Alastair D Green, Srividya Vasu*, R Charlotte Moffett, Peter R Flatt

Co-culture of clonal beta cells with GLP-1 and glucagon-secreting cell line impacts on beta cell insulin secretion, proliferation and susceptibility to cytotoxins

Alastair D Green, Srividya Vasu*, R Charlotte Moffett, Peter R Flatt

SAAD Centre for Pharmacy & Diabetes, University of Ulster, Coleraine, Northern Ireland, UK

* Corresponding author: Dr. Srividya Vasu, SAAD Centre for Pharmacy & Diabetes, University of Ulster, Cromore Road, Coleraine BT52 1SA, Northern Ireland, United Kingdom.

E-mail:

Tel: +44 (0) 28 70124419

Short title: GLP-1 and glucagon

Key words: GLP-1, glucagon, MIN6, GLUTag, αTC1.9

Abstract

We investigated thedirect effects on insulin releasing MIN6 cells of chronic exposure to GLP-1, glucagon or a combination of both peptides secreted from GLUTag L-cell andαTC1.9alpha-cell lines in co-culture.MIN6, GLUTag and αTC1.9 cell lines exhibited high cellular hormone content and release of insulin, GLP-1 and glucagon, respectively. Co-culture of MIN6 cells with GLUTag cellssignificantly increasedcellular insulin content, beta-cellproliferation, insulin secretory responsesto a range of established secretogogues and afforded protection against exposure cytotoxic concentrations of glucose, lipid, streptozotocin or cytokines. Benefits of co-culture of MIN6 cells with αTC1.9 alphacells were limited to enhanced beta-cell proliferation with marginal positive actions on both insulin secretion and cellular protection. In contrast, co-culture of MIN6 with GLUTag cells plus αTC1.9cells, markedly enhanced both insulin secretory responses and protection against beta-cell toxins compared with co-culture with GLUTag cells alone. These data indicate important long-term effects of conjoint GLP-1 and glucagon exposure on beta-cell function. This illustrates thepossible functional significance of alpha-cell GLP-1 production as well asdirect beneficial effects of dual agonism at beta-cell GLP-1 and glucagon receptors.

Introduction

Glucagon-like peptide 1 (GLP-1) is an enteroendocrine incretin hormone secreted in response to feeding that exerts multiple effects of benefit in T2DM [1 – 3]. At the level of the pancreatic beta cell, these include stimulation of insulin biosynthesis, insulin secretion, beta cell proliferation and protection against apoptosis[4 – 6]. The peptide also inhibits glucagon (GCG) secretion and exerts extrapancreatic actions to inhibit gastric emptying and feeding, resulting in significant weight loss [5, 7]. Outside of blood glucose control,GLP-1 has been linked to cardioprotection, improved cognition and prevention of bone loss [8 – 11].

The beneficial actions of GLP-1 have been harnessed by the development of enzyme resistant forms of the peptide, including exenatide and liraglutide, and development of dipeptidyl-peptidase IV (DPPIV)inhibitor drugs which extend the half-life of endogenous GLP-1 and gastric inhibitory polypeptide (GIP) by inhibiting their degradation [12 – 14]. These therapies are now in widespread use for treatment of T2DM and further research efforts are ongoing to develop new agents which can provoke GLP-1 release from intestinal L-cells as a therapy in combination with DPPIV inhibition [15].A further line of intense effort relates to the discovery that dual agonism of GLP-1 and GCG receptors using oxyntomodulinor other dual auctioning agonist peptides provide an additional therapeutic option.

Glucagon is counterregulatory hormone to insulin which circulates at elevated levels in T2DM and is considered to contribute significantly to diabetic hyperglycaemia [16]. Thus it is unexpected that dual agonism of GLP-1R and GCGR would have beneficial effects [17, 18].However, it appears that the hyperglycaemic action of glucagon is more than overpowered by the beta cytotropic actions of GLP-1 and that the other actions of glucagon such as inhibition of feeding and stimulation of lipolysis add significantly to the beneficial actions of GLP-1[19, 20].

