Xenin-25 Amplifies GIP-Mediated Insulin Secretion in Humans With Normal and Impaired Glucose Tolerance but Not Type 2 Diabetes

All protocols were approved by Washington University’s Human Research Protection Office and the Food and Drug Administration (IND 103,374) and are registered with ClinicalTrials.gov (NCT00798915). Studies were performed in the Clinical Research Unit of the Institute of Clinical and Translational Sciences of Washington University after obtaining written informed consent. Male and female subjects with NGT, IGT, and mild T2DM were studied (Table 1). Glucose tolerance was defined by the 2-h plasma glucose level during a 75-g oral glucose tolerance test using diagnostic criteria of the American Diabetes Association (15). Baseline characteristics were determined during a screening visit after a 10-h fast. HbA_1c levels were required to be ≤9.0 in all subjects. Subjects treated with insulin were excluded. Subjects treated with oral antidiabetic medications were enrolled if the agent(s) could be safely discontinued for 48 h preceding each study visit. These selection criteria were designed to exclude T2DM subjects with advanced β-cell failure and to identify participants with residual insulin secretion who have the potential to respond to incretin peptides. Women of childbearing potential were required to use birth control. Subjects were excluded if they 1) had a history of chronic pancreatitis and/or risk factors for chronic pancreatitis, 2) had a history of gastrointestinal disorders, 3) were taking nondiabetes medications known to affect glucose homeostasis, and 4) had significant systemic illness including heart, kidney, liver, inflammatory, or malignant disease.

Studies were performed after a 10-h overnight fast. One intravenous catheter was placed into a hand vein. This hand was kept in a thermostatically controlled box (50–55°C) to facilitate venous sampling and to provide arterialized venous blood (16,17). A second intravenous line was inserted for administration of glucose/peptides. Subjects with a fasting blood glucose ≥120 mg/dL were given boluses of intravenous human insulin (∼0.01 units/kg) at 30-min intervals as needed to decrease the blood glucose level to 100–120 mg/dL to limit variability of initial glucose levels. Blood glucose was stable for 20 min before starting the graded glucose infusion (GGI).

All subjects were administered four separate GGIs, each separated by at least 2 weeks. During the respective GGIs, along with glucose, subjects also received an intravenous infusion with GIP plus xenin-25, GIP alone, albumin alone (no peptide), or xenin-25 alone (subjects were blinded to treatment). From 0 to 40, 40 to 80, 80 to 120, 120 to 160, 160 to 200, and 200 to 240 min, the glucose infusion rates were 1, 2, 3, 4, 6, and 8 mg ⋅ kg⁻¹ ⋅ min⁻¹, respectively. Peptides were administered by primed-continuous intravenous infusion. Infusions rates from 0 to 3, 3 to 7, 7 to 10, and 10 to 240 min were 10.8, 7.7, 5.6, and 4.0 pmol ⋅ kg⁻¹ ⋅ min⁻¹.

GIP and xenin-25 were custom synthesized under good manufacturing practice conditions (Bachem, Torrance, CA), dissolved in water (United States Pharmacopeial Convention, injectable), filter sterilized, aliquoted into cryovials as single doses, and stored at −80°C. Representative vials were analyzed for identity, purity, concentration, sterility, and endotoxins. Peptides were compounded in normal saline containing 1% Flexbumin (Baxter Healthcare Corp., Westlake Village, CA) immediately before infusions.

GIP administered by an intravenous infusion at a dose of 4 pmol ⋅ kg⁻¹ ⋅ min⁻¹ increases insulin release during a hyperglycemic clamp in healthy individuals approximately fourfold (7,8) but is ineffective in persons with T2DM (7). Similar results were obtained using GIP at a dose of 2.4 pmol ⋅ kg⁻¹ ⋅ min⁻¹ (6). Only one human clinical trial with xenin-25 has been published (18). The authors assessed the effects of xenin-25 infused at a dose of 4 pmol ⋅ kg⁻¹ ⋅ min⁻¹ on intestinal motility. This dose increased plasma xenin-25 levels and gut motility (18). Effects on insulin secretion and glucose concentrations were not reported. On the basis of these studies, GIP and xenin-25 were each infused at a dose of 4 pmol ⋅ kg⁻¹ ⋅ min⁻¹.

