Comparison of independent and combined effects of the neurotensin receptor agonist, JMV-449, and incretin mimetics on pancreatic islet function, glucose homeostasis and appetite control

A B S T A Background: Neurotensin receptor activation augments the biosctivity of glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP). JMV-449, a C-terminal neurotensin-like fragment with a reduced peptide bond, represents a neurotensin receptor agonist. Methods: The present study assessed the actions of JMV-449 on pancreatic beta-cells alone, and in combination with GIP and GLP-1. Further studies examined the impact of JMV-449 and incretin mimetics on glucose ho- meostasis and appetite control in mice. Results: JMV-449 was resistant to plasma enzyme degradation and induced noticeable dose-dependent insulin- releasing actions in BRIN-BD11 beta-cells. In combination with either GIP or GLP-1, JMV-449 augmented ( P < 0.05) the insulinotropic actions of both hormones, as well as enhancing ( P < 0.001) insulin secretory activity of both incretin peptides. JMV-449 also increased beta-cell proliferation and induced significant benefits on beta-cell survival in response to cytokine-induced apoptosis. JMV-449 (25 nmol/kg) inhibited ( P < 0.05 – P < 0.001) food intake in overnight fasted lean mice, and enhanced ( P < 0.01) the appetite supressing effects of an enzymatically stable GLP-1 mimetic. When injected co-jointly with glucose, JMV-449 evoked glucose lowering ac- tions, but more interestingly significantly augmented ( P < 0.05) the glucose lowering effects of established long-acting GIP and GLP-1 receptor mimetics. In terms of glucose-induced insulin secretion, only GIP receptor sig- nalling was associated with increases in insulin concentrations, and this was not enhanced by JMV-449. Conclusion: JMV-449 is a neurotensin receptor agonist that positively augments key aspects of the biological action profile of GIP and GLP-1. General significance: These observations emphasise the, yet untapped, therapeutic potential of combined neu- rotensin and incretin receptor signalling for diabetes.

Taken together, it appears that the biological activity of both incretin hormones could be enhanced through concurrent administration of neurotensin related peptides, which would have particular importance in the diabetes clinic. Indeed, the intracellular signalling pathways stimulated by GIP and GLP-1 receptor activation within pancreatic betacells are very similar, in that both are initially mediated by activation of a specific G-protein coupled receptor with subsequent increases of intracellular cAMP [11]. This results in comparable benefits on beta-cell function, growth and survival, but some variations in downstream signalling may account for small efficacy differences between the two peptides [11]. Unfortunately, a significant hindrance to the therapeutic applicability of neurotensin peptides relates to a short biological half-life [12]. However, metabolic stability of neurotensin can be improved through appropriate structural modification [10]. As such, introduction of a reduced peptide bond, CH 2 NH, between Lys 8 and Lys 9 in neurotensin(8-13)-like peptide, generates an enzymatically stable and biologically active peptide that acts as a powerful neurotensin receptor agonist [10]. Similarly, enhanced stability and bioactivity of the C-terminal hexapeptide of xenin has also been demonstrated through introduction of a reduced pseudopeptide bond between the two N-terminal amino acid residues, to yield Ψ-xenin-6 [13]. More interestingly, Craig and colleagues have recently confirmed that Ψ-xenin-6 significantly augments the bioactivity of GIP [6]. Furthermore, when Ψ-xenin-6 was administered to high fat fed mice in combination with the dipeptidyl peptidase-4 (DPP-4) inhibitor drug sitagliptin, clear additive antidiabetic effects were noted [14]. Together, these observations confirm the applicability of metabolically stable C-terminal hexapeptides of either neurotensin or xenin. Since the receptor activation profile of xenin and related fragment peptides is not yet fully elucidated [7], neurotensinbased peptides may offer a more therapeutically attractive option to augment the biological action of incretin hormones and improve their antidiabetic efficacy. In this regard, JMV-449 is a recognised biologically active neurotensin fragment peptide, with a reduced pseudopeptide CH 2 NH bond between Lys 8 and Lys 9 [10]. In the current study, we initially aimed to characterise the effects of JMV-449 at the level of the pancreatic betacell. Further studies examined the ability of JMV-449 to positively interact with GIP and GLP-1 in terms of both in vitro pancreatic beta-cell induced insulin secretion, as well as glucose homeostasis, insulin secretion and appetite control in mice.

