Ligands on Brown Adipocyte Activation
Brown pre-adipocytes were seeded, grown to confluence and differentiated in 6-well plates as described in section General. Cell starving was performed for 1 h using unsupplemented, glucosefree DMEM. Thirty minutes prior to [ 18 F]FDG incubation, cells were pre-treated with the adrenergic beta-3 agonist (2 µM CL 316,243), adrenergic beta antagonist (2 µM (S)-propranolol), the MCHR1 ligands (20 µM SNAP-7941 or FE@SNAP), or vehicle (baseline). [ 18 F]FDG uptake was performed for 50 min at 37 • C (humidified atmosphere, 5% CO 2 ). Supernatant was taken off, cells were washed with ice-cold PBS and finally trypsinized. Radioactive cell fractions were gamma counted (2480 Wizard 2 , PerkinElmer) and normalized to percentage uptake per well. [ 18 F]FDG uptake under baseline conditions refers to 100% uptake. Three independent experiments were performed in triplicates. Statistical analysis was performed using an unpaired, two-tailed t-test.

Investigation of Specific [ 11 C]SNAP-7941 Binding to Brown Adipocytes
Brown pre-adipocytes were seeded, grown to confluence and differentiated in six-well plates as described above. Thirty minutes prior to the experiment, maintenance medium (cf. 2.1) was discarded and replaced with respective serum-and additivefree medium to avoid plasma protein binding of the PET-tracer. Cell starving, as performed for [ 18

MCHR1 Expression in Murine Brown Adipocytes
Two 6 weeks old male BALB/c mice were sacrificed and organs (brain, lung, and spleen) were removed to serve as reference tissues in Western Blot and real-time quantitative PCR (qPCR) experiments. Organs were divided and tissue samples were subjected to protein extraction and RNA extraction. Murine brown adipocytes were treated as described above (cf. 2.2) and both protein and RNA were extracted from identical subcultures on day 5 and 6 after induction.
RNA extraction of murine brown adipocytes and reference tissues was performed with TRIzol reagent (ambion R ). In the case of the adipocytes, an additional centrifugation step ensured for the separation of the fatty layer. RNA concentration and purity was measured using NanoDrop (Thermo Fisher Scientific). Reverse transcription was carried out with the qScript R cDNA Synthesis Kit (Quantabio). Subsequently, twostep real-time qPCR was run on CFX96 TM Real-Time System (Bio-Rad). All reactions were performed in duplicates, using the Luna R Universal qPCR Kit (BioLabs). ß-actin or Rpl27 were used as endogenous controls. Sequences of the primers (eurofins) are noted in the Supplementary Material. Data was analyzed with the corresponding CFX Manager 3.1 (Bio-Rad) and Excel 2013 (Microsoft R Office). Mean Ct values of all four qPCR runs were generated and relative mRNA expression was calculated with the formula log (2 Ct ), compared to negative control (spleen).
Protein extraction was performed using commercially available radioimmunoprecipitation assay (RIPA) buffer. Protease inhibitor cocktail was added according to the manufacturer's instructions. Cell and tissue lysates were centrifuged until a clear (lipid-free) supernatant was obtained (12,000 rpm, 4 • C). Protein concentration was determined using a bicinchoninic acid kit (BCA kit, Thermo Scientific) and 20 µg protein were loaded each. After gel electrophoreses (TGX TM precast gels, Bio-Rad, Laboratories, Inc.) and subsequent semi-dry blotting, Ponceau S staining was performed to ensure equal protein loading and transfer.
Frontiers in Endocrinology | www.frontiersin.org  A second application of 20 µM MCHR1 ligand after 120 min led to a decrease in cell-associated radioactive signal, which was more pronounced after a third addition of unlabeled compound (Figure 1)

Histological Staining and Metabolic Activity of Brown Adipocytes
Morphological characterization of brown adipocytes was performed via Oil Red O staining. Mature brown adipocytes were obtained 4-6 days after induction. Fully differentiated brown adipocytes showed typical morphologic characteristics (spherical shape, multilocular lipid vesicles, central nucleus, Figure 2) and were thus visually distinguishable from preadipocytes (fibroblast-like shape  (Figure 4).

