Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Apr 21.
Published in final edited form as: Cell Metab. 2025 Jul 14;37(8):1698–1714.e5. doi: 10.1016/j.cmet.2025.06.009

Localized GLP-1 receptor pre-internalization directs pancreatic alpha cell to beta cell communication

Jason CL Tong 1, Charlotte Frazer-Morris 1, Ali H Shilleh 1, Katrina Viloria 1, Anne de Bray 1, Adithya Muraleedaran Nair 1, Paul RV Johnson 1,2, Rebecca Spiers 1,2, Ahmad Kobiita 2, Oladapo E Olaniru 3, Shanta J Persaud 3, Robert Hauffe 4, André Kleinridders 4, Carsten Schultz 5, C Bruce Verchere 6, Canqi Cui 7, Jonathan E Campbell 7,8,9, Malgorzata Cyranka 10, Alexey Epanchintsev 10, Carina Ämmälä 10, Johannes Broichhagen 11, David J Hodson 1,*
PMCID: PMC12276844  NIHMSID: NIHMS2095799  PMID: 40664215

SUMMARY

Pancreatic alpha cells modulate beta cell function in a paracrine manner through the release of glucagon. However, the detailed molecular architecture underlying alpha-to beta-cell regulation remains poorly characterized. Here, we show that glucagon-like peptide-1 receptor (GLP1R) is enriched as nanodomains on beta cell membranes that contact alpha cells, in keeping with increased single-molecule transcript expression. At low glucose, beta cells next to alpha cells directly sense micromolar glucagon release by pre-internalizing GLP1R. Pre-internalized GLP1R is associated with earlier beta cell Ca2+ responses to high glucose, which are then propagated across the islet. Beta cells adjacent to alpha cells are more secretory than beta cells next to other beta cells. Localized GLP1R signalling occurs in vitro and in vivo, is operative in the postprandial state, and GLP1R contacts decrease between beta cells and alpha cells during metabolic stress. Thus, we detail a regulated pathway through which glucagon modulates insulin release.

Keywords: GLP1R, alpha cell, beta cell, insulin, glucagon, islet, signalling, cAMP, Ca2+

Graphical Abstract

graphic file with name nihms-2095799-f0001.jpg

eTOC blurb

Tong et al. utilize advanced imaging approaches to reveal that beta cells located next to alpha cells are specially adapted as glucagon sensors by virtue of their high expression levels of GLP1R. These beta cells respond first to rising glucose, triggering robust responses across the pancreatic islet.

INTRODUCTION

Pancreatic islets are complex micro-organs comprised of endocrine, neuronal, vascular and supporting cells1. Interactions between the various cell types are necessary for cell differentiation/specification, cell survival and hormone release2–5. Amongst endocrine cells, paracrine communication is a well-established mechanism for the proper regulation of beta cell activity and insulin release (reviewed in6,7). In response to high glucose concentration, beta cells increase cytosolic ATP/ADP ratios via oxidative phosphorylation and the mitochondrial phosphoenolpyruvate cycle, leading to the closure of ATP-sensitive K+ channels, Ca2+ influx and insulin release8–11. Cell-intrinsic beta cell responses are however shaped by external cues such as those from neighboring alpha cells and delta cells12–15.

At low glucose concentration, alpha cells release glucagon and acetylcholine, which have a strong priming or amplifying effect on beta cell function16–19. Whilst alpha cell regulation of beta cell function has been long established16,17, the pathways involved are still being pieced together. Multiple studies have shown that glucagon promiscuously signals at the glucagon-like peptide-1 receptor (GLP1R)17,20–24, which is expressed on the beta cell but not alpha cell surface25,26. GLP1R is a class B G protein-coupled receptor (GPCR) and member of the secretin family along with glucagon receptor (GCGR). GLP1R antagonism or conditional GLP1R deletion from the beta cell compartment leads to a large (~80%) reduction in the ability of glucagon to amplify insulin release20,21,23,24, with a contribution from GCGR27–29. Implicating a role for glucagon in this effect, deletion of proglucagon-derived peptides from alpha cells impairs insulin secretion, which can be rescued by application of GLP1 or glucagon21. Whilst glucagon has ~10–1000-fold less potency for GLP1R versus GCGR21,22,24, concentrations achieved at the beta cell surface are likely to be higher than in circulation, and circulating islet concentrations of glucagon are able to generate cAMP and stimulate insulin secretion at the endogenous GLP1R24,30.

Following ligand stimulation, GLP1R rapidly (milliseconds) engages Gαs subunit, which activates adenylate cyclase and generates cAMP, a potent amplifier of Ca2+-dependent insulin exocytosis31,32. G protein-coupled receptor (GPCR) kinase phosphorylates activated GLP1R at C-terminal serine and threonine residues, leading to beta-arrestin recruitment and desensitization, followed by clathrin-dependent endocytosis33,34. Internalized GLP1R either signals from the endosomal compartment, is recycled back to the cell surface, or is sent to the lysosome for degradation35,36. The exact molecular and cellular events through which glucagon influences beta cell activity and insulin release in primary islets, however, remain largely unknown.

We and others have recently shown that GLP1R is enriched at the beta cell surface, but undetectable in alpha cells25,26. We also showed that GLP1R form nanodomains (i.e. clustered aggregations) at the single molecule level, which are important for signalling25,37. Lastly, we showed that GLP1R signalling within the islet is heterogeneous, with beta cell subpopulations recruited into GLP1R trafficking37. Together with imaging studies showing close associations between beta cells and alpha cells38,39, we hypothesize that enrichment of GLP1R at the beta cell-alpha cell interface might underlie glucagon-priming of insulin release. In the present study, we set out to provide further molecular details for alpha cell to beta cell communication.

RESULTS

GLP1R is enriched at the beta cell-alpha cell interface

GLP1R is a low abundance surface protein, which is notoriously difficult to (specifically) detect (reviewed in40). We thus optimized immunohistochemistry in mouse pancreas slices, using a knockout-validated antibody with known specificity for mouse GLP1R41. By Nyquist-sampling and generating large 3D volumes, followed by GPU-accelerated deconvolution, we were able to obtain highly detailed images with excellent signal-noise-ratio for quantification (Figure 1A). In this manner, GLP1R expression could be carefully examined at the beta cell-beta cell, beta cell-alpha cell and beta cell:delta cell interface. Using our high-resolution imaging approach, we first confirmed expression of GLP1R at the beta cell but not alpha cell membrane (Figure S1A and S1B). We were also unable to detect robust GLP1R expression in delta cells (Figure S1C and S1D), confirming previous FACS studies showing the presence of Glp1r transcript but not protein in somatostatin-positive fractions26.

Figure 1: GLP1R is enriched at the beta cell-alpha cell interface.

Figure 1:

A) Tissue processing, staining and 3D imaging/deconvolution workflow.

B) Specific immunostaining reveals GLP1R expression in the beta cell compartment (left panels), with GLP1R enriched on beta cell membranes close to alpha cells (zoom-in, right panels) (inset shows cell outlines and GLP1R enrichment zone in orange) (representative images from n = 32 islets, 5 mice).

C) As for B), but showing no obvious GLP1R enrichment on beta cell membranes close to delta cells (representative images from n = 20 islets, 5 mice). Pancreas sections in Figure 1A and B were taken from 8–12 wk old mice, and the GCG data also serve as control for aged (81 wk old) mice shown in Figure 2K and L.

D) GLP1R antibody ab218532 is non-specific in human islets (n = 2 donors).

E, F) GLP1R is enriched on beta cell membranes close to alpha cells in human islets (E) (n = 5 islets, 2 donors) but not stem cell-derived islets (SC-islets) (F) (n = 6 SC-islets, 2 differentiations).

G, H) Intersection distance (G) and distance to peak (H) are longer for GLP1R versus SST than GLP1R versus GCG (Mann-Whitney test) (n = 15 islets from 5 mice).

I, J) Intersection distance (I) and distance to peak (J) for GLP1R versus GCG are longer in SC-islets compared to human islets (Mann-Whitney test) (n = 2 differentiations or 2 donors).

Scale bar = 20 μm. Bar and line graphs show mean ± SEM. ns, non-significant; *P<0.05; ***P<0.001. GLP1R, glucagon-like peptide-1 receptor; INS, insulin; GCG, glucagon; SST, somatostatin. Where shown, nuclei are labelled with Hoechst 33342 (blue).

See also Figure S1 and Table S1.

GLP1R enrichment/contacts were quantified by measuring the intensity peaks of the various cell type markers (SST, GCG, INS) and their distance to the GLP1R intensity peak (Figure S2A). At the high axial resolution achieved here, enrichment of GLP1R could clearly be seen at the beta cell-alpha cell interface (Figure 1B and C), confirmed by the short intersection distance between peak GLP1R and GCG signals (Figure 1G and H) (Figure S2A). Discrete contacts were present between GLP1R-expressing beta cells and adjacent alpha cells (Figure 1B, G and H) (Figure S2B). GLP1R contacts were also detected between beta cells and delta cells, but were ~ 3-fold fewer in number than those between beta cells and alpha cells (Figure 1C, G and H). Almost 60% of GLP1R contacts were found to overlap with glucagon, 25% with insulin and 15% with somatostatin (Figure 1B, C, G and H). The image analysis workflow is shown in Figure S2A, and these and all other analyzed images are provided via the Mendeley Data repository (http://dx.doi.org/10.17632/bm9jsyg2f2.1).

GLP1R enrichment is conserved between rodents and humans

Studies were repeated in human islet sections to understand whether any changes in islet arrangement might influence GLP1R enrichment at the beta cell-alpha cell interface. For immunofluorescent staining, we tested two different antibodies with cross-reactivity against human GLP1R, Abcam ab218532 and Iowa DSHB 3F52. While ab218532 showed excellent specificity in mouse tissue (Figure 1B and C, also41), significant overlap was detected with insulin in human islets suggesting cross-reactivity with non-GLP1R epitopes (Figure 1D). Iowa DSHB 3F52 showed staining of the beta cell membrane in human islets, in line with known GLP1R localization (Figure 1E) as well as published data42,43. Surprisingly, however, GLP1R staining was much more variable in human versus mouse islets, with weaker GLP1R expression, and some islet preparations or sections failing to show any staining at all. We put this down to a combination of GLP1R expression levels, donor heterogeneity, as well as processing/isolation time/age where GLP1R is expected to decline, as also noted by others43. Differences in surface:internalized GLP1R ratio are unlikely to feature here, since antibody should stain extracellular and intracellular compartments in permeabilized preparations, and neither were we able to easily detect GLP1R using LUXendins, our specific fluorescent antagonists25. Nonetheless, in a number of donors, we were able to detect GLP1R fluorescence at the beta cell membrane (Figure 1E), assisted by using highly-sensitive silicon-based SilVIR emission detectors. As for rodent islets (Figure 1A–C), GLP1R was enriched on beta cells adjacent to alpha cells (Figure 1E, I and J).

While stem cell (SC)-islets consistently showed robust surface GLP1R expression (2 separate differentiations shown here), enrichment was not readily observed at the beta cell-alpha cell interface (Figure 1F, I and J). Hence, proper GLP1R remodelling in SC-islets might require a tertiary step post-differentiation such as further cell maturation and/or in transplantation.

