Identification of the Thinner allele. The Thinner phenotype is observed among third-generation (G3) C57BL/6J mice heterozygous or homozygous for mutations induced by ENU. Thinner mice have decreased body weight, decreased fat mass, and slightly decreased lean mass compared with WT mice (Figure 1, A–C). The Thinner phenotype was mapped as a quantitative trait. AMM (6) implicated a missense allele of Gpr75 as the causative mutation, displaying the strongest linkage in an additive model of inheritance (Figure 1, D–F). The Thinner mutation was a single nucleotide transition from T to C, causing substitution of a lysine for a proline at position 144 (L144P) in the GPR75 protein (Supplemental Figure 1A; supplemental material available online with this article; https://doi.org/10.1172/JCI182121DS1). These data suggest an association between GPR75 and leanness in mice.

Identification and mapping of the Thinner allele. (A–C) Relative body weight (A), fat weight (B), and lean weight (C) phenotypic data plotted versus genotype at the Gpr75 mutation site. Mean (μ) and SD (σ) are indicated. Ref, homozygous for the reference allele; Het, heterozygous for the reference allele and for the Thinner allele; Var, homozygous for the Thinner allele. Raw weight data were compared with the predicted weight of mice based on age and sex to calculate the relative values, minimizing the effects of age and sex differences in G3 mice. (D–F) Manhattan plots showing P values calculated using an additive model of inheritance about relative body weight (D), fat weight (E), and lean weight (F). The −log10P values (y axis) are plotted versus the chromosomal positions of 53 mutations (x axis) identified in the G1 founder of the pedigree. Horizontal dark red and pink lines represent thresholds of P = 0.05 with and without Bonferroni’s correction, respectively.

GPR75-deficient mice exhibit a lean phenotype. The Thinner mutation (L144P) did not affect the stability of GPR75 as revealed by expression levels similar to those in the WT protein in 293T cells (Supplemental Figure 1B). We suspected that the mutation affects the function of GPR75 protein. By CRISPR/Cas9 gene targeting, we introduced a null allele of Gpr75, encoding the first 6 aa followed by 9 aberrant aa and a termination codon, into the germline of C57BL/6J mice (Supplemental Figure 1, A and C). Homozygous Gpr75-knockout (Gpr75–/–) mice had no reproductive or developmental defects, and both male and female Gpr75–/– mice were fertile. There was no difference in body weights, fat weights, or lean weights between WT and Gpr75–/– mice at 6 weeks of age (Figure 2, A–C). However, after only 2 weeks of HFD feeding, Gpr75–/– mice began to exhibit significantly decreased fat weight compared with WT mice (Figure 2B). After 4 weeks of HFD feeding, Gpr75–/– mice had decreased body weight (Figure 2A). No difference in lean weight was observed between WT and Gpr75–/– mice over the course of 8 weeks of HFD feeding (Figure 2C). The lean phenotype of Gpr75–/– mice fed a HFD for 8 weeks was obvious, with a smaller size of epididymal white adipose tissue (eWAT), interscapular white adipose tissue (iWAT), interscapular brown adipose tissue (iBAT), and liver (Figure 2, D and E). Additionally, Gpr75–/– mice had decreased liver weight and liver triglycerides (Figure 2, F and G). H&E staining further revealed smaller adipocyte size in the adipose tissues and reduced fat content in the liver of Gpr75–/– mice (Figure 2H). The HFD used in our study contained high levels of saturated fatty acids (lard), which are known to rapidly induce obesity in susceptible C57BL/6J mice. To explore whether different types of fatty acids influence the phenotype of Gpr75–/– mice, we fed mice a HFD rich in unsaturated fatty acids (safflower oil). Interestingly, Gpr75–/– mice also exhibited decreased fat weight compared with WT mice, with no change in body weight or lean weight (Figure 2, I–K). Unlike on a HFD, Gpr75–/– mice on a regular chow diet displayed very small body weight change and fat weight change compared with WT mice. Only a slight difference was observed in the fat weight of WT and Gpr75–/– mice at 16 weeks of age (Figure 2, L–N).

