Equine glucagon-like peptide-1 receptor physiology

eGLP-1R gene expression

For gene expression studies, archived samples that had been stored at −80 °C after being collected and snap-frozen using liquid nitrogen immediately following the euthanasia of nine healthy, Standardbred horses (Equus caballus, <15 years old) were used (The University of Queensland animal ethics approval number: SVS/013/08/RIRDC), in addition to abattoir sourced specimens from twelve horses. The collection of all abattoir specimens was approved (Queensland University of Technology tissue use approval number: 1400000039) and all horses were examined by a veterinarian prior to euthanasia. The samples included pancreas, left ventricle of the heart, kidney, digital lamellae, tongue, skeletal muscle and duodenum. Tissue samples (50–100 mg) were prepared for total RNA extraction either by being pulverised with a hammer (lamellae) or homogenised (all other tissues; Omni International, Kennesaw, GA, USA). Trizol reagent (1 mL/100 mg tissue) was used to extract the RNA according to the manufacturer’s instructions (Invitrogen, Scoresby, Vic, Australia) and genomic DNA was eliminated with RNAse-free DNAse I (Invitrogen, Scoresby, Vic, Australia). The RNA concentration and integrity were determined with a 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) prior to cDNA synthesis.

The Tetro cDNA Synthesis Kit (Bioline, Alexandria, NSW, Australia) was used to synthesise cDNA from Dnase-treated RNA. Each reaction (20 µL) contained 1 µg of cDNA, 1 µL Oligo (dT)₁₈, 4 µL of 5× reverse transcription buffer, 1 µL RiboSafe Rnase inhibitor, 1 µL Tetro reverse transcriptase (200 µg/µL) and DEPC-treated water. The cDNA was stored at −80 °C until polymerase chain reaction (PCR) analysis. The eGLP-1R primer (forward: 5′-AGCGCATCTTCAGGCTCTAT-3′, reverse: 5′-GGGATGAGTGTCAGTGTGGA-3′) was designed using Primer-BLAST software (Ye et al., 2012) and primer concentration (250 nm) and PCR conditions (below) were optimised prior to use. No template, negative controls containing water instead of cDNA were included. The selected reference gene was GAPDH (GenBank accession number: AF157626).

After optimisation, PCR was performed using the OneTaq Hot Start Quick-Load 2×  Master Mix protocol (New England BioLabs, Ipswich, MA, USA). Each 25 µL reaction contained 50 ng cDNA, 12.5 µL OneTaq Master Mix Buffer, 10 pM/µL of each primer (forward and reverse), and 10.5 µL deionised water (dH₂O). The PCR cycle protocol involved 94 °C initial denaturation for 30 s, 94 °C denaturation for 30 s, 60 °C annealing for 30 s, 72 °C extension for 15 s and 72 °C final extension for 5 min. After cycling, the PCR products were visualised with 2% agarose gel electrophoresis. The PCR products were purified using an ISOLATE II PCR and Gel Kit (Bioline, Alexandria, NSW, Australia) and sequenced using a standard protocol on an ABI 3500 sequencing platform (Applied Biosystems, Carlsbad, CA, USA). Sequencing confirmed amplification of the desired target and reference genes.

