Abstract
Activation of the free fatty acid receptor FFAR4 in pancreatic islets potentiates insulin and inhibits somatostatin (SST) secretion. Here, we investigated the relative roles of δ and β cells in the insulinotropic effect of FFAR4 activation in mouse and human islets. The effects of the FFAR4 agonists Compound A (Cpd A) and AZ13581827 on insulin and SST secretion were investigated in islets from mice following ablation of δ cells and deletion of SST and the G protein Gαz (Gnaz−/−), in purified mouse β and δ cells, in mouse β cell-only pseudo-islets, and in human islets and EndoC-βH5 cells. Ca++ dynamics were measured in δ cells from Gnaz−/− mouse islets and β cells from dissociated mouse islets and human islets. The insulinotropic effect of FFAR4 activation was lost in δ cell-ablated and SST-deficient mouse islets, purified mouse β cells, and β cell-only pseudo-islets. FFAR4 activation did not increase β-cell Ca++ transients in dissociated mouse islets. Gαz deletion in mice partially blunted Cpd A inhibition of SST secretion but not the potentiation of insulin release. Cpd A diminished Ca++ transients in mouse δ cells, an effect that was partially lost in Gαz deficient islets. In human islets, Cpd A increased insulin secretion and intracellular Ca++ without affecting SST secretion. AZ13581827 potentiated insulin secretion in human EndoC-βH5 cells. We conclude that FFAR4 activation stimulates insulin secretion from mouse islets indirectly, in part via Gαz-coupled inhibition of SST secretion from δ cells, while in human islets, it directly stimulates insulin release from β cells.
Keywords: FFAR4, GPR120, insulin, somatostatin, Gαz, islets
G protein-coupled receptors (GPCRs) are important drug targets due to their role in many physiological processes and extensive involvement in human pathophysiology (1). GPCRs transduce signals by engaging heterotrimeric G proteins, composed of Gα, Gβ, and Gγ subunits, as well as arrestins (2). There are 4 major families of Gα subunits, namely, Gαs, Gαi/o, Gαq/11, and Gα12/13. The long-chain fatty acid (FA) receptor FFAR4 (previously referred to as GPR120) is a class A GPCR considered a potential therapeutic target for the treatment of type 2 diabetes, as agonists of this receptor exert numerous beneficial effects on glucose and energy homeostasis in preclinical models (3). FFAR4 activation alleviates chronic systemic inflammation and insulin resistance in obese mice (4, 5), inhibits white adipose tissue lipolysis (6), increases adipogenesis (7-9) and brown adipose tissue thermogenesis (10, 11), suppresses neuroinflammation (12), and regulates food intake (13, 14). In entero-endocrine cells, FFAR4 activation stimulates glucagon-like peptide-1 (15-17), glucose-dependent insulinotropic polypeptide (18, 19), and cholecystokinin (20) secretion, and inhibits ghrelin (21-23) and somatostatin (SST) (24) secretion. In pancreatic islets, FFAR4 activation prevents inflammation (25) and apoptosis (26); inhibits SST secretion (27-29); and promotes insulin (28-34), glucagon (28, 30, 35), and pancreatic polypeptide (36) secretion.
Pancreatic islets are endocrine organs predominantly composed of β, α, and δ cells that, respectively, secrete the major glucoregulatory hormones insulin and glucagon, and locally acting SST. Although glucose is the primary insulin secretagogue, an array of metabolic, hormonal, and neuronal signals modulate glucose-stimulated insulin secretion (GSIS) (37), either directly via GPCR on β cells or indirectly via intra-islet paracrine signaling (38). The importance of local cell-to-cell communication between islet endocrine cell types in the control of insulin levels is illustrated by its disruption in type 2 diabetes (38). Until recently, how islet FFAR4 activation promotes insulin and glucagon secretion was unknown. A possible indirect role was suggested by the demonstration of elevated Ffar4 expression in SST-secreting δ cells compared to other islet endocrine cells in mouse islets (27-29, 36, 39-41). In a previous study using δ-cell specific Ffar4 knockout mice, we revealed that the insulinotropic and glucagonotropic effects of FFAR4 activation are mediated by inhibitory signaling in δ cells (28). We showed that FFAR4 activation in mouse islets leads to a reduction in intracellular Ca++ signaling and forskolin-induced cyclic AMP levels in δ cells (28). Our previous results are consistent with a model whereby FFAR4 signaling inhibits SST secretion, which in turn alleviates the paracrine inhibitory action of SST and results in an increase in insulin and glucagon secretion. In contrast, Wu et al (30) and McCloskey et al (33) showed that FFAR4 agonists stimulate insulin secretion in rodent insulin-secreting cell lines, suggesting a direct effect at least in these β cell lines. Therefore, our field has not yet reached consensus regarding the precise site(s) of action of FFAR4 in islets. Our previous study established a role for SST in FFAR4 action in mouse islets by showing that the nonselective SST receptor antagonist cycloSST (cSST) partially blocked the insulinotropic and glucagonotropic effects of FFAR4 agonists (28). However, there remains the possibility of additional, direct mechanisms of action on α or β cells. Of note, a direct effect to stimulate insulin or glucagon secretion would imply coupling to a different G protein—in contrast to the inhibitory actions of FFAR4 on δ cells. To this point, FFAR4 signaling exhibits ligand-dependent differential coupling with Gαi/o, Gαs, Gαq, and β-arrestin in heterologous cells (42). In mice, FFAR4 stimulates pancreatic polypeptide secretion via coupling to Gαq (36) and inhibits gastric ghrelin secretion via coupling to Gαi/o (23). Surprisingly, inactivation of inhibitory Gαi/o proteins using pertussis toxin in mouse islets did not prevent FFAR4 inhibition of SST secretion (28), warranting further investigations into its coupling mechanisms. Finally, although Ffar4 is predominantly expressed in δ cells in mouse islets, RNA-sequencing studies in human islets identified significant FFAR4 expression in both β and δ cells (41, 43, 44) suggesting potential species-related differences that would be important to resolve as FFAR4 is considered a potential therapeutic target for metabolic diseases.
Given that our understanding of the mechanisms of action of FFAR4 in islets is still fragmented, this study aimed to determine the relative roles of δ and β cells in the insulinotropic effect of FFAR4 in mouse and human islets.
Materials and methods
Reagents and solutions
RPMI 1640 and fetal bovine serum (FBS) were obtained from Life Technologies Inc (Burlington, ON, Canada). Penicillin/streptomycin was from Multicell Wisent Inc (Saint-Jean Baptiste, QC, Canada). FA-free BSA was obtained from Equitech bio (Kerrville, TX). Compound A (Cpd A) was from Cayman Chemical (Ann Arbor, MI). AZ13581827 was generously provided by AstraZeneca (Gothenburg, Sweden). cSST was from Tocris Bioscience (Minneapolis, MN). Insulin radioimmunoassay (RIA) kits were obtained from MilliporeSigma (Cat# RI-13K, RRID:AB_2884035). SST RIA kits were obtained from Alpco Diagnotics (Cat# 38-RB306RUO, RRID:AB_3754169). Poly-D-Lysine was from Life Technologies Inc. (Burlington, ON, Canada). Prodo Islet Media and AB serum were from Prodo Laboratories (Aliso Viejo, CA). EndoC-βH5 cells (RRID:CVCL_C7TU), coating solution (βCOAT), media (ULTIβ1 and ULTI-ST), and KREBS buffer solution (βKREBS) for EndoC-βH5 cell culture were obtained from Human Cell Design (Toulouse, France). All other reagents and solutions were from MilliporeSigma unless otherwise specified.
