Dysfunction of the adhesion G protein-coupled receptor latrophilin 1 (ADGRL1/LPHN1) increases the risk of obesity

Abstract

Obesity is one of the diseases with severe health consequences and rapidly increasing worldwide prevalence. Understanding the complex network of food intake and energy balance regulation is an essential prerequisite for pharmacological intervention with obesity. G protein-coupled receptors (GPCRs) are among the main modulators of metabolism and energy balance. They, for instance, regulate appetite and satiety in certain hypothalamic neurons, as well as glucose and lipid metabolism and hormone secretion from adipocytes. Mutations in some GPCRs, such as the melanocortin receptor type 4 (MC4R), have been associated with early-onset obesity. Here, we identified the adhesion GPCR latrophilin 1 (ADGRL1/LPHN1) as a member of the regulating network governing food intake and the maintenance of energy balance. Deficiency of the highly conserved receptor in mice results in increased food consumption and severe obesity, accompanied by dysregulation of glucose homeostasis. Consistently, we identified a partially inactivating mutation in human ADGRL1/LPHN1 in a patient suffering from obesity. Therefore, we propose that LPHN1 dysfunction is a risk factor for obesity development.

Fig. 1

Lphn1-deficient mice display an increased weight and fat accumulation. a Lphn1 knockout (KO) mice are bigger than wild-type (WT) littermates. Representative images of 30-week-old female mice on a cm scale. b Lphn1 knockout mice gain more weight over time than wild-type littermates. Mice were weighed weekly. Weight differences become visible after 12 weeks in females as well as males and increase with age. Data are given as means ± SD; n ≥ 8; *p < 0.05; **p < 0.01; ***p < 0.001. c The body length of Lphn1 knockout mice is indifferent from the one of wild-type littermates, but tails are shorter. Body and tail lengths were measured in mice 30–34 weeks of age. Given are means ± SD; n = 5-6; ***p < 0.001. d Lphn1 knockout mice have more fat, but similar lean body mass. Body composition of 30–34-week-old mice was determined using EchoMRI. Given are means ± SD; n ≥ 10; **p < 0.01; ***p < 0.001. e Organ mass of Lphn1 knockout and wild-type mice are similar. Only the livers are significantly heavier in Lphn1 knockout animals (30–34-weeks of age). Data are means ± SD; n ≥ 8; **p < 0.01. f Representative Oil red O staining of liver sections of 30–34-week-old male mice used to directly visualize the stored triglycerides in liver cells (red). The staining reveals high numbers of large lipid droplets within Lphn1 knockout livers. g Livers of Lphn1 knockout mice contain more fat than wild-type livers. Liver fat and lean body mass of 30–34-week-old animals was determined by EchoMRI. Data are given as means ± SD; n ≥ 5; **p < 0.01; ***p < 0.001

Fig. 2

Loss of Lphn1 increases food intake in male and female mice and alters energy metabolism in male mice. a Food intake during both the light and the dark phase is increased in Lphn1 knockout (KO) mice compared to wild-type (WT) littermates in male and female animals. b Analysis of food intake using body mass as a covariate shows that loss of LPHN1 rather than increased body weight is causative (ANCOVA, female: p = 0.09, male: p = 0.03). c 30-34-week-old Lphn1 male and female knockout mice show similar levels of movement activity than wild-type (WT) littermates independently of light or dark phase. d Wheel running activity shows a trend towards reduction in Lphn1 knockout animals. e Energy expenditure does not show significant differences in Lphn1 knockout mice compared to wild-type littermates. f The overall energy balance (food intake minus energy expenditure) is significantly increased in male Lphn1 knockout mice during both the light and dark phases. g, h The respiratory exchange rate in male Lphn1 KO mice is significantly higher than in WT controls during light and dark phase whereas no change is obvious in female animals. Data in are given as means ± SD of three to five 30–34-week-old animals; *p < 0.05; **p < 0.01; ***p < 0.001