In the light of these findings, significant effort has been devoted also to the development of stable analogues of oxyntomodulin which naturally activate GLP-1 and glucagon receptors [21 – 24] and designer dual acting GLP-1 and GCG agonists [18, 25]. Further, triple acting agonists have been devised which also exploit the incretin actions of GIP [26 – 29]. When administered to animal models of diabetes, these drugs have been shown to promote body weight loss, lower blood glucose, improve glucose tolerance and enhance insulin release and beta cell preservation [28 – 30].

Despite functionally important effects on isletsin vivo, including α-cell GLP-1 production [31, 32], rather little is known about the prolonged direct effects of dual activation of GLP-1R and GCGR at the level of the beta cells. Are effects due to GLP-1 activation alone, glucagon, a combination of bothpeptides or merely a reflection of improved blood glucose control? In the present study, we have utilised co-culture of insulin secreting MIN6 cells with GLP-1 releasing GLUTag cells and/or glucagon releasing alpha-TC1.9 cells to explore the chronic actions of these peptides on beta cell insulin secretion, proliferation and survival.This physiological approach to peptide exposure has the advantage of ensuring continual exposure to active peptide, avoiding exclusive exposure to their truncated degradation products.

Materials and Methods

2.1 Cell culture: The generation, culture characteristics and suitability of each of the three model cell lines - murine MIN6, GLUTag and αTC1.9, are detailed elsewhere [33 – 36].MIN6 cells and αTC1.9 cells were routinely cultured in Dulbecco’s Modified Eagle Medium (DMEM) containing 25 mM glucose and 2 mM L-glutamine supplemented with 10 % (v/v) FCS and antibiotics (100 U.ml-1 penicillin and 0.1 g/l streptomycin). GLUTag cells were routinely cultured in DMEM containing 5.5 mM glucose and with the same supplements. MIN6 cells and αTC1.9 cells were used in experiments to a maximum passage number of 40, while GLUTag cells were used up to a passage number of 25.

2.2 Co-culturing of MIN6 cells with GLUTag and αTC1.9 cells: To evaluate the effects of GLUTag and αTC1.9 cell secretions on MIN6 cells, GLUTag cells, αTC1.9 cells or a combination of the two were seeded into hanging inserts with 0.4 μm pore semi-permeable membranes (Millipore, MA, USA) at a total cell density of 5 x 104 per insert (2.5 x 104 of each cell type for combined group). These inserts were then placed into normal flat-bottomed 24 well plates seeded with 1 x 105 MIN6 cells per well. This allowed MIN6 cells to be cultured separately from the culture cells while allowing free exchanges of hormones and other cell secretions between the two compartments. Cells were co-cultured using DMEM containing 5.5 mM noted abovein this manner for 24 h to adapt the cells to basal glucose conditions before experimentation.Cells in inserts were imaged using an Olympus IX51 phase contrast microscope (Olympus, Essex, UK) connected to a Spot JR camera and imaging software (Diagnostic Instruments Inc, Sterling Heights, USA).

2.3 Measurement of acute insulin release: For acute tests, MIN6 cells cultured for 24 h using hanging inserts as described above were first pre-incubated for 40 mins at 37 °C in Krebs-Ringer bicarbonate buffer (KRBB) (115 mmol/l NaCl, 4.7 mmol.l KCl, 1.28 mmol/l CaCl2, 1.2 mmol.l MgSO4 10 mmol/lNaHCO3, 20 mmol/l Hepes) containing 1.1 mmol/l glucose supplemented with 0.1% w/v bovine serum albumin (BSA) (Gibco® Invitrogen, Paisley, UK), before being incubated for a further 60 min in KRBB supplemented with (v/v) 0.1% BSA and a range of concentrations of glucose and modulators of insulin secretion as described in the Figures. Following acute tests, supernatants were stored at -20 °C until insulin analysis [37].To help assess changes in insulin secretory responsiveness, insulin secretion in response to various agents was expressed as stimulation index calculated from insulin released (ng/million cells) evoked by each secretagogue divided by insulin release (ng/million cells) recorded in its absence or, in the case of 16.7 mM glucose, at 5.6 mM glucose.

