Chronic exposure to incretin metabolites GLP-1(9 (cid:0) 36) and GIP(3 (cid:0) 42) affect islet morphology and beta cell health in high fat fed mice

The incretin hormones, glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP), are rapidly degraded by dipeptidyl peptidase-4 (DPP-4) to their major circulating metabolites GLP-1(9 (cid:0) 36) and GIP(3 (cid:0) 42). This study investigates the possible effects of these metabolites, and the equivalent exendin molecule Ex(9 (cid:0) 39), on pancreatic islet morphology and constituent alpha and beta cells in high-fat diet (HFD) fed mice. Male Swiss TO-mice (6 – 8 weeks-old) were maintained on a HFD or normal diet (ND) for 4 months and then received twice-daily subcutaneous injections of GLP-1(9 (cid:0) 36), GIP(3 (cid:0) 42), Ex(9 (cid:0) 39) (25 nmol/kg bw) or saline vehicle (0.9% (w/v) NaCl) over a 60-day period. Metabolic parameters were monitored and excised pancreatic tissues were used for immunohistochemical analysis. Body weight and assessed metabolic indices were not changed by peptide administration. GLP-1(9 (cid:0) 36) significantly (p < 0.001) increased islet density per mm 2 tissue, that was decreased (p < 0.05) by HFD. Islet, beta and alpha cell areas were increased (p < 0.01) following HFD and subsequently reduced (p < 0.01-p < 0.001) by GIP(3 (cid:0) 42) and Ex(9 (cid:0) 39) treatment. While GLP-1(9 (cid:0) 36) did not affect islet and beta cell areas in HFD mice, it significantly (p < 0.01) decreased alpha cell area. Compared to ND and HFD mice, GIP(3 (cid:0) 42) treatment significantly (p < 0.05) increased beta cell proliferation. Whilst HFD increased (p < 0.001) beta cell apoptosis, this was reduced (p < 0.01-p < 0.001) by both GLP-1(9 (cid:0) 36) and GIP (3 (cid:0) 42). These data indicate that the major circulating forms of GLP-1 and GIP, namely GLP-1(9 (cid:0) 36) and GIP (3 (cid:0) 42) previously considered largely inactive, may directly impact pancreatic morphology, with an important protective effect on beta cell health under conditions of beta cell stress.


Introduction
The incretin hormones glucagon-like peptide-1 (GLP-1) and glucosedependent insulinotropic polypeptide (GIP), constitute a crucial regulatory system in glucose metabolism, exerting profound effects on pancreatic islet function and metabolic homeostasis [1].GIP accounts for almost half of the insulin response to oral glucose, whilst GLP-1 accounts for 30 %, and the remaining 20 % is triggered by glucose alone [2].Pathophysiological conditions like obesity and type 2 diabetes mellitus (T2DM) affect the secretion and/or action of GLP-1 and GIP, leading to an impaired incretin effect [3].Earlier findings indicate that GIP secretion is enhanced in obesity and T2DM [4], while its ability to stimulate insulin is compromised [5].On the other hand, sub-optimal GLP-1 secretion and signaling is observed in obesity, that may be linked to both onset and maintenance of the obesity phenotype [6].In the context of T2DM, the status of GLP-1 secretion remains a matter of controversy with conflicting opinions on whether it is impaired [3], conversely the insulinotropic effects of GLP-1 are confirmed to be preserved in T2DM [7].This has encouraged the development of a wide range of anti-diabetic drugs based on GLP-1 receptor (GLP-1R) agonism [8].
GLP-1 and GIP are expressed and secreted from intestinal L-and Kcells in response to feeding, although small amounts are evidenced in pancreatic alpha cells under conditions of islet stress [15].Within the whole body, the incretin hormones exert multiple effects at extrapancreatic sites [1] but within islets, incretin receptors on both beta, delta and possibly alpha cells mediate blood-borne, autocrine and paracrine effects on local secretory function [16,17].This is mediated by binding to specific G s receptors activating adenylyl cyclase and elevating cytosolic cAMP to potentiate insulin secretion [18].Both hormones also have proliferative and anti-apoptotic effects on beta cells, linked to effects on downstream signaling targets such as extracellular signal-regulated kinase 1/2 (ERK1/2) and the serine/threonine kinase AKT/PKB [19,20].
While both the incretin metabolites have been studied for their effects on insulin secretion, there is still much to discern on how they impact pancreatic islet morphology and beta cell health, especially under conditions of islet stress such as that observed in diet-induced obesity.Therefore, we have investigated changes in islet morphology, composition and cellular turnover following long-term administration of GLP-1(9− 36), GIP (3− 42), and Ex(9− 39) to mice fed a high-fat diet (HFD).Since alpha-and beta-cells constitute for over 90% of islet mass [38], our investigations focused on these cell types.However, other islet cells, such as delta and PP cells, may also exert important paracrine effects to help regulate overall islet morphology and beta cell health.