Animals
Animals were handled in accordance with the Canadian Council on Animal Care guidelines for the care and use of laboratory animals. Mice were housed under controlled temperatures on a 12-h light/dark cycle with unrestricted access to water and standard laboratory chow. B6N.Cg-Ssttm2.1(cre)Zjh/J (Sst+/Cre) on a C57Bl/6N background were purchased from The Jackson Laboratory (RRID:IMSR_JAX:018973). Experimental animals (Sst+/+ and SstCre/Cre) were generated from crosses between Sst+/Cre mice. C57BL/6-Gt(ROSA)26Sortm1(HBEGF)Awai/J (ROSA26+/iDTR) mice on a mixed C57Bl/6JN background were purchased from Jackson laboratory (RRID:IMSR_JAX:007900) and backcrossed for 9 generations to C57Bl/6N mice. Experimental animals (Sst+/Cre;ROSA26+/iDTR) were generated from crosses between Sst+/Cre and ROSA26iDTR/iDTR mice. Diphtheria toxin (DT) (126 ng/200 μL 0.9% saline, List Biological Laboratories, Catalog # 150) or vehicle (0.9% saline) was injected intraperitoneally on days 0, 3, and 4. On day 5, pancreata were processed for immunochemistry or islets isolated for secretion assays. Gnaztm1Iah (RRID:MGI:3575891) (Gnaz+/−) on a C57Bl/6N background were kindly provided by Kimple et al (45). Experimental animals (Gnaz+/+, Gnaz+/−, and Gnaz−/−) were generated from crosses between Gnaz+/− mice. Constitutive GCaMP6s Ca++ indicator mice were generated by crossing heterozygous lox-stop-lox GCaMP6s mice (B6;129S6-Gt(ROSA)26Sortm96(CAG-GCaMP6s)Hze/J, RRID:IMSR_JAX:024106) with a heterozygous β-actin-Cre mouse (B6N.FVB-Tmem163TG(ACTB-Cre)2Mrt/CjDswJ, RRID:IMSR_JAX:019099), resulting in germline removal of the stop cassette as described previously (46). Genotyping primers are listed in Table S1 (47). Animals were born at the expected Mendelian ratios. C57Bl/6N male mice were purchased from Charles River Laboratories (Saint-Constant, QC, Canada).
Islet isolation and cell culture
Mouse islets were isolated at 10 to 12 weeks of age by collagenase digestion and dextran density gradient centrifugation as described (48) and recovered overnight in RPMI 1640 supplemented with 10% (v/v) FBS, penicillin/streptomycin (100 U/mL), and glucose (11 mM).
Upon reception, human islets were recovered overnight in Prodo Islet Media containing 5% (v/v) Prodo AB serum serum, 1% (v/v) glutamine/glutathione mixture, and penicillin/streptomycin (100 U/mL).
EndoC-βH5 cells were cultured as described (49). Frozen cells were thawed in a 37 °C water bath, washed with ULTIβ1 media, seeded on βCOAT coated 24-well plastic plates (175 000 cells/well), and cultured in ULTIβ1 media. On day 6, the media was replaced with ULTI-ST media for 24 hours prior to experimentation.
Flow cytometric sorting of mouse β and δ cells
Mouse β and δ cells were purified from intact islets by flow cytometry based on the expression of the cell surface markers CD24, CD49f, CD71, and CD81 as described (50) with minor modifications. Batches of 400 islets were dispersed by gentle pipetting for 10 minutes at 37 °C in Accutase (500 μL, Innovative Cell Technologies, Inc., San Diego, CA). Dispersed islet cells were incubated with antibodies diluted in Hank's balanced salt solution (HBSS) supplemented with 2% FBS (Staining medium) for 15 minutes at 4 °C. Antibodies and dilutions are indicated in Table S2 (47). Prior to flow cytometric sorting, cells were labeled with LIVE/DEAD Fixable Aqua vivid (405 nm) Dead Cell Stain (BD Biosciences, San Jose, CA) diluted in staining medium. Cell sorting was performed using a FACSAria III Cell Sorter (BD Biosciences, RRID:SCR_016695) and data analyzed using FlowJo 9.9.4 (RRID:SCR_008520). Cells sorted for quantitative (q)PCR were collected directly in RLT buffer (Qiagen, Hilden, Germany). Cells sorted for pseudo-islet generation were collected into RPMI 1640 supplemented with 50% (wt/vol) FBS, penicillin/streptomycin (100 U/mL), and glucose (11 mM).
Beta cell-only pseudo-islets
Sorted β cells (1500) from wild-type mice were placed in each well of a 96-well plate (Revvity, Waltham, MA) for aggregation. After 24 hours, half of the culture medium was replaced. Pseudo-islets were collected after 72 hours for secretion studies.
Static incubations for hormone secretion
Isolated islets
Mouse and human islets were incubated in Krebs–Ringer bicarbonate HEPES (KRBH) buffer (pH 7.4) supplemented with 0.1% (w/v) FA-free BSA (KRBH/BSA) and 2.8 mM glucose for 20 minutes. Then, triplicate batches of 20 islets each were dispersed into a 24-well plate and incubated for an additional 20 minutes in KRBH/BSA and 2.8 mM glucose, followed by a 1-hour static incubation in KRBH/BSA in the presence of 2.8 or 16.7 mM glucose and the synthetic FFAR4 agonists Cpd A (20 μM) and AZ13581827 (10 μM) or vehicle (ethanol [EtOH]), as indicated in the figure legends. Cpd A (5) and AZ13581827 (16) were chosen among the available FFAR4 agonists for their selectivity toward FFAR4.
Sorted β and δ cells
Sorted β (20 000/well) and δ (1000/well) cells were suspended in RPMI 1640 supplemented with 10% (wt/vol) FBS, penicillin/streptomycin (100 U/mL), and glucose (11 mM), plated in poly-D-lysine coated 24-well plates and recovered overnight. For insulin secretion, each well of sorted β cells was preincubated with KRBH/BSA and 2.8 mM glucose for 1 hour followed by a 1-hour static incubation in the presence of 2.8 or 16.7 mM glucose and Cpd A (20 μM) or vehicle (EtOH). For SST secretion, each well of sorted δ cells was preincubated with KRBH/BSA and 2.8 mM glucose for 1 hour followed by 1 hour-sequential static incubations in the presence of 2.8 mM glucose, 16.7 mM glucose, and finally 16.7 mM glucose and Cpd A (20 μM).
Beta cell-only pseudo-islets
For insulin secretion, each well containing 10 β cell-only pseudo-islets was preincubated with KRBH/BSA and 2.8 mM glucose for 1 hour followed by sequential 1-hour static incubations in the presence of 2.8 mM glucose, 16.7 mM glucose and vehicle (EtOH), and finally 16.7 mM glucose and Cpd A (20 μM) or AZ13581837 (10 μM). To account for possible β-cell exhaustion due to repeated glucose stimulation, in half of the experiments the order of the agonist and vehicle were reversed.
EndoC-βH5 cells
EndoC-βH5 insulin secretion assays were performed as described (49) with minor modifications. Cells were preincubated in βKREBS with 0.1% (w/v) FA-free BSA (βKREBS/BSA) and 2.8 mM glucose for 1 hour. Then, sequential 40-minute static incubations in βKREBS/BSA and 2.8 mM glucose followed by 16.7 mM glucose and Cpd A (20 μM) or AZ13581827 (10 μM) were performed as indicated in the figure legends.
Hormone measurements
Secreted insulin and SST were measured in the supernatant by RIA. Intracellular insulin content was measured after acid-alcohol extraction.
Perifusions
Human islet perifusions were performed as described (51) with minor modifications. Batches of 100 islets were placed in chambers and perifused in KRBH/BSA containing sequentially 2.8 mM glucose for 20 minutes, 16.7 mM glucose with Cpd A (20 μM), AZ13581827 (10 μM) or vehicle (EtOH) for 40 minutes, followed by 2.8 mM glucose for 20 minutes and finally KCl (30 mM) for 20 minutes. Samples were collected at 2 minute intervals and secreted insulin and SST measured by RIA. Intracellular insulin content was measured at the end of the perifusion after acid-alcohol extraction.
RNA extraction and quantitative RT-PCR
Total RNA was extracted from 200 islets, 60 000 sorted β, and 1000 sorted δ cells, using RNeasy micro kits (Qiagen). RNA was quantified by spectrophotometry using a NanoDrop 2000 (Life Technologies, RRID:SCR_018042) and 50 ng of RNA was reverse transcribed. Real-time PCR was performed using a QuantiTect SYBR Green PCR kit (Qiagen). Data were normalized to cyclophilin A (Ppia) mRNA levels. qPCR primer sequences are listed in Table S3 (47).
Immunohistochemistry of pancreatic sections
Pancreata were isolated, fixed for 4 hours in 4% paraformaldehyde, and cryoprotected overnight in 30% sucrose. The tissues were embedded in optimal cutting temperature (OCT) compound, frozen, sectioned at 8 μm, and mounted on Superfrost Plus slides (Thermo Fisher Scientific, Waltham, MA). Antigen retrieval was performed using sodium citrate buffer (pH = 6). Sections were stained for NKX6.1 and SST, and nuclei were stained with Hoechst 33342 (Thermo Fisher Scientific). Primary and secondary antibodies and dilutions are listed in Table S4 (47). Slides were mounted using Vectashield mounting medium (Vector Laboratories Inc, Newark, CA), and images acquired using a Zeiss AX10 fluorescence microscope (Oberkochen, Germany).