GLP1R enrichment is plastic with high fat diet and ageing

To understand whether enrichment of GLP1R on alpha cell-adjacent beta cells is a stable islet feature, or plastic with metabolic stress, we conducted studies in tissue from mice fed either standard chow (SC) or high fat diet (HFD). Beta cell GLP1R protein expression levels were found to be almost two-fold higher in islets from HFD-fed versus SC-fed mice (Figure 2A–D), suggesting either a compensatory response or a reduction in in vivo GLP1 tone at the time of tissue harvest. Despite higher GLP1R expression, organization of GLP1R was different in HFD-fed versus SC-fed mice, with loss of specific GLP1R contacts at the beta cell-alpha cell interface (Figure 2A, B, I and J). No differences were seen in GLP1R contacts with delta cells in HFD-fed versus SC-fed animals (Figure 2C) (Figure S2B and S2C).

Figure 2: GLP1R remodels during high fat diet and aging.

Figure 2:

A-D) GLP1R is more enriched at the beta cell-alpha cell interface in islets of mice fed standard chow (SC) (A) versus high fat diet (HFD) (B), but remains absent at the beta cell-delta cell interface (C), despite increased GLP1R expression during HFD (D) (unpaired t-test) (n = 10 islets, 5 mice).

E-H) GLP1R is more enriched at the beta cell-alpha cell interface in islets of young mice (E) versus aged mice (F), but remains absent at the beta cell-delta cell interface (G), despite increased GLP1R expression in aged mice (H) (unpaired t-test) (n = 41 islets, 5 mice).

I, J) Intersection distance (I) and distance to GLP1R peak (J) are longer for GLP1R versus GCG in SC-fed versus HFD-fed mice (Mann-Whitney test) (n = 10 islets, 5 mice).

K, L) Intersection distance (K) and distance to GLP1R peak (L) are longer for GLP1R versus GCG in young (8–12 wk old) versus aged mice (81 wk old) (Mann-Whitney test) (n = 20 islets, 10 mice). The same GCG-GLP1R control (8–12 wk old) is shown here and in Figure 1F and G to allow easy comparison.

Note that fluorescence intensity of the zoom panels has been equalized between images. Scale bar = 20 μm. CTCF, corrected total cell fluorescence; GLP1R, glucagon-like peptide-1 receptor; INS, insulin; GCG, glucagon; SST, somatostatin. Where shown, nuclei are labelled with Hoechst 33342 (blue).

See also Figure S2.

Experiments were repeated in another model of beta cell dysfunction, namely ageing. Again, GLP1R expression levels were ~ two-fold higher in islets of aged (81 weeks) versus young (8–12 weeks) mice (Figure 2E–H). Suggesting a specific effect of metabolic stress on GLP1R contact formation, GLP1R enrichment was also reduced at the beta cell-alpha cell interface in aged versus young mice (Figure 2E, F, K and L). As for HFD, no differences were seen in GLP1R contacts with delta cells in aged versus young animals (Figure 2G).

Thus, beta cells preferentially enrich GLP1R against the alpha cell membrane, with evidence of physical contacts that decrease in number/extent during high fat diet and aging. We have previously shown that GLP1R signalling is impaired in HFD rodent islets as well as human islets exposed to free fatty acid excess44, which might be in part due to GLP1R remodelling at the beta cell-alpha cell interface.

GLP1R forms nanodomains at the beta cell-alpha cell interface

Since light microscopy is diffraction-limited, we instead turned to dSTORM nanoscopy to allow single molecule detection and quantification of GLP1R at the beta cell-alpha cell interface (Figure 3A and B). For this purpose, islets were used from SC-fed GLP1RSNAP/SNAP mice, which have a SNAP-tag enzyme self-label knocked into the GLP1R N-terminus (so that endogenous GLP1R can be reported)37. Since GLP1R is a beta cell specific marker25,26, GLP1R positivity and negativity can be used to identify beta cells and alpha cells, respectively. We also cross-confirmed cell identity by overlaying STORM images of glucagon (Figure S3). Islets were labelled with BG-SulfoCy5, a sulfonated membrane-impermeable label with optimal performance for dSTORM nanoscopy45. Reconstruction of GLP1R localizations over ~40,000 frames revealed GLP1R at the beta cell-alpha cell interface in unprecedented detail (Figure 3C). Confirming results with light microscopy, but now extending findings from ~150 nm to ~30 nm, GLP1R was enriched at the membrane between beta cells and alpha cells (Figure 3C). DBSCAN analysis was used to identify clusters possessing ≥ 8 localizations, considered to be individual GLP1R, since localizations from the same point are filtered out. In this way, increased GLP1R nanodomain formation could be detected at the beta cell-alpha cell interface (Figure 3D–F), a feature of GPCR higher organization and signalling37. However, the number of localizations (i.e. GLP1R) per cluster was similar between the beta cell-alpha cell and beta cell-beta cell interface, most likely due to tissue/cell limitations on maximum cluster size (Figure 3G).

Figure 3: Single-molecule quantification of GLP1R nanodomains at the beta cell-alpha cell interface.

Figure 3:

A) Schematic showing SulfoCy5 and LUXendin645 binding to SNAP-tag or the GLP1R orthosteric site, respectively, as well as photoblinking requirements for dSTORM.

B) dSTORM nanoscopy resolves endogenous GLP1R at the single molecule level.

C) GLP1R nanodomains are more prominent at the beta cell-alpha cell membrane versus beta cell-beta cell membrane, identified using GCG staining and/or lack of GLP1R.

D-G) DBSCAN cluster analysis confirms GLP1R nanodomain enrichment at the beta cell-alpha cell membrane (D-F), although cluster size does not change (G) (unpaired t test).

H) Enrichment of GLP1R nanodomains at the beta cell-alpha cell membrane is lost in the antagonized state.

I-L) DBSCAN analysis shows loss of GLP1R nanodomain enrichment at the beta cell-alpha cell membrane following antagonism, with preserved localizations per cluster (unpaired t test).

Scale bar = 2000 nm. Bar graphs show mean ± SEM. ns, non-significant; *P<0.05. n = 44 cells, 4 islets, 3 mice (SulfoCy5), n = 38 cells, 4 islets, 3 mice (LUXendin645).

See also Figure S3.

To determine the contribution of alpha cell -> beta cell signalling to GLP1R nanodomain formation, experiments were repeated with the Cy5-based fluorescent GLP1R antagonist, LUXendin64525,46 (i.e. to block GLP1R). Using LUXendin645, GLP1R could also be seen at the nanoscale level (Figure 3H) in its antagonized/non-stimulated state, confirming our previous results25. However, no differences in clustering were seen in LUXendin645-treated beta cells bordering alpha cells versus other beta cells (Figure 3H–L), suggesting that alpha cell secretory products bind to and in part drive localized GLP1R nanodomain formation.

GLP1R nanodomains are an intrinsic cell property

To understand whether GLP1R nanodomains are an intrinsic cell property, or form secondary to cell morphology/location, we performed single-molecule FISH (smFISH) for Glp1r transcripts (Figure 4A), and quantified expression of the ubiquitous membrane protein E-cadherin. smFISH probe validation showed good concordance between Ins1 and Glp1r transcripts (Figure S4A), with little overlap between Ins1 transcript, Glp1r transcript and GCG protein (Figure S4B and C). Analysis of Glp1r on islet slices revealed a ~20% increase in transcript expression in beta cells adjacent to alpha cells versus those next to other beta cells (Figure 4B and C), with smFISH being quantitative (Figure 4D). Control probes did not show any differences in expression of Ins1 between beta cells in different locations in two out of three mice examined (Figure 4B and E). In a single mouse, increased Ins1 expression was detected in beta cells next to alpha cells (Figure 4E). To rule out an effect of membrane size on GLP1R nanodomain formation, E-cadherin expression was quantified. E-cadherin was found to be consistently expressed across all beta cells examined- whether or not they bordered alpha cells- shown by representative images (Figure 4F and G) and distance to fluorescence peak measures (Figure 4H and I).

Figure 4: GLP1R nanodomain formation is a cell intrinsic property.

Figure 4:

A) Schematic showing workflow for multiplexed smFISH and immunostaining on the same sample.

B) Single-molecule FISH (smFISH) shows Glp1r transcript expression largely restricted to Ins1+ cells, although transcripts occupy different locations within the cytoplasm (zoom-in, right panel) (representative images from n = 22 islets, 4 mice).

C, D) Representative images (C) (zoom-in of B) and bar graphs (D) showing that Glp1r transcript is enriched in beta cells bordering alpha cells, shown across three different preparations (unpaired t test) (n = 98 cells, 3 islets, 3 mice).

E) Ins1 transcript expression is not influenced by beta cell-alpha cell position in 2 out of 3 preparations examined (unpaired t test).

F, G) E-cadherin staining showing similar membrane morphology/thickness irrespective of beta cell position in relation to alpha cells (representative images from n = 8 islets, 4 mice).

H, I) Intersection (H) and peak (I) distance to alpha cell membrane is shorter for GLP1R versus E-cadherin (n = 8 islets, 4 mice). Pancreas sections were cut from the same FFPE blocks as in Figure 1, hence GCG-GLP1R control (Figure 1F and G) is shown again here for easy comparison.

Scale bar in A-C = 10 μm; scale bar in E, F = 20 μm. Bar and line graphs show mean ± SEM. ns, non-significant; *P<0.05; ****P<0.0001. GLP1R/Glp1r, glucagon-like peptide-1 receptor; INS, insulin; GCG, glucagon; E-Cad, E-Cadherin. Where shown, nuclei are labelled with Hoechst 33342 (blue).

See also Figure S4, Table S2 and Table S3.

Thus, GLP1R enrichment in beta cells close to alpha cells likely represents a discrete Glp1r transcriptomic/expression state rather than membrane thickening and/or aggregation.

Alpha cell activity stimulates GLP1R trafficking in neighboring beta cells in vitro

We next set out to understand whether the enrichment of GLP1R at the beta cell-alpha cell interface was associated with preferential signalling responses in beta cells neighboring alpha cells. By taking advantage of GLP1RSNAP/SNAP islets, we could readout GLP1R trafficking and hence signalling in response to endogenous ligand-over dozens of minutes in response to various stimuli (Figure 5A). Using this approach, we found that glucagonotropic low (0.5 mM) glucose was associated with GLP1R internalization in beta cells in the immediate vicinity of alpha cells (Figure 5B and C), identified by lack of GLP1R. Beta cells adjacent to other beta cells did not respond to low glucose with GLP1R internalization, with receptor remaining at the membrane (Figure 5B and C). Zonation of GLP1R responses was specific, since pre-application of the GLP1R antagonist, Ex9, prevented GLP1R internalization in beta cells next to alpha cells (Figure 5B and D). No differences in GLP1R internalization were seen between the various islet zones following application of high (17 mM) glucose (Figure 5B and E). Conversely, application of 10 nM glucose-dependent insulinotropic polypeptide (GIP) to near-maximally stimulate glucagon secretion was associated with GLP1R internalization in all beta cells examined throughout the islet (Figure 5B and F).

Figure 5: Alpha cell activity engages local GLP1R signalling in vitro.

Figure 5:

A) Schematic showing workflow to interrogate localized GLP1R signalling in the islet.

B) Low glucose (0.5 mM, G0.5) stimulates GLP1R internalization only in beta cells close to alpha cells, an effect prevented by pre-application of 500 nM exendin9 (Ex9), a GLP1R antagonist. GLP1R remains at the cell surface in the presence of high glucose (17 mM, G17), whereas GIP stimulates widespread GLP1R internalization.

C-F) Summary bar graphs and line graph showing effects of G0.5, G0.5 + Ex9, G17 and GIP on GLP1R internalization at the beta cell-alpha cell versus beta cell-beta cell interface (unpaired t test) (n = 330 cells, 4 islets, 4 mice).