Plasma glucose was measured bedside by the glucose oxidase technique using a YSI analyzer (Yellow Springs Instruments, Yellow Springs, OH). Insulin, C-peptide, glucagon, complete metabolic profiles, and HbA_1c levels were determined by Washington University’s Core Laboratory for Clinical Studies. Insulin and C-peptide were measured using a solid-phase, two-site chemiluminescent immunometric assay (Siemens Medical Solutions Diagnostics, Los Angeles, CA). Glucagon was determined by radioimmunoassay (Millipore Corporation, St. Charles, MO). Immunoreactive (IR)-GIP and IR-xenin levels were determined using an enzyme-linked immunosorbent assay (ELISA) for total GIP (Millipore Corporation) and a custom ELISA developed in our laboratory for IR-xenin (10). Lipase and total amylase were determined by the Barnes Hospital Laboratories.

Insulin secretion rates (ISRs) were derived by stochastic deconvolution of the peripheral C-peptide concentrations as previously described using population-based estimates of C-peptide clearance kinetics (19–21).

Physiologic data were analyzed using mixed-effects models with subject as a random effect and peptide as a fixed effect. Pairwise comparisons were limited to evaluating the effects of 1) xenin-25 alone versus albumin, 2) GIP alone versus albumin, and 3) GIP plus xenin-25 versus GIP alone. One-tailed t tests were used when testing the hypotheses that ISRs were greater during infusion with 1) GIP plus xenin-25 compared with GIP alone and 2) GIP alone compared with albumin. Two-tailed t tests were used when testing the hypothesis that ISRs were similar during infusions with albumin alone and xenin-25 alone. All glucagon data were analyzed using two-tailed t tests.

Groups were well-matched and as anticipated, the 2-h and fasting plasma glucose and HbA_1c levels progressively increased in groups with IGT and T2DM compared with NGT (Table 1). Subjects with T2DM did not have gastroparesis or clinically evident peripheral neuropathies. Five patients with T2DM were taking metformin and two were also on a sulfonylurea. No subjects were receiving incretin-based therapies. Three subjects with T2DM received insulin to lower basal glucose levels before one, one, and two of their four respective GGIs.

Plasma total amylase levels (salivary plus pancreatic) were unaffected by any of the infusions (data not shown). On the basis of qualitative surveys taken before, during, and after each infusion, peptides were not associated with nausea, vomiting, chest pains, dizziness, heart palpitations, shortness of breath, fever, chills, blurred vision, or changes in salivation, sweating, or frequency of urination. Mild diarrhea and/or loose stools were observed in 57, 14, 21, and 0% of subjects who received GIP plus xenin-25, GIP alone, xenin-25 alone, or albumin alone, respectively. When experienced, diarrhea occurred within several hours of the infusion and was noted in all three groups. Stools did not contain blood, and diarrhea did not require pharmacologic or medical intervention. Infusion with GIP alone reduced mean arterial blood pressure 10–15 mmHg and increased resting heart rate ∼8 bpm in all three groups. In contrast, xenin-25, either alone or with GIP, had no effect on either parameter. Thus, xenin-25 does not appear to affect autonomic function.

In subjects with NGT, IGT, and T2DM (Fig. 1; some data not shown), fasting, preinfusion plasma levels of IR-GIP were <10 pmol/L, and fasting IR-xenin levels were below the limit of detection (typically <2 pmol/L). Infusion of GIP and xenin-25 increased steady-state levels of plasma IR-GIP and IR-xenin to ∼400 and ∼125 pmol/L, respectively. These levels were unaffected by coinfusion with the other peptide. Peak plasma IR-GIP levels after mixed-meal ingestion were ∼80 pmol/L, and plasma IR-xenin was undetectable (<2 pmol/L). Thus, pharmacologic levels of GIP and xenin-25 were maintained throughout the GGIs. On the basis of IR-xenin levels measured after peptide infusions were terminated (n = 6), the circulating half-life of IR-xenin is 2.5 ± 0.3 min (data not shown).