Ca++ imaging in mouse and human islets
Recombinant adenoviruses for Ca++ imaging Ad5-CMV-jRGECO1b and Ad5-CMV-GcaMP6s-P2A-NLS-mKate2 were generated using the AdEasy protocol (50). Isolated Gnaz+/+ and Gnaz−/− mouse and human islets (Table S5 (47)) were transduced with Ad5-CMV-jRGECO1b and Ad5-CMV-GcaMP6s-P2A-NLS-mKate2, respectively, at approximately 100 multiplicity of infection. For imaging, microfluidics chambers were bonded to 35-mm dishes with a glass bottom (MatTek, Ashland, MA). Islets were allowed to adhere to the glass during overnight culture. For dissociation, islets were handpicked into 15 mL conical vials, gravity sedimented and triturated in 500 μL 0.25% Trypsin, then suspended in RPMI. Pelleted (5 minutes, 500×g) and dissociated cells were seeded at 30 000 cells/imaging chamber. Beta cell identity after imaging was confirmed via post hoc immunofluorescence as described (46). Continuous perfusion of KRBH at a rate of 200 μL/min was maintained using the Elveflow microfluidics system, with different treatments adjusted using the Mux distributor. Ca++ responses were imaged using a Nikon ECLIPSE Ti2 inverted microscope (RRID:SCR_021068) using a 60× oil objective for mouse islet imaging or a 4× objective for human islet imaging. Regions of interest were drawn around individual mouse δ cells or whole human islets on NIS-Elements v.5.02.01 (RRID:SCR_014329). Fluorescence intensity in each region of interest was measured to determine sensor activity over time. In mouse islet data, individual cell intensity values were corrected using a rolling window of the baseline response over the 3 minutes before and after a time point to account for baseline changes in intensity associated with z-drift. This approach uses a rolling quantile detrending approach where the lowest 20% of values within the 3-minute window on either side of a time point determine the baseline. Using only the lowest 20% allows sharp Ca++ transients that reflect real biological events of different frequency at different phases of the Ca++ trace from skewing the baseline upwards while application of a rolling window isolates these sharp Ca++ transients from gradual changes in intensity that may reflect minor z-drift or photobleaching (52). Ca++ peaks were identified when Ca++ data exceeded 1 standard deviation of the rolling baseline at that time point. Peak identification was performed using code generated in Python v3.8 (code available at https://github.com/Huising-Lab).
Whole-mount staining of intact islets following live imaging was carried out as described previously (53). Primary and secondary antibodies and dilutions are listed in Table S6 (47). Whole-mount islets were imaged on a Nikon ECLIPSE Ti2 inverted microscope (RRID:SCR_021068) using a 60× lens with oil.
Statistical analyses
Data are expressed as mean ± SEM. Significance was tested using standard 1-way ANOVA, Brown-Forsythe and Welch ANOVA, and corrections, in cases of variance heterogeneity, 2-way ANOVA with post hoc adjustment for multiple comparisons, and paired and unpaired Student's t-test, as appropriate, using GraphPad (RRID:SCR_000306). Tukey or Dunnett post hoc tests were performed as indicated in figure legends. P < .05 was considered significant.
Study approval
All procedures involving animals were approved by the Institutional Committee for the Protection of Animal at the Centre Hospitalier de l’Université de Montréal. Islets from nondiabetic human donors were obtained from the National Institute of Diabetes and Digestive and Kidney Diseases-sponsored Integrated Islet Distribution Program (RRID:SCR_014387), NIH grant 2UC4DK098085; the Alberta Diabetes Institute Islet Core at the University of Alberta (Edmonton, AB, Canada), and Prodo Laboratories Inc. (Aliso Viejo, CA). The use of human islets was approved by the Institutional Ethics Committee of the Centre Hospitalier de l’Université de Montréal (protocol # MP-02-2019-7880).
Results
Delta-cell ablation prevents the potentiation of insulin secretion by Cpd A in mouse islets
Previously, by knocking out Ffar4 in SST-expressing cells, we demonstrated that FFAR4 signaling in δ cells is necessary for the insulinotropic effects of Cpd A in isolated mouse islets (28). As FFAR4 signaling in δ cells inhibits SST secretion (27, 28) and SST is a paracrine inhibitor of insulin secretion (54), our results were consistent with the possibility that FFAR4 agonism alleviates SST inhibition of insulin secretion. Yet, SST receptor blockade only partially reduced FFAR4 agonist-dependent potentiation of insulin secretion (28), leaving open the possibility of a role of FFAR4-dependent δ cell-derived signals other than SST. To test whether ablation of δ cells prevents the potentiation of insulin secretion by Cpd A, δ-cell deficient mice were generated by injecting DT into Sst+/Cre;ROSA26+/iDTR mice which express the DT receptor (DTR) in SST-expressing cells (53). DT injection of Sst+/Cre;ROSA26+/iDTR mice was previously shown to specifically deplete islet δ cells up to 3 months postinjection with no impact on α or β cell numbers (53). We confirmed the near-complete absence of SST staining in pancreatic sections of DT-injected-Sst+/Cre;ROSA26+/iDTR mice (Fig. S1 (47)). In 1-hour static incubations, high-glucose concentrations (16.7 mM) significantly increased insulin secretion in islets from both saline- and DT-injected-Sst+/Cre;ROSA26+/iDTR male mice compared to the low (2.8 mM) glucose condition (Fig. 1A). Glucose increased SST secretion in islets from saline-injected-Sst+/Cre;ROSA26+/iDTR mice (Fig. 1B). As expected, SST secretion from DT-injected-Sst+/Cre;ROSA26+/iDTR mouse islets was near the lower limit of detection of the assay (Fig. 1B). As shown previously (28), the FFAR4 agonist Cpd A (20 μM) significantly potentiated GSIS and inhibited glucose-stimulated SST secretion (GSSS) in control islets (Fig. 1A and 1B). In contrast, the effect of Cpd A on GSIS was no longer significant following δ-cell ablation (Fig. 1A). As shown previously (53), insulin content was not affected by δ-cell ablation (1484 ± 157 ng/mL vs 1506 ± 135 ng/mL for control; n = 5-6; not significant [NS]).
Figure 1.

Delta-cell ablation abolishes the potentiation of GSIS by Cpd A in mouse islets. Insulin secretion presented as a percentage of the 16.7 mM glucose condition (A) and SST secretion (B) measured in parallel in 1-hour static incubations in response to 2.8 or 16.7 mM glucose with Cpd A (20 μM) or vehicle (EtOH) in isolated islets from Sst+/Cre;ROSA26+/iDTR male mice injected with saline (control [CTL]) or DT. Data are expressed as mean ± SEM from 5 to 6 independent experiments. *P < .05, **P < .005, ***P < .0005, ****P < .0001 vs 16.7-EtOH condition using 2-way ANOVA with Tukey's post hoc adjustment for multiple comparisons. L.o.d., limit of detection.
Loss of SST expression prevents the potentiation of insulin secretion by Cpd A in mouse islets
To then specifically ascertain the role of SST in the action of Cpd A in islets, we took advantage of the absence of SST expression in islets from SstCre/Cre mice (53, 55). We confirmed that homozygous SstCre/Cre mice have no detectable SST protein expression in islets (Fig. S2 (47)). In 1-hour static incubations, high-glucose concentrations (16.7 mM) increased insulin secretion in islets from both Sst+/+ and SstCre/Cre male (Fig. 2A) and female (Fig. 2B) mice compared to the control 2.8 mM glucose condition. Glucose also increased SST secretion in islets from Sst+/+ male (Fig. 2C) and female (Fig. 2D) mice. As expected, SST secretion from SstCre/Cre mouse islets was near the lower limit of detection of the assay (Fig. 2C and 2D). The potentiation of GSIS by Cpd A was statistically significant in female (Fig. 2B) but not male (Fig. 2A) Sst+/+ mouse islets, likely due to a large variability of the data in these experiments. Nevertheless, Cpd A inhibited GSSS in both male (Fig. 2C) and female (Fig. 2D) Sst+/+ mouse islets. As in δ-cell-ablated islets (Fig. 1), loss of SST expression abrogated the stimulation of GSIS by Cpd A (Fig. 2A and 2B). Insulin content was not affected by SST deletion in either male (728 ± 108 ng/mL in SstCre/Cre vs 774 ± 79 ng/mL in Sst+/+; n = 9; NS) or female (1027 ± 109 ng/mL in SstCre/Cre vs 1249 ± 137 ng/mL in Sst+/+; n = 6; NS) islets.
Figure 2.