G) As for A), but showing that low glucose (2.7 mM; G2.7) similarly internalizes GLP1R in beta cells adjacent to alpha cells, an effect prevented by pre-application of Ex9, or the endocytosis inhibitors dynasore (dyna) and dyngo4a.

H-J) Summary bar graphs showing effects of G2.7, G2.7 + Ex9, G2.7 + dyna and G2.7 + dyngo4a on GLP1R internalization at the beta cell-alpha cell versus beta cell-beta cell interface (unpaired t test) (n = 778 cells, 5 islets, 3 mice).

K) As for A), but showing that alanine (Ala) stimulates GLP1R internalization at the beta cell-alpha cell interface at medium (8 mM; G8) but not high (17 mM; G17) glucose, whereas glucagon (GCG) stimulates widespread GLP1R internalization.

L-N) Summary bar graphs showing effects of G5 (5 mM glucose), G8, G8 + Ala, G17, G17 + Ala and G0.5 + GCG on GLP1R internalization at the beta cell-alpha cell versus beta cell-beta cell interface (unpaired t test) (n = 635 cells, 7 islets, 3 mice).

Scale bar = 26.5 μm. Bar and line graphs show mean ± SEM. ns, non-significant; *P<0.05, ****P<0.0001. GIP, glucose-dependent insulinotropic polypeptide.

See also Figure S4.

Experiments were repeated at less stimulatory 2.7 mM glucose with similar results to 0.5 mM glucose (Figure 5G and H). The dynamin inhibitors, Dynasore and Dyngo4a, prevented low glucose from internalizing GLP1R in beta cells next to alpha cells (Figure 5G, I and J), suggesting a role for dynamin-dependent endocytosis in the observed responses. Localized GLP1R internalization was detected at 5 mM glucose but not 8 mM glucose, unless alanine was co-applied i.e. to mimic glucagon release during the postprandial state (Figure 5K and L). No localized GLP1R internalization was detected in response to 17 mM glucose + alanine, suggesting that either prevailing insulin levels or glucagon tone influence the response (Figure 5K and M). As expected, acute application of exogenous glucagon led to GLP1R internalization across the beta cell population (Figure 5K and N), much higher than that observed with 0.5 mM glucose-alone (Figure 5B, C).

Together, these data show that activation of alpha cells by low (0.5 – 5 mM) glucose as well as 8 mM glucose + alanine leads to pre-internalization of GLP1R in neighboring beta cells.

Alpha cell activity stimulates GLP1R trafficking in neighboring beta cells in vivo

Hormone concentration, hormone release patterns and paracrine communication routes might all differ in vitro and in vivo due to the combined effects of blood flow rate, blood flow pattern, extravasation and vascular permeability47–49. Whilst alpha cell -> beta cell communication has been shown to be active in vivo21, we wanted to confirm whether the same is true for GLP1R pre-internalization. To do this, mice were fasted for 4–6 hrs, before injection of insulin to moderately lower blood glucose concentration and stimulate physiological glucagon release. Pancreases were retrieved at t = 30 min for morphometric analysis.

Using this in vivo model of glucagon release, loss of GLP1R membrane labelling was detected in beta cells adjacent to alpha cells, but not in beta cells next to other beta cells, confirming in vitro longitudinal imaging results using GLP1RSNAP/SNAP islets (Figure 6A–E). No differences were seen between beta cells:alpha cells and beta cells:beta cells following injection of saline (Figure 6A–E). As expected, mice responded to 0.75 U/kg insulin injection with a decrease in fasted blood glucose levels (Figure 6F). While saline injection led to a small increase in blood glucose, this was not beyond that expected from the handling needed for IP administration.

Figure 6: Alpha cell activity engages local GLP1R signalling in vivo.

Figure 6:

A) Schematic showing experimental protocol to induce alpha cell activity in vivo.

B) GLP1R is retained at the beta cell membrane in saline-injected mice.

C) As for B), but showing that insulin injection (to provoke glucagon secretion) only triggers internalization of GLP1R in beta cells located close to alpha cells.

D, E) Insulin but not saline induces localized GLP1R internalization, as shown by fluorescence intensity along a line spanning the beta cell-alpha cell or beta cell-beta cell membrane (D), as well as summary bar graphs (E) (Mann-Whitney test) (n = 4 islets, 3 mice).

F) Blood glucose levels at 0- and 30-min post-saline or insulin injection (n = 4 islets, 3 mice).

G) GLP1RSNAP/SNAP internalization (left panel) in response to increasing glucagon (GCG) concentration (4-parameter fit), with interpolation showing GCG levels achieved at the beta cell-alpha cell membrane during incubation with low (0.5 mM) glucose (right panel) (n = 15 islets, 13 mice).

H) Schematic showing principal behind proximity ligation assay (PLA).

I) PLA shows interactions (40 nm) between endogenous glucagon and GLP1R proteins in pancreases from saline-injected (fasted) mice (arrows show lack of PLA signal in INS-negative regions).

J) As for (H), but random-fed mice.

K) Anti-glucagon antibody (Sigma-Aldrich, cat no. G2654) used for PLA does not cross-react with proglucagon-derived peptides, determined using α-cellPCSK2−/− islets that lack glucagon, but not GLP1 and oxyntomodulin (n = 16 islets, 6 mice).

Scale bar = 20 μm. ns, non-significant; ****P<0.0001. GLP1R, glucagon-like peptide-1 receptor; INS, insulin; GCG, glucagon; SST, somatostatin. Where shown, nuclei are labelled with Hoechst 33342 (blue).

See also Figures S5 and S6.

These experiments demonstrate that alpha cells engage GLP1R signalling at the beta cell-alpha cell interface in vivo, excluding a major effect of blood flow and the vasculature on the in vitro findings here.

Glucagon directly binds GLP1R in vitro and in vivo

Glucagon binding to GLP1R has been largely inferred from mouse genetics, receptor antagonism and pharmacology experiments17,20–24. However, demonstrating direct binding of glucagon to GLP1R is more challenging. To provide details on glucagon-GLP1R interactions in real-time, we employed surface plasmon resonance (SPR). Using the GLP1R ectodomain immobilized via an Fc-tag, glucagon was perfused onto the chip at 1.45–200 nM without detectable signal (Figure S5A and B). Demonstrating that the sole GLP1R ectodomain is suitable for high affinity binding analysis, exendin4(1–39) gave clear SPR responses with a dissociation constant (Kd) ~3 nM (Figure S5C and S5D). Glucagon concentrations were then increased from 0.78 – 100 μM, leading to high ON/OFF rates that were too fast for fitting of kinetic constants (Figure S6A and S6B). Calculated Kd was ~15 μM, which although outside of instrument limits, indicates that glucagon concentrations > 0.5 μM are required for GLP1R interactions (Figure S6A and S6B). To confirm glucagon concentrations achieved at the beta cell-alpha cell interface with spatiotemporal precision, GLP1R internalization was measured in response to increasing glucagon concentration. From this, a standard curve could be plotted from which to interpolate concentrations of locally released glucagon. Glucagon was found to internalize GLP1R with an EC50 of 1.05 μM (95% CI 0.88–1.28 μM) (Figure 6G), with concentrations at the beta cell-alpha cell interface of 1.04 μM (95% CI 0.92–1.17 μM) during stimulation with 0.5 mM glucose, in keeping with the SPR measures.

To demonstrate direct binding of glucagon to GLP1R in the cell/tissue setting, we turned to proximity ligation assay (PLA). This technique relies upon rolling circle DNA synthesis that occurs when two antibody-DNA-oligo-labelled proteins are within 40 nm of each other, leading to signal amplification and visualization of even single proteins (Figure 6H). By applying PLA to pancreases from saline-injected fasted mice (i.e. with elevated glucagon levels), we could show the presence of bright punctate spots in the beta cell cytoplasm (Figure 6I). A similar labelling pattern was seen in pancreases from random-fed mice (Figure 6J). Due to the specificity of the glucagon and GLP1R antibodies, these spots likely represent glucagon directly interacting with the GLP1R. Very few PLA+ spots were detected in insulin-negative cells, assumed here to be alpha or delta cells in which GLP1R expression is very low (Figure 6I and J). Some spots were seen in exocrine tissue, which expresses low levels of GLP1R50 (Figure 6I and J).

To exclude cross-reactivity of the anti-glucagon antibody with proglucagon-derived peptides present in the islet or circulation, immunostaining was performed in alpha cell-specific PCSK2 knockout islets in which glucagon is absent but GLP1 and oxyntomodulin are present. Confirming that PLA results represent specific glucagon-GLP1R interactions, no signal could be detected in alpha cell-specific PCSK2 knockout islets stained with antibody against glucagon (Sigma-Aldrich, cat no. G2654) (Figure 6K).

Beta cells close to alpha cells propagate signals across the islet

We next wanted to know how beta cells neighboring alpha cells are able to influence global signals via pre-internalized GLP1R. To do this, GLP1RSNAP/SNAP islets were labelled to visualize GLP1R, before bulk-loading of the first 2–3 cell layers with the Ca2+ dye, Fluo8, and rapid imaging over 3–4 mins to capture beta cell activation.

Using this approach, GLP1R labelling could be used to identify beta cells (GLP1R+) and alpha cells (GLP1R-), as well as GLP1R pre-internalization, allowing correlation with Ca2+ signals (Figure 7A and B). Delta cells (GLP1R-) were excluded from analysis by their lack of activity at low glucose. During increases from low (2.7 mM) glucose to high (17 mM) glucose, beta cells close to alpha cells responded with earlier Ca2+ rises than beta cells close to other beta cells, with an average lag time to peak of −25 sec (Figure 7C and F) (full traces shown in Figure S7A). No differences were detected in Ca2+ area-under-the-curve (AUC) in beta cells adjacent to alpha cells versus other beta cells (Figure 7C and F). Pre-incubation with the GLP1R antagonist Ex9 abolished this spatiotemporal relationship, with two broad response patterns being detected: 1) beta cells responding with Ca2+ rises at similar times following introduction of high glucose (Figure 7D and G); or 2) beta cells responding heterogeneously to the same stimulus (Figure 7E and G). Ex9 also reduced the Ca2+ AUC of beta cells adjacent to alpha cells versus other beta cells (Figure 7G), suggesting that GLP1R signalling/internalization influences response time as well as extent of Ca2+ fluxes. To exclude any effects of the Ca2+ indicator per se, experiments were repeated with GCaMP6m, with similar results for all conditions examined (Figure S7B–S7G).

Figure 7: Beta cells close to alpha cells are functionally optimized.

Figure 7:

A, B) Overlay of GLP1RSNAP/SNAP and Fluo8 labelling showing location of alpha cells and beta cells for analysis in vehicle (A) and exendin9 (Ex9)-treated (B) samples.

C-E) Representative Ca2+ traces (~120 secs) showing that beta cells respond earlier to glucose at the beta cell-alpha cell interface (C), an effect blocked by Ex9, which leads to more homogenous (D) or heterogeneous (E) responses.

F, G) Δt peak lag and AUC analysis showing on average earlier responses in beta cells close to alpha cells, a relationship disrupted by Ex9 (grey boxes show more heterogeneous responses) (unpaired t test) (n = 11 islets, 4 mice).

H) Representative Ca2+ images showing location of early responding beta cells, and effects of Ex9 upon this. Scale bar = 20 μm.

