Endogenous Glucose-Dependent Insulinotropic Polypeptide Contributes to Sitagliptin-Mediated Improvement in β-Cell Function in Patients With Type 2 Diabetes

Dipeptidyl peptidase 4 (DPP-4) is a widely expressed enzyme that degrades the incretin hormones glucagon-like peptide 1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP) (1). DPP-4 inhibitors are efficient, safe, and frequently used as second-line glucose-lowering agents in the management of type 2 diabetes (2). Because the insulinotropic effect of exogenous GIP is severely reduced or even absent in patients with type 2 diabetes (3,4), DPP-4 inhibition is generally believed to act mainly through preventing GLP-1 degradation. However, reduced GIP-induced insulin secretion can, to some extent, be restored following optimization of glycemic control in patients with type 2 diabetes (5,6). Recently, using the highly selective GIP receptor (GIPR) antagonist GIP(3-30)NH₂ (7), we demonstrated that endogenous GIP 1) exerts greater potentiation of glucose-stimulated insulin secretion compared with endogenous GLP-1 in healthy individuals (8,9) and 2) contributes significantly to postprandial insulin secretion in patients with type 2 diabetes (10). Furthermore, in preclinical studies, inhibition of the GLP-1 receptor (GLP-1R) by antagonism or receptor knockout does not eliminate the glucose-lowering effect of DPP-4 inhibition (11,12), whereas the effect is lost in double–incretin receptor knockout mice and GLP-1–negative mice treated with GIPR-antagonizing antibodies (13–15). Clinical studies using the GLP-1R antagonist exendin(9-39)NH₂ also have shown that endogenous GLP-1 accounts for only approximately one-half of the glucose-lowering and insulinotropic effects of DPP-4 inhibition in patients with type 2 diabetes; thus, it was speculated that the remaining effect of DPP-4 inhibition is mediated by GIP (16,17). In the current study, we infused the GIPR antagonist GIP(3-30)NH₂ or saline during treatment with the DPP-4 inhibitor sitagliptin in patients with type 2 diabetes to test the hypothesis that endogenous GIP contributes to the insulinotropic and glucose-lowering effects of DPP-4 inhibitor treatment.

In this randomized, double-blind, placebo-controlled, crossover study, patients were randomly assigned to receive treatment with sitagliptin (Januvia, Merck, Branchburg, NJ) 100 mg once daily or placebo in two separate 13- or 14-day treatment periods with an interposed washout period of 1–3 weeks. Study medication was add-on therapy to a continuing stable dose of metformin. Study medication was purchased through the Capital Region Pharmacy (Herlev, Denmark), which also performed the blinding of active treatment and placebo treatment by encapsulating sitagliptin and placebo tablets to obtain identical appearance. Randomization was performed by otherwise uninvolved personnel using an online random sequence generator (https://www.random.org). All participants and study personnel were blinded to the type of treatment. The study was approved by the health research ethics committees of the Capital Region of Denmark (identification no. H-18040916) and by the Danish Data Protection Agency (journal no. VD-2019-193) and was performed in accordance with the Declaration of Helsinki. Written informed consent was obtained from all patients before inclusion.

Participants were recruited through local advertisement and from local general practitioners. Inclusion was based on a screening visit, which included a clinical examination and blood and urinary tests. Inclusion criteria were age ≥18 years, type 2 diabetes for >3 months, ongoing stable treatment with lifestyle interventions and metformin, HbA_1c <9% (75 mmol/mol), and BMI >27 kg/m². The exclusion criteria were any medication that could not be paused for 12 h, plasma ALT >210 units/L, renal impairment (estimated glomerular filtration rate <60 mL/min/1.73 m²), complicating heart disease, New York Heart Association heart failure functional classification group III or IV, anemia (hemoglobin <8.3 mmol/L for men and <7.3 mmol/L for women), special dietary preferences or planned weight loss, and pregnancy or lactation.

