Transgenic Expression of Glucagon-Like Peptide-1 (GLP-1) and Activated Muscarinic Receptor (M3R) Significantly Improves Pig Islet Secretory Function

Abstract

Porcine islets show notoriously low insulin secretion levels in response to glucose stimulation. While this is somehow expected in the case of immature islets isolated from fetal and neonatal pigs, disappointingly low secretory responses are frequently reported in studies using in vitro-maturated fetal and neonatal islets and even fully differentiated adult islets. Herein we show that β-cell-specific expression of a modified glucagon-like peptide-1 (GLP-1) and of a constitutively activated type 3 muscarinic receptor (M3R) efficiently amplifies glucose-stimulated insulin secretion (GSIS). Both adult and neonatal isolated pig islets were treated with adenoviral expression vectors carrying sequences encoding for GLP-1 and/or M3R. GSIS from transduced and control islets was evaluated during static incubation and dynamic perifusion assays. While expression of GLP-1 did not affect basal or stimulated insulin secretion, activated M3R produced a twofold increase in both first and second phases of GSIS. Coexpression of GLP-1 and M3R caused an even greater increase in the secretory response, which was amplified fourfold compared to controls. In conclusion, our work highlights pig islet insulin secretion deficiencies and proposes concomitant activation of cAMP-dependent and cholinergic pathways as a solution to ameliorate GSIS from pig islets used for transplantation.

Introduction

In the context of islet transplantation, aspects of porcine islets often overlooked are their poor secretory activity and their incapacity to respond to stimuli in a fashion similar to that seen with healthy islets from nonhuman primates (NHPs) or humans. Indeed, most, if not all, studies tackle the issue of graft immune rejection and a great majority of genetically engineered islets are designed in an attempt to reduce host immune reaction. The final goal of islet transplantation being the restoration of controlled insulin secretion in diabetic patients, it seems crucial to ensure that transplanted islets secrete enough insulin in response to stimuli to enable adequate control of host glycemia. In the case of preclinical porcine islet xenotransplantation, the low level of insulin production is often compensated by transplanting extremely large numbers [>50,000 islet equivalents (IEQ)/kg] of islets^1-3. Previous work from several groups has shown that porcine islets do exhibit a biphasic pattern of glucose-stimulated insulin secretion (GSIS) in in vitro perifusion assays^4-7. However, in comparison to human islets, isolated porcine islets secrete less insulin during the first (x6) and second (x3) phase after stimulation with 15 mM glucose despite similar insulin content in both islets. Unfortunately, our knowledge of porcine β-cell physiology remains limited since most β-cell physiologists use more easily obtainable rodent islets in their studies. However, there are some data showing that glucose transport, phosphorylation, and utilization in pig islets are similar to what are seen in rat islets⁸. Much less is known about stimulus–secretion coupling in pig islets, and although there is general agreement that GSIS is significantly increased in the presence of 3′,5′-cyclic adenosine monophosphate (cAMP)-increasing agents such as theophylline, forskolin (Fsk), or 3-isobutyl-1-methylxanthine (IBMX)^6,9,10, no one has explored this or other amplifying pathways to improve pig islet secretory function. Insulin granule exocytosis triggered by glucose metabolism and the ensuing rise in cytosolic calcium concentration is further regulated by two major amplifying pathways^11,12: (1) a cAMP-dependent pathway, activated physiologically by binding of glucagon-like peptide-1 (GLP-1) to its G proteincoupled receptor on β-cells and mimicked by cAMP-increasing drugs, and (2) a cholinergic pathway, activated by binding of acetylcholine or cholecystokinin to a type 3 muscarinic receptor and mimicked by carbachols or partially by direct protein kinase C ligands such as phorbol esters. Both of these pathways increase the number of readily releasable insulin granules in β-cells¹³ and result in a greater secretory response to glucose stimulation. Permanently inducing these changes in porcine β-cells by means of genetic engineering might be a novel and helpful approach to increase insulin secretion from isolated pig islets bringing their secretory function closer to that of human islets and rendering them more efficient in controlling host glycemia in both preclinical and clinical diabetes trials without the need to transplant extremely high numbers of islets.