I, J) Schematic of glucose-dependent changes in INS-sfGFP:C-peptide-mCherry ratio (I), and representative TIRF-M image showing C-peptide-mCherry and INS-sfGFP in a single beta cell (J).

K-M) Representative images of RINS1-transduced islets showing C-peptide-mCherry, INS-sfGFP and INS-sfGFP:C-peptide-mCherry ratio (i.e. secreted) 120 sec post-stimulation with G17 + Veh (K), G17 + Ex9 (L), and G2.7 + KCl (M). Note that the ratioed image is inverted (i.e. mCherry:sfGFP) to show secretion as an increase from blue-yellow-white.

N) Analysis of mCherry:sfGFP ratio showing increased exocytosis in beta cells close to alpha cells, an effect blocked by Ex9 (KCl is positive control) (n = 403 cells, 17 islets, 6 mice).

Bar and line graphs show mean ± SEM. ns, non-significant; *P<0.05. G2.7, 2.7 mM; G17, 17 mM glucose.

See also Figure S7.

Beta cells next to alpha cells have a highly secretory phenotype

The ratiometric insulin sensor, RINS1, was used to map beta cell secretory phenotypes as a function of their localization within the islet51. RINS1 is a double fluorescent proinsulin fusion construct, which releases insulin-sfGFP whilst retaining C-peptide-mCherry within the secretory granules.

Using RINS1 combined with TIRF microscopy to resolve the submembrane compartment, insulin release can be monitored across the beta cell population using the sfGFP:mCherry ratio (Figure 7I and J). A decrease in sfGFP:mCherry ratio represents an increase in insulin secretion. Two minutes post-stimulation with high (17 mM) glucose, beta cells next to alpha cells released almost 2-fold more insulin than those next to other beta cells, consistent across all islets examined (Figure 7K, L and N). Pre-incubation with the GLP1R antagonist, Ex9, abolished this secretory zonation, with all beta cells responding similarly to high glucose with moderate insulin release, as also shown using Zn2+ as a surrogate for insulin granule dissolution52 (Figure 7K, L and N). Similar results were observed with KCl, which stimulated high levels of insulin secretion across the beta cell population (Figure 7M and N).

DISCUSSION

We show that beta cells located next to alpha cells are enriched for GLP1R surface nanodomains, which specifically pre-internalize in response to low glucose and glucagon input. Pre-internalized GLP1R is associated with earlier Ca2+ responses in beta cells exposed to high glucose, with signals propagating across the beta cell population. Beta cells close to alpha cells respond early with increases in insulin secretion. During aging and high fat diet, GLP1R expression increases, but GLP1R nanodomains are lost through remodelling. In SC-islets, GLP1R nanodomains fail to model properly in the first place. We thus provide a detailed paracrine mechanism for the amplifying effects of glucagon upon beta cell function, and more broadly alpha cell to beta cell communication within the islet.

Suggesting that GLP1R is the principal mechanism underlying signalling at the beta cell-alpha cell interface, are the following observations: 1) GLP1R was seen to internalize and signal in beta cells close to alpha cells in vitro and in vivo; 2) Ex9 blocked GLP1R pre-internalization as well as beta cell priming in terms of Ca2+ signals; and 3) SPR and PLA showed direct interactions between glucagon and GLP1R. Given its low affinity and high ON/OFF rate for glucagon, it is likely that GLP1R is well calibrated for such a local communication role, since it would only respond to the high levels of glucagon that would build up at the beta cell membrane facing an alpha cell. The GCGR likely plays a wider shaping role in alpha cell -> beta cell communication due to its much higher affinity for glucagon27–29, expected to lead to GCGR signalling in the majority of beta cells where its transcripts are expressed at similar levels to GLP1R53.

Paracrine communication is clearly important for islet function, with regulatory loops described between alpha cells, beta cells and delta cells, including those targeted by GLP1R/GIPR dual (ant)agonists15,18,28,54. What is less clear are the molecular underpinnings for paracrine communications, since cells must have discrete states to allow such cross-talk. Whilst beta cell and alpha cell heterogeneity have been described at the transcript and functional levels55–58, the relevance of this for paracrine communication is not fully understood. Here, beta cells were shown to be defined by their expression of a surface protein as well as transcript, which combined with their location, lends them to their role as a paracrine relay. It will be interesting to understand whether other cell-type specific GPCRs, known to be critical for hormone release, might be enriched at homotypic and heterotypic cell interfaces, e.g. SST2R, AVP1R, GHSR1A59–62.

Previous studies have shown that acetylcholine from human alpha cells provides a strong priming signal for beta cell function during the transition from low to high glucose18,19 Since acetylcholine is co-released with glucagon, engagement of pre-internalized GLP1R signalling might provide a parallel mechanism to ensure that insulin release is able to rapidly respond to an increase in blood glucose level, as well as during the postprandial state. It makes sense that redundancy exists in a molculear mechanism that is essential to rapidly counteract hyperglycemia. It will be interesting in the future to understand how muscarinic and GLP1 receptors interact at the cell surface to regulate beta cell function, as previously suggested by synergistic effects of ACh and GLP1 on insulin release63. A recent study has also shown that glucagon acts upon GLP1R and GCGR to amplify glucose-induced microtubule re-arrangement in neighboring beta cells64, a necessary step in insulin exocytosis65. Accordingly, islets with higher alpha cell:beta cell ratios secrete more insulin, which fits with our data here.

We and others have shown the existence of heterogeneous functional beta cell subpopulations such as hubs, leaders and first responders55,58,66. Recent studies have also shown that alpha- to beta-cell communication contributes to coordinated beta cell function67,68. The data here provide further clues as to the molecular mechanisms underlying coordinated subpopulation activity, and suggest that paracrine signalling via GLP1R nanodomains is capable of ordering and amplifying beta cell responses to glucose.

There has been constructive debate about whether paracrine mechanisms shown in isolated islets exist in vivo, where blood supply and exocrine interactions are intact69,70. The work here provides a detailed molecular mechanism for alpha cell -> beta cell communication, as shown in multiple studies of isolated islets as well as in animal models16–21,24,30. Whilst much of the current work is in vitro, it would be unusual that such an intricate mechanism, tuned for local hormone sensing, would be an artefact of the islet isolation procedure or loss of blood flow. It is more likely that blood flow in vivo provides an additional layer of regulation for access of circulating nutrients/stimuli to islet cells, as well as exit of hormone from the pancreas, as shown by us and others47–49. We were able to confirm key elements of the paradigm in vivo using a model of physiological hypoglycemia and glucagon release, including showing proximity binding of glucagon to GLP1R.

The current work focuses on alpha cell -> beta cell GLP1R signalling during hypoglycemia. However, glucagon tone is present under normoglycemia, and can be modified by inputs such as amino acid and exercise. These features of glucagon secretion support the GLP1R pre-internalization data shown here. At high glucose, the concentration of glucagon is likely under the threshold to trigger GLP1R internalization, given its lower affinity/potency for glucagon versus GLP1. However, GLP1R is a highly amplified system, and can still signal at the membrane, as inferred from studies showing the contribution of alpha cell -> beta cell communication to insulin secretion in the fed state. Localized GLP1R internalization becomes more apparent as glucose levels descend past 8 mM glucose (Figure 5L). We can thus propose a model whereby GLP1R pre-internalization provides the initial trigger needed to engage beta cells into Ca2+ signalling during the transition from low to high glucose. As glucose levels normalize and glucagon levels decline, beta cell activity is maintained through GLP1R surface signalling, with stimulatory inputs (i.e. amino acid/exercise) re-engaging localized GLP1R pre-internalization in a glucose-dependent manner (schematic in Figure S7H).

In summary, we show that beta cells adjacent to alpha cells are equipped as glucagon sensors by virtue of GLP1R enrichment, and disproportionately contribute to alpha cell regulation of beta cell function.

Limitations of the study

There are a number of limitations to the present studies. Firstly, we were unable to confirm that human beta cells preferentially internalize GLP1R in response to low glucose. We and others observe significant heterogeneity in GLP1R expression between human islet preparations43, which complicates mechanistic examination. Secondly, we were unable to disrupt GLP1R nanodomains in vivo, but can infer their importance from data including intraperitoneal insulin tolerance tests in aged animals as well as disrupted GLP1R signalling during metabolic stress44. Thirdly, while SPR recordings indicate that glucagon interacts with GLP1R at concentration > 0.5 μM, we could not accurately fit glucagon-GLP1R kinetic or dissociation constants within the instrument limits due to the high ON/OFF rates and low affinity. In any case, it is highly likely that glucagon directly binds GLP1R based upon PLA and SPR, as well as previously published pharmacology, KO and antagonist experiments21–24. Cryo-EM of glucagon-bound GLP1R would be helpful going forward, although such studies are not of the endogenous receptor, and low binding affinity might be an issue in resolving any structure.

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, David J. Hodson (david.hodson@ocdem.ox.ac.uk).

Materials availability

This study did not generate any new unique reagents.

Data and code availability

  • Source data are provided with this paper (Data S1).

  • Individual microscopy images and localization counts used for analysis are provided in Mendeley Data (http://dx.doi.org/10.17632/bm9jsyg2f2.1).

  • All other data reported in this paper will be shared by the lead contact upon request.

  • This paper does not report original code.

STAR★Methods

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Mice

Mice were socially-housed in specific-pathogen free conditions under a 12 hour light-dark cycle with ad libitum access to food and water. For high fat diet studies, C57BL6/J mice were fed either 60% calories from fat (Research Diets, cat no. D12492) or standard chow for 8 weeks before pancreas harvest. For ageing studies, pancreases were retrieved from C57BL6/J mice at 81 weeks.

Animal studies were regulated by the Animals (Scientific Procedures) Act 1986 of the U.K. (Personal Project Licence PP1778740). Approval was granted by the University of Oxford Animal Welfare and Ethical Review Body (AWERB). Tissues from aged mice were taken in accordance to German laws for the protection of animals and in compliance with protocols approved by local government authorities (State Agency of Environment, Health and Consumer Protection, State of Brandenburg, Germany) and the University of Potsdam Animal Welfare Office.

Human

Human pancreases were retrieved from Donors after Brain-Stem Death (DBD) with appropriate consent and ethical approval under 09/H0605/2 (REC: Oxfordshire Rec B). Islets were isolated in the DRWF Oxford Human Islet Isolation Facility using established isolation methods.

Studies with human islets and pancreata were approved by the University of Oxford Ethics Committee, as well as the National Research Ethics Committee (REC 16/NE/0107, Newcastle and North Tyneside, UK). Human donor characteristics are reported in Table S1.

METHODS DETAILS

Mouse pancreas and islet isolation

For islet isolation, male and female 8–16 weeks old C57BL6/J or GLP1RSNAP/SNAP 37 mice were used. Mice were humanely culled using a schedule-1 method and the common bile duct injected with Serva NB8 (1mg/mL) collagenase, before digestion and islet isolation using Histopaque-1119 and 1083 (Sigma-Aldrich) gradients. Islets were cultured at 37°C, 5% CO2 in RPMI supplemented with 10% fetal bovine serum (Gibco), 100 units/mL penicillin, and 100 μg/mL streptomycin (Sigma-Aldrich). Islets were used up to four days after isolation.