Synthetic human GIP(3-30)NH₂ (custom synthesized by CASLO ApS, Lyngby, Denmark), with a purity of 98.55% confirmed by mass spectrometry, was dissolved in sodium hydrogen carbonate with 0.2% human albumin (CSL Behring, Marburg, Germany), tested for sterility and endotoxins by the Capital Region Pharmacy, and stored at −20°C. Under sterile conditions, the peptide was diluted to a volume of 1,000 mL in sodium chloride (9 mg/mL; Fresenius Kabi, Uppsala, Sweden) with 0.2% human albumin. Placebo infusions were 1,000 mL sodium chloride with 0.2% human albumin with an appearance identical to the GIP(3-30)NH₂ infusions.

At the end of each treatment period, two randomized study days, each including a standardized liquid mixed-meal test with a concomitant infusion of either GIP(3-30)NH₂ (1,200 pmol/kg/min) or saline (placebo) were performed. All study days were preceded by a 10-h overnight fast. The study drug sitagliptin or placebo was administered 2 h before ingestion of the liquid mixed-meal test, and all other medications (including metformin) were discontinued 12 h before each study day. Two cannulas were inserted in a cubital vein bilaterally, one for blood sampling and one for peptide/placebo infusion. The meal was ingested evenly over 5 min from time 0 min. The meal consisted of 480 kcal/200 mL Nutridrink Compact (energy content: 240 kcal/100 mL, 9.6 g protein, 9.3 g fat, and 29.7 g carbohydrate [15.0 g glucose]) to which was added a solution of 1.5 g acetaminophen in 100 mL water (for assessment of acetaminophen absorption rate as a proxy for gastric emptying) (18). The infusion was initiated 20 min before meal ingestion and was continued for 5 h. At time −25, 15, 75, 150, and 245 min, duplicate blood pressure and heart rate measures were obtained, and a mean value was calculated. Participants evaluated sensations of appetite, satiety, thirst, fullness, prospective eating, well-being, and nausea on a visual analog scale at time −30, 0, 15, 30, 60, 90, 120, 180, 240, and 300 min. Blood was drawn at fixed intervals (time −25, −20, 0, 15, 30, 45, 60, 90, 120, 180, 240, and 300 min). For bedside analysis of plasma glucose, blood was distributed into sodium fluoride–coated tubes and centrifuged immediately. For GIP, GLP-1, GIP(3-30)NH₂, proinsulin, and glucagon, blood was distributed into chilled tubes containing EDTA and a specific DPP-4 inhibitor (valine pyrrolidide 0.01 mmol/L, a gift from Novo Nordisk, Måløv, Denmark) and immediately cooled on ice. For insulin, C-peptide, and acetaminophen analysis, blood was collected in dry tubes containing separator gel and serum clot activator (silica particles) and left to coagulate at room temperature for 20 min. Samples in EDTA and dry tubes were centrifuged for 20 min at 2,000g and 4°C and then kept on ice while plasma/serum was transferred to chilled storage tubes also kept on ice and then frozen. Plasma and serum samples were stored at −20°C until batch analysis.

Plasma glucose was measured at bedside using the glucose oxidase method (YSI model 2900 STAT Plus Analyzer; Yellow Springs Instruments, Yellow Springs, OH). Intact biologically active plasma GIP was analyzed by an in-house radioimmunoassay using an N-terminally directed antiserum as previously described (19). Intact biologically active plasma GLP-1 was analyzed by a previously described in-house ELISA (20). Plasma levels of total GIP (C terminus) (9), total GLP-1 (C terminus) (21), and GIP(3-30)NH₂ (22) were measured by in-house radioimmunoassays as previously described. According to the manufacturer’s instructions, plasma glucagon and proinsulin were analyzed by ELISAs (Mercodia, Uppsala, Sweden). Serum insulin and C-peptide were analyzed by sandwich immunoassays using direct chemiluminescence (ADVIA Centaur XP; Siemens, Munich, Germany).