Sst inactivation prevents potentiation of GSIS in response to Cpd A in mouse islets. Insulin secretion presented as a percentage of the control 16.7 mM glucose condition (A and B) and SST secretion (C and D) measured in parallel in 1-hour static incubations in response to 2.8 or 16.7 mM glucose with Cpd A (20 μM) or vehicle (EtOH) in isolated male (A and C) and female (B and D) islets from Sst+/+ and SstCre/Cre mice. Data are expressed as mean ± SEM from 3 to 14 independent experiments. *P < .05, **P < .005, ***P < .0005 vs 16.7-EtOH condition using 2-way ANOVA with Tukey's post hoc adjustment for multiple comparisons. L.o.d., limit of detection.
Cpd A has no effect on insulin secretion in purified mouse β cells or β cell-only pseudo-islets
The data presented above suggest that the increase in insulin secretion in response to FFAR4 agonism in mouse islets is predominantly due to inhibition of SST secretion from δ cells. To further exclude a possible direct effect of Cpd A in mouse β cells, we purified live α, β, and δ cells from wild-type male mouse islets using a fluorescence activated cell sorting strategy adapted from Berthault et al (50) (Fig. S3 (47)). Compared to whole islets, fluorescence activated cell sorting-purified CD24lowCD81− cells were enriched by ∼2 fold in Ins1 and depleted by ∼50 fold in Gcg and ∼14 fold in Sst mRNA (Fig. 3A-3C). In contrast, CD24highCD49f− cells were depleted by ∼100 fold in Ins1 and ∼2.5 fold in Gcg, and enriched by ∼70 fold in Sst mRNA (Fig. 3A-3C). CD24lowCD81− and CD24highCD49f− cells are hereafter referred to as sorted β and δ cells, respectively. As expected, Cpd A significantly potentiated GSIS in intact islets (Fig. 3D). In contrast, sorted β cells were unresponsive to Cpd A (Fig. 3E). However, the interpretation of the observation of the absence of Cpd A effect in this setting is complicated by the well-known poor glucose responsiveness of dispersed β cells (56), also observed here (Fig. 3E). Therefore, we reconstituted pseudo-islets containing only β cells and exposed them sequentially for 1 hour to 2.8 or 16.7 mM glucose with or without Cpd A or the structurally dissimilar FFAR4 agonist AZ13581837 (Fig. 3F). Insulin secretion responded robustly to glucose in β cell-only pseudo-islets (Fig. 3F), albeit with a lower magnitude than intact islets. However, Cpd A had no effect on insulin secretion in β cell-only pseudo-islets, confirming the lack of functional FFAR4 signaling in mouse β cells. The inhibition of SST secretion in response to Cpd A observed in intact islets (Fig. 3G) was also observed in purified δ cells (Fig. 3H), confirming that FFAR4 is expressed and functional in mouse δ cells.
Figure 3.

Cpd A does not potentiate GSIS in dispersed or reaggregated mouse β cells. Ins1 (A), Gcg (B) and Sst (C) mRNA levels in intact male mouse islets and flow cytometry-sorted CD24lowCD81− (β), CD24lowCD81+/CD49f− (α), and CD24highCD49f− (δ) cells as measured by qPCR and normalized to cyclophilin. Data are presented as the fold change over intact islets. Insulin (D, E, and F) and SST (G and H) secretion presented as a percentage of the 16.7 mM glucose condition as assessed in isolated islets (D and G), sorted β cells (E), β cell-only pseudo-islets (F), and sorted δ cells (H) in 1-hour static incubations (D, E, and G) and sequential 1-hour static incubations (F and H) in response to 2.8 or 16.7 mM glucose with Cpd A (20 μM), AZ13581837 (10 μM), or vehicle (EtOH). Representative plots showing Ca++ traces of isolated mouse β cells and the average response of all β cells (I) under low glucose (5.5 mM) or stimulated by high glucose and amino acids with or without Cpd A (10 µM) or epinephrine (100 nM). Data are expressed as mean ± SEM from 3 to 5 (A–E, G, and H) or 7 to 15 (F) independent experiments and as mean for 316 dissociated β cell Ca++ traces (I). *P < .05, **P < .005, ***P < .0005 vs intact islets (A–C) or 16.7-EtOH condition (D–H) using 1-way ANOVA with Tukey's post hoc adjustment for multiple comparisons. L.o.d., limit of detection.
To further exclude a possible direct effect of Cpd A on mouse β cells, we dissociated islets from mice constitutively expressing the Ca++ reporter GCaMP6s into single cells to disrupt paracrine signaling (Fig. 3I). These experiments were performed at 11 mM glucose plus 2 mM each of alanine, arginine, and glutamine to ensure a robust and constant Ca++ response on top of which a change in Ca++ induced by Cpd A would be detectable. Beta cells were identified by exposure to epinephrin, which robustly inhibits Ca++. We did not observe an increase in Ca++ transients in response to Cpd A, contrary to its effects in intact islets (28). If anything, Cpd A exposure led to a momentary decrease in β-cell Ca++ fluxes. These results confirm a lack of direct stimulatory effect of Cpd A on mouse β-cell Ca++ signaling.
Cpd A inhibition of SST secretion is partially lost upon deletion of the inhibitory G protein Gαz in mouse islets
FFAR4 was shown to couple to the G proteins Gαs, Gαq, and Gαi/o in a ligand specific manner in heterologous cells (42). In our previous study, we showed that blocking Gαi/o signaling with pertussis toxin does not prevent Cpd A-mediated inhibition of GSSS in mouse islets (28). As Gαz mediates inhibitory signals in islets and is insensitive to pertussis toxin (57), we investigated FFAR4-Gαz coupling in 1-hour static incubations of male Gnaz+/+, Gnaz+/−, and Gnaz−/− mouse islets (45). Glucose (16.7 mM) increased SST secretion in all genotypes (Fig. 4A), with a comparable fold increase between Gnaz+/+ and Gnaz−/− mouse islets (Fig. 4B). In contrast to Gnaz+/+ and Gnaz+/− islets, Cpd A only partially suppressed GSSS in Gnaz−/− islets (Fig. 4A and 4C). High glucose increased insulin secretion in all genotypes (Fig. 4D), with a comparable fold increase between Gnaz+/+ and Gnaz−/− mouse islets (Fig. 4E). Cpd A potentiated GSIS in Gnaz+/+ and Gnaz+/− islets, as expected (Fig. 4D). Surprisingly, despite the diminished Cpd A-mediated repression of SST secretion, GSIS was still potentiated by Cpd A in Gnaz−/− islets (Fig. 4D) at a comparable level to Gnaz+/+islets (Fig. 4F). Islet insulin content was not different between the genotypes (Gnaz+/+, 998 ± 95 ng/mL; Gnaz+/−, 1011 ± 70 ng/mL; Gnaz−/−, 1112 ± 84 ng/mL; n = 7; NS). This suggests that the remaining ability of Cpd A to lower SST secretion in the absence of Gαz is sufficient to preserve the increase in insulin secretion.
Figure 4.

Gnaz deletion mitigates the inhibition of SST secretion by Cpd A in male mouse islets. SST secretion (A) and insulin secretion measured in parallel and presented as a percentage of the 16.7 M glucose condition (D) was assessed in 1-hour static incubations in response to 2.8 or 16.7 mM glucose with Cpd A (20 μM) or vehicle (EtOH) in isolated islets from Gnaz+/+, Gnaz+/−, and Gnaz−/− male mice. Fold change for SST (B and C) and insulin (E and F) secretion in response to glucose (16.7/2.8 mM glucose, B and E) and Cpd A (16.7 + Cpd A/16.7, C and F) has been calculated for Gnaz+/+ and Gnaz−/− mice. Data are expressed as mean ± SEM from 6 to 8 independent experiments. *P < .05, **P < .005, ***P < .0005, ****P < .0001 vs 16.7-EtOH condition following 2-way ANOVA with Tukey's post hoc adjustment for multiple comparisons. L.o.d., limit of detection.
We previously reported that Cpd A decreases δ-cell Ca++ signaling in mouse islets (28). To assess the role of Gαz, Gnaz+/+ and Gnaz−/− mouse islets were infected with an adenovirus expressing the Ca++ reporter jRGECO1b (Fig. 5A). As jRGECO1b was expressed in all islet cells, δ cells were identified by post hoc SST staining (Fig. 5A) and exposure to ghrelin, which stimulates SST secretion (39). These experiments were performed at 5.5 mM glucose to avoid the confounding effect of high β-cell Ca++ activity on the signal and to diminish the direct coupling from β to δ cells (52). Cpd A diminished Ca++ transients in Gnaz+/+ δ cells at 5.5 mM glucose, as expected (Fig. 5B, 5D, and 5F, and Fig. S4 [47]). In contrast, homozygous deletion of Gnaz diminished, but did not eliminate, the inhibitory effect of Cpd A on δ cell Ca++ transients (Fig. 5C, 5E, and 5F, and Fig. S4 (47)). Ghrelin increased Ca++ in both genotypes, confirming the identity of the cells as δ cells (Fig. 5B-5E and Fig. S4 [47]).