Insulin tolerance test

Mice were fasted for 4 to 6 hours, before intraperitoneal injection of 0.75 U/kg body weight insulin (Humulin S, 100 U/mL, Lilly). Blood glucose was measured 0, 15 and 30 mins after insulin injection using a Contour glucometer. Mice were humanely culled and pancreases retrieved for immunohistochemistry.

iPSC-derived islet-like structures

ALSTEM #iPS11 human-induced pluripotent stem cells (iPSC) were cultured on iPSC qualified Matrigel in mTeSR+ media (STEMCELL Technologies #05826). Differentiation to islet-like structures (SC-islets) was carried out in a suspension-based, magnetic CELLSPIN bioreactor system (PFEIFFER). Differentiation media was changed daily by letting spheres settle by gravity for 3–10 min. sBC differentiation was based on published protocols53,71.

Immunohistochemistry and image processing/analysis

Formalin-fixed paraffin-embedded (FFPE) mouse pancreases or human islets were sliced on a vibratome at 5 μm, deparaffinized and rehydrated. Sections were incubated overnight at 4°C with primary antibodies diluted in PBS containing 2% BSA and 0.2% Triton X-100. Sections were washed in PBS-Tween before incubation in secondary antibodies, diluted as above, for 2 h at room temperature, followed by washing in PBS-Tween. Where required, nuclear labelling was performed with 1 μg/mL Hoechst 33342 (Thermo Fisher Scientific, cat no. H3570) for 15 min. Sections were then mounted with ProLong Glass Antifade Mountant (Thermo Fisher Scientific, cat no. P36982) and stored at 4°C pending imaging. SC-islets were formalin-fixed before embedding in optimal cutting temperature compound, and slicing at 5 μm on a cryostat at − 20 °C.

Primary antibodies used were: guinea pig anti-insulin (Dako, cat no. A05640–12-2; RRID: AB_2617169), mouse monoclonal anti-glucagon (Sigma-Aldrich, cat no. G2654; RRID:AB_259852), mouse monoclonal anti-somatostatin (Thermo Fisher Scientific, cat no. 14–9751-82; RRID: AB_2572982), and mouse monoclonal anti-E-cadherin (BD Biosciences, cat no. 610181, RRID:AB_397580). Anti-GLP1R primary antibodies used were: rabbit recombinant anti-GLP1R (Abcam, cat no. ab218532, RRID: AB_2864762) and mouse monoclonal anti-GLP1R (DSHB, cat no. Mab 3F52, RRID: AB_2618100).

Secondary antibodies used were donkey anti-rabbit Alexa Fluor 488 (Thermo Fisher Scientific, cat no. A-21206, RRID: AB_2535792), goat anti-mouse Alexa Fluor 488 (Thermo Fisher Scientific, cat no. A-11001, RRID: AB_2534069), goat anti-guinea pig Alexa Fluor 568 (Thermo Fisher Scientific, cat no. A-11075, RRID: AB_2534119), goat anti-rabbit Alexa Fluor 647 (Thermo Fisher Scientific, cat no. A-21244, RRID: AB_2535812) and goat anti-mouse Alexa Fluor 647 (Abcam, cat no. ab150119, RRID: AB_2811129).

Sections were imaged using: 1) Olympus FV3000 confocal microscope equipped with high-sensitivity spectral detectors and 40x, 1.25 NA and 60x, 1.30 NA UPLSAPOXS silicone objectives; and 2) Olympus FV4000 confocal microscope equipped with next generation SilVIR spectral detectors and 40x, 0.95 NA and 60x, 1.42 NA UPLSAPO oil objectives. Excitation and emission wavelengths for Hoechst 3342, Alexa Fluor 488, Alexa Fluor 568 and Alexa Fluor 647 were λex = 405 nm/λem = 406–461 nm, λex = 488 nm/λem = 499–520 nm, λex = 561 nm/λem = 579–603 nm, λex = 640 nm/λem = 649–700 nm respectively.

Images were captured at 1024×1024px and Nyquist-sampled in the axial plane, before iterative deconvolution using Imaris ClearView and a maximum likelihood estimator, with GPU acceleration provided by an AMD-Radeon Pro W5500 with 8 GB GDDR6 VRAM and 7 nm RDNA architecture. GLP1R signal was measured using corrected total cell fluorescence (CTCF), which is the integrated density corrected for the product of cell area and mean background fluorescence72,73.

GLP1R contacts analysis

GLP1R contacts were measured by calculating the distance between intersection points of normalized GLP1R and GCG/SST fluorescence peaks, with a zero-distance denoting co-localization. GLP1R internalization was quantified using either the full width at half maximum (FHWM) along a line transecting the beta cell-beta cell or beta cell-alpha cell interface, or single cell normalized integrated density.

dSTORM nanoscopy

GLP1RSNAP/SNAP and GLP1RWT/WT islets were labelled with BG-SulfoCy5 (500 nM) and LUXendin645, respectively, for 60 mins at 37 °C before fixation in 4% formalin. Islets were placed on a cavity slide, submerged in freshly reconstituted STORM buffer (Abbelight), and sealed using a 170 μm coverslip + EcoSil dental resin (Picodent, cat no. 1300 6100). Samples were imaged in HILO mode on an Evident/Abbelight SAFe 180 system, using a 100x, 1.5 NA Olympus UPLAPO100XOHR objective. SulfoCy5 and LUXendin645 were pumped to the dark state using an Oxxius laser combiner and 500 mW 640 nm diode laser, before initiation of photoblinking. A 405 nm laser was slowly ramped up during the acquisition to increase transition back to the dark state. Single molecule events were recorded using an LP650 filter with an integration time of 50 ms on a Hamamatsu ORCA-Fusion sCMOS for 20000–40000 frames. Localizations were extracted, and images reconstructed, using Abbelight NEO software. Density-Based Spatial Clustering of Applications with Noise (DBSCAN) was used to determine localization clustering, implemented in Abbelight NEO software, with ε = 25 nm (the average precision of the data), and minPts = 8.

smFISH

To determine localization of individual RNA transcripts, cryosections were incubated in 2X SSC Buffer (Thermo Fisher Scientific, cat no. AM9763), followed by FISH wash buffer containing 1X SSC and deionized formamide in RNAse free water. smFISH probes were hybridized in a humidity chamber for 18 hours at 30°C in buffer containing dextran sulfate, ribonucleoside vanadryl complexes, yeast tRNA and formamide, before washing in FISH wash buffer and nuclear labelling with Hoechst 33342. Sections were then mounted in ProLong Glass antifade mounting medium. smFISH probes used were against Ins1 and Glp1r and conjugated to Quasar 670 and CAL Fluor Red 590 (LGC Biosearch Technologies), respectively. smFISH probe sequences are provided in Tables S2 and S3 (Ins1 probes taken from74).

Where sequential antibody labelling was required, probe hybridization was followed by incubation with 2% BSA and 0.2% Triton X-100 in PBS treated with RNAsecure (Thermo Fisher Scientific, cat no. AM7005). Sections were then incubated overnight at 4°C with primary antibodies diluted in PBS containing 2% BSA and 0.2% Triton X-100. Sections were washed in PBS-Tween treated with RNAsecure, before incubation in secondary antibodies, diluted as above, for 2 h at room temperature, followed by wash in PBS-Tween treated with RNAsecure. Nuclear labelling was performed with Hoechst 33342. Sections were then mounted in ProLong Glass antifade mounting medium.

Sections were imaged using an Olympus FV4000 confocal microscope and 40x, 0.95 NA and 60x, 1.42 NA UPLSAPO oil objectives. Excitation and emission wavelengths for Hoechst 33342, Alexa Fluor 488, CAL Fluor Red 590, and Quasar 670 were λex = 405 nm/λem = 406–461 nm, λex = 488 nm/λem = 499–520 nm, λex = 561 nm/λem = 579–603 nm, λex = 640 nm/λem = 680–720 nm respectively. Images were deconvolved as described above.

RNA transcripts were quantified on a single-cell level by calculating the mean background-subtracted intensity. Background signal was determined by subtracting the median of the lower 50 percentile nuclear voxels of all cells from the median intensity of all nuclear voxels.

Proximity ligation assay

Proximity ligation assay was performed using commercially available reagents (Sigma-Aldrich, cat nos. DUO92001; DUO92005; DUO92013; DUO82049).

FFPE or OCT-embedded mouse pancreases were sliced on a vibratome at 5 μm, deparaffinized and rehydrated. Antigen retrieval was performed in buffer consisting of 10 mM Tris base, 1 mM EDTA, and 0.05% Tween-20 in PBS, at pH 9.0, for 25 min at 95°C in a water bath. Sections were quenched by incubation with 0.1 M glycine + 0.1M NH4Cl for 15 min at room temperature, followed by blocking in DuoLink blocking solution for 1 hour at 37°C. Primary antibodies against glucagon (Sigma-Aldrich, cat no. G2654; RRID:AB_259852) and GLP1R (Abcam, cat no. ab218532, RRID: AB_2864762) were diluted in DuoLink antibody diluent and incubated with the sections overnight at 4°C. Where required, primary antibody against insulin was also incubated in this way. Sections were then washed with DuoLink PLA Wash Buffer A, followed by incubation with PLA Probe PLUS and PLA Probe MINUS dissolved in DuoLink antibody diluent for 60 min at 37°C. Where required, secondary fluorescent antibody to label insulin was also incubated in this way. Samples were washed in PLA Wash Buffer A, followed by ligation for 60 min at 37 °C, then washing in PLA Wash Buffer A. Amplification was performed for 100 min at 37 °C, followed by washing in PLA Wash Buffer B, and where required, incubation with 1 μg/mL Hoechst 33342. Sections were then mounted with ProLong Glass antifade mounting medium pending imaging.

Mouse monoclonal anti-glucagon (Sigma-Aldrich, cat no. G2654; RRID:AB_259852) was tested for cross-reactivity against proglucagon-derived peptides using FFPE pancreas sections of alpha cell-specific PCSK2 knockout mice (Gcg-CreERT2;PCSK2fl/fl)75, and staining protocols as in “Immunohistochemistry”. Recombinant anti-GLP1R (Abcam, cat no. ab218532, RRID: AB_2864762) has previously been validated in GLP1R knockout islets41.

Sections were imaged using an Olympus FV4000 confocal microscope and 60x, 1.42 NA UPLSAPO oil objectives. Excitation and emission wavelengths for Hoechst 33342, Alexa Fluor 568, and PLA-Cy5 were λex = 405 nm/λem = 406–461 nm, λex = 561 nm/λem = 579–603 nm, λex = 640 nm/λem = 680–720 nm respectively.

Surface plasmon resonance

Binding kinetics of GLP-1R ectodomain and Glucagon (Tocris, cat. no. 6927) were determined using a Biacore™ T200 surface plasmon resonance system (Cytiva). Exendin4(1–39) served as a positive control. A Human Antibody capture kit (Cytiva) was used to immobilize Anti-Human IgG (Fc) on a Series S Sensor Chip CM5 according to the manufacturer’s instructions. For all measurements 10 mM Hepes pH 7.4, 150 mM NaCl, 0,05% Tween 20 was used as a running buffer at a temperature of 25 °C. Human GLP1R with a C-terminal Fc-tag (Sino Biological, cat. no. 13944-H02H) was captured on the chip with concentrations ranging from 0.5 μg/ml to 10 μg/ml and contact times of 45–60 seconds such that the final response with analytes did not exceed 50 RU. 120 μL of Ex4 or glucagon in running buffer were injected in varying concentrations and the response was recorded. Analyses were performed at 25 °C with a flow rate of 30 μl/min with 240 sec for association and 1000 sec for dissociation. After each run, surfaces were regenerated with 90 μl of 3 M Magnesium chloride at a flow rate of 30 μl/min for 180 seconds. The Biacore T200 evaluation software v3.2.1 was used to calculate association and dissociation rate constants using a 1:1 Langmuir fitting model.