Data are presented as mean ± SEM unless otherwise stated. Data from blood sample analyses are presented as baseline, peak, time to peak, area under the curve (AUC), and baseline-subtracted AUC (bsAUC) values as prespecified in the protocol. When available, baseline values were calculated as a mean of time −30-, −15-, and 0-min values. According to our sample size calculation, 12 participants were needed to detect a 30% difference in bsAUC for C-peptide with a power of at least 80% and a two-sided significance level of 5% based on values reported for a similar population (mean ± SD C-peptide levels 9.1 ± 2.8 ng/mL × h) (16). The primary end point was differences in AUC for plasma glucose concentrations during GIP(3-30)NH₂ infusion and saline infusion during sitagliptin treatment. Key secondary end points were the contribution of endogenous GIP to the insulinotropic effect of sitagliptin, postprandial changes in serum insulin and C-peptide, serum C-peptide/plasma glucose ratio, insulin secretion rate (ISR), plasma glucagon, and blood pressure. AUC and bsAUC were calculated by the trapezoidal method. We evaluated differences using a two-sample Student t test (two-tailed). Repeated-measures one-way ANOVA was applied to analyze variations among more than two estimates. Using the deconvolution method, we calculated ISR based on detected serum C-peptide concentrations and population-based variables for C-peptide kinetics in combination with weight, height, sex, and age (23). A two-sided P < 0.05 was chosen to indicate statistically significant differences.

As expressed in this equation, we calculated an estimate of the absolute sitagliptin-mediated impact of GIP on β-cell function as the insulinogenic index during sitagliptin treatment plus saline infusion minus the insulinogenic index during sitagliptin plus GIP(3-30)NH₂. This estimate was expressed relative to the maximal potential contribution of GIP to the effect of sitagliptin (100%), defined as the difference between the full sitagliptin treatment effect, including actions mediated by GIP (sitagliptin + saline), and the physiological response minus any contribution by GIP [placebo treatment + GIP(3-30)NH₂]. Mean arterial pressure (MAP) was calculated as (systolic blood pressure + [2 × diastolic blood pressure])/3.

The data generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.

We included four women and eight men with metformin-treated type 2 diabetes (duration of diabetes 11 ± 5 years [mean ± SD], age 67 ± 5 years, BMI 27.4 ± 2.6 kg/m², fasting plasma glucose 165 ± 52 mg/dL [9.2 ± 2.9 mmol/L], HbA_1c 7.1 ± 1.4% [54 ± 15 mmol/mol]) (Table 1). All recruited participants completed the study, and their data were analyzed.

Infusion with the GIPR antagonist GIP(3-30)NH₂ resulted in mean ± SD steady-state concentrations of 94 ± 17 nmol/L during placebo treatment and 102 ± 16 nmol/L during sitagliptin treatment before ingestion of the liquid mixed meal (Fig. 1G). The result was an antagonist/agonist ratio >1,100-fold previously shown to significantly inhibit GIP(1-42)-induced actions (24). We observed no adverse reactions to the infusions.

Ten days of placebo treatment reduced fasting plasma glucose by −0.5 ± 0.5 mmol/L (P = 0.335), and sitagliptin treatment lowered fasting plasma glucose by −1.6 ± 0.6 mmol/L (P = 0.016), corresponding to a between-treatment reduction of fasting plasma glucose of −1.1 ± 0.2 mmol/L (P < 0.001). During the mixed meal, sitagliptin reduced the AUC of the glucose excursion and the peak plasma glucose concentration compared with placebo, but bsAUC was unchanged by sitagliptin (Fig. 1A–C and Table 2). During placebo treatment, GIP(3-30)NH₂ increased postprandial plasma glucose AUC by 7.3 ± 2.8% (P = 0.017) and increased peak plasma glucose concentration by 0.8 ± 0.2 mmol/L (P = 0.003) compared with saline infusion. During sitagliptin treatment, GIP(3-30)NH₂ did not change postprandial plasma glucose AUC or peak glucose concentration compared with saline infusion. Compared with placebo, sitagliptin delayed the time to peak plasma glucose by 20 ± 6.1 min (P = 0.008), but the time to peak plasma glucose was unaffected by GIP(3-30)NH₂ infusion during both placebo and sitagliptin treatment (P = 0.731 and P = 0.220, respectively).