MCHR1 Expression in Murine Brown Adipocytes
In addition to radioligand binding, Mchr1 mRNA and MCHR1 protein were determined in murine brown adipocytes. Highest Mchr1 mRNA levels were found in mouse brain (positive control, mean Ct 26.4), whereas low expression was detected in mouse lung (mean Ct 35.5) and murine brown adipocytes (mean Ct 33.1). No Mchr1 mRNA expression was found in mouse spleen (mean Ct 38.5) that served as a negative control (Figure 6). A representative figure of the melt curves is included in the Supplementary Material. In line with this, MCHR1 protein was detected in mouse brain, lung and murine brown adipocytes. No MCHR1 protein expression was found in mouse spleen (Figure 7).

DISCUSSION
For affinity determination toward ADRB3, a competitive binding assay using CHO-K1-ADRB3 cell membranes was established and tested using known ADRB3 ligands. The obtained K i values for (S)-propranolol and pindolol differed slightly from literature due to different experimental settings (applied radioligand, whole cells vs. membranes, etc.), which emphasizes the importance of testing reference compounds for distinct experimental settings. Furthermore, whole cell binding experiments using mature brown adipocytes were performed to investigate the complexity of binding interactions apart from binding to a single receptor site, as in membrane preparations of CHO-K1-ADRB3 cells. Generally, experiments using whole cells account for unspecific processes including diffusion, nonspecific, and offtarget binding-processes that are often overlooked when applying cell membranes. Thus, pre-adipocytes were cultivated and differentiated to mature brown adipocytes. Morphologic (histological staining of lipids) and metabolic ([ 18 F]FDG uptake measurements) characteristics of the BAT phenotype were achieved 4-6 days after induction, thus, experiments using brown adipocytes were performed in this time frame. [ 18 F]FDG uptake by mature brown adipocytes followed linear kinetics mirroring the unique trapping mechanism of [ 18 F]FDG: it is transported via glucose transporters and intracellularly phosphorylated by hexokinase leading to trapping of the respective metabolite [ 18 F]FDG-6-phosphate, as further metabolism is hindered (34)(35)(36). [ 18 F]FDG uptake by undifferentiated pre-adipocytes was only minor indicating fibroblast-like functionality.
ADRB3 stimulation by the agonist CL 316,243 in brown adipocytes led to enhanced [ 18 F]FDG uptake demonstrating brown adipocyte activation. This experiment served as the positive control, as this phenomenon is well described for brown adipose tissue in rodents (37,38). Accordingly, the adrenergic beta receptor antagonist (S)-propranolol exerted no BAT activating effect. These results prove the expression of functional ADRB3 on the cultivated brown adipocytes used within this study. Interestingly, opposite effects (significant decrease in [ 18 F]FDG uptake) are observed for pharmacological doses of SNAP-7941 and FE@SNAP. The role of MCHR1 in brown adipocytes, especially the involvement in glycolytic activity and glucose transporter expression is yet unexplored.
As accumulation of the MCHR1 PET-tracers was found in brown adipose tissue of naïve rats, specific binding to brown adipocytes was investigated in vitro. In competition binding experiments, the MCHR1 antagonist led to ∼23% blocking of [ 11 C]SNAP-7941 accumulation evincing specific binding to brown adipocytes. This finding points at MCHR1 expression in brown adipocytes and a subsequent involvement of the melaninconcentrating hormone system in brown fat depots. Western Blot analysis and qPCR finally showed the expression of the MCHR1 in murine brown adipocytes. In addition, as MCHR1 expression in lung is often discussed, we included mouse lung into our analysis. Our qPCR confirmed the abundance of Mchr1 mRNA in mouse lung, which is in accordance with Kokkotou et al. (39). Furthermore, protein expression was detected in Western Blot analysis. Since neither qPCR nor Western Blot showed MCHR1 expression in spleen, we suggest it to be a legitimate negative control. This was further described by Kokkotou et al. (39). Although reference tissues were harvested from only two individuals, qPCR and Western Blot data are considered as representative. Further experiments are required to determine MCHR1 expression in rodent BAT with respect to age, sex, and diet.  