GLP1RSNAP/SNAP imaging

To image GLP1R internalization, islets were isolated from homozygous GLP1RSNAP/SNAP mice and labelled with 1.5 μM BG-Sulfo646 for 30 mins, followed by wash in PBS. The islets were then incubated in the treatment condition for 30 min at 37°C, before imaging using an Olympus FV3000 microscope and UPLSAPO 60x / 1.30 NA silicone UPLSAPO objective. Excitation was delivered at λ = 640 nm using a Coherent OBIS diode laser, with emitted signals detected using a high-sensitivity spectral detector at λ = 650 to 750 nm. Z-stacks were acquired with slices acquired at 0.55 μm intervals.

GLP1RSNAP/SNAP-positive cells were classified as beta cells:alpha cells or beta cells:beta cells. GLP1R internalization was then calculated by measuring single cell normalized integrated density within a standardized region of interest in each cell. To determine glucagon concentration achieved at the beta cell-alpha cell interface, a four-parameter logistic (4PL) dose-response curve was plotted for GLP1RSNAP/SNAP internalization in response to known-glucagon treatment concentrations. Glucagon concentration values were then interpolated from the 4PL curve.

Calcium and RINS1 imaging

For Ca2+, islets were loaded with Fluo8 (AAT Bioquest) in pluronic acid and 1.5 μM BG-Sulfo646 for 30 mins, or transduced with AdV-GCaMP6m, before live imaging using an Olympus FV3000 microscope and 40x/1.4 NA UPLSAPO oil objective. Excitation was delivered at λ = 488 and 640 nm using a Coherent OBIS solid state laser, with emitted signals detected using a high-sensitivity spectral detector at λ = 500 to 540 nm and λ = 650–750 nm for Fluo8 and BG-Sulfo646, respectively. Glucose was puff-applied directly into the imaging cuvette at the indicated concentration and time, and Ca2+ traces captured for 3–5 mins. HEPES-bicarbonate buffer was used containing in mM: 120 NaCl, 4.8 KCl, 24 NaHCO3, 0.5 Na2HPO4, 5 HEPES, 2.5 CaCl2, 1.2 MgCl2, and 1–17 d-glucose. Time from glucose application to first cell response was 55.9 ± 10.23 secs. Time to peak was measured for beta cell-beta cell (t1) and beta cell-alpha cell (t2) subpopulations, before calculating lag (Δt = t2 – t1) normalized to the mean time to peak for t1.

For RINS1, islets were transduced with adenovirus harboring the insulin probe RINS151. Islets were pre-treated at low glucose for 30 min followed by stimulation under the indicated conditions for 300 sec, before fixation for 15 mins in 2% formalin. Islets were transferred to cavity slides, submerged in TIRF CitiFluor imaging medium (CitiFluor, cat no. AF3–100), and secured using a 170 μm coverslip + dental resin. Islets were imaged in TIRF mode on an Evident/Abbelight SAFe 180 system using a 100x, 1.5 NA Olympus UPLAPO100XOHR objective. sGFP was excited at λ = 488 nm, with emitted signals detected at λ = 498–555 nm. mCherry was excited at λ = 561 nm, with emitted signals detected at λ = 577–623 nm. Illumination was provided by an Oxxius L4CC laser combiner, with detection provided by a Hamamatsu ORCA-Fusion sCMOS.

Traces are displayed as F/F0–10 where F = fluorescence at any time point and F0–10 = average fluorescence of the first 10 frames. RINS1 data are calculated as INS-sfGFP:C-peptide-mCherry ratio, which is then inverted (i.e. mCherry:sfGFP) so that insulin secretion is represented by increasing values.

QUANTIFICATION AND STATISTICAL ANALYSIS

Islet isolation is a nuisance variable and as such data are taken from independent islet preparations. Samples and animals were allocated to treatment groups in a randomized manner. D’Agostino-Pearson normality test was used to assess data distribution. Pairwise comparisons were made using either unpaired t test (parametric) or Mann-Whitney test (non-parametric). Individual datapoints are shown in bar graphs. Unless otherwise stated in the figure legend, all error bars represent mean ± S.E.M. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. GraphPad Prism 10 was used for all statistical analyses.

Supplementary Material

1
2

Table S1: Human islet donor characteristics, related to Figure 1.

3

Table S2: smFISH Ins1 probe sequences, related to Figure 4.

4

Table S3: smFISH Glp1r probe sequences, related to Figure 4.

5

Key resources table

REAGENT or RESOURCE SOURCE IDENTIFIER
Antibodies
Guinea pig anti-insulin Dako cat no. A056401-2-2; RRID: AB_2617169
Mouse monoclonal anti-glucagon Sigma-Aldrich cat no. G2654; RRID:AB_259852
Mouse monoclonal anti-somatostatin Thermo Fisher Scientific cat no. 14-9751-82; RRID: AB_2572982
Mouse monoclonal anti-Ecadherin BD Biosciences cat no. 610181, RRID:AB_397580
Rabbit recombinant anti-GLP1R Abcam cat no. ab218532, RRID: AB_2864762
Mouse monoclonal anti-GLP1R Iowa DSHB cat no. Mab 3F52, RRID: AB_2618100
Donkey anti-rabbit Alexa Fluor 488 Thermo Fisher Scientific cat no. A-21206, RRID: AB_2535792
Goat anti-mouse Alexa Fluor 488 Thermo Fisher Scientific cat no. A-11001, RRID: AB_2534069
Goat anti-guinea pig Alexa Fluor 568 Thermo Fisher Scientific cat no. A-11075, RRID: AB_2534119
Goat anti-rabbit Alexa Fluor 647 Thermo Fisher Scientific cat no. A-21244, RRID: AB_2535812
Goat anti-mouse Alexa Fluor 647 Abcam cat no. ab150119, RRID: AB_2811129
Biological samples
Human cadaveric islets University of Oxford, Oxford Consortium for Islet Transplantation https://www.nds.ox.ac.uk/research/oxcit
Chemicals, peptides, and recombinant proteins
Glucagon Tocris cat. no. 6927
Glucagon Sigma-Aldrich G2044
GIP Bachem 4030658
Exendin 9-39 Sigma-Aldrich E7269
Dynasore Abcam ab120192
Dyngo4a Abcam ab120689
BG-Sulfo646 Johannes Broichhagen (Leibniz-FMP) 10.1038/s41467-022-35716-1 and 10.1039/d1ob02216d
GLP1R with a C-terminal Fc-tag Sino Biological cat. no. 13944-H02H
Critical commercial assays
Duolink Proximity Ligation Assay Sigma-Aldrich cat nos. DUO92001; DUO92005; DUO92013; DUO82049
Deposited data
Microscopy images and localization counts used for GLP1R nanodomain analysis Mendeley Data http://dx.doi.org/10.17632/bm9jsyg2f2.1
Source Data Tong et al Data S1 - Source Data
Experimental models: Cell lines
Human iPSC Line Alstem Bio Cat no. iPS11
Experimental models: Organisms/strains
GLP1RSNAP/SNAP mice David Hodson (University of Oxford) 10.1038/s41467-022-35716-1
C57BL/6JOlaHsd mice Envigo RMS (UK) Ltd Model code 057
Gcg-CreERT2 mice Klaus Kaestner (University of Pennsylvania) RRID:MMRRC_042277-JAX
PCSK2-flox’d mice Bruce Verchere
Oligonucleotides
See Table S2 for smFISH Ins1 probe sequence LGC Biosearch Technologies N/A
See Table S3 for smFISH Glp1r probe sequence LGC Biosearch Technologies N/A
Recombinant DNA
Ad5-RINS1 insulin sensor Carsten Schultz (OHSU) 10.1016/j.chembiol.2017.03.001
Ad5-GCaMP6 Ca2+ indicator Vector Biolabs Cat no. 1909
Software and algorithms
GraphPad Prism 10.4.1 GraphPad software https://www.graphpad.com/features
Fiji NIH https://imagej.net/software/fiji/
Imaris Clearview Oxford Instruments https://imaris.oxinst.com/products/clearviewgpu-deconvolution
CellSens Evident Scientific https://www.olympus-lifescience.com/en/software/cellsens/
Abbelight NEO Abbelight https://www.abbelight.com/solutions/
Biacore T200 evaluation software Biacore v3.2.1
Other
Evident FV3000 confocal microscope Evident Scientific RRID:SCR_017015
Evident FV4000 confocal microscope Evident Scientific N/A
Abbelight SAFe 180 TIRF/dSTORM Abbelight N/A
High fat diet Research Diets Cat# D12492

HIGHLIGHTS.

  • GLP1R is enriched as nanodomains on the beta cell membrane facing alpha cells.

  • GLP1R nanodomains are disrupted with aging and metabolic stress.

  • Beta cells next to alpha cells pre-internalize GLP1R at low glucose.

  • GLP1R pre-internalization allows earlier & more robust responses to high glucose.

ACKNOWLEDGEMENTS

D.J.H. was supported by MRC (MR/S025618/1) and Diabetes UK (22/0006389) Grants. This research was funded in whole, or in part, by the UKRI [EP/X026833/1] (to D.J.H.). This project has received funding from the European Union’s Horizon Europe Framework Programme (deuterON, grant agreement no. 101042046 to J.B.). This work was supported on behalf of the “Steve Morgan Foundation Type 1 Diabetes Grand Challenge” by Diabetes UK and SMF (grant number 23/0006627 to D.J.H. and J.B.). A.H.S. was supported by a Novo Nordisk – Oxford Fellowship. O.E.O was supported by an MRC cell and oligonucleotide therapy fellowship with AstraZeneca (MR/X004643/1). A.K. was supported by the Deutsche Forschungsgemeinschaft (DFG) grant project KL 2399/8-1. C.S. was supported by a Helen Jo and Bill Whitsell endowment. The research was funded by the National Institute for Health Research (NIHR) Oxford Biomedical Research Centre (BRC). The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health. C.B.V. was supported by a Canadian Institutes of Health and Research grant (PJT-153156). J.E.C. was support by grants from the NIH (DK142423, DK141090, DK125353, and DK046492). The project involves an element of animal work not funded by the NIHR but by another funder, as well as an element focussed on patients and people appropriately funded by the NIHR. The DRWF Oxford Human Islet Isolation Facility was funded by the Diabetes Research and Wellness Foundation (DRWF) and Breakthrough T1D. We are a grateful to Dr Andreas Oder (Leibniz-Forschungsinstitut für Molekulare Pharmakologie) for help with Surface Plasmon Resonance. We thank Dr David d’Alessio (Duke University) for providing feedback on the research during various discussions. Graphical abstract was created in BioRender. Tong, J. (2025) https://BioRender.com/fuzvh0a. Figure 1A was created in BioRender. Tong, J. (2025) https://BioRender.com/zpnvhje. Figure 4A was created in BioRender. Tong, J. (2025) https://BioRender.com/htuhvd7. Figure 6H was created in BioRender. Tong, J. (2025) https://BioRender.com/5o0xdzt. Figure 7I was created in BioRender. Tong, J. (2025) https://BioRender.com/m98w369. Figure S7H was created in BioRender. Tong, J. (2025) https://BioRender.com/rwr2vdx. For the purpose of Open Access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission.