Serum acetaminophen concentrations were undetectable at baseline. Postprandial serum acetaminophen concentrations reached similar peak concentrations during all four interventions (Fig. 1D and Table 2). During sitagliptin treatment and saline infusion, the time to peak for serum acetaminophen was slowed by −19 ± 9.3 min (P = 0.068) compared with placebo and saline infusion. The time to peak concentration of serum acetaminophen was not affected by infusion of the GIPR antagonist.

Fasting levels of insulin and C-peptide were similar during both treatment periods. During placebo treatment, GIP(3-30)NH₂ caused a significant −20 ± 8% decrease in mean bsAUC for insulin and a −19 ± 6% decrease in mean bsAUC for C-peptide concentrations after the mixed meal (Table 2). In addition, during placebo treatment, GIP(3-30)NH₂ infusion lowered the C-peptide/glucose ratio by −35 ± 6%. During sitagliptin treatment, GIP(3-30)NH₂ caused a significant decrease in mean bsAUC for insulin (−6 ± 13%, P = 0.028) despite lower glycemia, but mean bsAUC for C-peptide was not significantly changed.

Sitagliptin improved the β-cell function assessed by ISR/plasma glucose ratio and serum C-peptide/plasma glucose ratio. This improvement was partially due to endogenous GIP, since GIP(3-30)NH₂ infusion lowered the serum C-peptide/plasma glucose ratio by −11 ± 7.3% during sitagliptin treatment (Fig. 2G and Table 2). The contribution of endogenous GIP to the insulinotropic effect of sitagliptin was found to be 34 ± 10% of the potential maximum based on serum C-peptide/plasma glucose ratios and 37 ± 12% of the potential maximum based on ISR/plasma glucose ratios (Fig. 2I).

Plasma proinsulin was similarly reduced by infusion of GIP(3-30)NH₂ during the two treatment periods (Fig. 3A–E and Table 2). The baseline values of the proinsulin/insulin ratio, a measure of β-cell stress, were nonsignificantly increased by 3% following sitagliptin treatment (Fig. 3D and Table 2). However, assessment of the proinsulin/insulin ratio controlled for the chronological order of the interventions revealed that patients undergoing sitagliptin treatment in the second treatment period (n = 5) had a −20% decrease in the proinsulin/insulin ratio (P = 0.005) following sitagliptin treatment compared with placebo treatment, reflecting reduced β-cell stress. Sitagliptin significantly lowered baseline values of the proinsulin/C-peptide ratio compared with placebo tablet treatment, and during sitagliptin treatment, compared with saline placebo, GIPR antagonism significantly reduced β-cell stress assessed by the proinsulin/C-peptide ratio (Fig. 3E and Table 2).

During placebo treatment, GIP(3-30)NH₂ infusion reduced the AUC for plasma glucagon compared with saline (Table 2 and Fig. 4A). Postprandial plasma glucagon concentrations reached similar peak concentrations during all four interventions. When taking into account the prevailing glucose levels, endogenous GIP significantly increased postprandial glucagon levels during placebo treatment but not during sitagliptin treatment (Fig. 4D–F and Table 2).

Visual analog scale evaluations of hunger, prospective food consumption, satiety, fullness, thirst, nausea, well-being, and tiredness were similar between the interventions (data not shown).