Peptides and animals
Peptides were purchased from GL Biochem Ltd. (Shanghai, China) and molecular masses verified in-house using matrix assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) as described previously [39].Male Swiss TO mice (6-8 weeks old; Harlan, Blackthorn, UK) were age-matched, divided into groups and housed individually in an air-conditioned room at 22 ± 2 • C with a 12 h light: 12 h dark cycle (08:00-20:00 h).One group of animals (n=4) had free access to standard rodent maintenance diet (ND -10 % fat, 30 % protein and 60 % carbohydrate; Trouw Nutrition, Northwick, UK), while four other groups (n=8) were placed on a high fat diet (HFD -45 % fat, 20 % protein and 35 % carbohydrate; Special Diets Service, Essex, UK) for four months.By this point, HFD mice exhibited significantly increased in body weight and circulating glucose, alongside impaired glucose tolerance when compared to ND mice.HFD mice then received twice-daily (10:00 and 16:00 h) subcutaneous injections of GLP-1 (9− 36), GIP (3− 42), Ex(9− 39) (each at a dose of 25 nmol/kg bw) or saline vehicle (0.9 % (w/v) NaCl) over a 60-day period.The choice of peptide dose was based on previous studies [9,40], but there is a possibility that the incretin metabolites exert distinct dose-related effects at respective receptors [26,36,41], that should be considered when interpreting the current findings.Mice on ND received twice-daily subcutaneous saline injections.All experiments were conducted under the UK Animals (Scientific Procedures) Act 1986 & EU Directive 2010/63EU and approved by the University of Ulster Animal Welfare and Ethical Review Body (AWERB).

Tissue processing
Pancreatic tissues were extracted from mice and fixed for 48 h in paraformaldehyde (4 % w/v in phosphate buffered saline (PBS) to preserve cellular architecture by cross-linking proteins.Tissues were then processed in an automated tissue processor (Leica TP1020, Leica Microsystems, Nussloch, Germany), which involved dehydrating tissues in 70-100 % ethanol, followed by xylene immersion to remove wax before paraffin embedding [42].Tissue blocks were sectioned (6 μm) using a Shandon Finesse 325 microtome (Thermo Scientific, Hemel Hempstead, UK) and picked for staining at intervals of 10 sections, placed on poly-L-lysine coated slides.

Immunohistochemistry
To assess immunoreactive staining for insulin, glucagon, Ki-67 and TUNEL as appropriate, pancreatic sections were dewaxed in histoclear for 30 mins, before being rehydrated with decreasing concentrations of ethanol.Sections were blocked with 2.5 % bovine serum albumin (BSA) and then incubated with designated primary antibody (Table 1) overnight.On day 2, sections were rinsed in PBS and incubated with suitable secondary antibody (Alexa Fluor® 594 for red and Alexa Fluor® 488 for green; Table 1) for 1 h at 37 • C.After PBS wash, slides were then incubated with DAPI for 15 mins at 37 • C [43].Finally, sections were mounted on coverslips using antifade mounting media before being viewed at 10× and 40× magnification using an Olympus IX51 inverted microscope and photographed using a DP70 digital camera system.

Image analysis
Image J software was used to analyse images to assess islet-, betaand alpha-cell areas as well as percentage of cells positive for antibody of interest in a blinded manner.To calculate number of islets (per mm 2 pancreas), the area of tissue was measured, and islet number counted within that area.Glucagon-positive cells located in the central core of islets, lacking direct contact with peripheral alpha cells, were counted as central alpha cells.For beta cell proliferation, insulin and Ki-67 copositive cells were counted whereas for apoptosis, insulin and TUNEL co-positive cells were counted, as described previously [42].

Biochemical analysis
Non-fasting plasma glucose was measured in blood taken from the cut tip on the tail vein of conscious mice using a hand-held Ascencia Contour blood glucose meter (Bayer Healthcare, Newbury, Berkshire, UK).At the end of the treatment regimen, a glucose tolerance was performed by evaluating plasma glucose after i.p. injection of glucose alone (18 mmol/kg body weight, AppliChem GmbH, Germany) in mice starved for 18 hours.