Figure 5.

Gnaz deletion diminishes the Cpd A-mediated decrease in intracellular Ca++ transients in male mouse δ cells. (A) Gnaz+/+ and Gnaz−/− islets were transduced with an adenovirus delivering the Ca++ sensor jRGECO1b. The δ cells were identified by post hoc immunofluorescent SST staining to mark δ cells (yellow arrow in A). (B and C) Representative plots showing Ca++ traces for each of 9 to 10 δ cells and the average response of all δ cells in Gnaz+/+ (B) and Gnaz−/− (C) islets. Islets were exposed to 5.5 mM glucose and Cpd A (10 µM) and Ghrelin (100 nM) added at the times indicated. (D and E) Representative plots from the analysis of Ca++ transients in both Gnaz+/+ (D) and Gnaz−/− (E) δ cells performed by identifying peaks (green dots) greater than 1 standard deviation away from a rolling baseline (red line) to account for cell drift. (F) The number of peaks per minute were quantified in Gnaz+/+ and Gnaz−/− δ cells in basal glucose (5.5 mM) and in the presence or absence of Cpd A. Stimulation with 100 nM ghrelin was used to confirm δ cell activity. Data are expressed as mean ± SEM from 9 to 25 δ cell from each of 2 independent experiments. **P < .01, ***P < .001 using 2-way ANOVA with Tukey's post hoc adjustment for multiple comparisons.
Taken together, these data indicate that FFAR4 couples partially to Gαz in δ cells to diminish Ca++ transients and SST secretion in response to Cpd A, but suggest that Gαz may not be solely responsible for Cpd A-mediated potentiation of GSIS.
FFAR4 agonists stimulate Ca++ transients and insulin secretion but do not impact SST secretion in human islets
To investigate the role of FFAR4 in human islets, we first assessed islet cell-type-specific FFAR4 expression from single-cell RNA-sequencing datasets (44). Interestingly, whereas Ffar4 is predominantly expressed in δ cells in mouse islets (28), significant FFAR4 expression was also detected in human β cells (Fig. 6A and 6B).
Figure 6.

Cpd A stimulates intracellular Ca++ transients in human islets. (A) Uniform manifold approximation and projection (UMAP) plot of 5 integrated human islet single-cell RNA-Seq studies as described in Ref. (44) showing FFAR4 expressing cells. (B) Violin plots showing FFAR4 expression level of individual α, β, and δ cells. Adapted from Ref. (44). (C) Human islets were transduced with an adenovirus delivering the Ca++ sensor GCaMP6s and exposed to 2.8 mM glucose followed by 11 mM glucose. Cpd A (10 μM) and cSST (250 nM) were added at the times indicated. Islets were imaged with a 4× wide field lens and the entire response of each islet used for quantification. (D) Representative plot showing Ca++ traces for each of 10 islets and the average signal. Arbitrary units. (E and F) Average intensity 10 minutes before (Baseline) and after treatment with Cpd A (E) and cSST (F). Data are expressed as mean ± SEM for 6 to 10 islets from each of 3 technical replicates from 1 donor (HP-24284-01). ****P < .0001 using a paired Student's t-test.
To investigate whether FFAR4 is active in human β cells, human islets were transduced with an adenovirus expressing the Ca++ reporter GcaMP6s (Fig. 6C). The AdEasy system was utilized to generate human adenovirus serotype 5 (Ad5) particles that were then transduced into human islets, as reported in Ref. (58). This virus and serotype exhibit selective tropism for mouse β cells, resulting in over 97% of transduced islet cells being insulin-positive (59). Both mouse and human β cells are enriched for the Coxsackie and adenovirus receptor (60), which Ad5 particles will interact with to deliver their payload. We decided to capture all β cell behavior from a finite resource by imaging human islets at the whole islet level because of the tropism of Ad5 particles for β cells. Increasing the glucose concentration from 2.8 to 11 mM stimulated Ca++ signaling and subsequent addition of Cpd A led to a further increase (Fig. 6D and 6E). Following agonist washout, the subsequent addition of the competitive SST receptor antagonist cSST increased Ca++ fluxes suggesting that δ cell-derived SST inhibits Ca++ signaling in β cells (Fig. 6D and 6F).
Perifusions of isolated human islets showed an increase in GSIS in response to Cpd A (Fig. 7A and 7B) and a variable response to AZ13581837 (Fig. 7A and 7C). In 1-hour static incubations, the effects of Cpd A and AZ13581837 were modest (Fig. 7D and 7E), significant only for Cpd A in female islets. In contrast to mouse islets, the increase in SST secretion in response to glucose was not inhibited by either Cpd A or AZ13581837 (Fig. 7F and 7G). Given the modest effect of the FFAR4 agonists on insulin secretion in static incubations (Fig. 7D and 7E) of isolated human islets, we performed similar experiments in human insulin-secreting EndoC-βH5 cells (49) (Fig. 7H). At the concentration tested, the effect of Cpd A did not reach statistical significance, but AZ13581837 significantly potentiated GSIS in EndoC-βH5 cells, confirming that FFAR4 signaling is active in human β cells and stimulates insulin secretion.
Figure 7.

FFAR4 agonists potentiate GSIS but do not inhibit SST secretion in human islets. (A) Insulin secretion (fold change vs baseline) from isolated human islets as assessed in perifusion in response to 16.7 mM glucose with or without Cpd A (20 μM) or AZ13581837 (10 μM). (B and C) Area under the curve from 10 to 50 minutes for insulin in response to Cpd A (B) or AZ13581837 (C). Male (black); female (orange). (D–G) Insulin (D and E) and SST (F and G) secretion from male (D and F) and female (E and G) human islets measured in parallel 1-hour static incubations in response to 2.8 or 16.7 mM glucose with or without Cpd A (20 μM), AZ13581837 (10 μM), or vehicle (EtOH). Insulin secretion is presented as the Stimulation Index (insulin at 16.7 mM glucose +/− agonist/2.8 mM glucose condition). The dashed line represented the 2.8 mM of glucose condition. Each human islet donor is represented with a different color. (H) Insulin secretion presented as a percentage of the 16.7 mM glucose condition as assessed in 1-hour static incubations in response to 2.8 or 16.7 mM glucose with Cpd A (20 μM), AZ13581837 (10 μM), or vehicle (EtOH) in EndoC-βH5 cells. Data are expressed as mean ± SEM from 4 to 5 independent experiments. *P < .05, **P < .005 vs 16.7-EtOH condition following 1-way ANOVA with Tukey's post hoc adjustment for multiple comparisons. L.o.d., limit of detection.
Overall, experiments in human islets and EndoC-βH5 cells lead us to conclude that in contrast to mouse islets, FFAR4 activation in human islets stimulates Ca++ fluxes and insulin secretion in the absence of detectable changes in δ-cell SST secretion.
Discussion
This study was designed to determine the relative roles of δ and β cells in the insulinotropic effect of FFAR4 activation in mouse and human islets. Our results show that (1) in the absence of δ cells or SST, FFAR4 agonists no longer potentiate GSIS in mouse islets; (2) FFAR4 agonists have no effect on insulin secretion from purified mouse β cells or mouse β cell-only pseudo-islets or Ca++ dynamics in dissociated mouse islets; (3) FFAR4 acts via partial coupling to Gαz in mouse δ cells; (4) in human islets, FFAR4 agonist stimulation of β cell Ca++ dynamics and insulin secretion is observed in the absence of detectable changes in SST secretion. (5) FFAR4 signaling stimulates insulin secretion from human EndoC-βH5 cells. Overall, our interpretation of these results is that FFAR4 stimulation of insulin secretion relies strictly on indirect signaling mediated via the attenuation of δ cell SST in mouse islets, whereas in human islets, a direct and stimulatory β-cell effect predominates.
The absence of a potentiating effect of Cpd A on GSIS when δ cells or Sst were removed from mouse islets indicate that the effect of FFAR4 agonism is largely indirect. Together with our earlier studies (28), these data using complementary models indicate that the insulinotropic effect of Cpd A is mediated by FFAR4-dependent inhibition of δ cell SST secretion. Of note, in our previous study, cSST only partially blocked the effects of Cpd A on insulin secretion in mouse islets (28). In contrast, in the current study, genetic deletion of Sst in SstCre/Cre mice completely abrogated the stimulation of insulin secretion by Cpd A. The most likely explanation for this discrepancy is that inhibition of SST signaling by cSST was only partial in our previous study.
