Gpr149 is involved in energy homeostasis in the male mouse

Generation of novel transgenic lines and mouse care

Tomato reporter mice (stock #007909) and male C57BL/6J mice 6–8 weeks of age were purchased from Jackson Laboratory. All studies were approved by the University of Texas (UT) Southwestern’s Institutional Animal Care and Use Committee (IACUC; APN# 2016-101590) and all mice were maintained in a barrier facility with controlled temperature (23 °C) on a 12 h light/dark cycle (lights on 0600–1800). Unless indicated otherwise for the purpose of a feeding study, animals were grouped housed and had unrestricted access to food and water. For enrichment and thermal comfort, nestlets and igloos were provided to all mice. Of note, euthanasia of animals strictly followed a protocol approved by our IACUC and the American Veterinary Medical Association. Animals with signs of visible distress, labored breathing, or excess weight loss (20% of initial weight) euthanized prior to the end of the experiment. Euthanasia of animals not needed for experiments was performed by carbon dioxide immediately followed by cervical dislocation. Animals needed for experiments were euthanized as described below (qPCR and tissue preparation sections). To the further extent possible, our manuscript followed the ARRIVE guidelines (Animal Research: Reporting of In Vivo Experiments) (Percie du Sert et al., 2020).

Transgenic mice were generated as follows: For the Gpr149 null allele: guides flanking a 2 kb region containing Gpr149 exon 1 were ordered from IDT: 5′ Guide ribonucleotides sequence: 5′-CUUAUAACUGGUCACCUAUGUGUUUUAGAGCUAUGCU-3′ 3′ ribonucleotide sequence: 5′- UUGGUAGUUAACGAGACCCCGUUUUAGAGCUAUGCU-3′. Guides, trcRNA, Cas9 protein were administered through a pronuclear injection in C57Bl/6N (Charles River) by the UT Southwestern Transgenic Technology Center. Founders were screened by PCR and sanger sequencing. One line generated through non-homologous end joining was selected and expanded by crossing with C57Bl/6N mice from Charles River. Mice were genotyped using the following primers: Gpr149 1: 5′-GCTGCTTGTAATGTGTGCAGAGAG-3′ Gpr149 2: 5′-GTCTACTCATGGCAGACCAAAGTAATGG-3′ Gpr149 3: 5′- GTCTCTTGGTGCTAGAGATGGGTG-3′ resulting in a WT band of 200 bp and a Gpr149 KO band of 350 bp. For the Cre-P2A-Gpr149 mouse, the endogenous Gpr149 locus was targeted with the following guide ordered from IDT: 5′-AAGUCAUAAUUCUACGGAGAGUUUUAGAGCUAUGCU-3′. Pronuclei were coinjected with an IDT Megamer TM (Integrated DNA Technologies) encoding 100 5′ arm 84 3′ homology arms and the 1,119 bp Cre-P2A at Gpr149’s ATG start sequence. Synonymous mutations were placed in the guide sequence to prevent recutting of the Cre-P2A insert. Sequences are also included in Table S1.

Quantitative PCR (qPCR)

On the day of sacrifice, mice were anesthetized with an overdose of chloral hydrate (500 mg/kg, i.p.), followed by decapitation. Tissue samples were collected according to methods detailed in our previously published studies (Bookout et al., 2006a, 2006b). Total RNA was isolated using the Quick-RNA microprep Kit using manufacturer’s instructions (Zymo, Irvine, CA, USA). cDNA was synthesized using the High-Capacity cDNA synthesis Kit following the manufacturer’s protocol (Applied Biosystems, Waltham, MA, USA). cDNA levels were measured in triplicates using a QuantStudio 5 Real-Time PCR System (Applied Biosystems by Thermo Fisher Scientific, Foster City, CA, USA). Pre-validated Taqman assays for Gpr149 (Mm00805216_m1), and 18s (Hs99999901_s1), were purchased from Thermo Fisher Scientific. The relative amount of transcript levels was calculated using the delta/delta CT method.