Materials and Methods

All experiments were conducted in accordance with the local ethics committee and carried out in accordance to EU Directive 2010/63/EU for animal experiments.

Pancreas Procurement and Islet Isolation

Pancreata from adult Belgian landrace pigs (>200 kg) were harvested aseptically at a local slaughterhouse (A. de Marbaix Center, Louvain-la-Neuve, Belgium), infused with 1 ml/g cold modified University of Wisconsin (UW) solution, and transported at 4°C in the same solution to the main laboratory in Brussels. Pancreata were digested with collagenase NB8 (4.4 U/g pancreas weight; Serva, Heidelberg, Germany). The digested tissue was filtered through a prepurification column before islet purification on a discontinuous Ficoll gradient (Mediatech, Miami, FL, USA). Pancreatic islets were then collected, washed, counted, and cultured in Roswell Park Memorial Institute (RPMI)-1640 medium (Life Technologies, Asse, Belgium) supplemented with 10% heat-inactivated fetal bovine serum (FBS), 1% penicillin, 1% streptomycin, and 5 mM glutamine (all from Westburg, Leusden, Netherlands) and 5 mM glucose.

For isolation of neonatal porcine islets (NPIs), we used a modification of a method originally described by Korbutt et al.¹⁴. Briefly, pancreata were harvested from exsanguinated piglets, cut into 1- to 3-mm³ pieces in cold Hank's balanced salt solution (HBSS; Life Technologies) supplemented with 0.25% bovine serum albumin (BSA; Roche, Basel, Switzerland), 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES; Westburg), 10 mM glucose, 1% penicillin, 1% streptomycin, and 0.1% amphotericin B (Westburg). Enzymatic digestion was done using 10 mg/pancreas collagenase V (Sigma-Aldrich, Darmstadt, Germany). Digested tissue was then filtered on a 500-μm mesh, washed three times in cold HBSS, and cultured in HAM-F10 (Life Technologies) supplemented with 0.5% BSA, 50 μM IBMX, 10 mM glucose, 2 mM glutamine, 10 mM nicotinamide (Sigma-Aldrich), 1% penicillin, and 1% streptomycin.

Adenoviral Transduction

Expression cassettes, encoding modified GLP-1 and activated type 3 muscarinic receptor (M3R), were synthesized by Delphigenetics (Gosselies, Belgium). Briefly, the 31-amino acid GLP-1 (7-37) sequence was modified by a mutation (A8S) substituting alanine at position 8 by a serine, thus rendering GLP-1Ser8 (serine-8) resistant to enzymatic degradation by dipeptidyl peptidases and prolonging its half-life¹⁵. Constitutive activation of M3R was achieved by a single point mutation (Q490L) as previously described¹⁶. These cassettes were then used to produce adenoviral expression vectors carrying GLP-1Ser8, activated M3R, both (bicistronic vectors), or green fluorescent protein (GFP; Genecust, Dudelange, Luxembourg). Islets from neonatal or adult pigs were transduced using these adenoviral vectors. Islet mRNA and protein extraction as well as in vitro insulin secretion tests were carried out 48-72 h following viral infection.