Statistical analysis
GraphPad PRISM (version 8.0) software was used to perform statistical analysis.Values are expressed as mean ± S.E.M. Comparative analyses between groups were carried out using a one-way ANOVA with Bonferroni post-hoc test.There was no inclusion and exclusion criteria applied.Groups of data were considered to be significant if p<0.05.

Effect of GIP(3− 42), GLP-1(9− 36) and Ex(9− 39) on beta and alpha cells
Representative images of islets stained for insulin and glucagon are shown in Fig. 2A.Analysis of percentage of beta and alpha cell numbers revealed no significant differences between the ND and HFD islets (Fig. 2B,C), with chronic administration of GIP(3− 42) and GLP-1(9− 36) in HFD mice also having no impact on these parameters (Fig. 2B,C).However, 60 days twice daily treatment with Ex(9− 39) did reduce (p<0.05)beta cell number and increase (p<0.05)alpha cell number when compared to ND mice (Fig. 2B,C).Although the percentage of islets with centrally stained glucagon-positive cells appeared slightly increased with HFD when compared to ND mice, this failed to reach significance (Fig. 2D).In addition, all treatment interventions had no significant impact on the percentage of islets with centrally located alpha cells (Fig. 2D).

Discussion
DPP-4 rapidly metabolizes native GLP-1 and GIP to GLP-1(9− 36) and GIP(3− 42), which represent the major circulating forms of the two incretin hormones [44].The biological significance of these metabolites is uncertain, although early work deemed them to be biologically inactive.Subsequent studies demonstrated their potential as weak antagonists at their respective receptors, although physiological significance appeared unlikely [23,34,45].In the present study, twice daily administration of high doses of the truncated metabolites GIP (3− 42) and GLP-1(9− 36) were examined in HFD-induced obese mice.Such HFD-fed animals are an established model of insulin resistance and can reproduce human obesity associated with development of impaired glucose homeostasis [46], although there are still obvious differences in pancreatic islet morphology between rodents and humans [47].In addition, impact of the GLP-1R antagonist Ex(9− 39) was also assessed in the same context [9], although this compound possibly also inhibits GIPR function [41].
As would be expected, HFD mice presented with increased body weight, non-fasting plasma glucose and overt glucose intolerance.Twice-daily administration of GIP (3− 42), GLP-1(9− 36) or Ex(9− 39) did not induce any significant alterations of these metabolic parameters in HFD mice.In good agreement with this, others have also found no obvious physiological alterations to metabolism following sustained injection of the incretin metabolites, or Ex (9− 39), in obese diabetic ob/ ob mice [9].This finding could also reflect the relatively large pool of 'resting' alpha-and beta-cells that has been observed previously [48,49].While previous research from the Habener laboratory has reported the ability of GLP-1(9− 36) to inhibit weight gain and diabetes development in diet-induced obese mice [25], these studies employed 60% high fat feeding and continuous peptide infusion.This could suggest effects of incretin metabolites on metabolism are dose-dependent or effected by continuous versus intermittent receptor interaction.In that respect, normal feeding patterns in humans would lead to natural diurnal fluctuations in exposure to incretins and their metabolites [50], similar to the current treatment regimen.
At the level of the endocrine pancreas, HFD mice presented with the anticipated adaptive response to insulin resistance resulting in enlarged islets as well as increased alpha and beta cell area, to help maintain near normal levels of glycemia [42,51].This effect was apparent despite an increase of beta cell death and decrease in cellular proliferation at the end of the study.However, our findings are in good agreement with previous observations, that found prolonged exposure to HFD to be associated with a prompt decline of beta cell proliferation and glucose tolerance [52].
GIP(3− 42) significantly altered islet morphology in HFD mice by decreasing islet, beta and alpha cell areas, restoring these parameters towards levels in ND mice.While the number of cells within islets remained unchanged, GIP(3− 42) exerted protective effects on beta cells by increasing proliferation and decreasing apoptosis.In contrast to our findings, Parker et al. found no changes in islet area or number with GIP (3− 42) treatment, a disparity that may be attributed to their use of the ob/ob mouse model and the brief once-daily administration of GIP (3− 42) for 14 days [9], as opposed to the twice-daily administration and 60-day duration utilized in our investigation.Existing research on GIP (3− 42) strongly supports the idea that it might not antagonize the GIPR [9,36], and can induce mild improvements in insulin sensitivity [9], as observed with GIPR agonism [53].GIP (3− 42) treated HFD mice also presented with a higher percentage of small islets, as observed in ND mice, that could suggest islet neogenesis.Furthermore, there is now substantial evidence [54] supporting early observations that both agonism and antagonism of the GIPR can prove beneficial to metabolic health [55].This complexity poses a challenge when attempting to extrapolate the actions of the native peptide to its metabolite and draw conclusive findings.