Sitagliptin treatment increased plasma levels of intact, active GIP and GLP-1 substantially during the mixed-meal test (Fig. 1F–I and Table 3) and the postprandial response of total GIP. Infusion of GIP(3-30)NH₂ did not change total GIP and GLP-1 during the mixed meal, but during sitagliptin treatment, the GIP(3-30)NH₂ infusion increased levels of intact GIP compared with the saline infusion.

During both placebo and sitagliptin treatment, infusion of GIP(3-30)NH₂ increased systolic and diastolic blood pressure, increased MAP, and decreased heart rate (Fig. 5). Differences were most pronounced during sitagliptin treatment, with significant increases of 8 ± 2, 4 ± 1, and 5 ± 1 mmHg in systolic pressure, diastolic blood pressure, and MAP, respectively, at time 150 min (P = 0.008, P = 0.015, and P = 0.001) (Fig. 5B and F and Supplementary Table 2). Moreover, MAP was increased by 5 ± 2 mmHg at time 245 min (P = 0.031). Infusion of GIP(3-30)NH₂ lowered heart rate by 5 ± 1 beats/min at time 15 min during sitagliptin (P = 0.009) and at time 75 min during both placebo and sitagliptin treatment (5 ± 1 [P = 0.007] and 5 ± 1 [P = 0.005], respectively) (Fig. 5A and E and Supplementary Table 2).

In this study, we used the selective GIPR antagonist GIP(3-30)NH₂ during a 5-h liquid mixed-meal test in 12 patients with type 2 diabetes treated with the DPP-4 inhibitor sitagliptin and placebo in a crossover design. We observed 1) an insulinotropic and glucose-lowering effect of endogenous GIP in patients with type 2 diabetes during placebo treatment and 2) that endogenous GIP contributes with 37 ± 12% of the sitagliptin-induced improvement in β-cell function.

The current study shows that during placebo treatment, endogenous GIP is insulinotropic and significantly increases both C-peptide and insulin secretion and consequently lowers plasma glucose in patients with type 2 diabetes. This effect was not associated with diabetes duration (before or after 10 years) or HbA_1c levels (less than or greater than 7.5% [58 mmol/mol]) (data not shown). We previously demonstrated an insulinotropic effect of endogenous GIP in 10 patients with type 2 diabetes, but the GIP(3-30)NH₂-induced increase of plasma glucose in that study did not reach statistical significance (10). In the current study of 12 patients with type 2 diabetes, we observed a significant glucose-lowering effect of endogenous GIP. Furthermore, endogenous GIP contributed to β-cell function assessed by serum C-peptide/plasma glucose ratio in these patients. During sitagliptin treatment, GIPR antagonism did not change plasma glucose, which could be due to increased GLP-1 effects blunting the effects of GIP, but further investigation is warranted to clarify this matter.

