Characterization of combined linagliptin and Y2R agonist treatment in diet-induced obese mice

Dipeptidyl peptidase IV (DPP-IV) inhibitors improve glycemic control by prolonging the action of glucagon-like peptide-1 (GLP-1). In contrast to GLP-1 analogues, DPP-IV inhibitors are weight-neutral. DPP-IV cleavage of PYY and NPY gives rise to PYY3-36 and NPY3-36 which exert potent anorectic action by stimulating Y2 receptor (Y2R) function. This invites the possibility that DPP-IV inhibitors could be weight-neutral by preventing conversion of PYY/NPY to Y2R-selective peptide agonists. We therefore investigated whether co-administration of an Y2R-selective agonist could unmask potential weight lowering effects of the DDP-IV inhibitor linagliptin. Male diet-induced obese (DIO) mice received once daily subcutaneous treatment with linagliptin (3 mg/kg), a Y2R-selective PYY3-36 analogue (3 or 30 nmol/kg) or combination therapy for 14 days. While linagliptin promoted marginal weight loss without influencing food intake, the PYY3-36 analogue induced significant weight loss and transient suppression of food intake. Both compounds significantly improved oral glucose tolerance. Because combination treatment did not further improve weight loss and glucose tolerance in DIO mice, this suggests that potential negative modulatory effects of DPP-IV inhibitors on endogenous Y2R peptide agonist activity is likely insufficient to influence weight homeostasis. Weight-neutrality of DPP-IV inhibitors may therefore not be explained by counter-regulatory effects on PYY/NPY responses.

www.nature.com/scientificreports/ In vitro assessment of linagliptin effect on NPY and PYY degradation. Human plasma was used as source of endogenous DPP-IV. Healthy adult individuals were enrolled after they provided written informed consent. The local Ethics Committee at the Hannover Medical School approved the protocol. All methods were carried out by observing the applicable legal provisions, including data protection regulations, as well as the principles of medical and professional ethics as laid down in the Declaration of Helsinki issued by the World Medical Association and the ICH/EU recommendations Note for Guidance on Good Clinical Practice (GCP recommendations). Blood samples were collected from the cubital vein into blood collection tubes (EDTA-plasma). Immediately after withdrawal, plasma was separated from cells by a two-step centrifugation procedure (10 min at 2000 × g followed by 15 min at 2500 × g, both centrifugation steps at room temperature). Plasma aliquots were transferred into 2 ml vials and stored − 80 °C. 30 pmol of human PYY  or NPY  was incubated in 85 µl PBS, and reaction was started by adding 5 µl plasma with or without 1 µM linagliptin and incubated for 8 h (NPY) or 18 h (PYY) at 37 °C. Parent NPY, PYY and corresponding fragments were determined by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) as described previously 34 .
In vitro assessment of PYY 3-36 analogue effect on NPY and PYY degradation. NPY and PYY were incubated in human EDTA-plasma in presence or absence of linagliptin and/or Y2R-selective PYY  analogue and subsequently subjected to MALDI-TOF-MS analysis to determine the rate of N-terminal cleavage by DPP-4. PYY (30 pmol) and NPY (30 pmol) were solubilised in PBS and added to 5 µl human plasma with or without linagliptin (1 µM) and/or Y2R-selective PYY  analogue at different concentrations ,000 fmol). The plasma was then incubated for 18 h at 37 °C. Subsequently the reaction was stopped, prepared via solid-phase extraction and subjected to mass spectrometry. After measurement, signal intensities were exported and conversion rates for NPY and PYY were expressed as product/(product + substrate). Statistical significance against controls was tested using Welch's t-test. Each incubation was carried out in duplicate and measured in duplicate at two different concentrations (n = 8). Data were pooled from three individual experiments.
Statistics. Data