2.4 Cell proliferation:Effect of co-culture on cell proliferation was determined by counting cells using trypan blue exclusion method and expressed as % of seeding density.

2.5 Cytotoxin treatments and assessment of metabolic cell viability: To investigate the effects of GLUTag and αTC1.9 co-culture on the cytoprotective capabilities of MIN6 cells, MIN6 cells that had previously been co-cultured in the presence and absence of GLUTag cells, αTC1.9 cells or both for a period of 24 h were incubated for 8 hr at 37 °C with cytotoxins at the concentrations indicated in the Figures. The cytotoxic agents studied were streptozotocin (STZ), glucose, palmitate and cocktails of the proinflammatory cytokines containing interleukin-1β (IL1-β), interferon γ (IFNγ) and tumor necrosis factor α (TNFα). Working solutions were prepared by diluting concentrated stocks appropriately in tissue culture media immediately before use, with the exception of glucose which was added directly to the culture media with no intermediate stock. Metaboliccell viability was then determined using the colorimetric (MTT) assay as described previously [38].

2.6 Determination of protein, hormone and glucose concentration: GLUTag and αTC1.9 cell extracts were prepared by overnight incubation at 4 °C with acid-ethanol (1.5 % (v/v) HCl, 70 % (v/v) ethanol).Protein contents were determined by Bradford assay. Insulin was determined by radioimmunoassay[37].Total GLP-1 was determined using specific enzyme linked immunoassay following manufacturer’s instructions (GLP-1 Total ELISA, EZGLP-1T-36K, Millipore, MA, USA). Glucagon was determined using glucagon chemiluminescent assay (EZGLU-30K, Millipore, MA, USA) following manufacturer’s instructions. These kits are considered specific by the suppliers. The GLP-1 assay does not discriminate between GLP-1 (7-36) and GLP-1 (9-36) but shows negligible reactivity with GLP-2, GIP, glucagon and oxyntomodulin. The glucagon kit cross-reacts <5% with oxyntomodulin and not at all with GLP-1 or the glucagon fragments (1-18) and (19-29). GLP-1 was not extracted from samples for the assays while glucagon was extracted from samples using acetonitrile and reconstituted in assay buffer.It would be interesting to know the levels of oxyntomodulin secreted during these experiments but this was not possible due to poor specificity of commercially available assays tested.

2.7 Statistics: Results are expressed as mean ± S. E. M. for a specific number of observations (n). Data sets were normally distributed and compared using Student’s unpaired t-test with two-tailed p-values and 95 % confidence intervals and one-way analysis of variance (ANOVA) with Bonferroni post-hoc test where applicable. Two groups of data were considered to be significantly different if p<0.05.

3. Results

3.1 Characteristics of GLUTag and αTC1.9 cells cultured in semi-permeable hanging co-culture inserts: GLUTag cells, αTC1.9 cells and combinations of the two were cultured in semi-permeable hanging inserts designed for the purpose of co-culturing different cell types (Figure 1A). GLP-1 and glucagon contents of GLUTag and αTC1.9 cells grown in inserts for 24 h contained relatively high concentrations of their correspondent hormones (Figure 1 B, D). GLUTag cells additionally contained small amounts of glucagon and αTC1.9 cells contained negligible amounts of GLP-1(Figure 1 B-E). Inserts that had been seeded with equal numbers of GLUTag and αTC1.9 cells contained hormone levels which were part way between the relative contents of each hormone in inserts seeded with individual cell types (Figure 1 B, D). These results were mirrored by levels of GLP-1 and glucagon released into the culture media during the 24 h period (Figure 1 C, E), indicating good functionality of the cells.Final concentrations of GLP-1 and glucagon measured in the media were 2-57 pM and 18-102 pM respectively.

