Optogenetic stimulation of vagal nerves for enhanced glucose-stimulated insulin secretion and β cell proliferation

Abstract

The enhancement of insulin secretion and of the proliferation of pancreatic β cells are promising therapeutic options for diabetes. Signals from the vagal nerve regulate both processes, yet the effectiveness of stimulating the nerve is unclear, owing to a lack of techniques for doing it so selectively and prolongedly. Here we report two optogenetic methods for vagal-nerve stimulation that led to enhanced glucose-stimulated insulin secretion and to β cell proliferation in mice expressing choline acetyltransferase-channelrhodopsin 2. One method involves subdiaphragmatic implantation of an optical fibre for the photostimulation of cholinergic neurons expressing a blue-light-sensitive opsin. The other method, which suppressed streptozotocin-induced hyperglycaemia in the mice, involves the selective activation of vagal fibres by placing blue-light-emitting lanthanide microparticles in the pancreatic ducts of opsin-expressing mice, followed by near-infrared illumination. The two methods show that signals from the vagal nerve, especially from nerve fibres innervating the pancreas, are sufficient to regulate insulin secretion and β cell proliferation.

Fig. 1. ChR2–YFP was expressed in vagal nerves of ChAT–ChR2 mice.

a, Graphical abstract of two oVNS methods. Left: ChAT–ChR2 mice implanted with a small optical fibre in the subdiaphragmatic oesophagus This fibre delivers light stimulation to the vagal nerves that are connected to the pancreas. Right: blue light-emitting LMPs in the pancreatic ducts of the ChAT–ChR2 mice. Mice are exposed to NIR light, which stimulates the vagal nerves connected to the pancreas without the need for fibre implantation. b, YFP-positive vagal trunk running along the oesophageal wall. Each white arrowhead denotes vagal trunk. c, YFP-positive parasympathetic ganglia adjacent to a pancreatic islet, counter-stained with anti-Tuj1 antibody (general neuronal marker). More than two ChAT–ChR2 mice were tested. Each white arrowhead denotes parasympathetic ganglia. DAPI, 4′,6-diamidino-2-phenylindole. d, YFP-positive parasympathetic fibres in contact with each of the islet cell types. More than two ChAT–ChR2 mice were tested. Each white arrowhead indicates YFP-positive nerves that make contact with CD31-positive vascular endothelial cells in islets. Each 3D islet image is reconstructed as Z-stacks of 70 images with a pitch of 0.3 μm. Ins, insulin-positive β cells; Gcg, glucagon-positive α cells; Sst, somatostatin-positive δ cells. e, Contact density between cholinergic fibres and islet cells, expressed as the proportion of the total volume of the shared areas between YFP-positive nerve fibres and islet cells/total volume of islet cells (n = 3 mice, 5 islets per mouse). Data are presented as mean ± s.e.m. Scale bars denote 100 μm (b), 50 μm (c) and 30 μm (d).

Fig. 2. Optical stimulation activated vagal nerves of ChAT–ChR2 mice.

a, The setting for electrophysiological measurement of ChAT–ChR2 mice subdiaphragmatic vagal nerve response to blue light stimuli (470 nm) under anaesthesia. b, Vagal nerve responses evoked by various pulse-width patterns. c,d, Persistent vagal nerve responses evoked by a train of pulses (25 ms, 20 Hz, 100 pulses, total 5 s) (c) and its repetition of 30 trains (10 s interval, total 5 min) (d). e, Only an initial short compound response, which did not persist, was produced by illumination for 5 s.

Fig. 3. LED-oVNS elicited acute and chronic vagal nerve activation in vivo and induced pancreatic β cell proliferation.