DECLARATION OF INTERESTS

M.C., A.E., C.A. are employees of Novo Nordisk A/S. J.E.C. has received funding for research from Eli Lilly, Novo Nordisk, Merck, Structure Therapeutics, Fractyl Therapeutics, Alticore, and Prostasis. J.E.C. has served as a consultant/advisory in the past 12 months with Arrowhead Therapeutics, Boehringer Ingelheim, Neurocrine Biosciences, Protagonist Therapeutics, Prostasis, and Structure Therapeutics. A.K. has received speakeŕs fees from Novo Nordisk A/S and Daiichi Sankyo. D.J.H. and J.B. have filed a patent on GLP1R and GIPR chemical probes. D.J.H. and J.B. receive licensing revenue from Celtarys Research for provision of GLP1R/GIPR chemical probes. D.J.H. has filed patents related to type 2 diabetes therapy and GLP1R agonism.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

REFERENCES

  • 1.Dolenšek J, Rupnik MS, and Stožer A (2015). Structural similarities and differences between the human and the mouse pancreas. Islets 7, e1024405. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Sasson A, Rachi E, Sakhneny L, Baer D, Lisnyansky M, Epshtein A, and Landsman L (2016). Islet Pericytes Are Required for β-Cell Maturity. Diabetes 65, 3008–3014. [DOI] [PubMed] [Google Scholar]
  • 3.Gan WJ, Do OH, Cottle L, Ma W, Kosobrodova E, Cooper-White J, Bilek M, and Thorn P (2018). Local Integrin Activation in Pancreatic β Cells Targets Insulin Secretion to the Vasculature. Cell Rep. 24, 2819–2826.e2813. [DOI] [PubMed] [Google Scholar]
  • 4.Adams MT, Dwulet JM, Briggs JK, Reissaus CA, Jin E, Szulczewski JM, Lyman MR, Sdao SM, Kravets V, Nimkulrat SD, et al. (2021). Reduced synchroneity of intra-islet Ca(2+) oscillations in vivo in Robo-deficient beta cells. Elife 10, e61308. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Jevon D, Deng K, Hallahan N, Kumar K, Tong J, Gan WJ, Tran C, Bilek M, and Thorn P (2022). Local activation of focal adhesion kinase orchestrates the positioning of presynaptic scaffold proteins and Ca2+ signalling to control glucose-dependent insulin secretion. eLife 11, e76262. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Caicedo A (2013). Paracrine and autocrine interactions in the human islet: more than meets the eye. Semin. Cell Dev. Biol 24, 11–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Huising MO (2020). Paracrine regulation of insulin secretion. Diabetologia 63, 2057–2063. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Rutter GA, Pullen TJ, Hodson DJ, and Martinez-Sanchez A (2015). Pancreatic beta-cell identity, glucose sensing and the control of insulin secretion. Biochem. J 466, 203–218. [DOI] [PubMed] [Google Scholar]
  • 9.Rorsman P, and Ashcroft FM (2018). Pancreatic beta-Cell Electrical Activity and Insulin Secretion: Of Mice and Men. Physiol. Rev 98, 117–214. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Lewandowski SL, Cardone RL, Foster HR, Ho T, Potapenko E, Poudel C, VanDeusen HR, Sdao SM, Alves TC, Zhao X, et al. (2020). Pyruvate Kinase Controls Signal Strength in the Insulin Secretory Pathway. Cell Metab 32, 736–750.e735. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Merrins MJ, Corkey BE, Kibbey RG, and Prentki M (2022). Metabolic cycles and signals for insulin secretion. Cell Metab 34, 947–968. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Rodriguez-Diaz R, Menegaz D, and Caicedo A (2014). Neurotransmitters act as paracrine signals to regulate insulin secretion from the human pancreatic islet. J. Physiol 592, 3413–3417. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Rodriguez-Diaz R, Molano RD, Weitz JR, Abdulreda MH, Berman DM, Leibiger B, Leibiger IB, Kenyon NS, Ricordi C, Pileggi A, et al. (2018). Paracrine Interactions within the Pancreatic Islet Determine the Glycemic Set Point. Cell Metab 27, 549–558.e544. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Holter MM, Saikia M, and Cummings BP (2022). Alpha-cell paracrine signaling in the regulation of beta-cell insulin secretion. Front. Endocrinol. (Lausanne) 13, 934775. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Huang JL, Pourhosseinzadeh MS, Lee S, Krämer N, Guillen JV, Cinque NH, Aniceto P, Momen AT, Koike S, and Huising MO (2024). Paracrine signalling by pancreatic δ cells determines the glycaemic set point in mice. Nat. Metab 6, 61–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Samols E, Marri G, and Marks V (1965). Promotion of Insulin Secretion by Glucagon. The Lancet 286, 415–416. [DOI] [PubMed] [Google Scholar]
  • 17.Moens K, Flamez D, Schravendijk CV, Ling Z, Pipeleers D, and Schuit F (1998). Dual Glucagon Recognition by Pancreatic β-Cells via Glucagon and Glucagon-Like Peptide 1 Receptors. Diabetes 47, 66–72. [DOI] [PubMed] [Google Scholar]
  • 18.Rodriguez-Diaz R, Dando R, Jacques-Silva MC, Fachado A, Molina J, Abdulreda MH, Ricordi C, Roper SD, Berggren PO, and Caicedo A (2011). Alpha cells secrete acetylcholine as a non-neuronal paracrine signal priming beta cell function in humans. Nat. Med 17, 888–892. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Molina J, Rodriguez-Diaz R, Fachado A, Jacques-Silva MC, Berggren PO, and Caicedo A (2014). Control Of Insulin Secretion By Cholinergic Signaling In The Human Pancreatic Islet. Diabetes 63, 2716–2726. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Zhu L, Dattaroy D, Pham J, Wang L, Barella LF, Cui Y, Wilkins KJ, Roth BL, Hochgeschwender U, Matschinsky FM, et al. (2019). Intraislet glucagon signaling is critical for maintaining glucose homeostasis. JCI Insight 4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Capozzi ME, Svendsen B, Encisco SE, Lewandowski SL, Martin MD, Lin H, Jaffe JL, Coch RW, Haldeman JM, MacDonald PE, et al. (2019). beta Cell tone is defined by proglucagon peptides through cAMP signaling. JCI Insight 4, e126742. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Chepurny OG, Matsoukas M-T, Liapakis G, Leech CA, Milliken BT, Doyle RP, and Holz GG (2019). Nonconventional glucagon and GLP-1 receptor agonist and antagonist interplay at the GLP-1 receptor revealed in high-throughput FRET assays for cAMP. J. Biol. Chem 294, 3514–3531. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Shuai H, Xu Y, Ahooghalandari P, and Tengholm A (2021). Glucose-induced cAMP elevation in beta-cells involves amplification of constitutive and glucagon-activated GLP-1 receptor signalling. Acta Physiol. (Oxf.) 231, e13611. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Cabrera O, Ficorilli J, Shaw J, Echeverri F, Schwede F, Chepurny OG, Leech CA, and Holz GG (2022). Intra-islet glucagon confers β-cell glucose competence for first-phase insulin secretion and favors GLP-1R stimulation by exogenous glucagon. J. Biol. Chem 298, 101484. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Ast J, Arvaniti A, Fine NHF, Nasteska D, Ashford FB, Stamataki Z, Koszegi Z, Bacon A, Jones BJ, Lucey MA, et al. (2020). Super-resolution microscopy compatible fluorescent probes reveal endogenous glucagon-like peptide-1 receptor distribution and dynamics. Nat. Commun 11, 467. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Gray SM, Xin Y, Ross EC, Chazotte BM, Capozzi ME, El K, Svendsen B, Ravn P, Sloop KW, Tong J, et al. (2020). Discordance between GLP-1R gene and protein expression in mouse pancreatic islet cells. J. Biol. Chem 295, 11529–11541. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Svendsen B, Larsen O, Gabe MBN, Christiansen CB, Rosenkilde MM, Drucker DJ, and Holst JJ (2018). Insulin Secretion Depends on Intra-islet Glucagon Signaling. Cell Rep 25, 1127–1134 e1122. [DOI] [PubMed] [Google Scholar]
  • 28.Zhang Y, Han C, Zhu W, Yang G, Peng X, Mehta S, Zhang J, Chen L, and Liu Y (2021). Glucagon Potentiates Insulin Secretion Via β-Cell GCGR at Physiological Concentrations of Glucose. Cells 10, 2495. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Suba K, Patel Y, Martin-Alonso A, Hansen B, Xu X, Roberts A, Norton M, Chung P, Shrewsbury J, Kwok R, et al. (2024). Intra-islet glucagon signalling regulates beta-cell connectivity, first-phase insulin secretion and glucose homoeostasis. Mol. Metab 85, 101947. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Siegel EG, and Creutzfeldt W (1987). Effect of low concentrations of glucagon on insulin release and cyclic AMP content in isolated rat islets. Metabolism 36, 953–957. [DOI] [PubMed] [Google Scholar]
  • 31.Drucker DJ, Philippe J, Mojsov S, Chick WL, and Habener JF (1987). Glucagon-like peptide I stimulates insulin gene expression and increases cyclic AMP levels in a rat islet cell line. Proc. Natl. Acad. Sci. U.S.A 84, 3434–3438. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Holz G.G.t., Leech CA, and Habener JF (1995). Activation of a cAMP-regulated Ca(2+)-signaling pathway in pancreatic beta-cells by the insulinotropic hormone glucagon-like peptide-1. J. Biol. Chem 270, 17749–17757. [PMC free article] [PubMed] [Google Scholar]
  • 33.Widmann C, Dolci W, and Thorens B (1996). Desensitization and phosphorylation of the glucagon-like peptide-1 (GLP-1) receptor by GLP-1 and 4-phorbol 12-myristate 13-acetate. Mol. Endocrinol 10, 62–75. [DOI] [PubMed] [Google Scholar]
  • 34.Roed SN, Wismann P, Underwood CR, Kulahin N, Iversen H, Cappelen KA, Schaffer L, Lehtonen J, Hecksher-Soerensen J, Secher A, et al. (2014). Real-time trafficking and signaling of the glucagon-like peptide-1 receptor. Mol. Cell. Endocrinol 382, 938–949. [DOI] [PubMed] [Google Scholar]
  • 35.Widmann C, Dolci W, and Thorens B (1995). Agonist-induced internalization and recycling of the glucagon-like peptide-1 receptor in transfected fibroblasts and in insulinomas. Biochem. J 310, 203–214. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Kuna RS, Girada SB, Asalla S, Vallentyne J, Maddika S, Patterson JT, Smiley DL, DiMarchi RD, and Mitra P (2013). Glucagon-like peptide-1 receptor-mediated endosomal cAMP generation promotes glucose-stimulated insulin secretion in pancreatic beta-cells. Am. J. Physiol. Endocrinol. Metab 305, 161–170. [DOI] [PubMed] [Google Scholar]
