A neuronal circuit driven by GLP-1 in the olfactory bulb regulates insulin secretion

Abstract

Glucagon-like peptide 1 (GLP-1) stimulates insulin secretion and holds significant pharmacological potential. Nevertheless, the regulation of energy homeostasis by centrally-produced GLP-1 remains partially understood. Preproglucagon cells, known to release GLP-1, are found in the olfactory bulb (OB). We show that activating GLP-1 receptors (GLP-1R) in the OB stimulates insulin secretion in response to oral glucose in lean and diet-induced obese male mice. This is associated with reduced noradrenaline content in the pancreas and blocked by an α₂-adrenergic receptor agonist, implicating functional involvement of the sympathetic nervous system (SNS). Inhibiting GABA_A receptors in the paraventricular nucleus of the hypothalamus (PVN), the control centre of the SNS, abolishes the enhancing effect on insulin secretion induced by OB GLP-1R. Therefore, OB GLP-1-dependent regulation of insulin secretion relies on a relay within the PVN. This study provides evidence that OB GLP-1 signalling engages a top-down neural mechanism to control insulin secretion via the SNS.

Fig. 1. PPG neurons and GLP-1R+ neurons are present in the OB.

a DAPI-stained mouse OB section. GCL, granular cell layer; MCL, mitral cell layer; EPL, external plexiform layer; GL, glomerular layer. Scale bar = 200 µm. Ten biological replicates were conducted, obtaining similar results. b Representative photomicrographs of a PPG-YFP mouse OB showing YFP labelling of PPG cells in the GCL. Scale bar = 100 µm (left), 50 µm (right). Four biological replicates were conducted, obtaining similar results. c Strategy to deliver fluorescent Ex4 in the OB of OB-cannulated C57BL6 mice. The creation of this figure includes elements inspired by the style used in Montaner et al.. d Photomicrograph of representative coronal OB sections from a fluorescent Ex4-injected RD mouse. Scale bar = 100 µm. Ten biological replicates were conducted, obtaining similar results. e-h RT-qPCR of Glp1r (e; n = 16 RD and 12 WD; p-value = 0.0097), Pcsk1 (f; n = 6 RD and 6 WD; p-value = 0.0099), Dpp4 (g; n = 6 RD and 6 WD; p-value = 0.0005) and Ppg (h; n = 10 RD and 6 WD) in the OB of RD and WD mice. Data are compared using unpaired Student’s t test. i–k RT-qPCR of Glp1r (i; n = 5 RD and 5 WD; p-value = 0.0066), Pcsk1 (j; n = 6 RD and 6 WD; p-value = 0.0413), Dpp4 (k; n = 5 RD and 6 WD) in the HT of RD and WD mice. Data are compared using two-sided unpaired Student’s t test. Data are given as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ns not significant. Source data are provided as a Source Data file.

Fig. 2. Pharmacological modulation of GLP-1R in the OB regulates glucose homeostasis.

a Weight gain of age-matched RD and WD mice from week 1 to 25 under WD (n = 9/diet group; p-value _diet = 0.0002). At week 17 mice undergo surgery. Data were analysed using Repeated Measures (RM) two-way ANOVA followed by Bonferroni post-hoc test. b Body weight, fat mass and lean mass of OB-cannulated RD and OB-cannulated WD mice at 16 weeks (n = 9/diet group, p-value Body Weight <0.0001; p-value Body Composition <0.0001). Data were analysed using two-sided unpaired Student’s t test. c Schematic illustration of bilateral OB-injections of GLP-1 (0.5 μg/μl), Ex4 (0.5 μg/μl) and Ex9 (12.5 μg/μl) followed by an OGTT or ITT in OB-cannulated RD and WD mice. The creation of this figure includes elements inspired by the style used in Montaner et al.. d–f OGTT tests combined with preceding OB injections of GLP-1 (d; n = 11; p-value OGTT _diet*drug < 0.0001; p-value AUC _RD = 0.0009; p-value AUC _WD < 0.0001), Ex4 (e; n = 10 RD, n = 9 WD; p-value OGTT _diet*drug = 0.018; p-value AUC _WD = 0.001) or Ex9 (f; n = 5 RD, n = 4 WD; p-value OGTT _diet*drug = 0.002; p-value AUC _RD = 0.007; p-value AUC _WD < 0.008) in OB-cannulated RD and WD mice after 20 weeks under WD (26-week-old). Black squares on X axis indicate drug delivery before the glucose challenge (t0 min). Inset above, right: incremental AUC of glycaemia. OGTT curves were analysed using RM three-way (diet, drug, time) mixed ANOVA; AUCs were analysed using RM two-way (diet, drug) mixed ANOVA followed by two-sided paired Student’s t tests (within diet) with Bonferroni correction. g–i Plasma insulin levels of OB-cannulated RD and WD mice measured during GLP-1 (g; n = 11; p-value t0 _WD = 0.0307; p-value t15 _WD = 0.0023), Ex4 (h; n = 10 RD, n = 9 WD; p-value t0 _WD = 0.0276; p-value t15 _WD = 0.0056; p-value t30 _WD = 0.007; p-value t60 _WD = 0.0114) or Ex9 (i; n = 5 RD, n = 4 WD) OB-injected OGTT tests (tests in d–f respectively). Data were analysed using RM three-way ANOVA followed by two-sided paired Student’s t tests (within diet) with Bonferroni correction. j–l AUC of ITT tests on mice previously OB-injected with GLP-1 (j; n = 9 RD, 12 WD; p-value _WD = 0.001), Ex4 (k; n = 9 RD, 7 WD; p-value _WD = 0.017), or Ex9 (l; n = 4 RD, n = 3 WD; p-value _WD = 0.033). WD mice were fed a WD for 24 weeks. The AUC is measured from baseline (before central perfusions, t-12) until t30 following insulin delivery (t0). Data were analysed using RM two-way (diet, drug) mixed ANOVA followed by two-sided paired Student’s t tests (within diet) with Bonferroni correction. Data are given as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ns not significant. Source data are provided as a Source Data file.