The phenotype of Gpr75–/– mice. (A–C) Body weight (A), fat weight (B), and lean weight (C) of male Gpr75–/– mice (n = 9) and WT littermates (n = 9) fed a HFD from 6 to 14 weeks of age. (D) Photograph of a 14-week-old male Gpr75–/– mouse and a WT (+/+) littermate fed a HFD for 8 weeks. (E) Representative photographs of eWAT, iWAT, iBAT, and liver from 14-week-old male mice fed a HFD for 8 weeks. (F and G) Liver weight (F) and liver triglyceride levels (G) of 14-week-old male mice fed a HFD for 8 weeks. (H) H&E stainings of sections from eWAT, iWAT, iBAT, and liver of 14-week-old male mice fed a HFD for 8 weeks. Scale bars: 100 μm. (I–K) Body weight (I), fat weight (J), and lean weight (K) of male Gpr75–/– mice (n = 5) and WT littermates (n = 5) fed an unsaturated HFD from 6 to 9 weeks of age. (L–N) Body weight (L), fat weight (M), and lean weight (N) of male Gpr75–/– mice (n = 7) and WT littermates (n = 7) fed a chow diet from 10 to 16 weeks of age. Data are presented as the mean ± SD. P values were determined by a mixed-effects model with Holm-Šidák’s multiple-comparison test (A–C and I–N) or 2-tailed, unpaired Student’s t test (F and G). *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001; NS, P > 0.05. Data points represent individual mice (F and G). Data are representative of 2 independent experiments.

To further check other metabolic profiles of Gpr75–/– mice, we measured the fasting serum of WT and Gpr75–/– mice after 4 weeks of HFD feeding. We observed no difference in blood glucose, insulin, cholesterol, or triglyceride levels between WT and Gpr75–/– mice (Supplemental Figure 2, A–D). However, Gpr75–/– mice had significantly decreased levels of leptin, which was likely caused by the decreased fat mass (Supplemental Figure 2E). Both the glucose tolerance test (GTT) and insulin tolerance test (ITT) were performed on WT and Gpr75–/– mice, and no notable differences were observed (Supplemental Figure 2, F and G). These data suggest that there was no change in glucose or insulin metabolism between WT and Gpr75–/– mice at the onset of the lean phenotype development.

The lean phenotype of Gpr75–/– mice is due to decreased food intake. Food intake of WT and Gpr75–/– mice on a chow diet or a HFD was monitored beginning at 6 weeks of age (Figure 3, A–D). On the chow diet, we noted no difference in food intake between Gpr75–/– mice and WT littermates (Figure 3, A and B). However, food intake was significantly decreased in Gpr75–/– mice compared with WT mice after HFD feeding for 10 days (Figure 3, C and D). To check if the decreased food intake was the cause of the lean phenotype in GPR75-deficient mice, WT mice were pair-fed with the same amount of a HFD as Gpr75–/– mice beginning at 6 weeks of age when the body weights, fat weights, and lean weights of the mice were similar (Figure 3, E–G). Three weeks after pair-feeding with a HFD, WT and Gpr75–/– mice gained similar amounts of body weight, fat weight, and lean weight (Figure 3, E–G), implying that the decreased food intake directly contributed to the development of the lean phenotype in GPR75-deficient mice.

Gpr75–/– mice have a decrease in food intake, which causes the lean phenotype. (A and B) Food intake of male Gpr75–/– mice (n = 7) and WT littermates (n = 7) on a chow diet was monitored from 6 to 8 weeks of age. Cumulative food intake (g) (A) and average food intake per mouse per day (g/day) during the 2-week period (B). (C and D) Food intake of male Gpr75–/– mice (n = 6) and WT littermates (n = 6) fed a HFD was monitored from 6 to 8 weeks of age. Cumulative food intake (g) (C) and average food intake per mouse per day (g/day) during the 2-week period (D). (E–G) Body weight (E), fat weight (F), and lean weight (G) of male Gpr75–/– mice (n = 5) and WT littermates (n = 5) before and after pair-feeding from 6 weeks of age. Data are presented as the mean ± SD. P values were determined by 2-tailed, unpaired Student’s t test (A–D) or a mixed-effects model with Holm-Šidák’s multiple-comparison test (E–G). *P ≤ 0.05; NS, P > 0.05. Data points represent individual mice (B, D), and data are representative of 2 independent experiments.