In vitro beta-cell proliferation and apoptosis
BRIN-BD11 beta-cells were used to investigate effects of JMV-449 (10 − 8 and 10 − 6 M) on beta-cell proliferation and protection against cytokine-induced apoptosis. GLP-1 was employed as a positive control for all studies. Ki-67 immunostaining was used to assess effects on proliferation. Briefly, cells were seeded onto coverslips (40,000 cells per coverslip) and cultured overnight (18 h;37 • C), in the presence of test peptides (10 − 8 and 10 − 6 M). Cells were then washed with PBS, and fixed using 4% paraformaldehyde. Following antigen retrieval with citrate buffer (90 • C for 20 min), tissues were blocked using 1.1% BSA for 30 min. Cells were then incubated with Ki-67 primary antibody (1:500; Abcam, ab15580), followed by Alexa Fluor® 488 secondary antibody (1:400, Invitrogen, A-11008). Coverslips were washed with PBS, mounted on slides for viewing using a fluorescent microscope (Olympus System Microscope) and photographed by DP70 camera adapter system. Proliferation frequency was expressed as percentage of total cells analysed. For analysis of the ability of JMV-449 to protect against cytokineinduced apoptosis, cells were seeded as above. However, cells were also exposed to a cytokine-cocktail (IL-1β 100 U/mL, IFN-γ 20 U/mL, TNF-α 200 U/mL, all purchased from Sigma-Aldrich) in the presence or absence of test peptides (10 − 8 and 10 − 6 M) for 18 h, with hydrogen peroxide as an additional control. TUNEL staining (Roche Diagnostics Ltd., UK) was employed to quantify beta-cell apoptosis, as previously described [18]. Apoptosis was expressed as percentage of total cells analysed. Approximately 150 cells were analysed per group.

Animals
Animal studies were carried out using male C57BL/6 mice (10 month old, Envigo Ltd., UK), all housed individually in an air-conditioned room at 22 ± 2 • C with a 12 h light: 12 h dark cycle. Animals were maintained on standard rodent chow diet (10% fat, 30% protein and 60% carbohydrate, Trouw Nutrition, UK) with ad libitum access to diet and water. All animal experiments were carried out in accordance with the UK Animal Scientific Procedures Act 1986 and approved by Ulster University Animal Welfare and Ethical Review Body (AWERB).

Acute effects on food intake, glucose tolerance and insulin secretion in mice
For in vivo studies, JMV-449 as well as fully characterised enzymatically stable forms of GIP and GLP-1 were employed, namely (D-Ala 2 ) GIP and exendin-4 [19,20]. For assessment of appetite control, cumulative food intake was assessed in overnight fasted (18 h) mice following i.p. injection of saline vehicle (0.9% w/v NaCl) or test peptide(s) (at either 2.5 or 25 nmol/kg bw, as appropriate), with food intake measured at 30 min intervals for 180 min. For assessment of glucose tolerance, blood glucose and plasma insulin concentrations were determined immediately prior to and 15, 30, 60 and 105 min after i.p. injection of glucose alone (18 mmol/kg bw) or in combination with test peptide(s) (each at 25 nmol/kg bw) in 18 h fasted mice.

Biochemical analysis
Blood samples were collected from the cut tip on the tail vein of conscious mice into chilled fluoride/heparin glucose micro-centrifuge tubes (Sarstedt, Numbrecht, Germany). Blood glucose was measured directly using a Contour blood glucose meter. For plasma insulin analysis, blood samples were collected into chilled fluoride/heparin glucose micro-centrifuge tubes (Sarstedt, Numbrecht, Germany) and immediately centrifuged using a Beckman microcentrifuge (Beckman Instruments, Galway, Ireland) for 1 min at 13,000g and stored at − 20 • C prior to insulin RIA [17].

Statistical analysis
Statistical analysis was completed using GraphPad PRISM (Version 5). Results are expressed as means ± SEM and data compared using repeated measures ANOVA followed by Student-Newman-Keuls post-hoc test. Unpaired Student t-test was used where appropriate. Incremental plasma insulin and glucose area under the curve (AUC) were calculated using the trapezoidal rule with appropriate baseline subtraction. Groups of data were considered significant if P < 0.05.

Peptide characterisation and plasma enzyme stability
MS analysis confirmed identity of JMV-449 (H-LYS-ψ -LYS-PRO-TYR-ILE-LEU-OH), with theoretical and experimental masses differing by less than 0.5 Da ( Table 1). The HPLC profile for JMV-449 confirmed 99.4% purity as stated by the manufacturer (data not shown). When incubated with murine plasma, JMV-449 remained 100% intact up to 8 h (Table 1).