Consistent with a lack of functional signaling of FFAR4 in primary murine β cells, Cpd A did not potentiate GSIS in dispersed or reaggregated mouse β cells. Yet, Wu et al (30) and McCloskey et al (33) demonstrated potentiation of GSIS in murine MIN6 cells in response to the FFAR4 agonists AZ13581837 and TUG891 and in rat BRIN-BD11 insulinoma cells in response to the FFAR4 agonists Cpd A and GSK137647, respectively. Although, Ffar4 expression is detectable in primary murine β cells (27-29, 36, 39, 40), one possibility is that differences in the level of FFAR4 protein between transformed insulin-secreting cell lines and primary β cells could account for these discrepancies. In addition, given that FFAR4 exhibits ligand-dependent conformational bias resulting in biased signaling (61), we cannot exclude that a minor, direct role of FFAR4 in rodent β cells in the control of insulin secretion may be detected in response to other ligands.
The relatively modest effect of SST removal—either by δ-cell ablation or genetic deletion of SST in SstCre/Cre mice on GSIS, is surprising but consistent with our previous observation in these models (53). It can be explained by the fact that these experiments were performed in response to glucose alone, without amino acids. SST is not a very strong inhibitor of β cell Ca++ influx induced by high concentrations of glucose because mouse β cells express only ciliary SSTR3 receptors, thus preventing SST from modulating cytoplasmic Ca++ levels (46). In the presence of amino acids, α cells are activated upon SST removal and contribute to β cell insulin secretion.
Previously, we found that pertussis toxin pretreatment does not prevent Cpd A-mediated repression of GSSS in mouse islets (28), suggesting that the inhibitory G proteins Gαi/o are not involved. We therefore considered the alternative possibility that the pertussis toxin-insensitive inhibitory G protein Gαz, which is expressed in mouse islets and mediates the inhibition of GSIS by prostaglandin E via the EP3 receptor (62), might be implicated. Such a possibility is also consistent with our previous observation that Cpd A reduces forskolin-induced cyclic AMP levels in mouse δ cells (28), in line with the known inhibition of adenylate cyclase activity by Gαz (62). By analyzing Gnaz−/− mice, we showed that Gαz partially mediates FFAR4-dependent inhibition of Ca++ signaling and SST secretion in δ cells. In contrast, Stone et al (27) provided evidence that repression of GSSS by the FFAR4 agonist Metabolex 36 is pertussis toxin-sensitive in isolated mouse islets. Both Gαz and Gαi/o are expressed in δ cells (39), and mouse FFAR4 has been shown previously to couple to Gαi/o (29). These discrepant results could again be explained by biased signaling (61), whereby Cpd A favors FFAR4-Gαz coupling whereas Metabolex 36 favors FFAR4-Gαi/o coupling in mouse δ cells. On the other hand, in heterologous cells, Cpd A promotes FFAR4-Gαq coupling (5) and in BRIN-BD11 cells, Cpd A promotes GSIS (33), suggesting possible context-dependent coupling with stimulatory G proteins (eg, Gαq or Gαs). Why Cpd A favors inhibitory FFAR4-Gαz coupling in δ cells where Gαq is also expressed (39) may be due to cell-type-specific factors, other than differential G protein expression, including interference from other δ-cell GPCR that modulate the availability of G protein isoforms at FFAR4, as suggested from GPCR competition studies (63).
Unexpectedly, we found that Cpd A still potentiates GSIS despite the reduced Cpd A-mediated repression of SST secretion in Gnaz−/− mouse islets. One possible explanation for this discrepancy is that Cpd A was still able to diminish SST secretion, albeit not statistically significantly, as well as Ca++ influx in δ cells in the absence of Gαz, suggesting additional coupling to other Gα isoforms. These data could be also explained by competitive coupling of FFAR4 to Gαz or Gαq in β cells whereby in wild-type β cells, neither inhibitory Gαz nor stimulatory Gαq signaling predominates in response to Cpd A leading to a net null effect on insulin secretion. In such a scenario, Gnaz−/− β cells Cpd A signals exclusively via FFAR4-Gαq coupling to potentiate GSIS. Further studies are required to test this possibility.
Previous studies demonstrated that the FFAR4 agonists AZ13581827 and TUG891 potentiate GSIS in human islets and pseudo-islets in static incubations (30). Similarly, we found that the FFAR4 agonist Cpd A stimulates Ca++ fluxes and insulin secretion in human islets; however, unlike in mouse islets, this occurred in the absence of detectable, agonist-regulated SST secretion. Although to our knowledge, ours is the first study addressing a role of FFAR4 agonists in SST secretion in human islets, Du et al (29) found that the FFAR4 inverse agonist AH7614 increases GSSS in human islets, suggesting that FFAR4 signaling, possibly due to endogenous islet-derived FFAR4 ligands, inhibits SST secretion. In light of these data and given the limitations of the static and perifusion assay to accurately detect small changes in SST levels, we cannot exclude a contribution of FFAR4-dependent regulation of SST secretion in the insulinotropic response to FFAR4 agonists in human islets. Nevertheless, our observation that FFAR4 agonists potentiate GSIS in EndoC-βH5 cells are also in line with a predominantly direct action on human β cells. We acknowledge that EndoC-βH5 cells do not faithfully replicate primary human β cells. We have attempted to purify human β cells from isolated islets using a monoclonal antibody against nucleoside triphosphate diphosphoyhydrolase-3 as described by Saunders et al (64). Unfortunately, in our hands, the purification protocol did not completely eliminate δ cells, as also shown by Saunders et al, due to nucleoside triphosphate diphosphoyhydrolase-3 expression in a small number of δ cells. We therefore elected to use EndoC-βH5 cells despite the inherent limitations of transformed cell lines, including the reported low levels of expression of SST in these cells (49).
Although FFAR4 exhibits promiscuous coupling with Gαi/o, Gαs, Gαq and β-arrestin in heterologous cells (42), the direct effect of Cpd A on GSIS suggests a coupling to stimulatory G proteins Gαs and/or Gαq in this context. Further studies will be required to establish the coupling of FFAR4 in human β cells.
In conclusion, our data suggest that FFAR4 controls insulin secretion by distinct regulatory mechanisms in rodent and human islets. FFAR4 activation stimulates insulin secretion from mouse islets indirectly via Gαz-coupled inhibition of δ-cell SST secretion, while in human islets, it stimulates insulin release via a direct effect on β cells. These species-related differences are important as FFAR4 is considered a potential therapeutic target for metabolic diseases.
Acknowledgments
We thank Maria Sörhede Winzell and Linda Sundström from Astra Zeneca for providing AZ13581837; Mélanie Éthier, Sabrina Sedano-Benitez, and Gaël Dulude from the CRCHUM for valuable technical assistance; and Raphaël Scharfmann and Alexis Fouque for sharing unpublished cell sorting protocols.
Contributor Information
Laura Reininger, University of Montreal Hospital Research Center (CRCHUM), Montreal, QC H2X 0A9, Canada.
Muhammad Rehman, Department of Neurobiology, Physiology & Behavior, College of Biological Sciences, University of California Davis, Davis, CA 95616, USA; Department of Physiology and Membrane Biology, School of Medicine, University of California Davis, Davis, CA 95616, USA.
Ryan Hart, Department of Neurobiology, Physiology & Behavior, College of Biological Sciences, University of California Davis, Davis, CA 95616, USA; Department of Physiology and Membrane Biology, School of Medicine, University of California Davis, Davis, CA 95616, USA.
Sarah Ferragne, University of Montreal Hospital Research Center (CRCHUM), Montreal, QC H2X 0A9, Canada.
Caroline Tremblay, University of Montreal Hospital Research Center (CRCHUM), Montreal, QC H2X 0A9, Canada.
Mélanie Ethier, University of Montreal Hospital Research Center (CRCHUM), Montreal, QC H2X 0A9, Canada.
Amélia Bouabcha, University of Montreal Hospital Research Center (CRCHUM), Montreal, QC H2X 0A9, Canada.
Michelle Kimple, University of Wisconsin School of Medicine and Public Health, Madison, WI 53705, USA; Research Service, William S. Middleton Memorial Veterans Hospital, Madison, WI 53705, USA.