For GLP-1(9− 36), we also employed the well-established antagonist Ex (9− 39) in an attempt to gain a more comprehensive understanding of its effects within islets.Surprisingly, Ex(9− 39) acted in a similar manner to GIP (3− 42) in reducing islet, beta and alpha cell areas, similar to that observed in ND mice.Thus, whilst it has been previously demonstrated that Ex(9− 39) possesses GIPR antagonistic actions, at least in animals [41], this molecule is generally considered as a specific GLP-1R antagonist [56].Considering the similarities of 60-days Ex(9− 39) or GIP (3− 42) treatment on the endocrine pancreas in HFD mice, it may be that the original observations of GIPR inhibitory effects of Ex(9− 39) deserve more consideration [41].In harmony with previous work in ob/ob mice, sustained treatment with Ex(9− 39) had no obvious impact on metabolism [56,57].However, Ex(9− 39) decreased beta cell while increasing alpha cells, as well as numbers of smaller sized islets, retuning these parameters towards levels seen in ND mice, which correlated with increased beta cell death.It is well known that Ex(9− 39) disrupts anti-apoptotic properties of GLP-1 [58] and stimulates glucagon secretion [59].However, the metabolite GLP-1(9− 36) had minimal effects on beta cells as their area remained comparable to HFD group.This could be attributed to the fact that the actions of the GLP-1 metabolite are suggested to predominantly occur through receptors other than the GLP-1R [24,29], likely due to its low GLP-1R binding affinity.In agreement, a glucose-lowering effect of GLP-1(9− 36) in human volunteers was shown to be independent of alterations in islet hormone secretion [27].
Unlike GIP (3− 42), GLP-1(9− 36) selectively reduced alpha cell area in HFD mice with no change in the number of alpha cells.This agrees with previous studies that report an inhibitory effect of GLP-1(9− 36) on glucagon secretion in-vitro and in-vivo, possibly via action at glucagon receptors [24,28].Further, we have also shown that GLP-1R agonism can reduce alpha cell area in HFD-fed mice, irrespective of alterations to beta cells [43].However, the possibility of the metabolite's interactions with GLP-1R cannot be dismissed, as only GLP-1(9− 36) led to increased islet density in the pancreas of HFD mice, being typically ascribed to GLP-1's capability to promote islet neogenesis through beta cell proliferation [60], although there was no obvious increase in the number of smaller sized islets in these mice when compared to HFD controls.That said, there were no observations of alterations in beta cell proliferation within the GLP-1(9− 36) group.We did however detect a rise in beta cell apoptosis with GLP-1(9− 36), providing support for the notion that it might indeed act through the GLP-1R to enhance the advantageous functions of native GLP-1 long after its degradation, as receptor engagement can be tissue specific [61].In accord with this, isolated human islets treated with exenatide exhibit enhanced glucose sensitivity and decreased apoptosis [62,63], but whether this extends to T2DM patients treated with exenatide or related GLP-1R agonists still needs to be determined.

Conclusion
These findings contribute to the understanding of the biological activity of incretin metabolites and their possible roles within islets.It seems unlikely that they have significant impact on whole body metabolism.However, the present results indicate the metabolites, that circulate at relatively high concentrations and can be produced locally within islets by alpha cells following rapid peptide degradation by DPP-4, may play a role in the maintenance of islet composition via paracrine interactions.Further studies are required to assess interactions of incretin metabolites with GLP-1, GIP and glucagon receptors on the various islet cell types together with their possible contribution to the overall control of islet function.

Funding
These studies were supported by Diabetes UK RD Lawrence Fellowship grant to RCM and Ulster University strategic funding.

Declaration of Competing Interest
All authors declare no conflict of interest.

Data availability
Data will be made available on request.A. Sridhar et al.

Table 1
Target, host and source of primary and secondary antibodies employed for immunofluorescent imaging experiments.