Analysis of Cultured Islets

Islet total RNA was extracted using the High Pure RNA Kit from Roche. Following DNAse treatment, cDNA was generated using a QuantiTect Reverse Transcription Kit (Qiagen, Hilden, Germany). For negative controls, the reverse transcription (RT) reaction was carried out in the absence of reverse transcriptase. RT products were amplified by polymerase chain reaction (PCR) using primer pairs specific for GLP-1Ser8 [5′-TCCTCGGCCTTTCACCAGCC-3′ and 5′-GTGATCTGACTTCTGGTCTC-3′; PCR product size, 179 base pairs (bp)] and for M3R (5′-CCCAATTGATGTACCCATAC-3′ and 5′-GTGATCTGACTTCTGGTCTC-3′; PCR product size, 873 bp). Islet cell lysates were prepared in radioimmunoprecipitation assay (RIPA) buffer containing 150 mM NaCl, 50 mM Tris, 0.5% sodium deoxycholate, 0.1% Triton, 0.1% sodium dodecyl sulfate (SDS), 1% protease inhibitor cocktail, and 0.1% phosphatase inhibitor (all from Merck, Darmstadt, Germany). Total islet proteins were electrophoresed and transferred to a 0.2-μm nitrocellulose membrane from Bio-Rad (Temse, Belgium). After blocking for 45 min, the membrane was incubated with mouse anti-GLP-1 (1:500) or mouse anti-hemagglutinin (HA) tag (1:200) and developed using horseradish peroxidase (HRP)-conjugated secondary antibodies. All primary antibodies were from Abcam (Cambridge, MA, USA), and secondary antibodies were from Dako (Carpinteria, CA, USA).

Electron Microscopy

Control and transduced adult islets from one single preparation were fixed by 2.5% (v/v) glutaraldehyde (Merck), washed, dehydrated, and embedded in Spun-resin (Sigma-Aldrich). Ultrathin sections were obtained with a Reichert ultramicrotome from Leica (Diegem, Belgium), collected on rhodanium grids, and contrasted with 3% uranyl acetate (Fisher Scientific, Merelbeke, Belgium) and lead citrate (Sigma-Aldrich). Grids were then washed and dried before examination in an FEI CM12 electron microscope (FEI, Eindhoven, Netherlands).

In Vitro Insulin Secretion Assays

For all experiments, the working medium was a bicarbonate-buffered solution containing 120 mM NaCl, 4.8 mM KCl, 2.5 mM CaCl₂, 1.2 mM MgCl₂, 24 mM NaHCO₃, 1 mg/ml BSA, and varying concentrations of glucose and test substances as indicated in the figures. Fsk was from Calbiochem (San Diego, CA, USA), and glucokinase activator 50 (GKA50) and PMA (phorbol myristate acetate) were from Sigma-Aldrich. For static incubation assays, batches of 100-200 IEQ were incubated for 2 h at 37°C in 1 ml of test medium, and then an aliquot was taken and diluted before insulin assay. For dynamic insulin secretion tests, batches of 200-1,000 IEQ were placed in perifusion chambers, covered with 8-μm cellulose filters, and sealed. Test solutions kept at 37°C and gassed continuously to stabilize pH at around 7.2 were pumped at a flow rate of 1 ml/min. Effluent fractions were collected at 2-min intervals and saved for insulin assay using a human insulin radioimmunoassay (RIA) kit (Millipore, Billerica, MA, USA). At the end of the experiments, islets were recovered, and their insulin content was determined after extraction in acid–ethanol (75% ethanol, 180 mM HCl from Merck).

Presentation of Results

All experiments have been performed with islets from 4 to 11 preparations. Results are presented as means±standard error of the mean (SEM). Data were analyzed by analysis of variance (ANOVA), followed by a Dunnett's test for comparisons of test groups with controls or a Newman–Keuls test for comparisons between test groups using GraphPad Instat version 3.0 (GraphPad Software, La Jolla, CA, USA). Differences were considered to be statistically significant at values of p<0.05.

Results

Islet Isolation Outcome

Porcine pancreatic islets are notoriously difficult to isolate because of their lack of a protective fibrous capsule found in human and rodent islets. Indeed, using our static digestion method, islet yield from adult pig pancreata was variable, and we consistently observed a 30%-50% loss in islet number after 24 h of culture. In contrast, islet isolation from neonatal piglets was reproducible, and the number of islets constantly increased throughout the 8-day culture period. Table 1 summarizes some of the parameters recorded before, during, and after islet isolation. In particular, islet yield per gram of pancreas from neonatal piglets was almost 10 times greater than adult pigs. Average insulin content was 175 versus 616 μU/TEQ in neonatal and adult islets, respectively.