During infusion with albumin alone, plasma insulin, C-peptide, and glucose levels, as well as ISRs (Fig. 2), progressively increased in response to the stepwise increases in glucose infusion rates. Maximal plasma glucose levels were higher in the group with T2DM (313 ± 13 mg/dL) versus IGT (240 ± 12 mg/dL) and NGT (186 ± 11 mg/dL) (P < 0.003 for IGT vs. NGT, P < 0.0001 for T2DM vs. NGT). In spite of the progressively higher plasma glucose levels, plasma insulin and C-peptide levels, as well as ISRs, tended to reach lower maximal levels (in pmol/min) in subjects with T2DM (838 ± 224) versus IGT (1,003 ± 118) and NGT (1,037 ± 141). As shown in Fig. 3, ISRs increased linearly as a function of plasma glucose level during infusion with albumin alone. However, the slope of the line relating ISR to glucose declined in the progression from NGT to IGT to T2DM (slopes in pmol/min per mg/dL are 9.0, 6.1, and 3.1, respectively), reflecting worsening β-cell insulin secretory responses to glucose with IGT and T2DM.

Compared with albumin, xenin-25 did not significantly affect plasma insulin, C-peptide, or glucose levels or ISRs in any group (Fig. 2). Small and transient, but nonsignificant, increases in insulin, C-peptide, and ISR were noted at the 10- and 20-min time points during infusion of xenin-25 alone (see below). Plots of ISRs versus plasma glucose levels in the presence and absence of xenin-25 were also nearly identical within each group (Fig. 3). Incremental areas under the curve (iAUCs) were calculated for ISR and plasma glucose for each individual during the 240-min GGI with and without infusion of the xenin-25. The ratio of these iAUCs represents a summary measure of β-cell sensitivity (ISR per glucose concentration) over the entire range of glucose levels. As shown in Fig. 3, the means of the ratios were not different within each group during infusions with albumin versus xenin-25 alone in subjects with NGT (8.7 ± 1.4 vs. 9.5 ± 1.4), IGT (6.2 ± 0.8 vs. 6.9 ± 0.8), and T2DM (3.6 ± 1.8 vs. 3.4 ± 1.8). Thus, infusion of xenin-25 alone does not significantly affect ISR over a broad range of plasma glucose levels in humans with NGT, IGT, or T2DM.

Subjects were next administered GGIs with GIP or GIP plus xenin-25. GIP plus or minus xenin-25 increased insulin secretory responses that varied as functions of time, plasma glucose level, and glucose tolerance (Figs. 2 and 3). At all glucose levels studied (Fig. 3), ISRs were greater within each group during GGIs with GIP and GIP plus xenin-25 compared with albumin alone or xenin-25 alone. Since the effects of GIP plus or minus xenin-25 on ISRs were not linear as a function of plasma glucose levels, iAUCs rather than slopes were used to quantify ISRs. Compared with albumin, infusion with GIP increased the 240-min ISR iAUC/glucose iAUC (Fig. 3) 1.7-, 1.5-, and 1.9-fold in subjects with NGT, IGT, and T2DM, respectively (all P values ≤0.007 vs. albumin), and these ratios were further increased during infusion with GIP plus xenin-25 in subjects with IGT (1.7-fold further increase vs. GIP alone; P = 0.002). In subjects with NGT and T2DM, the differences were not significant. Since the 240-min summary measures are taken over differing ranges of glucose levels between groups, these fold increases do not directly compare peptide effects between the three groups.