eGLP-1R immunohistochemistry

Immunohistochemistry was undertaken to examine the immunolocalisation of the GLP-1 hormone, and also separately of the eGLP-1R, in the pancreas. Following harvest at the abattoir, fresh pancreatic tissues from four horses were immediately placed in 10% formalin, prior to embedding in paraffin and cutting into 5 µm thick sections. Sections were deparaffinised with xylene and then rehydrated through decreasing concentrations of ethanol (100, 95 and 70% for 1 min each) to water. Slides were then incubated in antigen retrieval solution (Target Retrieval Solution; DakoCytomation Inc., CA, USA) at 95 °C for 30 min and quenched with 0.3% hydrogen peroxide. After blocking with normal horse serum (from a fasted Standardbred horse with very low levels of serum GLP-1), the primary antibodies (GLP-1 (7-36) rabbit anti-human polyclonal; Merck, Kenilworth, NJ, USA; dilution 1:50 and GLP-1 receptor rabbit anti-human polyclonal; Lifespan Biosciences, Seattle, WA, USA; LS-A1208; dilution 1:50 [20 ug/mL]) were applied separately to duplicate sections and incubated overnight in a moist environment at 4 °C. Although not equine-specific, the immunogen sequence identity of the GLP-1 receptor antibody to Equus caballus was 100%. The antibody has no homology with other proteins, except weakly (50%) with the ubiquitous protein PRPSAP2, thus cross reactivity is unlikely. The slides were then washed and incubated at room temperature for 2 h with secondary antibody (mouse anti-rabbit HRP-conjugated; Santa Cruz Biotechnology, Dallas, TX, USA). Staining was developed using DAB substrate-chromogen solution (Sigma-Aldrich, Castle Hill, NSW, Australia) and the slides were counterstained with hematoxylin. Slides were then dehydrated using alcohol and xylene and mounted using DPX mounting medium (Sigma-Aldrich, Castle Hill, NSW, Australia).

Equine GLP-1 secretion in vitro

Horse duodenal tissue (10 cm) was collected 30–50 cm distal to the pylorus immediately within 10 min of euthanasia of three, mixed-breed horses (at a local abattoir). The tissues were rinsed with ice-cold saline (0.9%, Baxter Healthcare, Old Toongabbie, NSW, Australia) and placed in ice-cold buffer (135 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 1.8 mM CaCl₂, 20 mM Hepes, and 0.05% (W/V) BSA at pH 7.4) on ice for 5 min. Smaller explants (5 mm × 5 mm) of mucosal tissue were then warmed to 30 °C in buffer (as above but with oxygenation (95% O₂/5% CO₂)) over 10 min. Individual explants were then removed and placed immediately into the wells of a 96 well reaction plate, and incubated in triplicate in 0.8 mL of buffer containing 0, 4 or 12 mM glucose for 30 min while agitating the plate every 5 min. The glucose concentrations of 4 and 12 mM were selected to approximate fasted and markedly elevated equine blood glucose concentrations, respectively (Bamford et al., 2015; de Laat, McGree & Sillence, 2016). Following the incubation, the explants were removed and blotted dry before being weighed and discarded. The buffer was collected separately, snap-frozen and stored at −80 °C until analysed for active GLP-1 concentration with an ELISA (MPEZGLPHS35K, Merck Millipore, Billerica, MA, USA) validated (de Laat, McGree & Sillence, 2016) for use in horses (inter-assay CV 6.8%). Absorbance was measured spectrophotometrically at 450 nm (GloMax; Promega, WI, USA). The active GLP-1 concentration was corrected for the weight (mg) of the intestinal tissue explant used in each reaction/well.

Stimulation of pancreatic islet eGLP-1Rs

Equine pancreatic islets were isolated from freshly obtained pancreas (following euthanasia of a further five healthy horses at a local abattoir, six replicates were performed per horse) as described previously (Kheder et al., 2017). Briefly, the islets were isolated using collagenase (11213857001; Roche, Mannheim, Germany) digestion of the tissue and separation of the islets through a density gradient (Histopaque 1.1 g/mL; Sigma Aldrich, Castle Hill, NSW, Australia). The freshly isolated islets from each horse (>80% live islets) were examined microscopically and then used immediately in a series of concentration–response experiments to determine insulin responses to a human peptide fragment (7-37 amide) of GLP-1 (G9416; Sigma Aldrich, Castle Hill, NSW, Australia) and a synthetic GLP-1 agonist (exendin-4, E7144; Sigma Aldrich, NSW, Australia) that has been demonstrated to be insulinogenic in other species (Kim et al., 2016). Separate aliquots were used for each experiment. The insulin secretion in each experiment was corrected for the basal insulin secretion and also normalised to protein concentration. Finally, a human GLP-1R antagonist (exendin 9-39, E7269; Sigma Aldrich, Castle Hill, NSW, Australia) was used to determine if any positive insulin responses to the agonists could be inhibited. Equine-specific products were not available.

