To the Editor: Many beta cell growth factors have been identified, one of the most promising being glucagon-like peptide-1 (GLP-1), a peptide secreted from intestinal endocrine L-cells in response to nutrient ingestion. GLP-1 is rapidly cleaved and inactivated in vivo by dipeptidyl peptidase-4 (DPP-4), a ubiquitous serine protease. DPP-4 inhibitors (DPP-4i) have been shown to raise circulating levels of active GLP-1 and thus increase and maintain effective concentrations of this peptide reaching target tissues [1]. Studies in humans with type 2 diabetes have shown that DPP-4i therapy improves glucose tolerance [2]. In addition, DPP-4i treatments can preserve pancreatic beta cell mass in animal models of type 2 diabetes [3] and stimulate beta cell regeneration in streptozotocin-induced diabetic rats [4].

Regarding deficits in pancreatic beta cell mass resulting from autoimmune (type 1) diabetes, the GLP-1 analogue, exendin-4, reversed diabetes in NOD mice; however, this required additional immunosuppressive therapy [5]. Addition of the gastrointestinal peptide hormone, gastrin, to GLP-1 therapy increased pancreatic beta cell mass and restored normoglycaemia in diabetic NOD mice without the use of immunosuppressants [6]. The objective of this study was to determine whether combination therapy with a DPP-4i to raise endogenous levels of GLP-1, together with a proton pump inhibitor (PPI) to raise endogenous levels of gastrin, could reverse diabetes in NOD mice. Here, we report that such a combination therapy increased circulating levels of GLP-1 and gastrin in acutely diabetic NOD mice, and that pancreatic insulin content, insulin secretion and normoglycaemia were restored.

NOD female mice, age 6-8 weeks, were purchased from Taconic (Germantown, NY, USA). NOD-severe combined immunodeficient (NOD-scid) female mice were purchased from Jackson Laboratory (Bar Harbor, ME, USA). The mice were housed and fed under specific pathogen-free conditions and cared for according to the guidelines of the Canadian Council on Animal Care. Diabetes onset was diagnosed by non-fasting blood glucose ≥10 mmol/l. Acutely diabetic NOD mice (blood glucose 10–16 mmol/l) were allocated to four groups within 4–7 days after diabetes onset and treated for 12 weeks with (1) the DPP-4i vehicle, 5 mg/ml methylcellulose in H2O given by oral gavage once daily and the PPI vehicle, PBS given by s. c. injection twice daily; (2) 10 mg/kg of a selective DPP-4i, 1-{[(3-hydroxy-1-adamantyl)amino]acetyl}-2-cyano-(S)-pyrrolidine (Dalton Chemical Laboratories, Toronto, ON, Canada) given by oral gavage once daily; (3) 30 mg/kg of a PPI, pantoprazole (Nycomed, Oakville, ON, Canada) given by s. c. injection twice daily; and (4) 10 mg/kg DPP-4i and 30 mg/kg PPI.

Acutely diabetic NOD mice treated with vehicle became progressively more hyperglycaemic (≥25 mmol/l) and were killed within 6 weeks (Fig. 1a). Treatment with DPP-4i restored normoglycaemia in three of eight mice (38%) (Fig. 1b) and PPI was effective in three of nine mice (33%) (Fig. 1c). In contrast, the combination of DPP-4i and PPI restored normoglycaemia in six of eight mice (75%) (Fig. 1d). The plasma active GLP-1 level was increased threefold by DPP-4i demonstrating protection of GLP-1 from degradation by DPP-4 enzyme; this increase in GLP-1 was not changed by additional treatment with PPI (Fig. 1e). The serum gastrin level was significantly increased by PPI demonstrating the expected elevation of gastrin production; however, addition of DPP-4i treatment did not change the serum gastrin level achieved by PPI treatment (Fig. 1f). Treatments with DPP-4i and PPI given as single agents did not significantly increase plasma C-peptide or pancreatic insulin content compared with vehicle treatment, whereas combined DPP-4i and PPI treatments significantly increased plasma C-peptide (Fig. 1g) and pancreatic insulin content (Fig. 1h). In a follow-up study with more acutely diabetic NOD mice (blood glucose 10–16 mmol/l), we found that normoglycaemia (3.5–7.0 mmol/l) was restored in 17 of 21 mice (81%) after 9 weeks of treatment with the combination of DPP-4i and PPI, compared with four of 14 mice (27%) treated with DPP-4i alone (p < 0.01), four of 14 mice (27%) treated with PPI alone (p < 0.01) and zero of 15 mice treated with vehicle (p < 0.01); analysis was by one-way ANOVA and Dunnett's multiple comparison tests.

Fig. 1
figure 1

Acutely diabetic NOD mice were treated with vehicle (a), 10 mg/kg DPP-4i once daily (b), 30 mg/kg PPI twice daily (c) and both 10 mg/kg DPP-4i and 30 mg/kg PPI (d) for 12 weeks. Blood glucose concentrations were measured weekly and are shown for individual mice for 12 weeks or until blood glucose was ≥25 mmol/l and the mice were killed. Normoglycaemia (blood glucose 3.5–7.0 mmol/l) was determined in NOD-scid mice. e Mouse active GLP-1-(7-36)amide in plasma was measured by an ELISA electrochemiluminescence assay (Meso Scale Discovery, Gaithersburg, MD, USA). Serum gastrin (f) was measured using a RIA kit for the major circulating forms of human or mouse gastrin (MP Biomedicals, Orangeburg, NY, USA). C-peptide in plasma (g) was measured using a RIA kit specific for mouse C-peptide (Linco Research, St Charles, MO, USA). Pancreatic insulin content (h) was measured in ethanolic extracts of pancreas using a RIA kit for mouse insulin (Linco). ef Mean ± SE values are shown for all eight to nine mice in each group at the end (≤12 weeks) of the treatments indicated. *p < 0.05, **p < 0.01 vs vehicle-treated mice (DPP-4i 0 mg/kg, PPI 0 mg/kg), analysed by one-way ANOVA and Dunnett's multiple comparison test

DPP-4 inhibition and consequent elevation of plasma GLP-1 levels was recently reported to partially decrease diabetes incidence in NOD mice [7]. Here, we found that DPP-4 inhibition after diabetes onset in NOD mice raised plasma GLP-1 levels threefold and restored normoglycaemia in some of the mice; however, elevation of the serum gastrin level by concomitant PPI therapy potentiated the effects of DPP-4i treatment and normoglycaemia was restored in most of the mice; also plasma C-peptide and pancreatic insulin content were significantly increased by the combination therapy.

DPP-4 inhibition increases endogenous levels of gastric inhibitory polypeptide (GIP) and GLP-1, both of which can mediate the glucoregulatory actions of DPP-4 inhibition [8]. It is possible, therefore, that GIP, as well as GLP-1, might have contributed to the reversal of diabetes by DPP-4i in this study. A recent report, however, suggests that GLP-1 and not GIP has the major role in mediating DPP-4i effects on beta cell survival and function in mice [9]. The beneficial effect of PPI therapy on beta cell function and glucoregulation observed here is likely to be due to elevation of serum gastrin induced by PPI, because PPI-induced endogenous hypergastrinaemia has been reported to regenerate the endocrine pancreas and improve glucose tolerance after partial pancreatectomy in rats [10], as well as after pancreatoduodenectomy in humans [11]. The results of this study suggest that DPP-4i and PPI combination therapy may offer a promising strategy for correction of the beta cell deficit and insulin deficiency in type 1 diabetes.