a, Illustration of optical fibre implantation surgery. b, An optical fibre-implanted mouse. c, A silicon cuff was wrapped around the subdiaphragmatic oesophagus. The other end of the fibre was passed through the back using a needle. d, The schedule of LED-oVNS experiments. e, Intrapancreatic vagal nerve responses evoked by blue light stimulation. ISI, inter-stimulus intervals. f,g, Plasma insulin (f) and glucose (g) levels of ChAT–ChR2 mice, which had undergone acute LED-oVNS, during GTTs (f: two-way repeated measures ANOVA followed by Bonferroni post hoc test; CC-ctrl versus CC-LED at 15 min, *P = 0.0190; CC-ctrl versus CC-LED at 30 min, ***P = 0.0019; CC-ctrl versus CC-LED at 60 min, *P = 0.0280, n = 5). h, β cell masses of ChAT–ChR2 mice after 2 weeks of chronic LED-oVNS (one-way ANOVA followed by Ryan’s method as a post hoc test; CC-ctrl versus CC-LED, *P = 0.0239; WT-LED versus CC-LED, *P = 0.0217, n = 5); representative images are shown in the right three panels. Scale bars denote 100 µm. Ins, insulin-positive areas. i, The ratio of BrdU-positive β cells to all β cells in the islets of ChAT–ChR2 mice after 2 weeks of chronic LED-oVNS (one-way ANOVA followed by Ryan’s method as a post hoc test; CC-ctrl versus CC-LED, *****P = 2.121 × 10⁻⁴; WT-LED versus CC-LED, *****P = 7.08 × 10⁻⁵, n = 5); representative images are shown in the right three panels. Each arrowhead denotes a BrdU-positive β cell. Scale bars denote 50 µm. Data are presented as mean ± s.e.m. DAPI, 4′,6-diamidino-2-phenylindole.

Fig. 4. Development of the NIR light-mediated oVNS method.

a, Illustration of the NIR-oVNS method, wherein blue light-emitting LMPs are injected into the pancreatic ducts of ChAT–ChR2 mice (referred to as CC-blue-NIR-mice). Following the injection, NIR light is externally applied from 20 cm above the mouse body to illuminate the bodies of freely moving mice. b, Excised pancreas collected 4 weeks after injection, containing red or blue LMPs, illuminated with NIR light. c, Intrapancreatic vagal nerve responses evoked by NIR illumination. d, The ratio of c-Fos-positive ganglion cells to all YFP positive ganglion cells adjacent to the islets of ChAT–ChR2 mice after 1 week of chronic NIR-oVNS (two-tailed unpaired t-test (CC-red-NIR versus CC-blue-NIR), ***P = 0.0016, n = 3); representative images are shown in the right two panels. An arrowhead denotes a c-Fos-positive ganglion cell. Scale bars denote 50 µm. DAPI, 4′,6-diamidino-2-phenylindole. e, NIR illumination using a rotator. The rotational speed was 3 s per cage every 36 s.

Fig. 5. NIR-oVNS elicited acute vagal nerve activation in vivo.

a,b, Plasma insulin (a) and glucose (b) levels of ChAT–ChR2 mice, which had undergone acute NIR-oVNS, during GTTs (a: two-way repeated measures ANOVA (CC-red-NIR versus CC-blue-NIR), *P = 0.0262 at 5 min, *P = 0.0152 at 15 min and **P = 0.0092 at 30 min; b: two-way repeated measures ANOVA (CC-red-NIR versus CC-blue-NIR), ***P = 0.0030 at 5 min, ***P = 0.0029 at 15 min, ****P = 0.000649 at 30 min and *P = 0.0475 at 60 min, n = 4 of CC-red-NIR, 5 of CC-blue-NIR). c, Plasma insulin (left) and glucose (right) levels of ChAT–ChR2 mice, which had undergone acute NIR-oVNS with or without subdiaphragmatic vagotomy (Vx), during GTTs (c: two-way repeated measures ANOVA followed by Bonferroni post hoc test; CC-red-NIR versus CC-blue-NIR plasma insulin at 15 min, ****P = 4.434 × 10⁻¹⁷; CC-blue-NIR versus CC-blue-NIR + Vx plasma insulin at 15 min, ^####P = 3.329 × 10⁻¹⁵; CC-red-NIR versus CC-blue-NIR plasma glucose at 30 min, ****P = 4.994 × 10⁻¹⁰; CC-red-NIR versus CC-blue-NIR + Vx plasma glucose at 30 min, ^††P = 0.00495; CC-blue-NIR versus CC-blue-NIR + Vx plasma glucose at 30 min, ^####P = 0.00000147, n = 5 of CC-red-NIR, 7 of CC-blue-NIR, 8 of CC-blue-NIR + Vx). d, Plasma insulin (left) and glucose (right) levels of ChAT–ChR2 mice, which had undergone acute NIR-oVNS with or without administration of the muscarinic acetylcholine receptor M3 antagonist 4-DAMP, during GTTs (d: two-way repeated measures ANOVA followed by Bonferroni post hoc test; CC-red-NIR + Veh versus CC-blue-NIR + Veh plasma insulin at 15 min, ****P = 6.461 × 10⁻¹⁰; CC-blue-NIR + Veh versus CC-blue-NIR + 4-DAMP plasma insulin at 15 min, ^####P = 1.658 × 10⁻⁷; CC-red-NIR + Veh versus CC-blue-NIR + Veh plasma glucose at 30 min, ****P = 0.0000147; CC-blue-NIR + Veh versus CC-blue-NIR + 4-DAMP plasma glucose at 30 min, ^####P = 0.00000425; CC-red-NIR + Veh versus CC-blue-NIR + Veh plasma glucose at 60 min, ****P = 0.00000833; CC-blue-NIR + Veh versus CC-blue-NIR + 4-DAMP plasma glucose at 60 min, ^####P = 0.00000219, n = 4). Data are presented as mean ± s.e.m.