Fig. 3

LPHN1 is present on hypothalamic neurons that regulate feeding behavior. a Meta-analysis of RNA-seq data from revealed expression of all three latrophilin homologs in mouse AGRP- and POMC neurons. The mRNA levels of Lphn1 and Lphn3 are comparable to those of some GPCRs (Gshr and Npy1R–Npy6R) known to be involved in the regulation of food intake. Eltd1, the fourth member of this aGPCR group, is not expressed. Given are mean transcript per million mapped reads (TPM) values ± SD of five individuals. b The expression levels of neuronal components involved in the regulation of food intake are not altered in obese Lphn1-deficient mice. The hypothalamus of 30–34-week-old mice were subjected to qPCR analysis using gene-specific primers for prohormone convertase 1 (gene Pcsk1), prohormone convertase 2 (gene Pcsk2), the orexigenic neuropeptides NPY, AGRP, galanin (gene Gal), pro melanin concentrating hormone (gene Pmch), and hypocretin (gene Hcrt) as well as the anorexigenic neuropeptides POMC, Cart, neurotensin (gene Nts), cholecystokinin (gene Cck), and oxytocin (gene Oxt). Data are given as means ± SD; n ≥ 3 normalized to β-microglobulin/-actin (20.60 ± 2.33)

Fig. 4

Lphn1 knockout mice show changes in hormone levels controlling metabolism and an altered glucose and insulin tolerance as well as lipolysis. a Several metabolism-related parameters are altered in Lphn1 knockout (KO) compared to wild-type (WT) littermates. Parameters were determined from blood serum samples of 30–34-week-old male mice. Serum levels are given as means ± SD; n = 5–8; *p < 0.05; ***p < 0.001. b Lphn1 knockout male and female mice have a reduced glucose tolerance. Glucose tolerance tests were performed by intraperitoneal injection of 1 mg glucose/g body weight in mice of 30–34 weeks of age. Blood glucose levels are given as means ± SD; n ≥ 13; *p < 0.05; **p < 0.01; ***p < 0.001. c In the absence of LPHN1, mice show higher glucose levels after insulin administration. Blood glucose was measured after intraperitoneal injection of 1 mU insulin/g body weight. Data are given as means ± SD; n ≥ 8; *p < 0.05; **p < 0.01. d Plasma insulin contents are higher in overnight fasted Lphn1 knockout mice than in wild-type littermates. Levels are given as means ± SD; n ≥ 7; *p < 0.05; **p < 0.01. e Serum NEFA levels in obese Lphn1 knockout mice are reduced. Data are given as means ± SD; n ≥ 10. f Adipose tissue of 30–34-week-old mice was subjected to qPCR analysis using gene-specific primers. Expression levels of the lipases adipose triglyceride lipase (Atgl), hormone sensitive lipase (Hsl), monoacylglycerol lipase (Mgl), and lipoprotein lipase (Lpl) as well as perilipin 4 (Plin4) were assessed in visceral adipose tissue of female and male mice. Here, we observed no expression changes in females whereas expression is significantly altered for Hsl and Lpl in male Lphn1-deficient mice. Data are given as means ± SEM; n ≥ 3 normalized to β-microglobulin/-actin (18.62 ± 0.39). Note that high ΔCt values correspond to low expression and vice versa. g Phosphorylation levels of different PKA substrates in visceral (vWAT) and subcutaneous (scWAT) white adipose tissue in Lphn1 KO mice, as assessed by a broad phospho-PKA-substrates antibody. In scWAT samples of male mice, a trend towards a reduced fraction of phosphorylated PKA substrates is visible. Quantification of Western blot images from Supplementary Fig. S2