In the current study, we used similar calculations to quantify the contribution of GIP to the glucose-lowering effects of DPP-4 inhibition as were used in two previous studies that used the GLP-1R antagonist exendin(9-39)NH₂ to quantify the contribution of GLP-1 to the glucose-lowering effects of DPP-4 inhibition. Aulinger et al. (17) investigated the contribution of GLP-1 to the insulinotropic effect of sitagliptin (initiated 1 day before the experimental day) during an oral glucose tolerance test in participants with type 2 diabetes (mean HbA_1c 6.2 ± 0.2%). Nauck et al. (16) investigated the contribution of GLP-1 to the insulinotropic effect of 9–10 days of vildagliptin treatment during a mixed-meal test in participants with type 2 diabetes (mean HbA_1c 7.2 ± 0.5%) and healthy control subjects. In the current study, we used GIP(3-30)NH₂ to investigate the contribution of GIP to the insulinotropic effect of 13 days of sitagliptin treatment during a mixed-meal test in participants with type 2 diabetes (mean HbA_1c 7.1 ± 1.4%). Unlike Aulinger et al. and Nauck et al., our study was randomized and double blinded. Nevertheless, despite some minor differences in study designs, the results and calculations should be comparable, and DPP-4 inhibition influenced incretin hormone responses similarly in all three. It is well known that the feedback inhibition of GIP secretion is less prominent compared with that of GLP-1 secretion during sitagliptin treatment; still, in the current study, sitagliptin increased levels of active GIP and lowered total GIP, likely reflecting a negative feedback mechanism acting on the GIP-secreting enteroendocrine K cells, as previously suggested (25). Like Aulinger et al. and Nauck et al., we focused on changes in the β-cell secretory response during DPP-4 inhibition. As sitagliptin lowers plasma glucose, the β-cell secretory response needs to be interpreted relative to prevailing glycemia (i.e., assessed by insulinogenic indices [serum C-peptide/plasma glucose and ISR/plasma glucose ratios]). In the current study, we confirm that DPP-4 inhibition increases insulinogenic indices in patients with type 2 diabetes. Using a similar way of calculating the contribution of an antagonized incretin receptor as Nauck et al., and as a key finding, we demonstrate for the first time in human physiology that endogenous GIP is responsible for 37 ± 12% of sitagliptin-induced improvement in β-cell function (Fig. 2I). Both Aulinger et al. and Nauck et al. found that approximately one-half of DPP-4 inhibitor–mediated improvements in β-cell function could be ascribed to endogenous GLP-1. Thus, we cannot rule out the existence of other mediators contributing to the effect of DPP-4 inhibitors, and a future study with simultaneous antagonism of both incretin receptors during DPP-4 inhibition could perhaps contribute to clarifying this matter further.

Sitagliptin has effects beyond improved postprandial insulin secretion, likely because of improved β-cell function postprandially and in the fasting state (26) and increased insulin-stimulated peripheral glucose utilization (27). In the current study, sitagliptin treatment resulted in lower fasting plasma glucose, as previously shown (28,29). Since DPP-4 inhibitors increase circulating levels of active incretin hormones, we expected reduced postprandial glucose excursion irrespective of the fasting plasma glucose levels due to increased postprandial insulin secretion. Thus, it is puzzling that the postprandial glucose excursions assessed by bsAUC were unchanged compared with placebo treatment. Some studies, however, have found that DPP-4 inhibitor treatment does not affect postprandial glucose tolerance as measured by bsAUC (16,30), while others have found that bsAUC of plasma glucose is slightly lower (17,31–33). The current study does not allow for a direct evaluation of an increased sitagliptin-mediated effect of GIP on fasting glucose and fasting insulin levels.

We found a secretory increase in insulin following just 13 days of sitagliptin treatment, confirming that DPP-4 inhibitor treatment increases insulin secretion early on (32). Glucotoxic stress during chronic hyperglycemia is proposed to impair β-cell responsiveness to GIP by downregulation of GIPR gene expression (34,35), and an extended sitagliptin treatment period with improved glycemic control to relieve β-cell stress might have improved the insulinotropic effect of GIP further (5). The proinsulin/insulin ratio is a highly sensitive marker of β-cell stress (36), and circulating proinsulin levels are increased in type 2 diabetes (37,38). The dual GIPR/GLP-1R agonist tirzepatide reduced the proinsulin/insulin ratio by 26–37% after 26 weeks of treatment (39). In the current study, we observed a 20% decrease in the proinsulin/insulin ratio after sitagliptin treatment with increased endogenous GIP and GLP-1 (intact, active forms), as previously described (40).

GIPR antagonism inhibited the glucagonotropic actions of endogenous GIP in patients with type 2 diabetes. Postprandial glucagon concentrations were reduced, which was most pronounced when plasma glucose was declining (3–5 h postmeal), confirming a glucagonotropic action of GIP during normal to low plasma glucose levels (41). The effect was most pronounced during placebo treatment when the prevailing glucose levels were considered in a glucagon/glucose ratio. On the contrary, the effect was blunted during sitagliptin treatment, which could be explained by increased levels of active GLP-1 inhibiting glucagon secretion (4).