To check whether there were other factors beyond the decreased food intake that contributed to the lean phenotype of GPR75-deficient mice, we performed metabolic cage experiments on WT and Gpr75–/– mice after 2 weeks of HFD feeding (Supplemental Figure 3). At this stage, Gpr75–/– mice exhibited decreased body weights and fat weights but similar lean weights compared with WT mice (Supplemental Figure 3, A–C). No significant differences were observed in energy expenditure, respiratory exchange ratio, or physical activities between WT and Gpr75–/– mice (Supplemental Figure 3, D–I). Furthermore, Gpr75–/– mice also demonstrated a comparable ability to maintain body temperature compared with WT mice during an acute cold stress experiment without food (Supplemental Figure 3J). To eliminate the effect of difference in food intake on the metabolic cage experiments, we pair-fed WT mice the same amount of a HFD as Gpr75–/– mice every day when conducting the metabolic cage experiments. Despite this standardization, we observed no differences between WT and Gpr75–/– mice (Supplemental Figure 4, A–I). Additionally, Gpr75–/– mice showed no deficiencies in food digestion and absorption as revealed by a similar fecal energy density compared with that of WT mice (Supplemental Figure 4J). Taken together, these findings strongly suggest that reduced food intake is the sole determinant contributing to the lean phenotype observed in GPR75-deficient mice.

GPR75 is predominantly expressed in the brain and interacts with Gαq to signal in the brain. To understand why GPR75-deficient mice experienced reduced food intake, we examined the mRNA levels of Gpr75 in various mouse tissues. The expression of Gpr75 mRNA was notably higher in the brain compared with all other tested tissues, indicating a potential role of GPR75 in the brain (Figure 4A). Using the single-cell RNA-Seq (scRNA-Seq) data (17) from the Allen Brain Cell Atlas, we thoroughly analyzed the expression of Gpr75 mRNA in different cells of the mouse brain. Among 4.04 million brain cells, Gpr75 was expressed in 0.327 million cells, accounting for 8.09% of all cells in the brain (Supplemental Figure 5A). The majority of Gpr75+ cells belong to neuronal classes (88.68%), with only small portions represented by granule and immature neuronal classes (4.57%) and non-neuronal classes (6.75%) (Supplemental Figure 5, B and C). Gpr75+ cells were widely distributed among all neuronal classes and various brain regions, with a relatively higher percentage among serotonergic neurons (36.60%) (Supplemental Figure 5, B–G).

GPR75 is highly expressed in the brain and directly binds to Gαq to regulate various downstream signaling pathways. (A) Relative Gpr75 mRNA levels in different mouse tissues normalized by Polr2a (n = 3 mice). (B) Generation of 3xFlag-tagged Gpr75-knockin mice by CRISPR. Ex1, exon 1; Ex2, exon 2. (C) Immunoblot analysis of 3xFlag-Gpr75 protein expression in different mouse tissues (8-week-old males) by immunoprecipitation (IP). GAPDH was used as a loading control. IB, immunoblot. (D) Immunoblot analysis of 3xFlag-Gpr75 protein expression in various brain regions (8-week-old males) by immunoprecipitation. GAPDH was used as a loading control. (E) Mass spectrometric identification of GPR75-interacting proteins from Flag immunoprecipitates of Gpr75-3xFlag–knockin brain lysates. (F) Immunoblot analysis of immunoprecipitates (top and middle) and lysates (bottom) of 293T cells expressing HA-tagged Gnaq and 3xFlag-tagged Gpr75. (G) Summary of significantly changed genes (FDR <0.05) from RNA-Seq of hypothalamus from 8-week-old male WT and Gpr75–/– mice fed a chow diet or a HFD for 2 weeks (n = 3 mice for each group). (H and I) Volcano plots of differentially expressed genes in the hypothalamus of Gpr75–/– versus WT mice on a chow diet (H) or a 2-week HFD (I). Differentially expressed genes (FDR <0.05) are colored in red and blue indicating upregulation and downregulation, respectively. (J) Manhattan-like plot of pathways significantly associated (P < 0.05) with the loss of GPR75 on HFD feeding identified from a pathway overrepresentation analysis that mapped significant genes to the Reactome and WikiPathways databases. (K) Visualization of significantly associated pathways (P < 0.05) from J. Data are presented as the mean ± SD. Data points represent individual mice in A, and data are representative of 3 independent experiments.