Tissue preparation

On the day of sacrifice, mice were anesthetized with an overdose of chloral hydrate (500 mg/kg, i.p.) and perfused transcardially with 0.9% saline followed by 10% neutral buffered formalin (Sigma–Aldrich, St. Louis, MO, USA). Brains and peripheral tissues were dissected and post-fixed in formalin for 24 h at 4 °C. Then, samples were incubated for 24 h at 4 °C in 20% sucrose made in 0.1 M phosphate-buffered saline (PBS). Free-floating coronal brain sections of 25 μm thickness were produced using a freezing microtome (Leica, Teaneck, NJ, USA), collected in PBS, and stored in a cryoprotectant at −20 °C. Sections from other tissues with a thickness of 14 μm were produced using a cryostat (Leica, Teaneck, NJ, USA), collected onto SuperFrost slides, and stored at −80 °C. For imaging native tdTomato, brain sections were rinsed in PBS, mounted onto SuperFrost slides, and coverslips with ProLong^™ Gold Antifade (Thermo Fisher, Waltham, MA, USA) was placed on the samples. A total of three Gpr149-Cre-tdTomato mice were used to survey native tdTomato brains and peripheral organs. For other applications, tissues were processed as explained below.

RNAScope in situ hybridization (ISH)

Free-floating brain sections were rinsed in PBS, incubated in a solution of H₂O₂ for 10 min, rinsed with PBS, carefully mounted on SuperFrost slides, and desiccated overnight at room temperature. The brain slides were then processed using a multiplex fluorescent kit (cat# 323110) or a chomogenic FastRed kit (Cat#322350) following the manufacturer’s instructions. The H₂O₂ step in the ACD protocol was omitted because it was performed the day before. The Gpr149 probe (cat# 318071) was applied at 40 °C for 2 h. Amplification steps were carried out using either Opal570 (cat# FP1488001KT; Akoya Biosciences, Japan) for fluorescence assays or FastRed (ACD) for chromogenic assays. Fluorescently labeled sections were counterstained with DAPI and mounted using ProLong medium. Chromogenically-labeled sections were counterstained with hematoxylin and mounted using Ecomount medium (Biocare medical, Concord, CA, USA). Fluorescent assays were applied on three entire brains from three C57BL/6J mice and three Gpr149-Cre-tdTomato mice. For validation, chromogenic assays were applied to the brain nodose ganglion, and pituitary gland of one Gpr149^−/− and one Gpr149^+/+ mouse. When fluorescent RNAScope analysis was performed on the brain sections of Gpr149-Cre-tdTomato mice, the same procedure was followed. Native tomato was bright enough to be seen without any obvious decline in fluorescence intensity. Tissues other than the brain samples already collected on slides were processed in the same manner. However, the H₂O₂ step was performed on the hybridization day rather than the day before.

Metabolic profiling

We assessed the long-term effects on body weight, food intake, and glucose homeostasis in cohorts of male mice maintained on either chow (No. 2016 Harlan, Teklad) or high fat diet (60% Fat; Research Diets D12492). We used a group size of 6−9 mice/group. Body weight for the chow mice was assessed weekly until week 13 and for the HFD until week 20. At the end of 20 weeks, body composition was analyzed by magnetic-resonance whole-body composition analyzer (EchoMRI). Upon the completion of the studies, tissues including liver, white and brown fat pads, heart, and skeletal muscles were weighted. Food intake was assessed in a cohort of fasted (16 h) mice. Upon refeeding, pellets of chow diets were weighted over a period of 4 h. A cohort was also used to assess daily food intake under ad libitum food access. Pellets of chow diets were weighted once a day over a period of four consecutive days (see Fig. S4 for diet composition). A separate cohort was used for insulin and glucose tolerance tests (ITTs and GTTs) on chow diet. Single housed mice were fasted 6 h prior to testing starting at 8:00 AM with water provided ad libitum. For the GTT mice received an intraperitoneal injection of glucose at 1 g/kg. For ITT, mice received 1 U/kg. Blood glucose from the tail vein was measured at five time points 0, 15, 30, 60, and 120 min using a handheld commercial glucometer (Bayer’s Contour Blood Glucose Monitoring System; Bayer, Leverkusen, Germany).

