Central and peripheral GLP-1 systems independently suppress eating

Abstract

The anorexigenic peptide glucagon-like peptide-1 (GLP-1) is secreted from gut enteroendocrine cells and brain preproglucagon (PPG) neurons, which, respectively, define the peripheral and central GLP-1 systems. PPG neurons in the nucleus tractus solitarii (NTS) are widely assumed to link the peripheral and central GLP-1 systems in a unified gut-brain satiation circuit. However, direct evidence for this hypothesis is lacking, and the necessary circuitry remains to be demonstrated. Here we show that PPG^NTS neurons encode satiation in mice, consistent with vagal signalling of gastrointestinal distension. However, PPG^NTS neurons predominantly receive vagal input from oxytocin-receptor-expressing vagal neurons, rather than those expressing GLP-1 receptors. PPG^NTS neurons are not necessary for eating suppression by GLP-1 receptor agonists, and concurrent PPG^NTS neuron activation suppresses eating more potently than semaglutide alone. We conclude that central and peripheral GLP-1 systems suppress eating via independent gut-brain circuits, providing a rationale for pharmacological activation of PPG^NTS neurons in combination with GLP-1 receptor agonists as an obesity treatment strategy.

Extended Data Fig. 1. PPG^NTS neurons selectively encode large meal satiation

(a) Experimental model and paradigm for metabolic phenotyping of PPG^NTS-DTA mice (DTA, n=8) or mCherry-transduced controls (mCh, n=7). n=8 (DTA) / 7 (mCh) animals for analyses presented in b-o. (b) Cumulative hourly food intake over 1 day, 2-way mixed-model ANOVA: Virus F_(1,13)=0.015, p=0.904. (c) Daily food intake by sex, 2-way mixed-model ANOVA: Virus F_(1,11)=0.012, p=0.914; Sex F_(1,11)=0.683, p=0.426. (d) Directed ambulatory locomotion (excluding fine movements) over 1 day, 2-way mixed-model ANOVA: Virus F_(1,11)=0.493, p=0.497. (e-i) Meal pattern and metabolic parameters over 1 day, unpaired 2-tailed t-test or Mann-Whitney U test: e) U=25, p=0.779; f) t₍₁₃₎=0.997, p=0.337; g) t₍₁₃₎=0.565, p=0.582; h) t₍₁₃₎=0.797, p=0.440; i) t₍₁₃₎=0.323, p=0.752. (j) Mean bodyweight over the 24h test period, unpaired 2-tailed t-test: t₍₁₃₎=0.883, p=0.393. (k-l) Food intake during dark and light phases, unpaired 2-tailed t-test: k) t₍₁₃₎=0.668, p=0.516; l) t₍₁₃₎=1.251, p=0.233. (m) 24h water intake, unpaired 2-tailed t-test: t₍₁₃₎=0.205, p=0.841. (n) Ensure liquid diet preload intake, unpaired 2-tailed t-test: t₍₈₎=0.219, p=0.832; and post-Ensure chow intake, Mann-Whitney U test: U=2, p=0.038. (o) Post-fast refeed intake, unpaired 2-tailed t-test: t₍₁₂₎=2.501, p=0.028. (p-q) Hourly and cumulative intakes over 1 day from ad libitum eating PPG^NTS-hM4Di mice (n=8 animals), 2-way within-subjects ANOVA: p) Drug F_(1,7)=0.241, p=0.639; q) Drug F_(1,7)=0.411, p=0.542. (r) Raster plot of chow pellet retrievals during the dark phase. Plots from the same mouse after saline and CNO injections presented adjacently. All data presented as mean ± SEM.