Acute in vivo effects on glucose tolerance and insulin secretion
Administration JMV-449 in combination with glucose to lean mice resulted in significantly (P < 0.05 to P < 0.01) decreased individual plasma glucose levels at 15 and 30 min post-injection when compared to glucose alone control, but this did not translate to significant reductions in 0-105 min AUC values (Table 2). However, plasma insulin AUC was significantly (P < 0.05) increased by JMV-449, despite lack of effect on individual glucose-stimulated insulin concentrations ( Table 3). As expected, (D-Ala 2 )GIP evoked significant glucose homeostatic and insulin secretory actions in mice (Tables 2 & 3). Notably, the overall glucoselowering effect of (D-Ala 2 )GIP was significantly (P < 0.05) enhanced by JMV-449 (Table 2). An essentially similar biological action profile was noted when JMV-449 was co-administered with exendin-4, with slightly less prominent effects on glucose-induced insulin secretion (Tables 2 & 3). As such, both compounds had impressive glucoselowering actions (Table 2), with JMV-449 significantly enhancing the glucose homeostatic action of exendin-4 (Table 2). Interestingly, whilst combined administration of (D-Ala 2 )GIP and exendin-4 exerted pronounced effects to reduce blood glucose levels (Table 2), this was not associated with obvious benefits on insulin secretion (Table 3). JMV-449 was unable to enhance the marked positive effects of combined (D-Ala 2 ) GIP and exendin-4 injection (Tables 2 & 3).