Results
Linagliptin promotes marginal weight loss while significantly increasing circulating active GLP-1 and GIP levels. In order to obtain maximal DPP-IV inhibition, DIO mice were administered linagliptin using a dose of 3 mg/kg/day (s.c.). Linagliptin monotreatment was initially characterized for effects on key metabolic parameters in DIO mice. Compared to vehicle controls, linagliptin promoted a marginal weight loss (2.8%, p < 0.05) following 14 days of treatment (Fig. 1A,B). Food intake was essentially unaffected by linagliptin treatment (Fig. 1C,D). Linagliptin treatment once daily resulted in high linagliptin exposure (~ 200 nM), as measured in plasma samples before (approximately 24hrs after the previous dose) and 4 h after administration of the last dose on treatment day 14 ( Fig. 2A). At both sampling times, almost complete inhibition of plasma DPP-IV catalytic activity was observed (Fig. 2B,C). Correspondingly, linagliptin treatment significantly elevated plasma levels of active GLP-1 (12-fold increase) and GIP (8-fold increase), as measured 4 h post-dosing ( Fig. 2D,E). Terminal plasma insulin levels were unaffected by linagliptin treatment (Fig. 2F). Except for significantly reduced resistin and TNF-α concentrations, other plasma markers, including total PYY and pancreatic polypeptide, were unaffected by linagliptin treatment (Supplementary Fig. 1).

Linagliptin does not augment Y2R agonist-induced food intake inhibition and weight loss.
To investigate potential implications of linagliptin-induced bidirectional regulation of GLP-1 and neuropeptide Y-family peptides, we characterized the effect of a Y2R-selective agonist in DIO mice with or without co-administration of linagliptin. Weight curves are depicted in Fig. 4A Equivalent improvement in oral glucose tolerance following linagliptin monotreatment and combined linagliptin-Y2R agonist administration. Oral glucose tolerance was determined one week prior to treatment start (day − 7) and on treatment day 12. On day − 7, fasted blood glucose levels (t = 0) were similar in all DIO mouse groups (group mean 9.1-9.7 mM, p > 0.05) as well as glucose excursion curves and AUC-glucose values (Fig. 5A,B). Compared to day -7, fasted blood glucose levels were significantly lower in all treatment groups (p < 0.05, paired t-test), including vehicle controls, on treatment day 12 (group mean 6.9-8.2 mM, Fig. 5C). Compared to vehicle dosing, linagliptin significantly improved glucose tolerance. Y2R agonist treatment also improved glucose tolerance, albeit with less robust effect on glucose excursions. Improvements in glucose tolerance following combined linagliptin and Y2R agonist administration were superior to Y2R agonist, but not linagliptin, monotreatment (Fig. 5C,D).

Linagliptin-induced DPP-IV inhibition is unaffected by co-administration of a Y2R-selective agonist.
The selected dose of linagliptin (3 mg/kg) completely suppressed plasma DPP-IV activity as measured 4 h post-dosing on the last dosing day (day 14, Fig. 6A). All linagliptin-treated groups showed equally elevated plasma levels of active GLP-1 and GIP on treatment day 14 (Fig. 6C,E). Residual DPP-IV inhibitory activity was detected 6 days after the last linagliptin dose (day 20), being most consistent for linagliptin-Y2R agonist co-treatment groups (Fig. 6B). In the treatment combination groups, this was also reflected by slight, however significantly, increased active GLP-1, but not GIP, levels on day 20 (Fig. 6D,F). The Y2R agonist had no effect on DPP-IV activity as well as active GLP-1 and GIP levels. Treatments did not influence plasma insulin levels ( Fig. 6G). Because linagliptin and Y2R agonist co-administration demonstrated stronger inhibitory effects on DPP-IV activity as compared to linagliptin administration alone, we investigated whether the Y2R agonist influenced DPP-IV activity in human EDTA-plasma. In the low fmol range (30-300 fmol), the Y2R agonist www.nature.com/scientificreports/ slightly reduced NPY conversion rate. Also, linagliptin and the Y2R agonist showed additive inhibitory effects on NPY conversion rate (Fig. 6H). PYY conversion rate was also slightly reduced by the Y2R agonist alone (30-1000 fmol), but unaffected by combined Y2R agonist and linagliptin application (data not shown). At higher Y2R agonist amounts (≥ 1000 fmol), the inhibitory effect of NPY/PYY conversion inhibition was diminished or absent, which was ascribed to reduced solubility of the Y2R agonist.