Table 1.

Summarized Characteristics of Donor Pigs and Pancreatic Islet Isolation

	Neonatal Piglets (n = 53)	Adult Pigs (n = 40)
Age	14-21 days	12-24 months
Body weight (kg)	5.6±0.2	232.3±6.7
Pancreas weight (g)	14±1.1	351.2±14.3
Warm ischemia time	<1 min	10-25 min
Cold ischemia time	<5 min	45-60 min
Islet isolation	Easy/reproducible	Difficult/variable
	Manual shacking for 20 min	Static incubation for 25-35 min
Islet yield (IEQ/g pancreas)	2,164±267	219±9
Insulin content (μU/IEQ)	175±7	616±80

Values are means±SE. Islet yield numbers represent islet counts 8 days after isolation for neonatal piglets and 1 day after isolation for adult pigs.

Effect of Drug-Mediated Activation of PKA and PKC on Pig Islet Insulin Secretion

In vitro insulin secretion tests are a practical means of evaluating secretory function of isolated islets in response to different stimuli. In addition, such tests could be carried out in the presence of desired substances to study particular pathways involved in insulin secretion regulation. Neonatal and adult porcine islets incubated for 2 h in 15 mM glucose Krebs medium secreted 1% and 1.8% of their insulin content, respectively (Fig. 1). In the presence of Fsk (1 μM), which mimics the effect of GLP-1 by activating adenylate cyclase, thus raising intracellular cAMP and activating protein kinase A (PKA), insulin secretion was increased by 4.3-fold for neonatal islets and 2.3-fold for adult islets compared to controls. PMA (20 nM), which partially mimics the effect of acetylcholine by directly activating protein kinase C (PKC), caused a 5.6-fold amplification of GSIS in neonatal islets and a 4-fold increase in adult islets. The greatest effect on both neonatal and adult islets (x12.5 and x6.1, respectively) was seen when both Fsk and PMA were present in the incubation medium, indicating an additive synergic effect when both pathways are activated (Fig. 1). Fsk alone, PMA alone, and the combination of both drugs had no significant effect on insulin secretion at low glucose.

Figure 1.

Effects of forskolin (Fsk) and phorbol myristate acetate (PMA) on insulin secretion from incubated porcine islets. Batches of 100-200 islets were incubated in 1 ml of Krebs medium containing 15 mM glucose (G15) alone or in combination with Fsk (1 μM), PMA (20 nM), or both Fsk and PMA. Insulin secretion was measured in the incubation media and expressed as a percentage of total insulin content of each batch of islets. The numbers above the columns represent the fold increase in insulin secretion in test groups compared to 15 mM glucose alone. *p < 0.05. Values are means±standard error of the mean (SEM) for n = 11-35 from nine different neonatal preparations and n = 5-21 from eight different adult preparations.

Transgene Expression

To test the effects of GLP-1Ser8 and M3R activation on pig islet insulin secretion, we infected neonatal and adult pig islets with recombinant, replication-defective adenoviruses encoding GLP-1Ser8 (Ad-GLP-1Ser8), activated M3R (Ad-M3R), both (Ad-GLP-1Ser8M3R), or GFP (Ad-GFP). Transgene expression (Fig. 2) was dependent on the multiplicity of infection (MOI) with more GFP⁺ cells detected in islets treated with 10⁴ MOI Ad-GFP. Since islet survival and function were unaffected by exposure to the vector at MOI = 10⁴ compared to untreated islets, we used this concentration of adenovirus in all following experiments. Total RNA prepared from batches of 10,000 IEQ was used in RT-PCR experiments to confirm transgene expression in islet cells. Using transgene-specific primers, we could detect the expression of GLP-1Ser8 in islets treated with Ad-GLP-1Ser8 and Ad-GLP-1Ser8M3R vectors. M3R was detected in Ad-M3R- and Ad-M3RGLP-1Ser8-treated islets, whereas none could be detected in control islets (Fig. 3A). At the protein level, we could detect GLP-1 and the M3R HA-tag in transduced islet extracts (Fig. 3B).