Infusion with GIP alone at the start of the GGI caused rapid and transient increases in insulin, C-peptide, and ISRs in all groups (Figs. 2 and 3). This initial response occurred with little change in plasma glucose levels and was absent during infusions with albumin alone. Since this transient response was not anticipated, a post hoc analysis of the first 40 min was conducted (Fig. 4). In subjects with NGT, this early insulin secretory response to GIP peaked within the first 10 min of the GGI. ISRs then declined until ∼40 min but remained greater than levels during infusion with albumin or xenin-25 alone. A similar response to GIP was noted in subjects with IGT and T2DM, except peak ISRs were not reached until 20 min into the infusions and then decreased until ∼60–70 min into the GGI. Within each group, the iAUCs for the ISR during the first 40 min of the GGIs (pmol insulin secreted/40 min) were greater during infusion with GIP alone compared with albumin (iAUCs increased 4.6-, 6.1-, and 8.7-fold in subjects with NGT, IGT, and T2DM, respectively; all P values ≤0.0005). Moreover, the incremental insulin secretory responses to GIP during the first 40 min of the GGI were remarkably similar in humans with NGT (7,998 ± 1,263), IGT (10,016 ± 1,524), and T2DM (10,062 ± 2,290).

Compared with albumin, infusion with xenin-25 alone had no statistically significant effect on ISRs or plasma glucose levels in any group (Figs. 2–4). In contrast, ISR iAUCs (pmol insulin secreted/40 min) during infusions with GIP plus xenin compared with GIP alone were increased during this period in subjects with NGT (10,185 ± 1,263 compared with 7,998 ± 1,263; P = 0.048) and IGT (15,199 ± 1,524 compared with 10,016 ± 1,524; P = 0.002). However, xenin-25 did not significantly increase the effects of GIP in subjects with T2DM (12,128 ± 2,290 with GIP plus xenin-25 compared with 10,062 ± 2,290 with GIP alone; P = 0.18). Because plasma glucose fell below basal levels for some subjects during the initial 40-min period, ISRs were normalized to glucose AUC rather than to glucose iAUC for this time period. As shown in Fig. 4D and E, the magnitudes of the early insulin secretory responses to GIP remained similar in subjects with NGT, IGT, and T2DM, and ISRs were still significantly amplified by xenin-25 in subjects with NGT and IGT but not T2DM.

Glucagon plays a central role in the regulation of insulin secretion and glucose metabolism. Fasting, basal plasma glucagon levels were similar in all three groups and progressively decreased in humans with NGT, IGT, and T2DM during the GGI when albumin alone was infused (Fig. 5). As expected, the ability of plasma glucose to suppress plasma glucagon levels was attenuated in subjects with T2DM since glucagon levels at the end of the GGI declined by 31, 34, and 21 pg/mL in subjects with NGT, IGT, and T2DM, respectively. Compared with albumin, infusion with xenin-25 alone had no statistically significant effect on glucagon levels in any group. However, compared with GIP alone, infusion with GIP plus xenin-25 increased plasma glucagon levels in subjects with NGT and IGT during the first 40 min and during the entire 240 min of the GGI. In contrast, xenin-25 failed to amplify the effects of GIP on plasma glucagon levels in subjects with T2DM during either time period. Thus, xenin-25 amplified the glucagon response to GIP in the subjects with NGT and IGT but not T2DM.

GIP is an incretin hormone that in people with NGT, significantly increases the insulin secretory response to glucose. However, since GIP-stimulated insulin release is reportedly blunted in humans with T2DM, this peptide has not been pursued as a therapy for T2DM. The current study was undertaken to further our understanding of attenuated β-cell responsiveness to GIP in T2DM and to elucidate the mechanisms involved in GIP signaling in the β-cell.

Previous studies from our laboratory show that xenin-25 potentiates GIP-mediated insulin release in two mouse models that exhibit blunted insulin secretory responses to GIP (10). In contrast, xenin-25 alone had no effect on insulin secretion in either mouse model. Moreover, xenin-25 did not act directly on islet β-cells. Rather, xenin-25 activated a cholinergic relay in the periphery that did not require neurons in the central nervous system. The results presented in this study extend these observations and show that during GGIs at the peptide and glucose levels studied, xenin-25 alone does not affect ISR in humans but does potentiate GIP-mediated insulin secretion in humans with NGT and IGT but curiously, not with T2DM (see below).