Fig. 6. Chronic NIR-oVNS induced pancreatic β cell proliferation.

a, Insulin immunostaining (brown) of pancreatic sections from ChAT–ChR2 mice after 2 weeks of chronic NIR-oVNS. Bottom two panels are magnified images of the inset. More than 12 ChAT–ChR2 mice were tested. Scale bars denote 2 mm (top two panels) and 200 µm (bottom two panels). Ins, insulin-positive areas. b, β cell masses of ChAT–ChR2 mice after 2 weeks of chronic NIR-oVNS (b: two-tailed unpaired t-test (CC-red-NIR versus CC-blue-NIR), *****P = 0.0005371, n = 5 of CC-red-NIR, 7 of CC-blue-NIR). c, The ratio of BrdU-positive β cells to all β cells in the islets of ChAT–ChR2 mice after 2 weeks of chronic NIR-oVNS (c: two-tailed unpaired t-test (CC-red-NIR versus CC-blue-NIR), *****P = 0.0001129, n = 4); representative images are shown in the right two panels. Each arrowhead denotes a BrdU-positive β cell. Scale bars denote 50 µm. Ins, insulin-positive β cells; DAPI, 4′,6-diamidino-2-phenylindole. d, β cell masses of ChAT–ChR2 mice after 2 weeks of chronic NIR-oVNS with or without subdiaphragmatic vagotomy (d: one-way ANOVA followed by Ryan’s method as a post hoc test; CC-red-NIR versus CC-blue-NIR, *P = 0.0104; CC-blue-NIR versus CC-blue-NIR + Vx, P = 0.0572, n = 4). e, Changes in cell cycle-related gene expressions in the islets from ChAT–ChR2 mice after 2 weeks of chronic NIR-oVNS (e: two-tailed unpaired t-test (CC-red-NIR versus CC-blue-NIR), for Foxm1, *P = 0.0313; for Cdk1, *P = 0.0498, n = 9 of CC-red-NIR, 11 of CC-blue-NIR). f, β cell masses of ChAT–ChR2 mice after 2 weeks of chronic NIR-oVNS with or without the administration of muscarinic acetylcholine receptor M3 antagonist 4-DAMP (f: one-way ANOVA followed by Ryan’s method as a post hoc test, CC-red-NIR versus CC-blue-NIR, *P = 0.0152; CC-blue-NIR versus CC-blue-NIR + 4-DAMP, *P = 0.0242, n = 4). Data are presented as mean ± s.e.m.