Figure 2.

Adenoviral transduction of neonatal pig islets. Green fluorescent protein (GFP) was expressed in piglet islet cells by exposing the islets to a recombinant adenoviral vector at multiplicity of infection (MOI) = 200 (A) or MOI = 10⁴ (B) for 6 h. Islets were then washed and cultured for 48 h before fluorescence microscopy observation.

Figure 3.

Transgene expression in transduced porcine islets. (A) Reverse transcription polymerase chain reaction (RT-PCR) analysis of GLP-1Ser8 [glucagon-like peptide-1 (GLP-1) sequence modified by a mutation-substituting alanine at position 8 by a serine] and/or type 3 muscarinic receptor (M3R) expression in control and transduced islets 48 h after viral transduction. Negative control samples that had not been treated with reverse transcriptase did not yield any products. 18S rRNA was amplified in all samples as an internal control. (B) Western blot analysis of GLP-1 and transgenic M3R expression in control and treated islets. Anti-GLP-1 antibody was used to evaluate GLP-1 expression. To check for transgenic M3R expression, an anti-hemagglutinin (HA) tag antibody was used. Ad-GLP-1Ser8, adenoviruses encoding GLP-1Ser8; Ad-M3R, activated M3R; Ad-GLP-1Ser8M3R, both Ad-GLP-1Ser8 and Ad-M3R.

Effect of GLP-1Ser8 and Activated M3R Expression on Pig Islet Insulin Secretion

To evaluate the secretory function of control and modified islets, we stimulated batches of 200 adult IEQ or 1,000 neonatal IEQ with 15 mM glucose during dynamic perifusion experiments and measured insulin secretion over a period of 40 min for each batch of islets after a 30-min stabilization period at 1 mM glucose. The perifusion medium contained low glucose (1 mM, G1) during the first 10 min (-10 to 0), then it was changed to high (15 mM, G15) glucose at t=0 and remained so for the next 30 min (0 to 30). As shown by the traces with full circles in Figures 4A and 5A, control islets responded to glucose stimulation within 5 min with a biphasic 2.3- and 4.4-fold increase in insulin secretion for adult and neonatal, respectively (Figs. 4A and B and 5A and B). To determine whether GLP-1Ser8 and/or M3R expression affects the dynamics of insulin secretion, we divided the secretory response to a first (2-10 min) and a second (10-30 min) phase. Adult islets expressing GLP-1Ser8 (Ad-GLP-1Ser8) secreted the same amount of insulin as control islets in response to glucose during the first and second phases (Fig. 4C and D) in a way that their stimulation index did not change significantly (2.3 vs. 2.6 in control and Ad-GLP-1Ser8, respectively) (Fig. 4B). This was also the case with neonatal islets (Fig. 5). GSIS was equally amplified during both phases (2- to 3-fold compared to controls) in islets expressing activated M3R (Ad-M3R) (Figs. 4A and 5A, full diamond traces). The stimulation index increased to 3.4 for adult islets and to 6.5 for neonatal islets. Interestingly, when both GLP-1Ser8 and M3R were expressed in porcine islets (Ad-GLP-1Ser8M3R), the increase in GSIS was greater than that seen when each of the transgenes was expressed alone. There was a 3.6- to 3.8-fold increase during both phases of the secretory response with adult and neonatal Ad-GLP-1Ser8M3R islets (Figs. 4C and D and 5C and D). As a consequence, islet stimulation index was increased to 5.5 and 6.4 in adult and neonatal islets, respectively (Figs. 4B and 5B). To determine whether increased GSIS in Ad-GLP-1Ser8M3R-transduced islets was accompanied by ultrastructural changes within porcine β-cells, we examined insulin granule localization by transmission electron microscopy. Cells containing numerous secretory granules with an electron-dense core surrounded by a halo characteristic of pancreatic β-cells were easily identifiable in porcine islets. In controls, these granules appeared to be evenly distributed across the cytosol with few granules in close vicinity (<200 nm) of the plasma membrane (Fig. 6A). In transduced islets, we could observe a redistribution of insulin granules toward the cell periphery, and we could identify several of them docked to the plasma membrane (Fig. 6B).