Fig. 3. Chemogenetic modulation of Glp1r-expressing cells in the OB increases insulin secretion in WD mice.

a Weight gain of Glp1r-Cre RD and WD mice from week 1 to 25 under WD (n = 5/diet group; p-value _diet < 0.0001). At week 17 mice undergo surgery. Data were analysed using RM two-way ANOVA followed by Bonferroni post-hoc test. b Body weight, fat mass and lean mass of Glp1r-Cre RD and WD mice at 16 weeks under WD diet (n = 5/diet group; p-value Body Weight = 0.0016; p-value Body Composition <0.0001). Body composition (fat and lean mass) was expressed as a percentage of body weight. Data were analysed using two-sided unpaired Student’s t test. c Schematic illustration of viral delivery in the OB of Glp1r-Cre RD and WD mice (left). SALB or CNO delivery prior to the onset of the OGTT (right). The creation of this figure includes elements inspired by the style used in Montaner et al.. d OGTT preceded by an IP injection of SALB in Glp1r^OBKORD 26-week-old. The red arrow indicates the SALB injection before the glucose gavage (t0) (left). AUC of glycaemia. OGTT data were analysed using RM two-way ANOVA followed by Bonferroni post-hoc test and AUCs by using two-sided paired Student’s t test (n = 5) (right). e Plasma insulin levels of Glp1r^OBKORD RD mice after SALB IP injection (RM two-way ANOVA followed by Bonferroni post-hoc test; n = 5; p-value _NaCl = 0.0027; p-value _KORD = 0.0162). f OGTT tests preceded by an IP injection of CNO in Glp1r^OBhM3Dq WD mice after 20 weeks under WD (26-week-old, n = 5; p-value OGTT _drug = 0.005; p-value AUC = 0.0222). The green arrow indicates the CNO injection before the glucose gavage (t0) (left). AUC of glycaemia (right). OGTT data were analysed using RM two-way ANOVA followed by Bonferroni post-hoc test and AUCs by using two-sided paired Student’s t test. g Plasma insulin levels of Glp1r^OBhM3Dq WD mice after CNO IP injection (RM two-way ANOVA followed by Bonferroni post-hoc test; n = 5; p-value t₁₅ = 0.0492). h Schematic of SALB or CNO delivery prior to the refeeding phase. Food intake measurements were performed 1 h after drug delivery. The creation of this figure includes elements inspired by the style used in Montaner et al. . i–k Food intake in Glp1r^OBKORD RD mice (n = 5; p-value = 0.0169) fed a chow diet (i) and in Glp1r^OBhM3Dq WD mice (n = 9) fed both a chow (j; p-value = 0.0403) and a HFHS (k; p-value = 0.0162) diet 1 h after IP injection of SALB and CNO, respectively. Data are analysed using two-sided paired Student’s t test. Data were given as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ns not significant. Source data are provided as a Source Data file.

Fig. 4. The stimulatory effect of GLP-1 in the OB on insulin secretion is mediated by a decrease in sympathetic nervous system activity.