Extended Data Fig. 2. PPG^NTS neurons suppress eating without behavioural disruption

(a) Non-cumulative hourly food intake over the circadian cycle from ad libitum eating PPG^NTS-hM3Dq mice (n=7 animals for analyses presented in a-b), 2-way within-subjects ANOVA: Drug × Time F_(23,138)=4.599, p<0.0001. (b) Dark phase food intake by sex, 2-way mixed-model ANOVA: Drug F_(1,5)=19.97, p=0.0066; Sex F_(1,5)=3.854, p=0.107. (c-d) Photomicrographs of co-localised cFos immunoreactivity and DIO-hM3Dq-mCherry in NTS of PPG-Cre:tdRFP mice perfused 3 hours after injection of saline or CNO (photomicrographs representative of independent experiments from 4/3 animals), cc: central canal. Scale=100μm (inset 50μm). (e) Proportion of mCherry-expressing neurons co-localised with cFos-ir in mice perfused after administration of saline or CNO (n=4 (SAL) / 3 (CNO) animals), unpaired 1-tailed t-test: t₍₅₎=13.94, p<0.0001. (f) Non-cumulative hourly food intake over 1 day from 18h fasted PPG^NTS-hM3Dq mice (n=7 animals for analyses presented in f-h), 2-way within-subjects ANOVA: Drug × Time F_(23,138)=3.745, p<0.0001. The behavioural satiety sequence (BSS) was analysed during the first 40 minutes of the dark phase. (g-h) Food intake and eating rate over 40 minute BSS test, paired 2-tailed t-test and Wilcoxon matched pairs test: g) t₍₆₎=4.088, p=0.0064; h) W=18, p=0.156. All data presented as mean ± SEM.

Extended Data Fig. 3. Glp1r-expressing VANs suppress eating and condition flavour avoidance

(a-c) Light phase food intake and metabolic parameters from ad libitum eating GLP-1R^Nodose-hM3Dq mice (n=7 animals for analyses presented in a-c), paired 2-tailed t-test: a) t₍₆₎=0.0141, p=0.989; b) t₍₆₎=0.952, p=0.378; c) t₍₆₎=0.0406, p=0.969. (d) Photomicrographs of cFos immunoreactivity (cFos-ir) in coronal NTS sections from GLP-1R-Cre × PPG-YFP mice bilaterally injected in nodose ganglia with dye (Control) or AAV9-DIO-hM3Dq-mCherry (hM3Dq) and administered saline or CNO (photomicrographs representative of independent experiments from 3/3 animals). Distance in mm posterior to Bregma in bottom left, cc: central canal. Scale=100μm. (e) cFos immunoreactive cells in the NTS (mean per section) of control) and hM3Dq mice (n=3 animals per group for analyses in e-f), unpaired 2-tailed t-test: t₍₄₎=2.981, p=0.0407. (f) PPG^NTS neurons co-localised with cFos immunoreactivity in the NTS. Mann-Whitney 2-tailed U-test: U=0, p=0.100. (g) Photomicrograph of nodose ganglion section from GLP-1R-Cre:tdRFP mouse injected with AAV encoding Cre-dependent channelrhodopsin and eYFP fluorescent reporter (DIO-CHR2-eYFP), and co-localisation of the tdRFP and eYFP reporters (photomicrographs representative of independent experiments from 4 animals). Scale=100μm. (h-i) Quantification of viral transduction specificity (h; co-localised cells as % (±SEM) of all eYFP+ cells) and efficiency (i; co-localised cells as % (±SEM) of all tdRFP+ cells), from a total of 374 tdRFP+ cells and 366 eYFP+ cells from the nodose ganglia of 3 mice. All data presented as mean ± SEM.

Extended Data Fig. 4. Oxtr rather than Glp1r VANs are the major vagal input to PPG^NTS neurons