Discussion
The present study characterized the metabolic effects of linagliptin with or without co-administration of a Y2Rselective PYY 3-36 analogue in DIO mice. The linagliptin dose (3 mg/kg, s.c.) promoted complete plasma DPP-IV inhibition (> 98%) maintained throughout the dosing interval. The subcutaneous dosing regimen results in more pronounced DPP-IV inhibition as compared to oral dosing in DIO mice (~ 80% inhibition) 28 . In the clinic, ≥ 80% DPP-IV inhibition by linagliptin administration is sufficient to improve glucose tolerance 35 . Sustained increases in circulating incretin levels were observed in DIO mice treated with linagliptin. Although both linagliptin and the Y2R-selective PYY 3-36 analogue improved glucose tolerance and Y2R receptor agonist treatment promoted significant weight loss, no synergistic metabolic effects were observed following combined treatment. Therefore, our study in DIO mice indirectly suggests that weight-neutrality of DPP-IV inhibitors is not explained by reduced endogenous Y2R agonist action, and lends further support to the notion that increasing circulating GLP-1 levels by DPP-IV inhibition is insufficient to promote satiety responses. DPP-IV inhibition is a therapeutic option to extend circulating half-life of GLP-1 and GIP and improve glycemic control in patients with T2D. DPP-IV inhibitors are generally weight-neutral 18 . Our preclinical study www.nature.com/scientificreports/ corroborates end extends previous studies on linagliptin administration in DIO mice 28,30,36 , by demonstrating that linagliptin induces marginal weight loss when applying a dosing regimen that promotes sustained linagliptin exposure, nearly complete inhibition of plasma DPP-IV activity and ≥ 10-fold increases in plasma GLP-1 and GIP levels. The mild weight loss attained by linagliptin administration in DIO mice may potentially associate to beneficial effects on white adipose tissue mass, liver fat and thermogenesis 28,30 .
Weight-neutrality of DPP-IV inhibitors contrasts the weight loss efficacy of long-acting GLP-1 analogues which are DPP-IV resistant and have slow elimination kinetics 37,38 . It has therefore been speculated that DPP-IV inhibitors are weight-neutral by regulating the biological activity of other peptide hormones involved in energy homeostasis 11 . NPY-family peptides may be potential candidates for promoting metabolic counter-regulatory responses to DPP-IV inhibitor treatment. DPP-IV shows greater affinity for PYY and NPY at physiological concentrations as compared to GLP-1 and GIP 20 . Studies on DPP-IV substrates other than GLP-1 and GIP are limited. Because reliable techniques are lacking to distinguish between intact and cleaved PYY and NPY immunoreactive forms there is little information on the distribution of circulating PYY and NPY cleavage products following DPP-IV inhibitor treatment. We therefore used MALDI-TOF-MS to determine PYY/NPY isoforms in human plasma following incubation with synthetic human full-length NPY and PYY. Linagliptin prevented N-terminal cleavage of PYY and NPY resulting in substantial lowering of PYY 3   www.nature.com/scientificreports/ of DPP-IV mediated PYY cleavage is further emphasized by the absence of anorectic effect of exogenous PYY  in rats carrying an inactivating DPP-IV point mutation 46 . Sitagliptin has been demonstrated to increase plasma PYY 1-36 :PYY 3-36 ratios in T2D patients, indicating that PYY is a physiological substrate for DPP-IV 47 . NPY is a neuronal-derived peptide that regulates many aspects of energy balance, including feeding behavior end energy expenditure 48 . NPY 1-36 is an orexigenic neuropeptide which acts by stimulating hypothalamic Y1R signaling 25 . The anorexigenic actions of PYY 3-36 are thought partially mediated by inhibiting arcuate NPY neuronal activity 49 .