Figure 4.

Comparison of glucose-stimulated insulin secretion (GSIS) from unmodified adult porcine islets and GSIS from islets expressing GLP-1Ser8 [glucagon-like peptide-1 (GLP-1) sequence modified by a mutation-substituting alanine at position 8 by a serine] and/or type 3 muscarinic receptor (M3R). Forty-eight hours following transduction with 10⁴ multiplicity of infection (MOI) Ad-GLP-1Ser8 (○), Ad-M3R (♦), or both (⋄), batches of 200 islet equivalents (IEQ) were perifused with Krebs medium to determine time course of insulin changes (A). Islets in 1 mM glucose (Gl) were stimulated with 15 mM glucose (G15), and insulin secretion from transduced islets was compared to controls (•). (B) Mean insulin secretion during low (Gl; −10 to 0 min) and high (G15; 2 to 30 min) glucose. (C, D) Insulin secretion integrated over 8 min for the first phase (2-10 min) and 20 min for the second phase (10-30 min). Significant differences (p<0.01) between resting (white columns) and stimulated (black columns) secretion are shown above pairs of columns (stimulation index) in (B). *p<;0.05, significant difference with controls. Fold increase compared to controls is shown above the columns in (C) and (D). Values are means±standard error of the mean (SEM) for six to eight experiments. Ad-GLP-1Ser8, adenoviruses encoding GLP-1Ser8; Ad-M3R, activated M3R; Ad-GLP-1Ser8M3R, both Ad-GLP-1Ser8 and Ad-M3R.

Figure 5.

Comparison of glucose-stimulated insulin secretion (GSIS) from unmodified neonatal porcine islets and GSIS from islets expressing GLP-1Ser8 [glucagon-like peptide-1 (GLP-1) sequence modified by a mutation-substituting alanine at position 8 by a serine] and/or activated type 3 muscarinic receptor (M3R). Forty-eight hours following transduction with 10⁴ multiplicity of infection (MOI) Ad-GLP-1Ser8 (○), Ad-M3R (♦), or both (⋄), batches of 1,000 islet equivalents (IEQ) were perifused with Krebs medium to determine time course of insulin changes (A). Islets in 1 mM glucose (G1) were stimulated with 15 mM glucose (G15), and insulin secretion from transduced islets was compared to controls (•). (B) Mean insulin secretion during low (G1; −10 to 0 min) and high (G15; 2 to 30 min) glucose. (C, D) Insulin secretion integrated over 8 min for the first phase (2-10 min) and 20 min for the second phase (10-30 min). Significant differences (p<0.01) between resting (while columns) and stimulated (black columns) secretion are shown above pairs of columns (stimulation index) in (B). *p<;0.05, significant difference with controls. Fold increase compared to controls is shown above the columns in (C) and (D). Values are means±standard error of the mean (SEM) for four to five experiments. Ad-GLP-1Ser8, adenoviruses encoding GLP-1Ser8; Ad-M3R, activated M3R; Ad-GLP-1Ser8M3R, both Ad-GLP-1Ser8 and Ad-M3R.

Figure 6.

Effect of GLP-1Ser8 [glucagon-like peptide-1 (GLP-1) sequence modified by a mutation substituting alanine at position 8 by a serine] and type 3 muscarinic receptor (M3R) expression on insulin granule localization in adult porcine islets. Transmission electron microscopy of control and Ad-GLP-1Ser8M3R-transduced islets. Scale bars: 1 μuM. Plasma membrane (PM) and insulin secretory granules (SG) are pointed out by black arrows. This is a representative image from one single adult islet preparation. Ad-GLP-1Ser8, adenoviruses encoding GLP-1Ser8; Ad-M3R, activated M3R; Ad-GLP-1Ser8M3R, both Ad-GLP-1Ser8 and Ad-M3R.