To check the relative expression of endogenous GPR75 protein, we generated 3xFlag-Gpr75–knockin mice using CRISPR-mediated homologous replacement (Figure 4B). The endogenous GPR75 protein expression was exclusively detected in the Flag immunoprecipitates of brain lysates, aligning with the Gpr75 mRNA expression profile (Figure 4C). Different parts of the brain have diverse functions, and the hypothalamus plays a crucial role in the regulation of food intake. Hence, we evaluated the expression of endogenous GPR75 protein in various brain regions. Consistent with the scRNA-seq data (Supplemental Figure 5, D and E), we found that GPR75 protein was expressed throughout different parts of the brain without clear regional differences (Figure 4D).

To investigate the molecular mechanism underlying the role of GPR75 in the regulation of food intake, we utilized our 3xFlag-Gpr75–knockin mice to pull down GPR75 in brain lysates for the identification of interacting proteins by mass spectrometry (Figure 4E). GNAQ (aka Gαq), a guanine nucleotide–binding protein, was identified as a GPR75-interacting protein (Figure 4E). We further confirmed the interaction between GPR75 and Gαq by immunoprecipitation assays (Figure 4F). Prior studies have indicated that Gαq mediates GPR75 signaling (15). Thus, it is plausible that GPR75 also functions through Gαq in modulating food intake centrally.

To gain insights into the role of GPR75 in the central regulation of feeding, we performed transcriptomics analysis (RNA-Seq) of the hypothalamus from WT and Gpr75–/– mice on a chow diet or on a 2-week HFD. For a FDR below 0.05, only a total of 8 genes were significantly changed in the hypothalamus of chow diet–fed Gpr75–/– mice, including 5 genes with increased expression and 3 genes with decreased expression (Figure 4, G and H). While on a HFD, a total of 30 genes were significantly changed in the hypothalamus of Gpr75–/– mice for a FDR below 0.05, including 5 genes with increased expression and 25 genes with decreased expression (Figure 4, G and I). It is worth noting that only 1 gene, proteolipid protein (myelin) 1 (Plp1), was markedly decreased in both chow diet– and HFD-fed Gpr75–/– mice. The major transcriptome change observed only in HFD-fed Gpr75–/– mice is consistent with the appearance of a strong lean phenotype of Gpr75–/– mice driven by the HFD. To facilitate interpretation and identify relevant signaling pathways associated with loss of GPR75 in the hypothalamus, we next performed overrepresentation analyses mapping significant genes to the Reactome and Wikipathways databases included in the ConsensusPathDB (18) (Figure 4, J and K). We found an overrepresentation of pathways with important roles in the regulation of energy homeostasis, including TGF-β signaling (19, 20) and signaling by receptor tyrosine kinases (21), and development and function of neurons, including extracellular matrix organization (22), signaling by neurotrophic tyrosine receptor kinases (NTRKs) (23), and glial cell differentiation. These data suggest that GPR75 regulated various signaling pathways after HFD feeding.