(a) Photomicrograph of coronal NTS section from PPG-Cre:tdRFP mouse transduced with DIO-TVA-mCherry + DIO-RabiesG, and subsequently with rabies virus-ΔG-GFP (RABV). Bilateral NTS injection of TVA+RabiesG and counterbalanced unilateral injection of RABV (4 mice / side) resulted in 40.5% (±5.5) of all PPG^NTS neurons being successfully transduced ‘starter’ neurons, identified by co-localization of mCherry (and/or tdRFP) and GFP (photomicrographs in a-h representative of independent experiments from 8 animals). Despite unilateral RABV injection, starter neurons were observed in left and right NTS in all mice, indicating significant viral spread and bilateral transduction. Scale=100μm. (b) Total RABV+ cells in left and right nodose ganglia (LNG / RNG; n=6 / 7 biologically independent samples), unpaired 2-tailed t-test: t₍₁₁₎=0.214, p=0.834. (c-d) Quantification of Glp1r and Oxtr co-localization in nodose ganglia (c), and proportions of dual-expressing Glp1r / Oxtr cells co-localised with RABV (d). This dual population comprises 24.7% of all Glp1r cells and 19.7% of all Oxtr cells. 9% of RABV+ vagal inputs to PPG^NTS neurons express both Glp1r and Oxtr, and 26.1% of dual-expressing Glp1r / Oxtr cells are RABV+ vagal inputs to PPG^NTS neurons. (e) Quantification of RABV and Glp1r co-localization in NG as proportions of all RABV+ cells and all Glp1r+ cells, including those Glp1r cells that also express Oxtr. (f) Quantification of RABV and Oxtr co-localization in NG as proportions of all RABV+ cells and all Oxtr+ cells, including those Oxtr cells that also express Glp1r. (g-h) Photomicrographs of left and right nodose ganglion sections showing rabies virus GFP expression (RABV) and Glp1r and Oxtr FISH. RABV+Glp1r co-localization shown by white arrows, RABV+Oxtr by green arrows and RABV+Glp1r+Oxtr by white-edged green arrow. Scale=100μm. All data presented as mean ± SEM.

Extended Data Fig. 5. PPG^NTS neurons are necessary for oxytocin-induced eating suppression

(a-b) Food intake and bodyweight change over 1 day in eGFP and DTA mice (n=5 (DTA) / 7 (eGFP) animals) administered oxytocin (0.4 mg/kg, i.p.), 2-way mixed-model ANOVA: a) Drug F_(1,10)=0.00474, p=0.947; b) Drug F_(1,10)=0.0989, p=0.760. (c-d) Photomicrographs of coronal NTS sections from PPG-Cre:GCaMP3 mice injected with eGFP control virus (c) or DTA virus (d). Note the complete absence of green (GCaMP3-expressing, amplified by immunostaining against the GFP antigen) PPG^NTS neurons in DTA-ablated tissue, and the extent of viral spread as demonstrated by constitutive expression of mCherry (photomicrographs representative of independent experiments from 7/5 animals). Distance in mm posterior to Bregma in bottom left, cc: central canal. Scale=100μm. All data presented as mean ± SEM.

Extended Data Fig. 6. PPG^NTS neurons are not a major synaptic target of area postrema Glp1r neurons

(a) Photomicrographs of coronal NTS section showing RABV expression, Glp1r FISH and TH-ir. RABV+Glp1r+TH-ir co-localization shown by white-edged green arrows (photomicrographs representative of independent experiments from 4 animals). Scale=100μm (inset 20μm). (b-c) Quantification of Glp1r and TH-ir co-localization in area postrema (b), and proportions of dual Glp1r / TH-ir cells co-localised with RABV (c). This dual population comprises 49.4% of all TH-ir cells and 31.2% of all Glp1r cells. 9.7% of RABV+ AP inputs to PPG^NTS neurons express Glp1r and are TH-ir, and 2.7% of dual Glp1r / TH-ir cells are RABV+ AP inputs to PPG^NTS neurons. All data presented as mean ± SEM.

Extended Data Fig. 7. Liraglutide and semaglutide suppress eating independently of PPG^NTS neurons