Effect of Glucokinase Activation on Pig Islet Insulin Secretion

Isolated pig islets exposed to gradually increasing concentrations of glucose show poor stimulation of insulin secretion with a slight nonsignificant increase when glucose concentration reaches 7 mM (Fig. 7, full circles). To verify whether this low secretory response is caused by impaired glucose phosphorylation in porcine β-cells, isolated pig islets were perifused in the presence of 5 μM GKA50. However, glucose-stimulated insulin secretion in the presence of GKA50 was similar to that seen in its absence (Fig. 7, open circles). In the presence of adenylate cyclase activator Fsk (1 μM), insulin secretion was stimulated (x3) at 5 mM glucose and further increased (x6.6) and reached a plateau at 7 mM glucose (Fig. 7, full diamonds). Glucokinase activation by GKA50 had an appreciable effect on insulin secretion only in the presence of Fsk. Islets perifused in the presence of GKA50 and Fsk combined show significantly stimulated insulin secretion in the presence of 3 mM glucose. Under these conditions, GSIS was further increased at 5 mM glucose but was no longer amplified by higher glucose concentrations (Fig. 7, open diamonds).

Figure 7.

Effect of glucokinase activation on glucose-stimulated insulin secretion (GSIS) from neonatal porcine islets under conditions of low and high 3′,5′-cyclic adenosine monophosphate (cAMP). The concentration of glucose was gradually increased (G in mM) as indicated on top of the figure in a control medium (•) or in the presence of 5 μM of the glucokinase activator 50 (GKA50) (○), 1 μM of the adenylyl cyclase activator forskolin (Fsk; ♦), or both GKA50 and Fsk (⋄). Values are means±standard error of the mean (SEM) for five different islet preparations.