GPR75 is localized in the primary cilia. Primary cilia are present in various cell types and play a crucial role in cell signaling. Dysregulation of cilia or ciliary proteins is closely linked to obesity (24). Many GPCRs are cilia-associated proteins, and their functions within the cilia are essential for regulating food intake and energy expenditure (25, 26). To explore whether GPR75 is a cilia-associated protein, we overexpressed 3xFlag-Gpr75 in mouse inner medullary collecting duct (mIMCD) 3 cells and observed its subcellular localization. Immunofluorescence staining results clearly indicated that the GPR75 protein was localized in the cilia (Figure 5A). TUB-like protein 3 (TULP3) is known to mediate the trafficking of GPCRs into the primary cilia (27, 28). Indeed, GPR75 interacted with TULP3 when expressed in 293T cells (Supplemental Figure 6). In Tulp3-knockout (Tulp3–/–) mIMCD3 cells, GPR75 protein failed to localize in the cilia, suggesting that the ciliary localization of GPR75 was dependent on TULP3 (Figure 5B). Considering the specific expression of GPR75 protein in the brain, we thought it would be interesting to check the ciliary localization of GPR75 in brain neuron–derived cell lines. Therefore, we used the mouse embryonic hypothalamic cell line N11 to examine the subcellular localization of GPR75. Similar to mIMCD3 cells, we found that GPR75 was exclusively localized in the cilia of N11 cells (Figure 5C). However, the L144P-mutant form of GPR75 identified in Thinner mice failed to localize in the cilia (Figure 5D), suggesting that ciliary localization is important for the role of GPR75 in energy homeostasis. Besides cell lines, we found that overexpressed GPR75 was also localized in the cilia of mouse primary hypothalamic neurons (Figure 5E). To assess the subcellular localization of endogenous GPR75 protein, we isolated primary hypothalamic neurons from WT control and 3xFlag-Gpr75–knockin mice. As shown in Figure 5, F and G, endogenous GPR75 protein colocalized with the cilia marker ADCY3, confirming its presence in the cilia. While human GPR75 protein is reported to be expressed on the cell surface and localize in the plasma membrane in human embryonic kidney (HEK) 293 cells, 2 loss-of-function mutations of GPR75 (p.Ala110fs and p.Gln234*) were unable to localize in the plasma membrane (14). However, whether human GPR75 is localized in the primary cilia remains unknown. Similar to mouse GPR75, we observed that human GPR75 was specifically localized in the cilia of N11 cells (Figure 5H). Additionally, 2 mutations of human GPR75 (p.Ala110fs and p.Gln234*) that are associated with lower BMI failed to localize in the cilia (Figure 5, I and J). In conclusion, we found that GPR75 was a cilia-associated protein whose localization in the primary cilia was crucial for its function in regulating body weight in both mice and humans.

GPR75 is located in the primary cilia. (A) mIMCD3 cells expressing 3xFlag-tagged Gpr75 were immunostained with Flag antibody (green), Ac-tubulin (Ac-Tub) (red), and Hoechst 33342 (blue). (B) Tulp3–/– mIMCD3 cells expressing 3xFlag-tagged Gpr75 were immunostained with Flag antibody (green), Ac-tubulin (red), and Hoechst 33342 (blue). (C and D) N11 cells expressing 3xFlag-tagged Gpr75 WT (C) or L144P (D) were immunostained with Flag antibody (green), Ac-tubulin (red), and Hoechst 33342 (blue). (E) Mouse primary hypothalamic neurons expressing 3xFlag-tagged Gpr75 were immunostained with Flag antibody (green), ADCY3 (red), and Hoechst 33342 (blue). (F and G) Primary hypothalamic neurons isolated from WT mice (F) or homozygous 3xFlag-Gpr75–knockin mice (G) were immunostained with Flag antibody (green), ADCY3 (red), and Hoechst 33342 (blue). (H–J) N11 cells expressing 3xFlag-tagged human GPR75 WT (H), p.Ala110fs (I), or p.Gln234* (J) were immunostained with Flag antibody (green), Ac-tubulin (red), and Hoechst 33342 (blue). Data are representative of 3 independent experiments. Scale bars: 10 μm.

Loss of GPR75 has no effect on the development of obesity in Lepob-mutant mice or Adcy3-mutant mice. Leptin signaling in the brain plays an important role in the regulation of food intake (29). We crossed Gpr75–/– mice with Lepob-mutant mice to determine whether GPR75 is involved in leptin signaling and whether loss of GPR75 would attenuate the obesity phenotype of Lepob-mutant mice. As expected, Lepob/+ and Lepob/ob mice had a greater increase in body weight, fat weight, and lean weight than did WT mice at 4 weeks of age (Figure 6, A–C). Complete knockout of Gpr75 did not reduce the increased fat and lean weight in either Lepob/+ or Lepob/ob mice (Figure 6, A–C). ADCY3 catalyzes the synthesis of cAMP, and its functions within the primary neuronal cilia are essential in regulating body weight (30–32). To assess the genetic interaction between GPR75 and ADCY3, we had to cross Gpr75–/– mice with Adcy3–/– mice. Unfortunately, Adcy3–/– mice are known to be anosmic and have a very high fatality rate within 48 hours of birth (33). During our genetic screening, we identified a viable hypomorphic Adcy3-mutant mouse (Adcy3L278H/L278H) with massive obesity. As shown in Figure 6, D–F, Adcy3L278H/L278H mice had increased body and fat weights compared with WT mice at 8 weeks of age. Loss of Gpr75 failed to reduce the development of obesity in Adcy3L278H/L278H mice (Figure 6, D–F). Taken together, these findings demonstrate that loss of Gpr75 did not attenuate the obesity phenotype of Lepob- or Adcy3-mutant mice.