a Pancreatic NA levels of RD and WD mice after Ex4 injection in the OB (n = 9 mice in total, p-value _treatment = 0.0054). The dataset includes four groups: NaCl OB RD (n = 2), NaCl OB WD (n = 3), Ex4 OB RD (n = 2) and Ex4 OB WD (n = 2). For NaCl OB RD, the values are as follows: minimum = 555.9, 25th percentile = 555.9, median = 665.2, 75th percentile = 774.5 and maximum = 774.5. For NaCl OB WD, the values are: minimum = 303.5, 25th percentile = 303.5, median = 376.8, 75th percentile = 469.6 and maximum = 469.6. For Ex4 OB RD, the values are: minimum = 285.4, 25th percentile = 285.4, median = 337.8, 75th percentile = 390.2 and maximum = 390.2. For Ex4 OB WD, the values are: minimum = 13.49, 25th percentile = 13.49, median = 25.98, 75th percentile = 38.46 and maximum = 38.46. Data were analysed using ordinary two-way ANOVA followed by Fisher’s LSD pairwise comparisons. b Schematic illustration of bilateral OB injections of Ex4 combined with IP administration of UK14304 followed by an OGTT in OB-cannulated WD mice. The creation of this figure includes elements inspired by the style used in Montaner et al. . c OGTT tests preceded by an IP injection of UK14304 and an OB injection of Ex4 in OB-cannulated WD after 20 weeks under WD (n = 8; p-value OGTT treatment = 0.0047) (left). AUC of glycaemia (right); p-value AUC NaCl IP - NaCl OB vs. NaCl IP - Ex4 OB = 0.0159; p-value AUC NaCl IP - Ex4 OB vs. UK14304 IP - Ex4 OB = 0.0490. OGTT data were analysed using RM two-way ANOVA and AUCs by using RM one-way ANOVA followed by Bonferroni. d Plasma insulin levels of OB-cannulated WD during OGTT tests preceded with OB-injected Ex4 and IP injected UK14304 (RM two-way ANOVA followed by Bonferroni; n = 8; p-value t15 NaCl IP - NaCl OB vs. NaCl IP - Ex4 OB = 0.0313; p-value t15 NaCl IP - Ex4 OB vs. UK14304 IP - Ex4 OB = 0.0324; p-value t30 NaCl IP - NaCl OB vs. NaCl IP - Ex4 OB = 0.0078; p-value t30 NaCl IP - Ex4 OB vs. UK14304 IP - Ex4 OB = 0.0055). e Strategy to deliver Ex4 or vehicle in the OB of OB-cannulated C57BL6 mice prior to cFos counting in the MCL (f) or LH (g). The creation of this figure includes elements inspired by the style used in Montaner et al. . f Representative photomicrographs of cFos immunoreactivity and automated cell counting in the MCL using CellProfiler (left), scale bar: 200 µm. cFos density in the MCL after Ex4 or vehicle injection in the OB of WD mice (right). Data are expressed as cFos+ nuclei/mean MCL area of all mice. Data were analysed using two-sided unpaired Student’s t test (n = 5; p-value = 0.0401). g Representative photomicrographs of cFos immunoreactivity in the LH (left). cFos density in the LH after Ex4 or saline injection in the OB of WD mice (right). Data are expressed as cFos+ nuclei/mean LH area of all mice. Data were analysed using two-sided unpaired Student’s t test (n = 4; p-value = 0.0227). h Schematic illustration of unilateral Alexa 488-conjugated cholera toxin B subunit (CTB_green) delivery in the LH for retrograde tracing of the primary olfactory cortex. The creation of this figure includes elements inspired by the style used in Montaner et al. . i Coronal sections of the LH after unilateral injection of CTB_green (left). CTB staining in the primary olfactory cortex (right). LH lateral hypothalamus, AHP anterior hypothalamus – posterior part, VMH ventromedial hypothalamus, 3V third ventricle, AON anterior olfactory nucleus, VTT ventral tenia tecta, VON ventral olfactory nucleus, PC piriform cortex, OT olfactory tubercle. Structure boundaries were drawn based on Franklin and Paxinos mouse brain atlas. Scale bar: 200 µm (left) and 100 µm (right). Four biological replicates were conducted, obtaining similar results. j Schematic illustration of viral delivery in the OB and bilateral cannula placement in the PVN of Glp1r-Cre WD mice followed by OGTTs. The creation of this figure includes elements inspired by the style used in Montaner et al.. k OGTT tests preceded with an IP injection of CNO in Glp1r^OBhM3Dq WD PVN-cannulated mice after 20 weeks under WD (26-week old; n = 10; p-value OGTT _treatment = 0.0019) (left). The black square and the green line indicate Bicuculline and CNO delivery, respectively, before the glucose gavage (t0). AUC of glycaemia (right; p-value NaCl PVN - NaCl IP vs. NaCl PVN - CNO IP = 0.0013; p-value NaCl PVN - CNO IP vs. Bic PVN - NaCl IP = 0.0213; p-value NaCl PVN - CNO IP vs. Bic PVN - CNO IP = 0.0446). OGTT data were analysed using RM two-way ANOVA and AUCs by using RM one-way ANOVA, both followed by Bonferroni post-hoc tests. l Plasma insulin levels of Glp1r^OBhM3Dq WD PVN-cannulated mice after CNO IP injection (RM two-way ANOVA followed by Bonferroni, n = 10; p-value t15 NaCl PVN - NaCl IP vs. NaCl PVN - CNO IP < 0.0001; p-value t15 NaCl PVN - CNO IP vs. Bic PVN - CNO IP < 0.0001). Data are given as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ns not significant. Source data are provided as a Source Data file.