Fig. 7. Chronic NIR-oVNS suppressed STZ-induced hyperglycaemia.

a, Time course of experiment. i.p., intraperitoneal injection; p.o., oral administration. b, Plasma insulin levels of ChAT–ChR2 mice after starting STZ administration (50 mg kg⁻¹ i.p., 5 consecutive days) accompanied by 3 weeks of chronic NIR-oVNS (two-way repeated measures ANOVA, CC-red-NIR + STZ versus CC-blue-NIR + STZ at week 2, ****P = 0.00006483; at week 3, ****P = 0.0002235, n = 5). c, Plasma glucose levels of ChAT–ChR2 mice after starting STZ administration (50 mg kg⁻¹ i.p., 5 consecutive days) accompanied by 2 months of chronic NIR-oVNS (two-way repeated measures ANOVA, CC-red-NIR + STZ versus CC-blue-NIR + STZ from day 6 to 56, ****P = 5.039 × 10⁻⁶, 2.345 × 10⁻⁶, 1.933 × 10⁻⁸, 3.061 × 10⁻⁸, 4.940 × 10⁻¹¹, 1.893 × 10⁻¹⁸, 1.335 × 10⁻¹¹, 5.361 × 10⁻¹⁰, 8.275 × 10⁻¹⁴, 1.886 × 10⁻¹⁰, 8.108 × 10⁻¹¹, 5.111 × 10⁻¹², 3.974 × 10⁻⁸, 5.265 × 10⁻⁸, 2.367 × 10⁻¹¹, 2.371 × 10⁻⁹, 1.458 × 10⁻¹⁰, 1.272 × 10⁻⁸, 1.026 × 10⁻¹¹, 2.273 × 10⁻¹⁰, 2.467 × 10⁻⁴ at days 6, 8, 9, 10, 13, 15, 17, 20, 22, 23, 24, 26, 28, 30, 34, 36, 38, 41, 44, 50 and 56, respectively, n = 5). d, The ratio of BrdU-positive β cells to all islet cells of ChAT–ChR2 mice after starting STZ administration (as stated above) accompanied by 3 weeks or 2 months of chronic NIR-oVNS (two-tailed unpaired t-test, CC-red-NIR + STZ versus CC-blue-NIR + STZ at 3 weeks, *****P = 0.0001868; CC-red-NIR + STZ versus CC-blue-NIR + STZ at 2 months, *****P = 7.858 × 10⁻⁶, n = 5, 6, 5 and 5, respectively). Representative images are shown in the right two panels. Each arrowhead denotes a BrdU-positive β cell. Scale bars denote 50 µm. DAPI, 4′,6-diamidino-2-phenylindole. e, β cell masses of ChAT–ChR2 mice after starting STZ administration (as above stated) accompanied by 3 weeks or 2 months of chronic NIR-oVNS (two-tailed unpaired t-test, CC-red-NIR + STZ versus CC-blue-NIR + STZ at 3 weeks, *P = 0.0444; CC-red-NIR + STZ versus CC-blue-NIR + STZ at 2 months, *P =0.0273, n = 8, 8, 5 and 5, respectively). Representative images are shown in the right two panels. Scale bars denote 100 µm. Data are presented as mean ± s.e.m. NS, not significant.

Extended Data Fig. 1. Development of the LED-oVNS system.

a, YFP-positive parasympathetic ganglia and fibres were detected in the stomach, duodenum, jejunum and ileum, but not in the liver or the portal vein. More than two ChAT-ChR2 mice were tested. Scale bars denote 100 µm. b, A plastic optical fibre (500 μm diameter) was attached to a silicon cuff. The scale bar indicates 1 cm. c, A needle was passed through the back wall of the abdominal cavity. Its hub was wrapped with a silicon tube. The scale bar indicates 1 cm. d, Body weights of ChAT-ChR2 mice during 4 weeks of chronic LED-oVNS (n = 5). e, Pancreatic blood flow change after LED-oVNS. (Left) representative organ blood flow images from a non-ChAT-ChR2 (ChAT-Cre(+); LSL-ChR2(−)) or a ChAT-ChR2 mouse before (‘Pre’) and during (‘Stim’) LED-oVNS, obtained by laser speckle flowmetry. Each area enclosed by the dotted line denotes the pancreas. We captured one image per second during each period (10 sec respectively). (Right) quantification of abdominal organ blood flow. Averaged blood flow levels were calculated from 10 images per period and normalised to the Pre period (two-tailed unpaired t-test, ChR2(−)-pancreas vs ChR2(+)-pancreas during the Stim period, *P = 0.0219; ChR2(−)-duodenum vs ChR2(+)-duodenum during the Stim period, *P = 0.0322, n = 5). a.u.: arbitrary units. Data are presented as means ± s.e.m. ns: not significant.