A xenin-25–specific receptor has not been identified, and there is strong evidence that the effects of xenin-25 are mediated by neurotensin receptor-1 (22–27). Similar to our results with xenin-25 alone, intravenous infusion of neurotensin alone in healthy humans caused no significant change in blood pressure or plasma levels of glucose, insulin, or glucagon (28). Thus, our results are consistent with those for another neurotensin receptor-1 agonist.

The GGI is designed to detect subtle changes in the insulin secretory response to glucose. As anticipated, ISRs during the GGIs with albumin alone increased linearly as a function of plasma glucose levels, but the slopes of lines relating ISR to plasma glucose were progressively reduced in the groups with IGT and T2DM compared with NGT. Infusion of xenin-25 alone did not significantly affect insulin secretion compared with the rates measured during control infusions (Figs. 2–4). In contrast, all subjects, including those with T2DM, exhibited a rapid and transient increase in ISR in response to GIP alone as soon as the peptide/glucose infusions were initiated. Xenin-25 significantly amplified this early response to GIP in subjects with NGT and IGT but not T2DM. After this initial response, ISRs progressively increased as a function of plasma glucose level and were greatest during infusions with GIP plus xenin-25. However, this later response was statistically significant only in subjects with IGT. Unfortunately, plasma glucose levels in the subjects with NGT remained lower during the GGI compared with the other groups. Since glucose-dependent increases in the incretin response were observed starting at ∼160 mg/dL glucose (see Fig. 3), the effects of GIP plus or minus xenin-25 may be underestimated in this group. Alternatively, GIP-mediated ISR may already be near maximal and cannot be further amplified by xenin-25 in subjects with NGT.

It is generally assumed that the insulinotropic effects of GIP are blunted in T2DM. However, many studies supporting this conclusion were conducted using hyperglycemic clamps (6,8,29). In a study of fasting nondiabetic humans, plasma insulin levels rapidly (5 min) and transiently increased immediately after GIP was administered as a single bolus in spite of the fact that plasma glucose remained euglycemic (30). This pattern of insulin secretion is very similar to that noted during the initial 40-min response during our GGIs in which GIP rapidly and transiently increased ISRs in subjects with NGT, IGT, and T2DM with little change in plasma glucose levels. Thus, it is possible that some effects of GIP are masked during hyperglycemic clamps. Of importance, even though the β-cell response to glucose (i.e., the slopes for ISR vs. plasma glucose during infusion with albumin alone) was progressively decreased in subjects with NGT, IGT, and T2DM (Fig. 3), the magnitudes of the early responses to GIP were similar in all groups (Fig. 4). Our results are thus consistent with the idea proposed by Meier and Nauck (31) that humans with T2DM may retain the ability to respond to GIP, but reduced GIP responses in these subjects may reflect worsening insulin secretory responses to glucose rather than defective GIP signaling. This is further supported by the fact that the effects of GIP on blood pressure and heart rate were not blunted in humans with T2DM. Moreover, the effects of GIP on ISRs remained elevated even after this initial transient response, suggesting that like the cephalic response (32), this early response to GIP may play an important role in amplifying subsequent insulin secretion.

As with ISR, xenin-25 also amplified the early and transient GIP-mediated increase in glucagon levels in humans with NGT and IGT but not T2DM. GIP rapidly and transiently increased plasma glucagon levels in healthy humans when administered during euglycemic conditions (33) but not during hyperglycemic clamps (34). The very early postprandial glucagon response is thought to prevent hypoglycemia as cells recruit Glut4 and amino acid transporters to the cell surface in anticipation of increased levels of nutrients in the blood (35). Thus, the rapid and transient amplification of GIP-mediated glucagon release by xenin-25 may reflect a normal islet secretory response to nutrients after an overnight fast. These results also suggest that in humans with T2DM, islets exhibit a blunted response to xenin-25 signaling.