Julien Ghislain, University of Montreal Hospital Research Center (CRCHUM), Montreal, QC H2X 0A9, Canada.
Mark O Huising, Department of Neurobiology, Physiology & Behavior, College of Biological Sciences, University of California Davis, Davis, CA 95616, USA; Department of Physiology and Membrane Biology, School of Medicine, University of California Davis, Davis, CA 95616, USA.
Vincent Poitout, University of Montreal Hospital Research Center (CRCHUM), Montreal, QC H2X 0A9, Canada; Department of Medicine, University of Montreal, Montreal, QC H3T 1J4, Canada.
Funding
This study was supported by the U.S. National Institutes of Health (R01-DK-132597 to M.O.H. and V.P. and R01 DK137505 to M.K.); U.S. Department of Veterans Affairs (I01 BX005804 and IK6 BX006804 to M.K.); and a Discovery Grant from the Natural Sciences and Engineering Research Council of Canada (RGPIN-2016-03952 to V.P.). L.R. was supported by a fellowship from the Société Francophone du Diabète and a travel award from the Quebec Cardiometabolic Health, Diabetes and Obesity Research Network. Human pancreatic islets were provided by the NIDDK-funded Integrated Islet Distribution Program (RRID:SCR_014387), National Institutes of Health Grant U24 DK098085.
Disclosures
The authors have no relevant conflict to disclose.
Data availability
Original data generated and analyzed during this study are included in this published article or in the data repository listed in Ref. (47).
References
- 1. Hauser AS, Attwood MM, Rask-Andersen M, Schiöth HB, Gloriam DE. Trends in GPCR drug discovery: new agents, targets and indications. Nat Rev Drug Discov. 2017;16(12):829‐842. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Hilger D, Masureel M, Kobilka BK. Structure and dynamics of GPCR signaling complexes. Nat Struct Mol Biol. 2018;25(1):4‐12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Ghislain J, Poitout V. Targeting lipid GPCRs to treat type 2 diabetes mellitus—progress and challenges. Nat Rev Endocrinol. 2021;17(3):162‐175. [DOI] [PubMed] [Google Scholar]
- 4. Oh DY, Talukdar S, Bae EJ, et al. GPR120 is an omega-3 fatty acid receptor mediating potent anti-inflammatory and insulin-sensitizing effects. Cell. 2010;142(5):687‐698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Oh da Y, Walenta E, Akiyama TE, et al. A Gpr120-selective agonist improves insulin resistance and chronic inflammation in obese mice. Nat Med. 2014;20(8):942‐947. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Husted AS, Ekberg JH, Tripp E, et al. Autocrine negative feedback regulation of lipolysis through sensing of NEFAs by FFAR4/GPR120 in WAT. Mol Metab. 2020;42:101103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Gotoh C, Hong YH, Iga T, et al. The regulation of adipogenesis through GPR120. Biochem Biophys Res Commun. 2007;354(2):591‐597. [DOI] [PubMed] [Google Scholar]
- 8. Song T, Zhou Y, Peng J, et al. GPR120 promotes adipogenesis through intracellular calcium and extracellular signal-regulated kinase 1/2 signal pathway. Mol Cell Endocrinol. 2016;434:1‐13. [DOI] [PubMed] [Google Scholar]
- 9. Hilgendorf KI, Johnson CT, Mezger A, et al. Omega-3 fatty acids activate ciliary FFAR4 to control adipogenesis. Cell. 2019;179(6):1289‐1305.e21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Quesada-Lopez T, Cereijo R, Turatsinze JV, et al. The lipid sensor GPR120 promotes brown fat activation and FGF21 release from adipocytes. Nat Commun. 2016;7(1):13479. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Schilperoort M, van Dam AD, Hoeke G, et al. The GPR120 agonist TUG-891 promotes metabolic health by stimulating mitochondrial respiration in brown fat. EMBO Mol Med. 2018;10(3):e8047. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Nakajima S, Demers G, Machuca-Parra AI, et al. Central activation of the fatty acid sensor GPR120 suppresses microglia reactivity and alleviates sickness- and anxiety-like behaviors. J Neuroinflammation. 2023;20(1):302. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Auguste S, Fisette A, Fernandes MF, et al. Central agonism of GPR120 acutely inhibits food intake and food reward and chronically suppresses anxiety-like behavior in mice. Int J Neuropsychopharmacol. 2016;19(7):pyw014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Velasco C, Conde-Sieira M, Comesana S, et al. The long-chain fatty acid receptors FFA1 and FFA4 are involved in food intake regulation in fish brain. J Exp Biol. 2020;223(Pt 17):jeb227330. [DOI] [PubMed] [Google Scholar]
- 15. Hirasawa A, Tsumaya K, Awaji T, et al. Free fatty acids regulate gut incretin glucagon-like peptide-1 secretion through GPR120. Nat Med. 2005;11(1):90‐94. [DOI] [PubMed] [Google Scholar]
- 16. Sundstrom L, Myhre S, Sundqvist M, et al. The acute glucose lowering effect of specific GPR120 activation in mice is mainly driven by glucagon-like peptide 1. PLoS One. 2017;12(12):e0189060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Owolabi AI, Corbett RC, Flatt PR, McKillop AM. Positive interplay between FFAR4/GPR120, DPP-IV inhibition and GLP-1 in beta cell proliferation and glucose homeostasis in obese high fat fed mice. Peptides. 2024;177:171218. [DOI] [PubMed] [Google Scholar]
- 18. Iwasaki K, Harada N, Sasaki K, et al. Free fatty acid receptor GPR120 is highly expressed in enteroendocrine K cells of the upper small intestine and has a critical role in GIP secretion after fat ingestion. Endocrinology. 2015;156(3):837‐846. [DOI] [PubMed] [Google Scholar]
- 19. Sankoda A, Harada N, Iwasaki K, et al. Long-chain free fatty acid receptor GPR120 mediates oil-induced GIP secretion through CCK in male mice. Endocrinology. 2017;158(5):1172‐1180. [DOI] [PubMed] [Google Scholar]
- 20. Tanaka T, Katsuma S, Adachi T, Koshimizu TA, Hirasawa A, Tsujimoto G. Free fatty acids induce cholecystokinin secretion through GPR120. Naunyn Schmiedebergs Arch Pharmacol. 2008;377(4-6):523‐527. [DOI] [PubMed] [Google Scholar]
- 21. Gong Z, Yoshimura M, Aizawa S, et al. G protein-coupled receptor 120 signaling regulates ghrelin secretion in vivo and in vitro. Am J Physiol Endocrinol Metab. 2014;306(1):E28‐E35. [DOI] [PubMed] [Google Scholar]
- 22. Lu X, Zhao X, Feng J, et al. Postprandial inhibition of gastric ghrelin secretion by long-chain fatty acid through GPR120 in isolated gastric ghrelin cells and mice. Am J Physiol Gastrointest Liver Physiol. 2012;303(3):G367‐G376. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Engelstoft MS, Park WM, Sakata I, et al. Seven transmembrane G protein-coupled receptor repertoire of gastric ghrelin cells. Mol Metab. 2013;2(4):376‐392. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Egerod KL, Engelstoft MS, Lund ML, et al. Transcriptional and functional characterization of the G protein-coupled receptor repertoire of gastric somatostatin cells. Endocrinology. 2015;156(11):3909‐3923. [DOI] [PubMed] [Google Scholar]
- 25. Wang Y, Xie T, Zhang D, Leung PS. GPR120 protects lipotoxicity-induced pancreatic β-cell dysfunction through regulation of PDX1 expression and inhibition of islet inflammation. Clin Sci. 2019;133(1):101‐116. [DOI] [PubMed] [Google Scholar]
- 26. Taneera J, Lang S, Sharma A, et al. A systems genetics approach identifies genes and pathways for type 2 diabetes in human islets. Cell Metab. 2012;16(1):122‐134. [DOI] [PubMed] [Google Scholar]
- 27. Stone VM, Dhayal S, Brocklehurst KJ, et al. GPR120 (FFAR4) is preferentially expressed in pancreatic delta cells and regulates somatostatin secretion from murine islets of Langerhans. Diabetologia. 2014;57(6):1182‐1191. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Croze ML, Flisher MF, Guillaume A, et al. Free fatty acid receptor 4 inhibitory signaling in delta cells regulates islet hormone secretion in mice. Mol Metab. 2021;45:101166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Du YQ, Sha XY, Cheng J, et al. Endogenous lipid-GPR120 signaling modulates pancreatic islet homeostasis to different extents. Diabetes. 2022;71(7):1454‐1471. [DOI] [PubMed] [Google Scholar]
- 30. Wu CT, Hilgendorf KI, Bevacqua RJ, et al. Discovery of ciliary G protein-coupled receptors regulating pancreatic islet insulin and glucagon secretion. Genes Dev. 2021;35(17-18):1243‐1255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Moran BM, Abdel-Wahab YH, Flatt PR, McKillop AM. Evaluation of the insulin-releasing and glucose-lowering effects of GPR120 activation in pancreatic beta-cells. Diabetes Obes Metab. 2014;16(11):1128‐1139. [DOI] [PubMed] [Google Scholar]