Experiment one: Stimulation experiments using GLP-1 (7-37)were performed using aliquots of pancreatic islets (1 mL; 2.75 mM glucose) pre-incubated for 10 min prior to the addition of GLP-1 at a range of concentrations (1 × 10⁻¹¹, 1 × 10⁻¹⁰, 1 × 10⁻⁹, 1 × 10⁻⁸ and 1 × 10⁻⁶ M). Each concentration was tested in triplicate over an incubation period of 60 min at room temperature (RT), before centrifugation (RT; 7 min; 16,400× g) to separate the supernatant from the cells. The incubation conditions have been optimised previously (Kheder et al., 2017). The supernatant was immediately frozen and stored at −80 °C ready for insulin analysis with a validated equine ELISA kit (Mercodia, Uppsala, Sweden; inter-assay CV 12.3%). The islets were solubilised in extraction buffer (PIE89901; Thermo Fisher Scientific, Scoresby, Vic, Australia) with a protease inhibitor cocktail (1:100 P8340; Sigma Aldrich, Castle Hill, NSW, Australia). The protein concentration of each replicate was determined (in triplicate; CV < 5%) using the bicinchoninic acid assay (PIE22323; Pierce, Rockford, IL, USA). Absorbance for both assays was measured spectrophotometrically (GloMax; Promega, WI, USA). Subsequently, the experiment was repeated in a subset of samples with a dipeptidyl peptidase-IV (DPPIV) inhibitor added (Diprotin B at 20 µM; BioVision, CA, USA), and the results compared. The enzyme DPPIV degrades GLP-1 (Baggio & Drucker, 2007).

Experiment two: The first experiment was repeated using increasing concentrations of the synthetic GLP-1 agonist exendin-4 (1 × 10⁻¹³, 1 × 10⁻¹², 1 × 10⁻¹¹, 1 × 10⁻¹⁰ and 1 × 10⁻⁸). Due to the improved performance of the assay when the DPPIV inhibitor was used in experiment one, all of the incubations in the second experiment contained Diprotin B. Islets from only three horses were measured due to unacceptably high islet death during incubation in two samples.

Experiment three: The ability of exendin 9-39 to antagonise GLP-1-stimulated insulin production was assessed under the same incubation conditions described for experiments one and two. The GLP-1 (7-37)was used at its EC₅₀ of 1 nM and exendin 9-39 was used at the same concentration (Gault et al., 2003). The antagonist was added to the incubation media 10 min prior to the addition of the islets. Due to a technical problem, islets from only four of the five horses were tested with exendin 9-39.

Experiment four: Considering the ability of GLP-1 to stimulate β-cell proliferation, the islets were analysed following experiment three for evidence of cellular proliferative and apoptotic activity. Caspase 3 and proliferating cell nuclear antigen (PCNA) were used as markers of apoptosis and proliferation, respectively. Colorimetric equine-specific ELISA kits (MyBioSorce, San Diego, CA, USA) were used to determine caspase 3 and PCNA concentrations (mean assay CVs < 10%). Absorbance was measured spectrophotometrically at 450 nm (VersaMax; Molecular Devices, CA, USA).