In general, our findings support the initial theory of Ito et al. assuming that the anti-obesity effect of MCHR1 antagonists is not limited to a central action, but MCHR1 antagonists may also act peripherally in BAT (21). Expression of functional MCHR1 in rat white adipocytes has previously been demonstrated and a peripheral role for MCH in adipocytes in addition to its centrally mediated effects was already proposed by Bradley et al. (19,40). Both, MCH and MCHR1 were found to be expressed in mouse and human pancreatic islets, thus, an autocrine role for MCH in the regulation of the hypothalamic-pancreatic axis was proposed (41). However, it was recently stated by Naufahu et al. that it is unknown whether MCH acts in a paracrine or autocrine manner or is released into the circulation (42). [ 11 C]SNAP-7941 binding to brown adipocytes was blocked to a minor proportion by both the selective ADRB3 agonist CL 316,243 and the non-selective antagonist (S)-propranolol, demonstrating the superior affinity of [ 11 C]SNAP-7941 toward MCHR1 compared to ADRB3. As (S)-propranolol displays even higher affinity for the adrenergic beta-1 and beta-2 receptor, [ 11 C]SNAP-7941 binding to adrenergic beta receptor subtypes other than beta-3 can additionally be excluded (43). However, relatively high non-displaceable binding was found, which may be due to unspecific accumulation in lipids within brown adipocytes.
Moreover, these in vitro results are in accordance with the previously performed in vivo experiments administering [ 18 F]FE@SNAP to conscious animals without the use of anesthesia, where significant blocking was achieved by a pharmacological dose SNAP-7941 (30). Contradictory results were obtained from µPET imaging, where binding enhancement of [ 11 C]SNAP-7941 in BAT was shown after administration of a pharmacological dose of the unlabeled compound. PET imaging in small animals requires anesthesia in order to prevent laboratory animals from moving during the examination. However, anesthesia bears the disadvantage of possible molecular alterations, which may affect the examination's outcome, as it is already well-established for brain PET imaging (44). Effects of isoflurane, the used anesthetic agent within the mentioned in vivo study, on ADRB3 and/or MCHR1 expression are not described, though potential effects cannot be excluded.
Investigation of specific binding to brown adipocytes was solely performed for [ 11 C]SNAP-7941, as extensive in vitro and in vivo investigations evinced the superior properties of [ 11 C]SNAP-7941 compared to the fluoroethylated analog regarding radiochemical yield, metabolic stability, target affinity, selectivity, and imaging contrast (29). Thus, [ 18 F]FE@SNAP was not further evaluated in this context.
The value of PET for BAT research was ultimately recognized when the radiolabeled glucose-analog, [ 18 F]FDG was introduced for imaging and quantification of BAT activation. Although primarily perceived as a clinical tool, PET employing specific tracers can also add to basic research. In this way, MCHR1 PET imaging provided first evidence for MCHR1 expression in BAT and will contribute to the elucidation of the MCHR1 axis in BAT.

CONCLUSION
The present study aimed at enhancing the understanding of the interplay between MCHR1 ligands and the pharmacologically complex brown adipose tissue. Uptake of [ 11 C]SNAP-7941 and [ 18 F]FE@SNAP in rat BAT gave the initial hint for MCHR1 expression in BAT. The herein performed in vitro binding studies, demonstrating specific [ 11 C]SNAP-7941 binding to brown adipocytes and superior selectivity toward MCHR1 compared to ADRB3, provide first evidence for a direct involvement of the MCH/MCHR1 system in brown adipose tissue. Additionally, MCHR1 expression in murine brown adipocytes was demonstrated on mRNA and protein level for the first time.

ETHICS STATEMENT
This study was carried out in accordance with the recommendations of the Institutional Animal Care and Use Committee of the Medical University of Vienna, Austria.