Extended Data Fig. 2. Other endocrine and exocrine hormone levels in LED-oVNS mice.

a, Plasma glucagon levels of ChAT-ChR2 mice, which had undergone acute LED-oVNS, during glucose tolerance tests (n = 5 of CC-ctrl, 6 of CC-LED). b, Plasma somatostatin and amylase levels of ChAT-ChR2 mice, which had undergone acute LED-oVNS, during glucose tolerance tests. Blood samples were collected at 0 and 15 min after starting LED-oVNS (n = 5 of somatostatin, 4 of amylase). c, Amylase concentrations in pancreatic-bile juice of ChAT-ChR2 mice before (‘Pre’ period) and after (‘Post’ period) starting LED-oVNS (n = 5). Pancreatic-bile juice was collected for 30 min during each period. d, Plasma glucose levels of ChAT-ChR2 mice, which had undergone acute LED-oVNS, during insulin tolerance tests (n = 5 of CC-ctrl, 6 of CC-LED). Data are presented as means ± s.e.m.

Extended Data Fig. 3. Chronic LED-oVNS did not affect oesophageal vagal nerve responses, insulin sensitivity, apoptosis of beta-cells, non-beta-cell proliferation or non-beta-cell size.

a, Vagal nerve responses to blue LED pulses were still stably detected after 2 weeks of LED-oVNS. b, Plasma glucose levels of ChAT-ChR2 mice, which had undergone 2 weeks of chronic LED-oVNS, during insulin tolerance tests concomitantly with or without LED-oVNS (n = 6 of CC-ctrl without LED illumination, 6 of CC-LED without LED illumination, 5 of CC-ctrl with LED illumination, and 6 of CC-LED with LED illumination). c, The ratio of TUNEL positive beta-cells to all cells in the islets of ChAT-ChR2 mice after 2 weeks of chronic LED-oVNS (n = 5, two-tailed unpaired t-test, CC-ctrl vs CC-LED); representative images are shown in the right two panels. Each arrowhead denotes a TUNEL-positive beta-cell. Scale bars denote 50 µm. d, The ratio of BrdU positive non-beta-cells (α, δ and exocrine acinar cells) to the total number of each cell type in the islets or exocrine tissues of ChAT-ChR2 mice after 2 weeks of chronic LED-oVNS (n = 5, two-tailed unpaired t-test, CC-ctrl vs CC-LED); representative images are shown in the bottom two panels. Each arrowhead denotes a BrdU-positive non-beta-cell. Scale bars denote 50 µm. e, The average sizes of non-beta-cells (α and δ cells) of ChAT-ChR2 mice after 2 weeks of chronic LED-oVNS (n = 5); respective representative images are shown in the right two panels (white boxes indicate α and δ cells). Cell size is expressed as the ratio of the hormone-positive area to cell number. Scale bars denote 50 µm. Data are presented as means ± s.e.m. ns: not significant. Ins: insulin-positive beta cells, Gcg: glucagon-positive α cells, and Sst: somatostatin-positive δ cells.

Extended Data Fig. 4. LED-oVNS promoted food intake and intestinal motility.

a, Food intake amounts of ChAT-ChR2 mice during 4 weeks of chronic LED-oVNS (two-way repeated measures ANOVA, CC-ctrl vs CC-LED, *P = 0.04781, 0.0009853, 0.01422, 0.007765, 0.007832, 0.01724, 0.001291, 0.007501, 0.00257, 0.000002757, 0.03505, 0.003765, at day 1~3, 3~5, 5~7, 7~9, 9~11, 11~13, 13~15, 17~19, 19~21, 21~23, 23~25, 25~27, respectively, n = 5). b, Intestinal motility change after LED-oVNS. As shown in Supplementary Video 2a, b, by tracing the trajectory of the mark on the intestine, total distance moved and maximum distance from the origin were measured for 10 sec before (‘Pre’) and after (‘Stim’) starting LED-oVNS (b: two-tailed unpaired t-test (Pre vs Stim), for the total moving distance, *P = 0.0313; for the maximum distance from the origin, *P = 0.0345, n = 5). Solid lines and dotted lines indicate average and individual values, respectively. Data are presented as means ± s.e.m.