Imaging

Fluorescence-labeled whole brains sections (RNAscope or tdTomato) were scanned by the Whole Brain Microscopy Facility at UTSouthwestern (see Acknowledgments) using a Zeiss Axioscan. Z1 and appropriate filters. We also used a confocal microscope (Zeiss LSM880) available at the Quantitative Light Microscopy Core (see acknowledgments) to image fluorescence-labeled tissues at higher magnifications. ImageJ (NIH, Bethesda, MD, USA) and Adobe Photoshop 2021 were used to uniformly adjust the resolution and contrast of all our digital images. Estimates of signal strengths were done by visually inspecting brain sections under epifluorescence microscopy (Leica, Teaneck, NJ, USA). Brain sites were visually identified and ranked by expression level with reference to the Franklin and Paxinos atlas (3^rd edition).

Data analysis

Quantitative data were presented as mean ±SEM. Differences between genotypes were compared with either a one-way (qPCR and physiological data) or two-way ANOVA analysis followed by the post-hoc Dunnett’s test (body weights curves). Statistical significance is accepted at a value of p < 0.05. The exact values are indicated above bar graphs. Statistical tests and their tabular results (degree of freedom, effect size, etc.) are included in Supplementary figures for all test with statistically significant results. Graphs of numerical data were produced using GraphPad Prism 9. Groups sizes are included either in our results section, or figure legends, or graphs. For statistical analysis, mice carrying wild-type alleles were considered as control groups.

Mapping of Gp149-expressing sites reveals neuronal enrichment

We mapped the expression of Gpr149 in mouse tissues by qPCR (Fig. 1). The strongest expression of Gpr149 is observed in the central nervous system (CNS) tissues including, among other examples, the striatum, hypothalamus, brainstem, and spinal cord (Fig. 1). The highest Gpr149 expression in a non-neuronal tissue is that in the pituitary gland (Fig. 1). Low levels of expression are observed for the gastrointestinal tract and female reproductive organs (Fig. 1). Other examined tissues showed very low or close-to-detection threshold expression. Raw data are included in Fig. S1.

The exact distribution of Gpr149 within the CNS was further explored by fluorescent RNAscope ISH. At least 80 brain regions express Gpr149 at varying levels. Figure 2 shows representative coronal brain sections with the highest Gpr149 expression and Table 1 summarizes the relative expression levels across the brain. The strongest expression of Gpr149 is observed in the islands of Calleja and surrounding nuclei, such as the olfactory tubercle, the ventromedial hypothalamus, the rostral interpeduncular nucleus, and a few select brainstem nuclei such as the sphenoid nucleus (Fig. 2; Table 1). Moderate-to-low expression is also shown in many brain sites across the basal forebrain, striatum, hypothalamus, brainstem, and spinal cord (Fig. 2; Table 1). Only low expression is observed in the cortical and subcortical regions such as the hippocampus. Gpr149 expression appears to be restricted to neurons rather than glia in the adult brain. In fact, white matter tracts, meninges, and epithelia presented no discernable signals. The above roadmap of Gpr149-expressing sites in the mouse brain is largely consistent with Gpr149 playing a role in regulating basic motivated behaviors, autonomic outflow, and sensory processes (Swanson, 2005). As shown later, probe specificity was validated in a novel knockout model.

Cre mouse lines are useful for a wide range of applications, including the manipulation and mapping of cell types (Madisen et al., 2010). Here, using CRISPR-Cas9, we generated a Gpr149-Cre-expressing mouse by knocking an Cre-P2A sequence into the 5′ UTR of Gpr149 (Fig. 3). This will allow for the coexpression of Cre and Gpr149 from the endogenous Gpr149 locus. Crossing this mouse to tdTomato reporter mice allowed us to further assess the expression of Gpr149. As a result, Gpr149-Cre-tdTomato mice present fluorescence across the entire brain following a pattern highly reminiscent of that of endogenous Gpr149 mRNA (Fig. 4A). Cells resembling neurons and their axons are observed across the entire brain including, among other examples, the striatum (Fig. 4B). We also noticed several brain regions that contained brightly labeled axonal fibers rather than cell bodies. These include the substantia nigra (Fig. 4D) and primary sensory areas of the brainstem (Fig. 4E). Lastly, tdTomato-labeled cells across the brain always resemble neurons, except for cells lining the aqueduct (Fig. 4F). These are the only non-neuronal cells observed.