Fig. 5

Variants of human LPHN1 discovered in cohort analysis show altered expression and signal transduction. a Different LPHN1 variants were previously linked to neurodevelopmental defects and are shown in black. Some of the patients also displayed overweight/obesity. Two novel variants identified in this study are located in the C terminus of the receptor and are shown in red. b Schematic depiction of the proteins resulting from the two LPHN1 variants identified in children of the Leipzig Childhood Adipose Tissue cohort. The identified two variants contain a point mutation yielding a premature stop codon (R1265*) and a base insertion causing a frame shift that leads to a premature stop codon at amino acid position 1324 (G1321fs), respectively. The child carrying the latter variant is suffering from overweight/obesity. Both variants are located in the C terminus of the receptor (red arrows), thereby truncating it. Wild-type full-length human LPHN1 contains various domains: RBL, rhamnose-binding domain; OLF, olfactomedin domain; HRM, hormone-binding domain; GAIN, GPCR-autoproteolysis-inducing domain; GPS, GPCR proteolytic site. c Cell surface expression of rLPHN1, hLPHN1 and its two variants in HEK293T and COS-7 cells determined by ELISA. All receptors are detectable on the cell surface in both cell lines. While in COS-7 cells, no difference between the variants and the full-length wild-type receptor are present, in HEK293T cells, cell surface levels of the two variants are reduced. Data are displayed as percentage of the ADP receptor P2Y₁₂ (positive control) and given as means ± SEM of at least six independent experiments, each performed in triplicate. The non-specific OD values (empty vector) are 0.001 ± 0.003 (HEK293T) and 0.025 ± 0.020 (COS-7) and the OD values of P2Y₁₂ are 0.75 ± 0.07 (HEK293T) and 0.70 ± 0.11 (COS-7) (set 100%). ns = not significant; **p < 0.01; ***p < 0.001. d To test for functional coupling of hLPHN1 and the variants, accumulation of the second messenger cAMP was determined as a measure for G_s activation. The rat LPHN1 homolog rLPHN1 served as positive control. Basal activity of hLPHN1 is even higher than rLPHN1 signals (in COS-7 cells), while LPHN1(R1265*) and LPHN1(G1321fs) signals are reduced in both, COS-7 and HEK293T cells. cAMP levels are displayed as means ± SEM of independent experiments, each performed in triplicate. Basal second messenger levels (250 ng empty vector) are 9.2 ± 6.3 nM cAMP (HEK293T) and 7.5 ± 5.9 nM cAMP (COS-7), *p < 0.05; ***p < 0.001, compared to the corresponding value of hLPHN1. e Schematic depiction of hLPHN1 with its tethered agonist (Stachel) sequence. This sequence, which also the sequence of the receptor-activating peptide pLPHN1, is located C-terminally of the cleavage site within the GPS and is identical to the one of rat LPHN1. scLPHN1 is a scrambled version of it. f The LPHN1 variants are activated by the Stachel sequence-derived peptide. The respective scrambled peptide of the same amino acid length and composition as the agonistic peptide served as negative control. HEK293T cells transfected with empty control vector (pcDps) or plasmid encoding the respective LPHN1 variant were incubated with 1 mM of different peptides, and subsequently, cAMP concentrations were quantified as a measure for Gs coupling. As a positive control, the mouse melanocortin receptor 4 (MC4R) was used, which was incubated with 10 µM α-MSH. The basal cAMP levels (empty vector, no peptide) are 94.7 ± 12.9 nM. Data are given as means ± SEM of four independent experiments, each performed in triplicate. *p < 0.05; **p < 0.01 compared to the respective unstimulated control. g CRE reporter gene assays reveal different activation of LPHN1 variants. HEK293T cells transfected with empty vector control (pcDps) or plasmid encoding the respective LPHN1 variant together with a vector carrying a CRE-luciferase were incubated with 1 mM of LPHN1-activating or scrambled peptide, respectively. Subsequently, luciferase activity was measured. Data is presented as fold changes over empty vector control without stimulation and given as means ± SEM of four independent experiments, each performed in triplicate. *p < 0.05; **p < 0.01: ***p < 0.001 compared to the respective unstimulated control if not indicated otherwise. Basal signals of the reporter assay (empty vector, no peptide) are 6,415 ± 182