(a-e) Cumulative food intake by virus at 1,2,4,6 and 21hr in eGFP and DTA mice (n=8 (DTA) / 7 (eGFP) animals for analyses presented in a-j) administered liraglutide (200 μg/kg, s.c.), 2-way mixed-model ANOVA: a) Drug F_(1,13)=0.246, p=0.628; b) Drug F_(1,13)=2.108, p=0.170; c) Drug F_(1,13)=37.44, p<0.0001, Virus F_(1,13)=0.836, p=0.377; d) Drug F_(1,13)=75.09, p<0.0001, Virus F_(1,13)=1.877, p=0.194; e) Drug F_(1,13)=154.9, p<0.0001, Virus F_(1,13)=1.272, p=0.280. (f-j) Cumulative food intake by virus at 1,2,4,6 and 21hr in eGFP and DTA mice administered semaglutide (60 μg/kg, s.c.), 2-way mixed-model ANOVA: f) Drug F_(1,13)=1.965, p=0.184; g) Drug F_(1,13)=17.1, p=0.0012; Virus F_(1,13)=0.630, p=0.442; h) Drug F_(1,13)=82.49, p<0.0001, Virus F_(1,13)=0.332, p=0.574; i) Drug F_(1,13)=98.21, p<0.0001, Virus F_(1,13)=0.840, p=0.376; j) Drug F_(1,13)=126.1, p<0.0001, Virus F_(1,13)=3.42, p=0.0873. (k-m) Representative photomicrographs of cFos immunoreactivity (cFos-ir) in arcuate nucleus of the hypothalamus (ARC) 4 hours after vehicle (VEH, n=4 animals) or semaglutide (SEMA, 60 μg/kg, s.c., n=4 animals) administration, and total cFos count, unpaired 1-tailed t-test: m) t₍₆₎=2.614, p=0.020. Scale=100μm. (n-p) Representative photomicrographs of cFos-ir in paraventricular nucleus of the hypothalamus (PVN) 4 hours after vehicle or semaglutide administration (n=4 / 4 animals), and total cFos count, unpaired 1-tailed t-test: p) t₍₆₎=5.109, p=0.0011. Scale=100μm. (q-t) Representative photomicrographs of cFos-ir in dorsal lateral and external lateral subdivisions of the parabrachial nucleus (dlPBN / elPBN) 4 hours after vehicle or semaglutide administration (n=3 / 4 animals), and total cFos count, unpaired 1-tailed t-tests: s) t₍₅₎=1.693, p=0.0756; t) t₍₅₎=3.57, p=0.0080. Semaglutide did not increase cFos-ir in the medial PBN, t₍₅₎=0.435, p=0.341. Scale=100μm. All data presented as mean ± SEM.

Extended Data Fig. 8. PPG^NTS neuron activation augments semaglutide-induced eating suppression

(a-d) Bodyweight change at 24 and 48 hours, and cumulative food intake at 48 and 72 hours (n=6 animals), 1-way within-subjects ANOVA: a) Drug F_(2.1,10.5)=61.61, p<0.0001; b) Drug F_(2.3,11.3)=102.7, p<0.0001; c) Drug F_(2.1,10.6)=24.38, p<0.0001; d) Drug F_(1.9,9.3)=40.35, p<0.0001. 72hr BW data not shown: Drug F_(2.0,10.2)=4.22, p=0.0454, no significant pairwise comparisons. (e) Photomicrographs of coronal NTS sections from PPG-Cre:GCaMP3 mice injected with AAV encoding Cre-dependent hM3Dq and mCherry fluorescent reporter (DIO-hM3Dq-mCherry), and co-localisation of the GCaMP3 (amplified by immunostaining against GFP antigen) and mCherry reporters (photomicrographs representative of independent experiments from 4 animals). Distance in mm from Bregma in bottom left, cc: central canal. Scale=100μm. (f-g) Quantification of viral transduction specificity (f; co-localised cells as % (±SEM) of all mCherry+ cells) and efficiency (g; co-localised cells as % (±SEM) of all GCaMP3+ cells), from a total of 410 mCherry+ cells and 391 GCaMP3+ cells from 4 mice. All data presented as mean ± SEM.