The N-terminal tyrosine residue of NPY/PYY is critical for efficient Y1R binding 50 . As for PYY  , NPY 3-36 has markedly reduced Y1R affinity but retained anorexigenic Y2R agonist activity 51 . In contrast to NPY, C-terminally truncated PYY variants were detected in human plasma which supports the notion that PYY is a substrate for several endopeptidases 52,53 . C-terminal cleavage renders the peptide Y2R inactive 54   www.nature.com/scientificreports/ it may be hypothesized that DPP-IV inhibitors are weight-neutral because lowering of circulating PYY  and NPY 3-36 levels shifts PYY/NPY activity towards Y1R signaling. As a consequence, food intake inhibitory responses subsequent to DPP-IV induced increases in active GLP-1 levels could be counterbalanced by lowered Y2R anorectic activity relative to Y1R orexigenic activity. We therefore sought to determine if pharmacological stimulation of Y2R activity could unmask potential contributory weight-lowering effects of linagliptin in DIO mice. The Y2R-selective agonist (compound 38) was selected based on a patent, demonstrating that compound 38 was one of the most efficacious PYY 3-36 analogues on food intake inhibition in db/db mice 31 . In the current study, food intake suppression was only observed during the first days of Y2R agonist administration while body weight continued to drop over the 14-day treatment period. Sustained weight loss by Y2R agonist treatment may be potentially explained by contributory effects from increased lipolysis and energy expenditure 55,56 . Linagliptin co-administration did not enhance the anorectic response or weight loss attained by Y2R agonist treatment. Following cessation of drug treatment, DIO mice showed compensatory overeating and gradually resumed    www.nature.com/scientificreports/ baseline body weight. The rate of weight regain was different in linagliptin and Y2R agonist treated mice which is likely explained by different pharmacokinetics. Co-administration of linagliptin and the Y2R agonist resulted in slightly stronger inhibitory effects on DPP-IV activity as compared to linagliptin administration alone. This effect likely explains the slightly higher plasma active GLP-1 levels observed in DIO mice receiving combination treatment. Additive DPP-IV inhibitory effects of linagliptin and the PYY 3-36 analogue were confirmed in vitro. Because NPY was added in 100-fold excess in comparison to the PYY 3-36 analogue, substrate inhibition seems less plausible. Therefore, future studies must aim to address the underlying mechanism for DPP-IV inhibition conferred by the PYY 3-36 analogue. Our findings contrast mounting evidence from pharmacological studies suggesting additive/synergistic effects of GLP-1 7-36 and PYY  . In infusion studies, the combination of PYY 3-36 and GLP-1 confers stronger suppression of food intake and hunger sensation compared to monotreatment in healthy and obese subjects [57][58][59] . Correspondingly, preclinical studies have demonstrated synergistic anorectic effects and robust weight loss following administration of a GLP-1 analogue in combination with native PYY  or PYY 3-36 analogue treatment [60][61][62][63] . Compared to the marked metabolic effects of pharmacological administration of GLP-1 and PYY 3-36 , weightneutral metabolic effects of DPP-IV inhibitors may therefore result from the modest changes in circulating levels of peptide hormones involved in appetite regulation and weight homeostasis. Both linagliptin and the Y2R agonist significantly improved oral glucose tolerance in DIO mice. As for body weight, combined linagliptin and Y2R agonist treatment showed no additive glycemic effects. In agreement, PYY 3-36 lacks insulinotropic properties and does not enhance insulin responses to GLP-1 7-36 administration 64 . Therefore, Y2R agonists are thought to improve glucose tolerance by increasing peripheral insulin sensitivity as result of weight loss 65 .

Conclusion
We report that linagliptin does not enhance anorectic and weight loss responses to Y2R agonist treatment in DIO mice. DPP-IV inhibitors may reduce circulating levels of PYY 3-36 and NPY 3-36 , however, any negative modulatory effect on Y2R activity may likely not contribute to shape the therapeutic profile of DPP-IV inhibitors. Our study lends support to the notion that weight neutrality of DPP-IV inhibitors is determined by active GLP-1 levels being insufficiently increased to elicit satiety responses.