Table 3
Effects of JMV-449 alone, and in combination with (D-Ala 2 )GIP or exendin-4, on plasma insulin concentrations in mice.  Plasma insulin concentrations were measured immediately before and 15, 30, 60 and 105 min after i.p. injection of glucose alone (18 mmol/kg bw) and test peptides (each at 25 nmol/kg bw) in 18 h fasted mice. Plasma insulin AUC values for 0-105 min post injection are also presented. Values represent mean ± SEM for 6 mice. *P < 0.05 and **P < 0.01 compared to glucose alone. related diabetes have not been fully translated to the clinic, this is of significant therapeutic interest. In that regard, bariatric surgery appears to be the only intervention that leads to sustained weight loss and potential remission of type 2 diabetes in humans [22], although recent observations using once-weekly subcutaneous injection of the GLP-1 mimetic semaglutide in obese men does offer some encouragement [23]. Interestingly, the most effective forms of bariatric surgery are associated with dramatic changes in the secretion and action of various gut-derived hormones, which is believed to play a major role in promotion of almost immediate resolution of diabetes post-surgery [24]. Several studies demonstrate increased circulating neurotensin concentrations following bariatric surgery in both rodents and humans [3,12,25,26], with surgical-induced modulation of GIP and GLP-1 levels already well accepted [24]. Thus, interplay between incretin and neurotensin receptor signalling likely has an appreciable contribution to the overall observed benefits of bariatric surgery. However, it should also be acknowledged that very low calorie diets do not alter circulating incretin hormone levels but can result in a relatively rapid remission of type 2 diabetes [27], although prolonged adherence to such strict diets appears to be particularly challenging [28]. As expected, JMV-449 was completely resistant to plasma enzymatic degradation, likely as a consequence of the reduced peptide bond present between Lys 8 and Lys 9 [10]. In keeping with the biological actions and insulin secretory profile of neurotensin receptor activation [29], JMV-449 evoked clear dose-dependent increases of insulin secretion from rodent BRIN-BD11 cells at basal glucose concentrations, which were less evident under hyperglycaemic conditions. Thus, this slight variability of insulin secretory profile is in complete harmony with established bioactivity of neurotensin-like peptides on pancreatic betacell function [29]. Risk of hypoglycaemia with JMV-449 would be unlikely given that activation of neurotensin receptors is known to induce glucagon secretion at low glucose concentrations [30], but this would need to be confirmed for JMV-449. Moreover, JMV-449 imparted independent benefits on beta-cell growth and survival [31], with potential translational benefits in terms of restoration of beta-cell mass and function in diabetes that require further study. These positive effects on pancreatic beta-cells are very reminiscent of the actions of both incretin hormones, GIP and GLP-1 [32]. Indeed, similar to observations with GIP and GLP-1 [33], neurotensin, and the related hormone xenin that likely functions in part as a neurotensin receptor agonist [8], are evidenced locally within the pancreas and believed to impart important autocrine or paracrine effects on islet function [31]. As such, we observed clear potentiating actions of JMV-449 on GIP-and GLP-1-induced insulin secretion at basal glucose concentrations, with similar effects previously reported following combined treatment of GIP with xenin [4][5][6][7][34][35][36]. Indeed, these positive GIP/xenin interactions ultimately led to development of a unimolecular dual-acting GIP/xenin hybrid peptide [37], which displays significant and sustained antidiabetic effects in rodent models of diabetes [38]. However, it should be noted that the incretin hormone potentiating actions of JMV-449 were much less obvious at elevated 16.7 mM glucose concentrations in the current setting.
In full harmony with in vitro findings, JMV-449 possessed prominent glucose homeostatic actions in lean healthy mice that were directly linked to elevated plasma insulin concentrations. Importantly, positive additive benefits were noted in terms of glycaemic status when JMV-449 was administered together with stable GIP or GLP-1 mimetics, although this was not associated with augmentation of glucose-induced insulin secretion. This is interesting, and could relate to reduced insulinotropic effectiveness of neurotensin at elevated glucose levels [29]. However, additive benefits of neurotensin and incretin hormones to improve glucose disposal could point towards insulin-independent glucoselowering actions. In this respect, GIP, GLP-1 and neurotensin all exert established, but somewhat distinctive, effects on glucagon secretion [12,39,40], which could be a factor here. In addition, upregulated incretin hormone signalling exerts well described extrapancreatic glucose-lowering actions [39,41], which may be augmented by neurotensin signalling. Both avenues are complex and as such require additional detailed study that is outside the scope of the current investigation. For example, recent studies with GIP and xenin suggest that neurotensin receptor activation can partially reverse the glucosedependent glucagonotropic action of GIP [6], whereas the impact of neurotensin on the glucagonostatic action of GLP-1 is unknown.
Neurotensin receptor activation exerts well documented effects to suppress appetite [42], with similar actions noted with GLP-1 [43] and more recently GIP [44]. These combined actions would be highly favourable in obesity-driven forms of diabetes, such as type 2 diabetes mellitus. Our studies confirmed inhibitory effects on food intake following individual administration of JMV-449, (D-Ala 2 )GIP or exendin-4 at a dose of 25 nmol/kg. Receptors for GIP, GLP-1 and neurotensin are located within the paraventricular nuclei region of the hypothalamus, known to critical for energy regulation and metabolism [45][46][47], and likely mediates these effects. The excellent efficacy of exendin-4 alone at a dose of 25 nmol/kg precluded any conceivable additive appetite suppressive effects. However, with a reduced dose of exendin-4 there was good evidence for JMV-449 additive benefits on appetite reduction in mice. As such, it has been suggested that subthreshold doses of GLP-1 are required to induced appetite and weight loss synergy with neurotensin [3]. JMV-449 was also unable to enhance (D-Ala 2 )GIP mediated reductions in food intake, which may be peptide dose and model related. However, studies with GIP and xenin in rodents also suggest that unlike effects of glucose homeostasis and insulin secretion, these peptides lack additive positive actions on appetite control [21].
Although potential benefits of combined incretin and neurotensin signalling have been previously suggested [3,5], this is the first study to directly evaluate possible interplay between both incretin hormones and neurotensin. It is reassuring to note that JMV-449 evoked clear in vitro insulinotropic and in vivo glucose-lowering potentiating benefits in combination with GIP and GLP-1, suggesting no obvious threshold for these benefits. Appropriate dose adjustment within the in vivo series of experiments may have allowed for observation of further additive benefits. With a view towards possible clinical application of our findings, the short sequence of JMV-449 is highly attractive, making it cheap and relatively easy to synthesise, as well as potentially allowing a noninjectable peptide drug administration [48]. In that regard, recent approval of oral delivery of the GLP-1 mimetic semaglutide [49], as well as confirmed oral bioavailability for DPP-4 inhibitor drugs [48], is encouraging. Moreover, advances in regulatory peptide drug design and development, leading to creation of functional unimolecular triple agonists, represents another potential route to clinical realisation of the benefits of combined GIP, GLP-1 and neurotensin receptor signalling.
In conclusion, these data establish that JMV-449 is an enzymatically stable and bioactive hexapeptide analogue of neurotensin, which possesses notable positive effects on pancreatic beta-cells. Furthermore, we clearly demonstrate the ability of JMV-449 to significantly potentiate the glucose homeostatic and insulin releasing actions of the incretin hormones, GIP and GLP-1. Further consideration is required to fully assess the clinical potential of upregulated neurotensin signalling in combination with approved incretin-based therapeutics.

Author contributions
NI, VAG and GH conceived/designed the study. NI and SLC drafted the manuscript. SLC and CES participated in the conduct/data collection and analysis and interpretation of data. All authors revised the manuscript critically for intellectual content and approved the final version of the manuscript.

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