Figure 1.. PPG^NTS neurons selectively encode large meal satiation

(a) Experimental model and paradigm for meal pattern analysis of post-fast refeeding in PPG^NTS-hM4Di mice using FED system. n=6 animals for analyses presented in b-f. (b) 4h dark phase food intake, 2-way within-subjects ANOVA: Drug × Time F_(3,15)=3.664, p=0.0367. (c) Raster plot of chow pellet retrievals over 1h dark phase. Plots from the same mouse after saline and CNO injections presented adjacently. (d) 1h intake by sex, 2-way mixed-model ANOVA: Drug F_(1,4)=29.09, p=0.0057. (e-f) Meal pattern parameters during 1h refeed, paired 2-tailed t-test or Wilcoxon matched-pairs test: E) t₍₅₎=2.757, p=0.040; F) W=−3, p=0.500. (g) Experimental model and paradigm for temporal analysis of Ensure intake in PPG^NTS-hM4Di mice. n=7 animals for analyses presented in h-k. (h) 1h Ensure intake, paired 2-tailed t-test: t₍₆₎=2.859, p=0.0288. Ensure intake was sex-independent, 2-way mixed-model ANOVA: Sex × Drug F_(1,5)=2.553, p=0.171. (i) Raster plot of Ensure drinking bouts over 1h dark phase. Plots from the same mouse after saline and CNO injections presented adjacently. (j-k) Temporal parameters of Ensure drinking, paired 2-tailed t-test: J) t₍₆₎=6.55, p=0.0006; K) t₍₆₎=1.263, p=0.254; M) t₍₆₎=4.784, p=0.0031. All data presented as mean ± SEM.

Figure 2.. PPG^NTS neurons suppress eating without behavioural disruption

(a) Experimental model and paradigm for ad libitum pellet eating from FED in PPG^NTS- hM3Dq mice. n=7 animals for analyses presented in b-d. (b) Daily food intake during 48h test, 2-way within-subjects ANOVA: Drug × Day F_(1,6)=14.52, p=0.0089. (c) Cumulative hourly food intake over two days, 2-way within-subjects ANOVA: Drug × Time F_(48,288)=6.481, p<0.0001. (d) 24h and 48h bodyweight change, 2-way within-subjects ANOVA: Drug F_(1,6)=10.41, p=0.018. (e) Experimental model and paradigm for BSS analysis in 18h fasted PPG^NTS-hM3Dq mice. n=7 animals for analyses presented in f-k. (f-g) Behavioural satiety sequences following saline and CNO injections. Satiation point/satiety onset (when duration inactive exceeds eating) shown by dotted lines. (h-k) Quantitative analysis of hM3Dq effect on BSS behaviours, 2-way with-subjects ANOVA: h) Drug × Time F_(7,42)=5.673, p=0.0001; i) Drug F_(1,6)=5.261, p=0.0616; j) Drug F_(1,6)=12.48, p=0.0123; k) Drug F_(1,6)=4.028, p=0.0915. All data presented as mean ± SEM.

Figure 3.. Glp1r-expressing VANs suppress eating and condition flavour avoidance

(a) Experimental model and paradigm for food intake and metabolic analysis of ad libitum eating GLP-1R^Nodose-hM3Dq mice. n=7 animals for analyses presented in b-e. (b) Cumulative hourly dark phase food intake, 2-way within-subjects ANOVA: Drug × Time F_(12,144)=2.078, p=0.0218. (c-e) Dark phase metabolic parameters and 24h bodyweight change, paired 2-tailed t-test: c) t₍₆₎=1.642, p=0.152; d) t₍₆₎=0.543, p=0.607; e) t₍₆₎=2.323, p=0.0296. (f) Experimental model and paradigm for optogenetically-evoked conditioned flavour preference and intake analysis in GLP-1R^Nodose-ChR2 mice. n=5 animals for analyses presented in h-i; n=5 (Ctrl) / 4 (ChR2) for analyses presented in k-l. (g) Z-projection photomicrograph of ChR2-mCherry expression in nodose ganglia (representative of 7 independent experiments). Scale=100μm. (h-i) Conditioned stimulus (CS+) preference and 0.5h food intake, paired 2-tailed t-test: h) t₍₄₎=3.216, p=0.0324; i) t₍₄₎=3.976, p=0.0165. (j) Photomicrographs of cFos immunoreactivity (cFos-ir) in coronal NTS sections from GLP-1R-Cre × PPG-YFP mice bilaterally injected in nodose ganglia with control virus (Control) or AAV9-DIO-ChR2-eYFP (ChR2) and exposed to blue light (photomicrographs representative of independent experiments from 4/5 animals). Distance in mm posterior to Bregma in bottom left, cc: central canal. Scale=100μm. (k) Total cFos immunoreactive cells in the NTS of control and ChR2 mice, unpaired 2-tailed t-test: t₍₇₎=4.122, p=0.0044. (l) PPG^NTS neurons co-localised with cFos immunoreactivity in the NTS. Mann-Whitney 2-tailed U-test: U=6, p=0.4127. All data presented as mean ± SEM.