3.2 Characteristics of MIN6 cells co-cultured with GLUTag cells, αTC1.9 cells or both: Co-culture of MIN6 cells with GLUTag cells significantly increased their insulin content (1.5 fold difference, p<0.05, Figure 2A). Co-culture with αTC1.9 cells and a combination of GLUTag and αTC1.9 cells had no effect on insulin content of MIN6 cells (Figure 2A) but all cell types significantly increased the proliferation rate of MIN6 cells, with rates increasing 2.0 fold, 1.4 fold and 1.9 fold in each respective group compared to controls (Figure 2B).

Co-culture of MIN6 cells with GLUTag cells or a combination of GLUTag and αTC1.9 cells significantly increased secretory responses to all agents tested compared to MIN6 only controls (1.2 – 1.9 fold for GLUTag co-culture groups and 1.4 – 2.1 fold responses for combined cell co-culture groups,Table 1). Combined cell co-culture significantly enhanced insulin secretion responses to 16.7 mM glucose, alanine, KCl and GIP compared to GLUTag co-culture (Table 1). αTC1.9 cell co-culture only significantly affected insulin secretion in response to GIP (Table 1).Both GLUTag and combined co-culture enhanced basal insulin secretion at 5.6 mM glucoseby 1.7 and 2 fold respectively(p<0.01) whereasαTC1.9 cell co-culture had no significant effect.

Metabolicviability of co-cultured MIN6 cells following exposure to a range of cytotoxins including streptozotocin, glucose, palmitate and cocktails of proinflammatory cytokines are shown in Figure 3. Streptozotocin, glucose, palmitate and cytokine cocktail decreased MIN6 metabolic cell viability (Figure 3). GLUTag, αTC1.9 and combined cell co-culture conferred a significant enhancement in the viability of MIN6 cells to STZ and palmitate compared to MIN6 cells cultured alone (1.4 – 1.7 fold increases, Figure 3). Both GLUTag and combined cell co-culture augmented protection against glucotoxicity compared to controls (2.0 and 1.5 fold increases respectively, Figure 3).αTC1.9 and combined cell co-culture afforded significant protection against a cytokine cocktail compared to controls (1.2 – 1.6 fold differences in MIN6 metaboliccell viability, Figure 3). Compared to both GLUTag and TC1.9 separately, combined cell co-culture significantly increased metaboliccell viability (1.3 – 1.7 fold increases, Figure 3). Protection was increased againstglucose and palmitate (1.2 – 1.4 fold increases in viability, Figure 3) compared to TC1.9 cells alone, but decreased (1.4 fold, Figure 3) compared to cells cultured with GLUTag cells.

Discussion

Dual activation of GLP-1 and glucagon receptors by oxyntomodulin or related peptides is being explored as a feasible therapeutic option for treatment of diabetes [22 – 24, 39 – 42]. This approach activates many beneficial pathways, including effects on feeding activity, gastric emptying, lipid metabolism and islet cell function [1]. The latter action has been ascribed to action at GLP-1R and the possible role of GCGR is unknown.This study has utilised murine beta, L- and alpha cell lines secreting insulin, GLP-1 and glucagon (MIN6, GLUTag and αTC1.9, respectively) to examine cellular actions of prolonged exposure to GLP-1 and glucagon on function of insulin secreting cells after 24 hours co-culture using flat bottomed plates with hanging inserts. Such an approach ensures a constant supply of freshly secreted active hormone, resulting in clear-cut effects at very low endogenous hormone concentrations. The cell lines have been extensively characterised previously as useful cell models [34, 43 – 45].

Each cell line thrived under culture conditions employed with high cellular hormone contents and release of GLP-1 and glucagon being recorded as expected. Co-culture of MIN6 cells with GLUTag cells significantly increased cellular insulin content and insulin secretory responses to all agents tested, consistent with positive effects onbeta-cells by GLP-1 aspreviously reported [46 – 49]. Despite release of high levels of glucagon, αTC1.9 cell co-culture was largely ineffective in altering cellular insulin or secretion responses from MIN6 cells but did evoke a significantly greater response to GIP. The reasons for this are unclear presently but the data overall suggest that intra-islet glucagon may have relatively little effect on beta cell secretory function. This view accords with our previous studies, comparing heterotypic and homotypic pseudoislets which showed little benefit on beta-cell function of combining αTC1.9 cells with MIN6 cells [50].