- 32. Zhang D, So WY, Wang Y, Wu SY, Cheng Q, Leung PS. Insulinotropic effects of GPR120 agonists are altered in obese diabetic and obese non-diabetic states. Clin Sci. 2017;131(3):247‐260. [DOI] [PubMed] [Google Scholar]
- 33. McCloskey AG, Miskelly MG, Flatt PR, McKillop AM. Pharmacological potential of novel agonists for FFAR4 on islet and enteroendocrine cell function and glucose homeostasis. Eur J Pharm Sci. 2020;142:105104. [DOI] [PubMed] [Google Scholar]
- 34. Croze ML, Guillaume A, Ethier M, et al. Combined deletion of free fatty-acid receptors 1 and 4 minimally impacts glucose homeostasis in mice. Endocrinology. 2021;162(3):bqab002. [DOI] [PubMed] [Google Scholar]
- 35. Suckow AT, Polidori D, Yan W, et al. Alteration of the glucagon axis in GPR120 (FFAR4) knockout mice: a role for GPR120 in glucagon secretion. J Biol Chem. 2014;289(22):15751‐15763. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Zhao YF, Li XC, Liang XY, et al. GPR120 regulates pancreatic polypeptide secretion from male mouse islets via PLC-mediated calcium mobilization. Endocrinology. 2020;161(10):bqaa157. [DOI] [PubMed] [Google Scholar]
- 37. Ahren B. Islet G protein-coupled receptors as potential targets for treatment of type 2 diabetes. Nat Rev Drug Discov. 2009;8(5):369‐385. [DOI] [PubMed] [Google Scholar]
- 38. Huising MO. Paracrine regulation of insulin secretion. Diabetologia. 2020;63(10):2057‐2063. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. DiGruccio MR, Mawla AM, Donaldson CJ, et al. Comprehensive alpha, beta and delta cell transcriptomes reveal that ghrelin selectively activates delta cells and promotes somatostatin release from pancreatic islets. Mol Metab. 2016;5(7):449‐458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Adriaenssens AE, Svendsen B, Lam BY, et al. Transcriptomic profiling of pancreatic alpha, beta and delta cell populations identifies delta cells as a principal target for ghrelin in mouse islets. Diabetologia. 2016;59(10):2156‐2165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Baron M, Veres A, Wolock SL, et al. A single-cell transcriptomic map of the human and mouse pancreas reveals inter- and intra-cell population structure. Cell Syst. 2016;3(4):346‐360.e344. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Mao C, Xiao P, Tao XN, et al. Unsaturated bond recognition leads to biased signal in a fatty acid receptor. Science (New York, NY). 2023;380(6640):eadd6220. [DOI] [PubMed] [Google Scholar]
- 43. Segerstolpe Å, Palasantza A, Eliasson P, et al. Single-cell transcriptome profiling of human pancreatic islets in health and type 2 diabetes. Cell Metab. 2016;24(4):593‐607. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Mawla AM, Huising MO. Navigating the depths and avoiding the shallows of pancreatic islet cell transcriptomes. Diabetes. 2019;68(7):1380‐1393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Kimple ME, Joseph JW, Bailey CL, et al. Galphaz negatively regulates insulin secretion and glucose clearance. J Biol Chem. 2008;283(8):4560‐4567. [DOI] [PubMed] [Google Scholar]
- 46. Hart RG, Lee JJ, Zhai K, et al. Somatostatin receptors shape insulin and glucagon output within the pancreatic islet in mice through direct and paracrine effects. Diabetologia. 2026;69(9):2569‐2584. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Reininger L, Rehman M, Hart R, et al. Supplementary data from: Free fatty acid receptor 4 agonism stimulates insulin secretion via different mechanisms in mouse versus human islets. Deposited in Figshare on September 4, 2026. doi: 10.6084/m9.figshare.33437338 [DOI] [PMC free article] [PubMed]
- 48. Latour MG, Alquier T, Oseid E, et al. GPR40 is necessary but not sufficient for fatty acid stimulation of insulin secretion in vivo. Diabetes. 2007;56(4):1087‐1094. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Blanchi B, Taurand M, Colace C, et al. EndoC-βH5 cells are storable and ready-to-use human pancreatic beta cells with physiological insulin secretion. Mol Metab. 2023;76:101772. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50. Berthault C, Staels W, Scharfmann R. Purification of pancreatic endocrine subsets reveals increased iron metabolism in beta cells. Mol Metab. 2020;42:101060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51. Fontes G, Zarrouki B, Hagman DK, et al. Glucolipotoxicity age-dependently impairs beta cell function in rats despite a marked increase in beta cell mass. Diabetologia. 2010;53(11):2369‐2379. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52. Pourhosseinzadeh MS, Huang JL, Shin D, et al. Heterogeneity in the coordination of delta cells with beta cells is driven by both paracrine signals and low-density Cx36 gap junctions. Proc Natl Acad Sci U S A. 2025;122(38):e2504151122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53. Huang JL, Pourhosseinzadeh MS, Lee S, et al. Paracrine signalling by pancreatic delta cells determines the glycaemic set point in mice. Nat Metab. 2024;6(1):61‐77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54. Hauge-Evans AC, King AJ, Carmignac D, et al. Somatostatin secreted by islet delta-cells fulfills multiple roles as a paracrine regulator of islet function. Diabetes. 2009;58(2):403‐411. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55. Viollet C, Simon A, Tolle V, et al. Somatostatin-IRES-Cre mice: between knockout and wild-type? Front Endocrinol (Lausanne). 2017;8:131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 56. Pipeleers D, in't Veld PI, Maes E, Van De Winkel M. Glucose-induced insulin release depends on functional cooperation between islet cells. Proc Natl Acad Sci U S A. 1982;79(23):7322‐7325. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 57. Kimple ME, Neuman JC, Linnemann AK, Casey PJ. Inhibitory G proteins and their receptors: emerging therapeutic targets for obesity and diabetes. Exp Mol Med. 2014;46(6):e102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58. Korobko MVSIV. An optimized electroporation-free method for generation of recombinant adenoviral plasmids. Mol Gen Microbiol Virol. 2019;34(4):228‐233. [Google Scholar]
- 59. Voznesenskaya A, Berggren P-O, Ilegems E. Sustained heterologous gene expression in pancreatic islet organoids using adeno-associated virus serotype 8. Front Bioeng Biotechnol. 2023;11:1147244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 60. Ifie E, Russell MA, Dhayal S, et al. Unexpected subcellular distribution of a specific isoform of the Coxsackie and adenovirus receptor, CAR-SIV, in human pancreatic beta cells. Diabetologia. 2018;61(11):2344‐2355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 61. Smith JS, Lefkowitz RJ, Rajagopal S. Biased signalling: from simple switches to allosteric microprocessors. Nat Rev Drug Discov. 2018;17(4):243‐260. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 62. Kimple ME, Nixon AB, Kelly P, et al. A role for G(z) in pancreatic islet beta-cell biology. J Biol Chem. 2005;280(36):31708‐31713. [DOI] [PubMed] [Google Scholar]
- 63. Tubio MR, Fernandez N, Fitzsimons CP, et al. Expression of a G protein-coupled receptor (GPCR) leads to attenuation of signaling by other GPCRs: experimental evidence for a spontaneous GPCR constitutive inactive form. J Biol Chem. 2010;285(20):14990‐14998. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 64. Saunders DC, Brissova M, Phillips N, et al. Ectonucleoside triphosphate diphosphohydrolase-3 antibody targets adult human pancreatic beta cells for in vitro and in vivo analysis. Cell Metab. 2019;29(3):745‐754.e4. [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.
Data Citations
- Reininger L, Rehman M, Hart R, et al. Supplementary data from: Free fatty acid receptor 4 agonism stimulates insulin secretion via different mechanisms in mouse versus human islets. Deposited in Figshare on September 4, 2026. doi: 10.6084/m9.figshare.33437338 [DOI] [PMC free article] [PubMed]
Data Availability Statement
Original data generated and analyzed during this study are included in this published article or in the data repository listed in Ref. (47).