Nauck et al. (16) and Aulinger et al. (17) reported that gastric emptying was not affected (oral glucose tolerance test) or slightly accelerated (mixed-meal test) by exendin(9-39)NH₂, whereas in our study, sitagliptin lowered the time to peak by ∼20 min with no effect of GIPR antagonism. This likely reflects elevated circulating levels of intact GLP-1 decelerating gastric emptying (42), whereas GIP exerts no influence on gastric emptying (43).

It has been hypothesized that postprandial elevated GIP levels increase mesenteric blood flow to facilitate nutrient digestion and utilization, resulting in decreased blood pressure and increased heart rate (44). Consistent with this hypothesis, and as previously reported (24,45,46), GIPR antagonism increased blood pressure and slowed the heart rate during the mixed-meal test in the current study.

It is uncertain whether GIP(3-30)NH₂ blocks GIPR completely (7,24). If not, this would lead to an underestimation of the effects of endogenous GIP, which is a limitation to the interpretation of the results.

Using GIPR antagonism, we demonstrate that endogenous GIP contributes to postprandial glucose homeostasis in patients with type 2 diabetes by increasing glucose-stimulated insulin secretion and glucagon secretion and is responsible for part of the improved β-cell function observed during DPP-4 inhibition by sitagliptin.

Acknowledgments. The authors thank all study participants for their participation. The authors also thank D. Baunbjerg (Center for Clinical Metabolic Research, Gentofte University Hospital, Gentofte, Denmark) for laboratory assistance during the clinical study and L.B. Albæk (Faculty of Health and Medical Sciences, Department of Biomedical Sciences, University of Copenhagen) for analytical assistance.

Funding. The study was funded by the European Foundation for the Study of Diabetes (EFSD)/Lilly European Diabetes Programme 2018 and the A.P. Møller Foundation.

Duality of Interest. L.S.G. has been a speaker for Eli Lilly and is cofounder of Antag Therapeutics ApS. B.H. is cofounder of Bainan Biotech ApS. T.V. has served on scientific advisory panels and/or speakers’ bureaus or has served as a consultant to and/or received research support from Amgen, AstraZeneca, Bristol-Myers Squibb, Boehringer Ingelheim, Eli Lilly, GlaxoSmithKline, MSD/Merck, Mundipharma, Novo Nordisk, Sanofi, and Sun Pharma. J.J.H. has served on scientific advisory panels for and/or has received speaker honoraria from Novo Nordisk and MSD/Merck. J.J.H. is one of the founders of Antag Therapeutics ApS. M.M.R. is one of the founders of Antag Therapeutics ApS, Bainan Biotech ApS, and Synklino ApS. M.B.C. is cofounder of and minority shareholder in Antag Therapeutics ApS. F.K.K. has served on scientific advisory panels and/or been part of speakers’ bureaus for, served as a consultant to, and/or received research support from Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Carmot Therapeutics, Eli Lilly, Gubra, MedImmune, MSD/Merck, Mundipharma, Norgine, Novo Nordisk, Sanofi, and Zealand Pharma and is a cofounder of and minority shareholder in Antag Therapeutics ApS. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. S.S. performed the clinical study. S.S., L.S.G., M.M.R., T.V., J.J.H., M.B.C., and F.K.K. designed the clinical study. S.S., L.S.G., and F.K.K. performed and are responsible for the data analysis and wrote the manuscript. J.J.H. and B.H. performed the radioimmunoassay and ELISA analyses. All authors contributed to the interpretation of data, reviewed and edited the manuscript, and approved the final version. S.S., L.S.G., and F.K.K. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were reported in abstract form at the 81st Scientific Sessions of the American Diabetes Association, Virtual, 25–29 June 2021, and the 57th European Association for the Study of Diabetes Annual Meeting, Virtual, 27 September–1 October 2021.