Extended Data Fig. 5. Development of the NIR-oVNS system.

a, Distribution of the LMPs injected into the pancreatic duct. LMPs were found in exocrine tissues (left 2 panels), lymphatic tissues (top right), and the pancreatic ducts (bottom right). More than two ChAT-ChR2 mice were tested. Each of the red arrows indicates LMPs. Scale bars denote 50 µm. b, NIR light intensities in the pancreas, from head to tail (b: two-tailed unpaired t-test (NIR(−) vs NIR(+)), ****P = 0.0003431, n = 4). c, Intrapancreatic vagal nerve responses of ChAT-ChR2 and WT mice, in which blue light-emitting LMPs had been implanted, to various wavelengths of light. d, Left: lack of oesophageal vagal nerve responses of ChAT-ChR2 mouse in which intrapancreatic vagal nerves had been activated by NIR-oVNS. Right: evident oesophageal vagal nerve responses of the same mouse in which oesophageal vagal nerves had been activated by LED-oVNS. Rec: recording site. Data are presented as means ± s.e.m.

Extended Data Fig. 6. Characterization of NIR-oVNS mice.

a, Plasma amylase levels of ChAT-ChR2 mice, which had undergone acute NIR-oVNS, during glucose tolerance tests. Blood samples were collected at 0 and 15 min after starting NIR-oVNS (n = 5 of CC-red-NIR, 4 of CC-blue-NIR). b–e, Body weights (b,d) and food intake amounts (c,e) of ChAT-ChR2 mice during 2 weeks of NIR-oVNS or the last 2 weeks of 2-month-NIR-oVNS (n = 5, two-way repeated measures ANOVA followed by Bonferroni post hoc test). f, Abdominal organ blood flow change after NIR-oVNS. (Left) Representative organ blood flow images of red or blue LMP-loaded ChAT-ChR2 mice before (‘Pre’) and after 3 seconds of NIR-oVNS (‘Stim’), values were obtained by laser speckle flowmetry. Area enclosed by the dotted line denotes the pancreas. We captured one image per second during each period (10 sec, respectively). (Right) Quantification of blood flow change in each organ. Averaged blood flow levels were calculated from 10 images per period and normalised to that of the Pre period (f: two-tailed unpaired t-test (CC-red-NIR pancreas vs CC-blue-NIR pancreas during Stim period), ****P = 0.000934, n = 4 of CC-red-NIR, 5 of CC-blue-NIR). a.u.: arbitrary units. g, Intestinal motility change after NIR-oVNS. As shown in Supplementary Video 3a, b, by tracing the trajectory of the mark on the intestine, total distance moved and maximum distance from the origin were measured for 10 sec before (‘Pre’) and after (‘Stim’) starting NIR-oVNS. Solid lines and dotted lines indicate average and individual values, respectively (n = 6, two-tailed unpaired t-test, Pre vs Stim). h, Body temperature of mouse after 2-day-NIR illumination (n = 4). Data are presented as means ± s.e.m. ns: not significant.

Extended Data Fig. 7. Other endocrine and exocrine hormone levels of NIR-oVNS mice.

a, Plasma glucagon and somatostatin levels of ChAT-ChR2 mice, which had undergone acute NIR-oVNS, during glucose tolerance tests. Blood samples were collected at 0 and 15 min after starting NIR-oVNS (n = 5). b, Amylase concentrations in pancreatic bile juice from ChAT-ChR2 mice before (‘Pre’ period) and after (‘Post’ period) starting NIR-oVNS (n = 5). Pancreatic-bile juice was collected for 30 min during each period. Data are presented as means ± s.e.m.

Extended Data Fig. 8. beta-cell proliferative effects of NIR-oVNS were maintained for 2 months.

a, The ratio of BrdU positive beta-cells to all beta-cells in the islets of ChAT-ChR2 mice after 2 weeks of chronic NIR-oVNS (two-tailed unpaired t-test (CC-red-NIR vs CC-blue-NIR), *****P = 0.00001875, n = 5 of CC-red-NIR, 6 of CC-blue-NIR); representative images are shown in the right two panels. Each arrowhead denotes a BrdU-positive beta-cell. Scale bars denote 50 µm. b, beta-cell masses of ChAT-ChR2 mice after 2 months of chronic NIR-oVNS (two-tailed unpaired t-test (CC-red-NIR vs CC-blue-NIR), ***P = 0.0029, n = 5 of CC-red-NIR, 6 of CC-blue-NIR); representative images are shown in the right two panels. Scale bars denote 100 µm.