Discussion

Porcine islets of Langerhans represent a promising alternative treatment for type 1 diabetes. Islets isolated from fetal, neonatal, and adult pigs have been considered as xenogeneic sources of β-cells for transplantation in humans, but the number of studies dedicated to evaluate porcine islet secretory function remains limited. However, most of these studies agree on the poor performance of these islets when they are challenged by glucose or other stimuli^4-7,17. Unlike most secretory cells that identify their stimuli through selective receptors, pancreatic β-cells do not express glucose receptors. Glucose must be metabolized for the β-cell to secrete insulin, and the rate of secretion is governed by the rate of metabolism through stimulus–secretion coupling pathways that have been extensively studied in rodent models^18,19. In addition to this proximal stimulus–secretion coupling, other distal amplifying pathways controlled by protein kinases and other metabolic messengers also contribute to insulin secretion regulation^20,21. The capacity of isolated islets or β-cell preparations to properly respond to a stimulus is often evaluated by measuring insulin secretion under resting and stimulation conditions. In the current study, we found stimulation indices (G15/G1 ratio) of 2.3 for adult islets and 4.4 for neonatal islets in dynamic perifusion assays. Moreover, when glucose concentration was increased stepwise as opposed to the abrupt increase from G1 to G15, there was barely an increase in insulin secretion at G7 and G10 (Fig. 7). This is of particular importance since transplanted islets are more likely to be exposed to such gradual increase in glucose concentration in the graft environment rather than the artificial one-step increase created in vitro. Our results show that glucose phosphorylation by glucokinase is not the limiting step in porcine islet stimulation since insulin secretion was unaffected and remained low in the presence of GKA50 (Fig. 7). In agreement with previous observations⁶, elevating cytosolic [cAMP] by Fsk allowed glucose to stimulate insulin secretion at 5 mM as expected and to further increase the response at 7 mM. Interestingly, we found that glucokinase activation in combination with elevated [cAMP] effectively shifted the glucose dose–response curve to the left as insulin secretion was greatly increased at substimulatory glucose concentration (3 mM). This observation highlights the prevalence of distal signals in the regulation of GSIS in porcine β-cells rather than defaults in glucose metabolism. While GLP-1 is known to increase stimulated insulin secretion by raising [cAMP] and activating PKA and exchange protein activated by cAMP (Epac) 2, acetylcholine amplifies GSIS in mouse and human islets mainly by activation of PKC²². We now show that direct activation of PKC by PMA increases insulin secretion stimulated by 15 mM glucose by four to six times in adult and neonatal porcine islets (Fig. 1) and that the effects of Fsk and PMA on these islets are additive (6- to 12-fold increase compared to controls). Interestingly, the effects of PKA and PKC on insulin granule pools were found not to be additive in rodent β-cells¹³. Cholinergic regulation of porcine insulin secretion has not been extensively studied, but there are few reports showing increased insulin secretion following electrical stimulation of the vagus nerves or acetylcholine infusion in whole perfused pig pancreas^23,24 and carbachol stimulation of perifused isolated islets¹⁷. Since our goal is to use porcine islets as xenografts to treat type I diabetes in a preclinical model²⁵, we checked whether we can reproduce the effects of PKA- and PKC-activating drugs on insulin secretion by means of genetic modification of porcine β-cells. Using adenoviral vectors, we expressed GLP-1Ser8 and/or activated type 3 muscarinic receptor in isolated pig islets. Surprisingly, and unlike the amplifying effect of Fsk, GLP-1 expression alone did not affect GSIS in adult and neonatal pig islets, while M3R expression significantly increased the secretory response to G15 during both phases of insulin secretion. The amplitude of the increase relative to control islets was greater in neonatal than in adult islets (compare Figs. 4C and D and 5C and D). Finally, and in accordance with our observation using a combination of Fsk and PMA, the greatest increase in GSIS was seen in islets coexpressing GLP-1Ser8 and M3R. In these islets, both the first and second phases of insulin secretion were increased by almost fourfold independent of the age of the animals. As a result, stimulation indices were increased from 2.3 to 5.5 for adult islets and from 4.4 to 6.4 for neonatal islets. Our preliminary observations from electron microscopy images of control and transduced islets suggest a change in insulin granule localization toward the plasma membrane in Ad-GLP-1Ser8M3R islets. The increased number of readily releasable granules in close contact with the plasma membrane has been previously suggested as one of the mechanisms by which amplifying signals exert their effect on GSIS. The comparison of GSIS from Ad-GLP-1Ser8M3R islets to that from human islets shows that coexpression of GLP-1Ser8 and M3R brings porcine islet secretory function closer to human islets. Indeed, modified pig islets secrete 30% less insulin than human islets during the first phase and 20% more during the second phase (Fig. 8). In total, the amount of insulin secreted during 30 min of stimulation with G15 (area under the curve) was 1.32% versus 1.27% of insulin content for modified pig islets and human islets, respectively. Basal insulin secretion from porcine islets is slightly higher (x1.4) than human islets and even more so in the case of modified pig islets (x2). Damage caused to porcine islets during isolation could explain this higher unstimulated secretion and thus lower the stimulation index. In conclusion, our work shows that pig islet responsiveness to glucose can be permanently improved by coexpression of modified GLP-1 and activated M3R in β-cells. Islets modified this way display normal time course of GSIS during in vitro perifusions with an increased stimulation index compared to unmodified islets. Generation of transgenic pigs with β-cell-specific expression of GLP-1Ser8 and M3R is currently underway. By combining pig line selection for low PERV (porcine endogenous retrovirus) animals, pathogen-free breeding and housing, and immunoisolation devices to our current findings, such pigs could become an ideal source of islets for transplantation in future clinical studies.

Figure 8.

Comparison of insulin secretion from perifused control (•) and Ad-GLP-1Ser8M3R (⋄) adult pig islets and human (▪) islets. Glucose (G) concentration was changed from 1 to 15 mM as indicated. Values are means±standard error of the mean (SEM) for four to five experiments. Human islet insulin secretion data are adapted from Dufrane et al.⁶.