Gpr75–/– does not inhibit the development of obesity in Lepob- orAdcy3-mutant mice. (A–C) Body weight (A), fat weight (B), and lean weight (C) of 5-week-old male mice. (D–F) Body weight (D), fat weight (E), and lean weight (F) of 8-week-old male mice. Data are presented as the mean ± SD. P values were determined by 1-way ANOVA with Holm-Šidák’s multiple-comparison test (A–F). *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001; NS, P > 0.05. Data points represent individual mice (A–F), and data are representative of 3 independent experiments.

Testing the ligands of GPR75. At present, CCL5 and 20-HETE, are reported to be ligands of GPR75 (15, 16). We conducted 2 distinct assays to test the effect of CCL5 and 20-HETE on GPR75. Initially, we generated a luciferase report construct that contained a multiple response element (MRE), a cAMP response element (CRE), a serum response element (SRE), and a luciferase gene. The MRE/CRE/SRE luciferase assay is capable of detecting agonist effects from Gi-, Gs-, and Gq-coupled receptors as well as the activities of most GPCRs (34, 35). Human GPR75 and the luciferase reporter construct were cotransfected in 293T cells to assess the luciferase activity with different concentrations of these ligands. The second assay we used was PRESTO-Tango assay, which is designed to identify ligands through the G protein–independent β-arrestin recruitment pathway (36). However, we did not observe strong activation of human GPR75 by CCL5 or 20-HETE, whether in the MRE/CRE/SRE luciferase assay or the PRESTO-Tango assay (Figure 7, A and B). Since 20-HETE is prone to oxidation, we used a stable synthetic analog, sodium 20-hydroxyeicosa-5(Z),14(Z)-dienoate (5,14-HEDE), to explore its potency in activating GPR75. In the MRE/CRE/SRE luciferase assay, we observed a modest induction of GPR75 with concentrations exceeding 1 μg/mL 5,14-HEDE (Figure 7C), whereas in the PRESTO-Tango assay, only a high concentration (≥1 μg/mL) of 5,14-HEDE appeared to activate GPR75 (Figure 7D). However, 5,14-HEDE did not increase intracellular cAMP level via GPR75 (Figure 7E), and a very high concentration (50 μg/mL) of 5,14-HEDE seemed to increase intracellular levels of inositol phosphate 1 (IP1) (Figure 7F). These findings do not conclusively establish CCL5, 20-HETE, and 5,14-HEDE as definitive ligands of GPR75. The pursuit of novel GPR75 ligands is worthwhile, particularly those that could potentially regulate food intake.

Exploring ligands of GPR75. (A) MRE/CRE/SRE luciferase assay. 293T cells expressing human GPR75 and pGL3-MRE/CRE/SRE-luciferase plasmids were treated with different concentrations of CCL5 and 20-HETE. (B) PRESTO-Tango β-arrestin recruitment assay in HTLA cells overexpressing GPR75-Tango constructs treated with different concentrations of CCL5 and 20-HETE. (C) MRE/CRE/SRE luciferase assay. 293T cells expressing human GPR75 and pGL3-MRE/CRE/SRE-luciferase plasmids were treated with different concentrations of 5,14-HEDE. (D) PRESTO-Tango β-arrestin recruitment assay in HTLA cells overexpressing GPR75-Tango constructs treated with different concentrations of 5,14-HEDE. (E) cAMP assay. 293T cells expressing human GPR75 were treated with different concentrations of 5,14-HEDE and forskolin. (F) IP1 assay. HTLA cells expressing human GPR75 were treated with different concentrations of 5,14-HEDE. Data are presented as the mean ± SD. P values were determined by 1-way ANOVA with Holm-Šidák’s multiple-comparison test (A–F). *P ≤ 0.05 and **P ≤ 0.01; NS, P > 0.05. Data points represent individual wells (A–F), and data are representative of 3 independent experiments.