Mouse studies indicate that pharmacologic doses of xenin-25 increase GIP-mediated insulin release via a cholinergic relay requiring muscarinic receptors in the periphery (10), and M3 muscarinic receptors in islets are required for both insulin and glucagon release (36). Results in this study suggest that at the dose required to amplify GIP-mediated insulin secretion in humans, xenin-25 does not increase parasympathetic activity since it had no effect on heart rate, blood pressure, salivation, or need to urinate. Thus, an analogous xenin-25–mediated cholinergic relay is the most likely mechanism for increasing insulin and glucagon release in humans. It is worth noting that in contrast to a previous report (37), we have been unable to detect endogenous IR-xenin in either mouse or human plasma prepared under fasting, fed, or postprandial conditions using an ELISA that can detect <2 pmol/L peptide, raising the possibility that xenin-25 is not a physiologic peptide. Moreover, we have not seen other reports where endogenous xenin has been measured in plasma or purified K cells. Alternatively, endogenous xenin-25 may be modified such that it is not detected by our ELISA. Characterization of xenin-25 purified from plasma or isolated K cells would be required to address this issue. If xenin-25 is a physiologic peptide, it is unclear if local levels (e.g., at nerve endings near the K cell) would be high enough to modify islet function in vivo. However, our results with exogenously administered xenin-25 provide important insights into β-cell function during the progression from NGT to IGT to T2DM even if xenin-25 is not a physiologic peptide or if physiologic levels do not modify islet function. These results also suggest that defects in cholinergic or neuronal signaling in the periphery may contribute to defective insulin secretion in humans with T2DM. For example, xenin-25–responsive cells involved in this relay may produce and/or release reduced amounts of acetylcholine, or islet endocrine cells may exhibit reduced sensitivity to acetylcholine. Alternatively, transmitters produced and released by xenin-25–responsive cells in humans with T2DM may be different from those in humans with NGT and IGT. Studies are underway to distinguish between these different mechanisms.

Portions of this research were supported by funds from the following: National Institutes of Health (NIH) grants 5RC1-DK-086163 and 5R01-DK-088126, a Clinical/Translational Research Award from the American Diabetes Association, Washington University Diabetes Research and Training Center Immunoassay Core Grant P60-DK-020579, Washington University Nutrition Obesity Research Center Grant P30-DK-056341 from the National Institute of Diabetes and Digestive and Kidney Diseases, Washington University Clinical and Translational Science Award UL1-RR-024992, the Biologic Therapy Core Facility of the Alvin J. Siteman Cancer Center at Washington University School of Medicine and Barnes-Jewish Hospital in St. Louis, MO (National Cancer Institute Cancer Center Support Grant P30-CA-91842), NIH National Center for Research Resources grants P41-RR-00954 and UL1-RR-024992, and the Blum Kovler Foundation.

Washington University is pursuing a patent related to the use of xenin-25 to treat T2DM. In the future, this could lead to personal financial benefit to B.M.W., K.S.P., and Washington University. No other potential conflicts of interest relevant to this article were reported.

B.M.W., D.N.R., B.W.P., and K.S.P. researched data, contributed to discussion, and wrote, reviewed, and edited the manuscript. H.D.T. and M.J.W. researched data and reviewed and edited the manuscript. D.L.C., J.H.L., and D.T.V. researched data, contributed to discussion, and reviewed and edited the manuscript. J.D. contributed to discussion and reviewed and edited the manuscript. B.M.W. is the guarantor of this work and, as such, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Parts of this study were presented in abstract form at the 71st Scientific Sessions of the American Diabetes Association, San Diego, California, 24–28 June 2011.

The authors thank Heather Robertson, Diane Salamon, and Erin Laciny of Washington University School of Medicine for recruiting, screening, and scheduling participants and the nurses of the Clinical Research Unit for performing the GGIs.