Extended Data Fig. 9. Effects of chronic NIR-oVNS on non-beta-cell proliferation, non-beta-cell size, insulin sensitivity and beta-cell apoptosis.

a, The ratio of BrdU-positive non-beta-cells to the total number of each cell type in the islets or exocrine tissues of ChAT-ChR2 mice after 2 weeks of chronic NIR-oVNS (a: two-tailed unpaired t-test (CC-red-NIR vs CC-blue-NIR for Amy+ BrdU+ cells), ***P = 0.0018, n = 4) ; representative images of Gcg stained cells (left, 2 panels), Sst stained cells (middle, 2 panels) and Amy stained cells (right, 2 panels). Each arrowhead denotes a BrdU-positive non-beta-cell. Scale bars denote 50 µm. b, The average size of non-beta-cells (α and δ cells) in ChAT-ChR2 mice after 2 weeks of chronic NIR-oVNS (n = 4); respective representative images of Gcg stained cells (top panels), Sst stained cells (lower panels). White boxes indicate α and δ cells. Cell size was expressed as the ratio of the hormone-positive area to cell number. Scale bars denote 50 µm. c, Plasma glucose levels of ChAT-ChR2 mice, which had undergone 2 weeks of chronic NIR-oVNS, during insulin tolerance tests conducted concomitantly with or without NIR-oVNS (n = 5 of CC-red-NIR without NIR illumination, 5 of CC-blue-NIR without NIR illumination, 5 of CC-red-NIR with NIR illumination, and 4 of CC-blue-NIR with NIR illumination). d, The ratio of TUNEL positive beta-cells to all cells in the islets of ChAT-ChR2 mice after 2 weeks of chronic NIR-oVNS (n = 4 of CC-red-NIR, 8 of CC-blue-NIR, two-tailed unpaired t-test, CC-red-NIR vs CC-blue-NIR); representative images are shown in the right two panels. Each arrowhead denotes a TUNEL-positive beta-cell. Scale bars denote 50 µm. e, beta-cell masses of non-ChAT-ChR2 (ChAT-Cre(+); LSL-ChR2(-)) mice after 2 weeks of chronic NIR illumination (two-tailed unpaired t-test, CC-red-NIR vs CC-blue-NIR, n = 6 of ChR2(−)-red-NIR, 5 of ChR2(−)-blue-NIR). Data are presented as means ± s.e.m. ns: not significant.

Extended Data Fig. 10. Effects of NIR-oVNS on STZ-treated ChAT-ChR2 mice.

a, beta-cell masses of ChAT-ChR2 mice before (‘Pre (day 0)’) and short after (‘+STZ (day 7)’) starting STZ administration (50 mg/kg i.p., 5 consecutive days) accompanied by chronic NIR-oVNS (n = 4). b, c, The ratios of TUNEL (b) and Iba-1 (c) positive cells to all cells in the islets of ChAT-ChR2 mice after starting STZ administration (as above stated) accompanied by 3 weeks of chronic NIR-oVNS (n = 8). Respective representative images are shown in the right two panels. Each arrowhead denotes a TUNEL-positive (b) or an Iba-1-positive (c) beta-cell. d, Body weights of ChAT-ChR2 mice after starting STZ administration (as stated above) accompanied by 2 months of chronic NIR-oVNS (two-way repeated measures ANOVA, CC-red-NIR+STZ vs CC-blue-NIR+STZ, *P = 0.02219, 0.01415, 0.008994, 0.003630, 0.002314, 0.01415, 0.03460, 0.005711, 0.003630, 0.001480, 0.001480, 0.0004012 at day 15, 20, 22, 23, 24, 34, 36, 38, 41, 44, 50, 56, respectively, n = 5). Scale bars denote 50 µm. Data are presented as means ± s.e.m.