  • 37.Ast J, Nasteska D, Fine NHF, Nieves DJ, Koszegi Z, Lanoiselee Y, Cuozzo F, Viloria K, Bacon A, Luu NT, et al. (2023). Revealing the tissue-level complexity of endogenous glucagon-like peptide-1 receptor expression and signaling. Nat. Commun 14, 301. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Cabrera O, Berman DM, Kenyon NS, Ricordi C, Berggren PO, and Caicedo A (2006). The unique cytoarchitecture of human pancreatic islets has implications for islet cell function. Proc. Natl. Acad. Sci. U. S. A 103, 2334–2339. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Bosco D, Armanet M, Morel P, Niclauss N, Sgroi A, Muller YD, Giovannoni L, Parnaud G, and Berney T (2010). Unique arrangement of alpha- and beta-cells in human islets of Langerhans. Diabetes 59, 1202–1210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Ast J, Broichhagen J, and Hodson DJ (2021). Reagents and models for detecting endogenous GLP1R and GIPR. EBioMedicine 74, 103739. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Bjornholm KD, Povlsen GK, Ougaard ME, Pyke C, Rakipovski G, Tveden-Nyborg P, Lykkesfeldt J, and Skovsted GF (2021). Decreased expression of the GLP-1 receptor after segmental artery injury in mice. J. Endocrinol 248, 289–301. [DOI] [PubMed] [Google Scholar]
  • 42.Pyke C, Heller RS, Kirk RK, Orskov C, Reedtz-Runge S, Kaastrup P, Hvelplund A, Bardram L, Calatayud D, and Knudsen LB (2014). GLP-1 receptor localization in monkey and human tissue: novel distribution revealed with extensively validated monoclonal antibody. Endocrinology 155, 1280–1290. [DOI] [PubMed] [Google Scholar]
  • 43.Gharib SA, Vemireddy R, Castillo JJ, Fountaine BS, Bammler TK, MacDonald JW, Hull-Meichle RL, and Zraika S (2024). Cystic fibrosis-related diabetes is associated with reduced islet protein expression of GLP-1 receptor and perturbation of cell-specific transcriptional programs. Sci. Rep 14, 25689. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Hodson DJ, Mitchell RK, Bellomo EA, Sun G, Vinet L, Meda P, Li D, Li WH, Bugliani M, Marchetti P, et al. (2013). Lipotoxicity disrupts incretin-regulated human beta cell connectivity. J. Clin. Invest 123, 4182–4194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Birke R, Ast J, Roosen DA, Lee J, Roßmann K, Huhn C, Mathes B, Lisurek M, Bushiri D, Sun H, et al. (2022). Sulfonated red and far-red rhodamines to visualize SNAP- and Halo-tagged cell surface proteins. Org. Biomol. Chem 20, 5967–5980. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Ast J, Novak AN, Podewin T, Fine NHF, Jones B, Tomas A, Birke R, Roßmann K, Mathes B, Eichhorst J, et al. (2022). Expanded LUXendin Color Palette for GLP1 R Detection and Visualization In Vitro and In Vivo. JACS Au 2, 1007–1017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Nyman LR, Ford E, Powers AC, and Piston DW (2010). Glucose-dependent blood flow dynamics in murine pancreatic islets in vivo. Am. J. Physiol. Endocrinol. Metab 298, E807–814. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Michau A, Hodson DJ, Fontanaud P, Guillou A, Espinosa-Carrasco G, Molino F, Peters CJ, Robinson IC, Le Tissier P, Mollard P, and Schaeffer M (2016). Metabolism Regulates Exposure of Pancreatic Islets to Circulating Molecules In Vivo. Diabetes 65, 463–475. [DOI] [PubMed] [Google Scholar]
  • 49.Dybala MP, Kuznetsov A, Motobu M, Hendren-Santiago BK, Philipson LH, Chervonsky AV, and Hara M (2020). Integrated Pancreatic Blood Flow: Bidirectional Microcirculation Between Endocrine and Exocrine Pancreas. Diabetes 69, 1439–1450. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Andersen DB, Grunddal KV, Pedersen J, Kuhre RE, Lund ML, Holst JJ, and Ørskov C (2021). Using a Reporter Mouse to Map Known and Novel Sites of GLP-1 Receptor Expression in Peripheral Tissues of Male Mice. Endocrinology 162, bqaa246. [DOI] [PubMed] [Google Scholar]
  • 51.Schifferer M, Yushchenko DA, Stein F, Bolbat A, and Schultz C (2017). A Ratiometric Sensor for Imaging Insulin Secretion in Single beta Cells. Cell Chem. Biol 24, 525–531 e524. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Peng X, Ren H, Yang L, Tong S, Zhou R, Long H, Wu Y, Wang L, Wu Y, Zhang Y, et al. (2024). Readily releasable β cells with tight Ca2+–exocytosis coupling dictate biphasic glucose-stimulated insulin secretion. Nat. Metab 6, 238–253. [DOI] [PubMed] [Google Scholar]
  • 53.Balboa D, Barsby T, Lithovius V, Saarimaki-Vire J, Omar-Hmeadi M, Dyachok O, Montaser H, Lund PE, Yang M, Ibrahim H, et al. (2022). Functional, metabolic and transcriptional maturation of human pancreatic islets derived from stem cells. Nat. Biotechnol 40, 1042–1055. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Moens K, Flamez D, Van Schravendijk C, Ling Z, Pipeleers D, and Schuit F (1998). Dual glucagon recognition by pancreatic beta-cells via glucagon and glucagon-like peptide 1 receptors. Diabetes 47, 66–72. [DOI] [PubMed] [Google Scholar]
  • 55.Johnston NR, Mitchell RK, Haythorne E, Pessoa MP, Semplici F, Ferrer J, Piemonti L, Marchetti P, Bugliani M, Bosco D, et al. (2016). Beta Cell Hubs Dictate Pancreatic Islet Responses to Glucose. Cell Metab 24, 389–401. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Segerstolpe Å, Palasantza A, Eliasson P, Andersson E-M, Andréasson A-C, Sun X, Picelli S, Sabirsh A, Clausen M, Bjursell MK, et al. (2016). Single-Cell Transcriptome Profiling of Human Pancreatic Islets in Health and Type 2 Diabetes. Cell Metab 24, 593–607. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Camunas-Soler J, Dai X-Q, Hang Y, Bautista A, Lyon J, Suzuki K, Kim SK, Quake SR, and MacDonald PE (2020). Patch-Seq Links Single-Cell Transcriptomes to Human Islet Dysfunction in Diabetes. Cell Metab 31, 1017–1031.e1014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58.Kravets V, Dwulet JM, Schleicher WE, Hodson DJ, Davis AM, Pyle L, Piscopio RA, Sticco-Ivins M, and Benninger RKP (2022). Functional architecture of pancreatic islets identifies a population of first responder cells that drive the first-phase calcium response. PLoS Biol 20, e3001761. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.DiGruccio MR, Mawla AM, Donaldson CJ, Noguchi GM, Vaughan J, Cowing-Zitron C, van der Meulen T, and Huising MO (2016). Comprehensive alpha, beta and delta cell transcriptomes reveal that ghrelin selectively activates delta cells and promotes somatostatin release from pancreatic islets. Mol. Metab 5, 449–458. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Adriaenssens AE, Svendsen B, Lam BYH, Yeo GSH, Holst JJ, Reimann F, and Gribble FM (2016). Transcriptomic profiling of pancreatic alpha, beta and delta cell populations identifies delta cells as a principal target for ghrelin in mouse islets. Diabetologia 59, 2156–2165. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 61.Kim A, Knudsen JG, Madara JC, Benrick A, Hill T, Abdul-Kadir L, Kellard JA, Mellander L, Miranda C, Lin H, et al. (2021). Arginine-vasopressin mediates counter-regulatory glucagon release and is diminished in type 1 diabetes. eLife 10, e72919. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62.Liu L, Dattaroy D, Simpson KF, Barella LF, Cui Y, Xiong Y, Jin J, König GM, Kostenis E, Roman JC, et al. (2021). α-cell Gq signaling is critical for maintaining euglycemia. JCI Insight 6, e152852. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 63.Miguel JC, Abdel-Wahab YHA, Green BD, Mathias PCF, and Flatt PR (2003). Cooperative enhancement of insulinotropic action of GLP-1 by acetylcholine uncovers paradoxical inhibitory effect of beta cell muscarinic receptor activation on adenylate cyclase activity. Biochem. Pharmacol 65, 283–292. [DOI] [PubMed] [Google Scholar]
  • 64.Ho K-H, Barmaver SN, Hu R, Yagan M, Ahmed HK, Kaverina I, and Gu G (2024). BioRxiv.
  • 65.Trogden KP, Lee J, Bracey KM, Ho K-H, McKinney H, Zhu X, Arpag G, Folland TG, Osipovich AB, Magnuson MA, et al. (2021). Microtubules regulate pancreatic β-cell heterogeneity via spatiotemporal control of insulin secretion hot spots. eLife 10, e59912. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 66.Salem V, Delgadillo Silva L, Suba K, Georgiadou E, Neda Mousavy S, Akhtar N, Martin-Alonso A, Gaboriau DCM, Rothery SM, Stylianides T, et al. (2019). Leader β cells coordinate Ca2+ dynamics across pancreatic islets in vivo. Nat. Metab 6, 615–629. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 67.Visa M, and Berggren P-O (2024). Sex-dependent intra-islet structural rearrangements affecting alpha-to-beta cell interactions lead to adaptive enhancements of Ca2+ dynamics in prediabetic beta cells. Diabetologia 67, 1663–1682. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 68.Balasenthilkumaran NV, Bombek LK, Ramirez D, Klemen MS, Leitgeb EP, Kerčmar J, Kopecky J, Zhang Y, Sethiya A, Takaoglu A, et al. (2025). bioRxiv.
  • 69.Caicedo A, Huising MO, and Wess J (2023). An Intraislet Paracrine Signaling Pathway That Enables Glucagon to Stimulate Pancreatic β-Cells. Diabetes 72, 1748–1750. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 70.Weir GC, and Bonner-Weir S (2023). Conflicting Views About Interactions Between Pancreatic α-Cells and β-Cells. Diabetes 72, 1741–1747. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 71.Rezania A, Bruin JE, Arora P, Rubin A, Batushansky I, Asadi A, O’Dwyer S, Quiskamp N, Mojibian M, Albrecht T, et al. (2014). Reversal of diabetes with insulin-producing cells derived in vitro from human pluripotent stem cells. Nat. Biotechnol 32, 1121–1133. [DOI] [PubMed] [Google Scholar]
  • 72.Gavet O, and Pines J (2010). Progressive activation of CyclinB1-Cdk1 coordinates entry to mitosis. Dev. Cell 18, 533–543. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 73.Viloria K, Nasteska D, Briant LJB, Heising S, Larner DP, Fine NHF, Ashford FB, da Silva Xavier G, Ramos MJ, Hasib A, et al. (2020). Vitamin-D-Binding Protein Contributes to the Maintenance of α Cell Function and Glucagon Secretion. Cell Rep 31, 107761. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 74.Farack L, Golan M, Egozi A, Dezorella N, Bahar Halpern K, Ben-Moshe S, Garzilli I, Toth B, Roitman L, Krizhanovsky V, and Itzkovitz S (2019). Transcriptional Heterogeneity of Beta Cells in the Intact Pancreas. Dev. Cell 48, 115–125 e114. [DOI] [PubMed] [Google Scholar]
  • 75.Ackermann AM, Zhang J, Heller A, Briker A, and Kaestner KH (2017). High-fidelity Glucagon-CreER mouse line generated by CRISPR-Cas9 assisted gene targeting. Mol. Metab 6, 236–244. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
2

Table S1: Human islet donor characteristics, related to Figure 1.

3

Table S2: smFISH Ins1 probe sequences, related to Figure 4.

4

Table S3: smFISH Glp1r probe sequences, related to Figure 4.

5

Data Availability Statement

  • Source data are provided with this paper (Data S1).

  • Individual microscopy images and localization counts used for analysis are provided in Mendeley Data (http://dx.doi.org/10.17632/bm9jsyg2f2.1).

  • All other data reported in this paper will be shared by the lead contact upon request.

  • This paper does not report original code.

RESOURCES