Statistical analyses

The data were normally distributed according to a Shapiro–Wilk test, except the PCNA data. The GLP-1 released from the intestinal explants in the absence of glucose was subtracted from the amount secreted at the two glucose concentrations tested, and these two values were compared using a paired t-test. The insulin concentration in islet samples containing GLP-1, with and without the antagonist, as well as the caspase 3 concentrations from the islet extracts, were also compared using a paired t-test. The PCNA data were examined with a Wilcoxon signed rank test and are reported as median [IQR]. All other data are reported as mean ± s.e.m. Significance was set at P < 0.05 with trends reported at P < 0.1. The EC₅₀ were calculated with Prism 7.0 (GraphPad software, La Jolla, CA, USA) and the remaining data analyses were performed with SigmaPlot v.13 (Systat software, San Jose, CA, USA).

eGLP-1R gene expression

The eGLP-1R gene was consistently expressed in the pancreatic tissue from all horses examined. The sequencing data demonstrated 98–100% sequence identity to the Equus caballus GLP-1R mRNA sequence available from GenBank (XM_014734596). The pancreas was subsequently used as a positive control for the electrophoresis of the PCR products from the other tissues. The eGLP-1R was expressed in all of the tissues tested: the heart, liver, kidney and duodenum, digital lamellae, tongue and gluteal skeletal muscle (Fig. 1). The GAPDH gene transcript was also amplified in all tissues (Fig. 1).

eGLP-1R immunohistochemistry

Following amplification of the eGLP-1R gene transcript in the pancreas, the location of the receptor within the pancreatic tissue was examined immunohistochemically. The GLP-1 hormone (Fig. 2A) and the eGLP-1R (Fig. 2B) were separately immunolocalised to be within the pancreatic islets. There was no evidence of the GLP-1 hormone or the eGLP-1R in other parts of the pancreas.

Equine GLP-1 secretion in vitro

The GLP-1 secretion from the intestinal explants under glucose stimulation (4 mM and 12 mM) was corrected for the incidental release when explants from the same animal were incubated with no glucose. At 4 mM glucose the explants secreted 3.8 ± 1.2 pM/mg of tissue of GLP-1 into the incubation media in 30 min (Fig. 3). The amount of GLP-1 secreted (per mg of tissue) trended toward being less (P = 0.07) when the explants were stimulated with 12 mM glucose (2.4 ± 0.75 pM/mg of tissue in 30 min).

Stimulation of pancreatic islet eGLP-1Rs

Experiment one: Insulin was secreted in a concentration-dependent manner in response to GLP-1 (Fig. 4A). The maximum insulin concentration (Cmax) recorded after the 60 min incubation occurred at a GLP-1 concentration of 10 nM in the presence of the DPPIV inhibitor. The GLP-1 Cmax was 484 µIU/mg of protein in the presence of the DPPIV inhibitor, and 336 µIU/mg of protein without the DPPIV inhibitor (Fig. 4A). Furthermore, the GLP-1 EC₅₀ (in the presence of a DPPIV inhibitor) was estimated to be 1 nM. The insulin secretion in the absence of GLP-1 (213 ± 58 µIU/mg protein over 60 min) was consistent with that expected in response to the glucose concentration contained in the incubation medium (Fig. 4, dashed line), based on optimisation experiments.

Experiment two: In contrast to human GLP-1 7-37, the synthetic analogue exendin-4 did not elicit an insulin response from the islets (Fig. 4B), above the baseline (237 ± 61 µIU/mg protein) at any concentration tested. Accordingly, exendin-4 was not tested in conjunction with the antagonist, exendin 9-39.

Experiment three: The amount of insulin released by individual explants in the presence of GLP-1 at its estimated EC₅₀ of 1 nM showed a statistical trend for being reduced (by 27%; P = 0.08) in the presence of 1 nM exendin 9-39 (Fig. 4C).

Experiment four: There was no significant difference in caspase 3 concentrations between GLP-1-stimulated islets when they were incubated with and without exendin 9-39 (45.8 ± 1.5 pM and 47.8 ± 1.19 pM, respectively). Further, the PCNA results were similar with no significant difference between stimulated and inhibited samples (2.5 [2.5–4.2] ng/mL and 3.7 [2.6–5.1] ng/mL, respectively).