Figure 4.. Oxtr rather than Glp1r VANs are the major vagal input to PPG^NTS neurons

(a) Experimental model for viral-mediated mapping of left and right Glp1r vagal afferent projections to the NTS, and photomicrographs of tdTomato expression in virus-injected nodose ganglia (NG) and non-injected contralateral NG (photomicrographs in a-c representative of independent experiments from 3 animals per injection side). Scale=100μm. (b-c) Photomicrographs of tdTomato-expressing terminal fields of L and R branch Glp1r vagal afferents along the rostro-caudal extent of the NTS (mm posterior to Bregma in bottom left) in PPG-YFP mice, cc: central canal. Scale=100μm. (d) Experimental model for rabies virus (RABV)-mediated monosynaptic retrograde tracing of vagal inputs to PPG^NTS neurons combined with RNAscope fluorescence in situ hybridization for GLP-1R (Glp1r) and oxytocin receptor (Oxtr) transcripts (photomicrographs in e-i representative of independent experiments from 8 animals). (e) Photomicrograph of nodose ganglion showing rabies virus GFP expression (RABV) and Glp1r FISH. RABV+Glp1r co-localization shown by white arrows, RABV+Glp1r+Oxtr by white-edged green arrow. Scale=100μm. (f) RABV and Glp1r co-localization as proportions of all RABV+ cells and all Glp1r+ cells, from 903 RABV+, 1188 Glp1r+ and 1460 Oxtr+ cells from L and R NG. (g) Photomicrograph of nodose ganglion showing RABV expression and Oxtr FISH. Scale=100μm. (h) RABV and Oxtr co-localization as proportions of all RABV+ cells and all Oxtr+ cells. RABV+Oxtr co-localization shown by green arrows, RABV+Glp1r+Oxtr by white-edged green arrow. (i) High magnification Z-projection of RABV, Glp1r and Oxtr cells in NG. Scale=20μm. All data presented as mean ± SEM.

Figure 5.. PPG^NTS neurons are necessary for oxytocin-induced eating suppression

(a) Experimental model for imaging of oxytocin-induced neuronal calcium dynamics in ex vivo brainstem slices from mice expressing GCaMP3 in PPG neurons. n=75 cells from 3 animals examined over 8 independent experiments for analyses in b-e. (b) Representative ΔF/F₀ traces from individual neurons and mean response (purple line) during bath application of oxytocin and glutamate. (c) Representative images of PPG^NTS:GCaMP3 neurons pseudocolored for fluorescence intensity under baseline conditions (aCSF) and responding to oxytocin (purple arrows) and glutamate (grey arrows). (d) Oxytocin-responsive PPG^NTS:GCaMP3 neurons as a proportion of all glutamate-responsive PPG^NTS neurons (i.e. healthy neurons with functional GCaMP3 expression). (e) Median AUC during exposure to oxytocin in oxytocin unresponsive and responsive PPG^NTS:GCaMP3 neurons, Mann-Whitney 2-tailed U-test: U=48, p<0.0001. (f) Experimental model and paradigm for oxytocin-induced eating suppression in PPG^NTS-DTA ablated mice or eGFP-transduced controls. n=5 (DTA) / 7 (eGFP) animals for analyses presented in g-i. (g-h) Cumulative 4h dark phase food intake in eGFP and DTA mice administered oxytocin (0.4 mg/kg, i.p.), 2-way within-subjects ANOVA: g) Drug × Time F_(2,12)=6.133, p=0.0146; h) Drug F_(1,4)=0.0117, p=0.919. (i) 4h food intake by virus, 2-way mixed-model ANOVA: Drug × Virus F_(1,10)=8.472, p=0.0155. All data presented as mean ± SEM except box plot in e, in which the box is centred on the median and bound at 25 and 75%, with whiskers at 5 and 95% and blue cross at the mean.