In the present study, secretory responses to stimulatory glucose, alanine, KCl, CaCl2 and GIP were all significantly increased in MIN6 cells that had been co-cultured with a combination of GLUTag and αTC1.9 cells compared with cells cultured with GLUTag alone. This was also associated withmodest enhancement of cellular insulin content. In contrast, responses to GLP-1 and glucagon were not significantly augmented in the combined cell co-culture group compared to the GLUTag alone, possibly reflecting down-regulation of respective receptors and signal transduction pathways[51 – 53]. These insulin secretion data suggest synergistic effects of the two hormones, indicating that dual agonism of GLP-1R and GCGR provides superior responses from pancreatic beta-cells to GLP-1R agonism alone. This presumably reflects stimulation of diverse signal transduction pathways by the two hormones leading to beta-cell insulin exocytosis.This enhanced effect of dual agonism of GLP-1R and GCGR is indeed reminiscent of previous studies using oxyntomodulin [23, 40, 54]and suggests that enhancement of pancreatic beta cell function by dual agonists in vivo[22 – 24] is not due to GLP-1R activation alone.Enhanced actions of GLP-1 and glucagon together are also likely to be significant under conditions of beta-cell stress when α-cells secrete both hormones as part of the islet adaptation process [31, 32, 55, 56].

A long-acting oxyntomodulin analogue has been shown recently to enhance islet morphology, stimulate beta-cell proliferation and protect against beta-cell apoptosis in mice with type 1 diabetes[24]. Co-culture with GLUTag cells and αTC1.9 cells both significantly stimulated MIN6 cell proliferation. This supports previous observations mediated via GLP-1R and GCGRusing both cells and transgenic mice [47, 49, 57 – 59]. GLUTag cells elicited a considerably greater response than αTC1.9 cells in this respect, likely due to the superior effects of GLP-1 over glucagon as observed previously [60]. MIN6 cells co-cultured with both GLUTag and αTC1.9 cells showed proliferation rates similar to those incubated with GLUTag cells alone. However, as half the number of GLUTag cells were seeded and taking into account the relatively lessened effects of αTC1.9 cells compared to GLUTag alone co-culture, it is likely that the difference is made up for by co-agonistic effects of both GLP-1 and glucagon.

Exposure of MIN6 cells to STZ, high glucose, palmitate or pro-inflammatory cytokines for 8 h caused decreases in mitochondrial metabolic viability in line with previous observations in pancreatic beta-cell lines and primary cells [61 – 68]. The potent cytoprotective effects of GLP-1 on beta-cells are well established [69 – 71]. Concordant with these observations, MIN6 cells incubated with GLUTag cells for 24 h prior to cytotoxin exposure showed significantly enhanced cytoprotective mechanisms against all the agents tested. Incubation with αTC1.9 cells also significantly enhanced protection against STZ, palmitate and a cocktail of pro-inflammatory cytokines compared to controls. This is corroborated by previous studies showing that under or over expression of GCGR in pancreatic beta-cells correlates with a respective decrease and increase in beta-cell mass [58, 59].As with decreased beta-cell apoptosis in type 1 diabetic mice treated with stable oxyntomodulin analogue [24], combined cell co-culture conferred enhanced protection to MIN6 cells against STZ, glucose and pro-inflammatory cytokines superior to that of GLUTag co-culture alone. This in vitro effect likely reflects synergistic protective actions of both hormones binding their respective receptors. Combined co-culture did not offer improved significantly improved resistance to palmitate compared to GLUTag co-culture, but viability was higher than in cells co-culture with only αTC1.9 cells, possibly due to the decreased activity of the more potent protective effects of GLP-1. Overall these data, combined with observations on beta cell proliferation, suggest that expansion of beta cell mass observed in diabetic animal models treated with oxyntomodulin or other agonist peptides [24]reflects actions at GCGR as well as GLP-1R [18, 41, 72].