We further validated the specificity and sensitivity of this new Cre line by assessing tdTomato in combination with RNAScope for Gpr149. As anticipated, the Gpr149 signals coincide very well with tdTomato-positive cell bodies across the brain. For example, in the basal forebrain around the island of Calleja, a site of high Gpr149 mRNA expression, tdTomato fluorescence is very intense (Figs. 5A and 5B). As another example, in the striatum and cortex tdTomato-positive neurons always express Gpr149 mRNA (Figs. 5C and 5D). However, the intensity of tdTomato fluorescence is not always correlated with the strength of mRNA expression. For instance, in the VMH, another site of high Gpr149 mRNA expression, the neurons are only faintly fluorescent for tdTomato (Figs. 5E and 5F). Coincidence between tdTomato and Gpr149 is also observed in the rest of the brain including the midbrain and brainstem (Figs. 5G and 5H). The only site with an apparent mismatch between tdTomato and Gpr149 is the cerebral aqueduct. Tomato-positive cells lining the aqueduct either present either low levels or undetectable Gpr149 signals (Fig. 5H). This may result from developmental expression of Gpr149-Cre. Overall, this novel Gpr149-Cre line is both sensitive and specific.

The distribution of tdTomato was further investigated in peripheral tissues. Many tdTomato-positive neurons are observed in the sensory ganglia (Figs. 6A and 6B). In the superior cervical ganglion, only tomato axons rather than cell bodies are observed (Fig. 6C).

The enteric nervous system is also labeled (Fig. 6C). In addition to enteric neurons, unidentified cells resembling vascular cells and interstitial muscular cells are seen in the gastrointestinal tract (Figs. 6D–6G). In agreement with our qPCR data, the pituitary gland anterior lobe also contains a dense network of cells resembling endocrine cells (Fig. 7A). The pancreas only contains a few axons, presumably originating from sensory neurons (Figs. 7B and 7C). Many other organs present sparsely distributed cells of unknown identity, including the spleen (Fig. 7D), liver (Fig. 7E), kidney (Fig. 7F), and heart (Fig. 7G). Because the latter cells are frequently around blood vessels, we deduced they may be (peri) vascular cells.

Novel Gpr149 knockout mice partially resist diet-induced obesity

Using CRISPR-Cas9, we generated a global knockout allele with mice lacking Gpr149 exon 1 (Fig. 8A). Gpr149^−/− mice were viable and born without overt phenotypes at expected mendelian ratios (Table 2). Loss of hypothalamic expression for Gpr149 expression was confirmed in the knockout model by qPCR (Fig. 8B). Raw qPCR data and experimental details are included in Fig. S2. In addition, the chromogenic (red) RNAscope^® system allowed us to further verify the loss of Gpr149 expression in three representative tissues with high Gpr149 expression, including the VMH (Fig. 9A), nodose ganglion (Figs. 9C and 9D), and pituitary gland (Figs. 9G and 9H). Gpr149 transcripts were virtually undetectable in the same tissues from the Gpr149^−/− mouse (Figs. 9B, 9E, 9F, 9I and 9J). Additional histological validation on the trigeminal ganglion and spinal cord is available in Fig. S3. While there is little reason to believe that Gpr149 would not be absent from tissues of knockout mice, it is important to note that our analysis solely focused on tissues with moderate and high expression, which is a caveat that should be acknowledged.

Next, a preliminary metabolic profile analysis of the Gpr149^−/− mouse was performed. (Fig. 10). When fed a standard chow diet, there is a trend toward a lower weight gain rate in the Gpr149^−/− mouse (Fig. 10A). When switched to HFD (60% fat), diet-induced obesity was significantly less pronounced in Gpr149^−/− mice (Fig. 10A). Gpr149^+/− mice show an intermediate phenotype. This is accompanied by a reduction in fat mass, as determined by NMR (Fig. 10B), independent of any differences in body length (Fig. 10C). Interestingly, organ weights were not statistically different between genotypes (Fig. 10D). Over a period of 4 h, separate cohorts of fasted refed mice of the three different genotypes ate the same amount of standard chow (Fig. 10E). Under conditions of ad libitum food access, the three different genotypes also ate the same amount of HFD on a daily basis (Fig. 10F). Glucose tolerance tests in a separate cohort fed on standard chow show no difference between genotypes (Fig. 10G). However, both Gpr149^−/− and Gpr149^+/− mice had greater sensitivity to insulin than the control mice (Fig. 10H). The above data indicate that GPR149 is involved in energy balance and glucose homeostasis. Raw physiological data, detailed statistical results, and diets composition are included in Fig. S4.