Figure 6.. PPG^NTS neurons are not a major target of area postrema Glp1r neurons

(a) Experimental model for rabies virus-mediated monosynaptic retrograde tracing of area postrema inputs to PPG^NTS neurons combined with FISH for Glp1r (photomicrographs in b-d representative of independent experiments from 4 animals). (b) Photomicrographs of coronal NTS section showing RABV expression and Glp1r FISH. RABV+Glp1r co-localization shown by white arrows. Scale=100μm (inset 20μm). (c) RABV and Glp1r co-localization as proportions of all RABV+ cells and all Glp1r+ cells, from 53 RABV+ and 549 Glp1r+ cells. (d) Photomicrographs of coronal NTS section showing RABV expression and TH-ir. Examples of RABV+TH-ir co-localization shown by green arrows. Scale=100μm (inset 20μm). (e) Quantification of RABV and TH-ir co-localization as proportions of all RABV+ cells and all TH-ir cells, from a total of 53 RABV+ and 341 TH-ir cells. All data presented as mean ± SEM.

Figure 7.. Liraglutide and semaglutide suppress eating independently of PPG^NTS neurons

(a) Experimental model and paradigm for GLP-1RA-induced eating suppression in PPG^NTS-DTA ablated mice or eGFP-transduced controls. n=8 (DTA) / 7 (eGFP) animals for analyses presented in b-g. (b-d) Cumulative food intake and bodyweight change over 1 day in eGFP and DTA mice administered liraglutide (200 μg/kg, s.c.), 2-way within-subjects or mixed-model ANOVA: b) Drug × Time F_(5,30)=35.35, p<0.0001; c) Drug × Time F_(5,35)=74.95, p<0.000; d) Drug F_(1,13)=33.17, p=0<0.0001, Virus F_(1,13)=1.198, p=0.294. (e-g) Cumulative food intake and bodyweight change over 1 day in eGFP and DTA mice administered semaglutide (60 μg/kg, s.c.), 2-way within-subjects or mixed-model ANOVA: e) Drug × Time F_(5,30)=51.83, p<0.0001; f) Drug × Time F_(5,35)=54.28, p<0.0001; g) Drug F_(1,13)=122.6, p=0<0.0001, Virus F_(1,13)=0.224, p=0.644. (h) Photomicrographs of cFos immunoreactivity (cFos-ir) in coronal NTS sections (mm posterior to Bregma in bottom left) from PPG-YFP mice perfused 4h after vehicle (VEH) or semaglutide (SEMA; 60 μg/kg, s.c.) administration (photomicrographs representative of independent experiments from 3/4 animals). n=3 (VEH) / 4 (SEMA) animals for analyses presented in i-k, cc: central canal. Scale=100μm. (i-j) Total cFos in NTS and AP of mice administered vehicle or semaglutide, unpaired 1-tailed t-tests: i) t₍₅₎=4.59, p=0.0029; j) t₍₅₎=2.66, p=0.0225. (k) PPG^NTS neurons co-localised with cFos immunoreactivity in NTS. Mann-Whitney 2-tailed U-test: U=0, p=0.0571. All data presented as mean ± SEM.

Figure 8.. PPG^NTS neuron activation augments semaglutide-induced eating suppression

(a) Experimental model and paradigm for semaglutide-induced eating suppression in PPG^NTS-hM3Dq mice and administered semaglutide (60 μg/kg, s.c.) and CNO (2 mg/kg, i.p.). n=6 animals for analyses presented in b-f. (b-f) Cumulative food intake at 1, 2, 4, 6 and 24 hours, 1-way within-subjects ANOVA: b) Drug F_(1.6,8.0)=11.94, p=0.0050; c) Drug F_(1.5,7.7)=22.12, p=0.0009; d) Drug F_(1.3,6.3)=35.35, p=0.0006; e) Drug F_(1.2,6.0)=40.72, p=0.0005; f) Drug F_(1.6,8.2)=125.8, p<0.0001. (g) Graphical representation of the core findings of this study and proposed model of central and peripheral GLP-1 system gut-brain satiation circuit architecture in the brainstem. All data presented as mean ± SEM.