Effect of Sitagliptin on Islet Function in Pancreatic Insufficient Cystic Fibrosis With Abnormal Glucose Tolerance

Abstract

Purpose

Impaired incretin secretion may contribute to the defective insulin secretion and abnormal glucose tolerance (AGT) that associate with worse clinical outcomes in pancreatic insufficient cystic fibrosis (PI-CF). The study objective was to test the hypothesis that dipeptidyl peptidase-4 (DPP-4) inhibitor-induced increases in intact incretin hormone [glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP)] concentrations augment insulin secretion and glucagon suppression and lower postprandial glycemia in PI-CF with AGT.

Methods

26 adults from Children’s Hospital of Philadelphia and University of Pennsylvania CF Center with PI-CF and AGT [defined by oral glucose tolerance test glucose (mg/dL): early glucose intolerance (1-h ≥ 155 and 2-h < 140), impaired glucose tolerance (2-h ≥ 140 and < 200 mg/dL), or diabetes (2-h ≥ 200)] were randomized to a 6-month double-blind trial of DPP-4 inhibitor sitagliptin 100 mg daily or matched placebo; 24 completed the trial (n = 12 sitagliptin; n = 12 placebo). Main outcome measures were mixed-meal tolerance test (MMTT) responses for intact GLP-1 and GIP, insulin secretory rates (ISRs), glucagon suppression, and glycemia and glucose-potentiated arginine (GPA) test-derived measures of β- and α-cell function.

Results

Following 6-months of sitagliptin vs placebo, MMTT intact GLP-1 and GIP responses increased (P < 0.001), ISR dynamics improved (P < 0.05), and glucagon suppression was modestly enhanced (P < 0.05) while GPA test responses for glucagon were lower. No improvements in glucose tolerance or β-cell sensitivity to glucose, including for second-phase insulin response, were found.

Conclusions

In glucose intolerant PI-CF, sitagliptin intervention augmented meal-related incretin responses with improved early insulin secretion and glucagon suppression without affecting postprandial glycemia.

Cystic fibrosis (CF) is a life-threatening autosomal recessive disorder caused by mutations affecting the gene encoding the CF transmembrane conductance regulator (CFTR) that result in severely impaired or absent function. CFTR normally functions as an anion channel transporting bicarbonate and chloride across epithelial membranes. Loss of CFTR function leads to exocrine pancreatic insufficiency, usually commencing in infancy, as well as to impaired respiratory secretion clearance, which contributes to increased susceptibility to sinopulmonary infections, pulmonary inflammation, and progressive lung function decline. CF-related diabetes (CFRD) affects 20% of adolescents and ~50% of adults with CF (1), and its development is associated with a reduction in nutritional status, pulmonary function, and survival (2). The increased risk for diabetes development in CF primarily affects individuals with pancreatic insufficiency (3) who have reduced β-cell secretory capacity compared to individuals with pancreatic sufficient CF (4). This contrast corroborates a reduction in functional β-cell mass in those affected by pancreatic exocrine disease. Additionally, worsening β-cell insulin secretory defects, manifesting as loss of early insulin secretion in the first 30-min following a meal, largely underlie the progression of abnormal glucose tolerance toward the development of CFRD (5), but the mechanisms underlying the progressive defects remain unknown.

Treatment with subcutaneous insulin is the mainstay of CFRD therapy, a recommendation motivated by the anabolic effects of insulin in a condition in which undernutrition has traditionally prevailed (6). Understandably, addition of multiple daily insulin injections to an already burdensome medical regimen is unacceptable for some patients, and oral therapy alternatives would be welcome. Moreover, therapies that may preserve or restore insulin secretion in people with pancreatic insufficient CF would be appealing for the possibility of early intervention that might prevent progressive deterioration of glucose tolerance in CF. The compromised insulin secretion in pancreatic insufficient CF may be explained by a combination of extension of pancreatic exocrine fibrosis and sclerosis to disrupt pancreatic islet structure and function (7), pancreatic islet inflammation (8), direct and indirect effects of abnormal CFTR function (9), genetic predisposition for type 2 diabetes (T2D) (10), as well as abnormal secretion of the incretin hormones glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP) (4,11,12). This last effect is best explained by exocrine pancreatic insufficiency that is incompletely addressed with pancreatic enzyme replacement therapy and, thus, residual maldigestion that impairs normal nutrient stimulation of incretin secretion. Compromised incretin secretion may be amenable to intervention using incretin-based therapeutics.

Both incretin hormones GLP-1 and GIP normally augment glucose-dependent insulin secretion following nutrient ingestion and are rapidly degraded following secretion due to N-terminal cleavage by the ubiquitous aminopeptidase protease dipeptidyl peptidase-4 (DPP-4) (13). GLP-1 agonists that resist inactivation by DPP-4 and inhibitors of DPP-4 that increase intact (active) GLP-1 and GIP concentrations are available as anti-hyperglycemic agents for the treatment of T2D. Sitagliptin (Januvia®), an oral DPP-4 inhibitor, was approved in 2006 for treatment of T2D (14,15). Daily dosing of 100 mg sustains ≥80% inhibition of DPP-4 and results in ≥2-fold increases in active GLP-1_{7-36 amide} and GIP_{1-42 amide} concentrations following meal ingestion (16,17). These actions are associated with enhanced glucose-dependent insulin secretion and glucagon inhibition as well as improvements in glucose excursion but without effects on satiety or gastric emptying that accompany subcutaneous administration of GLP-1 agonists (18). Because gastrointestinal symptoms and undernutrition are important considerations in people with pancreatic insufficient CF (19), inhibition of DPP-4 is an attractive approach to incretin-based therapy aimed at improving defective postprandial insulin secretion.

To test the hypothesis that chronic incretin-based therapy may improve glucose-dependent insulin secretion in pancreatic insufficient CF with abnormal glucose tolerance including early CFRD, we performed a 6-month randomized, placebo-controlled, double-blind clinical trial of oral sitagliptin and examined effects on meal-related incretin and islet hormone secretion, second-phase insulin secretion, β-cell secretory capacity, and α-cell suppression by glucose.

Participants and Methods

All participants were age ≥18 years on the date of consent, had a confirmed diagnosis of CF by CFTR mutation analysis or positive sweat test per the Cystic Fibrosis Foundation diagnostic criteria (20), and had pancreatic insufficiency requiring pancreatic enzyme supplementation. Pancreatic enzymes were dosed clinically between 1000 and 2500 lipase IU per kg per meal with all participants’ body weight stable and without symptoms of malabsorption. Participants were required to have had a 75-gram oral glucose tolerance test indicating abnormal glucose tolerance (5), defined here as early glucose intolerance (1-h glucose ≥ 155 mg/dL with 2-h glucose < 140 mg/dL), impaired glucose tolerance (2-h glucose ≥ 140 mg/dL and < 200 mg/dL), or CFRD (2-h glucose ≥ 200 mg/dL or previously confirmed diagnosis) without fasting hyperglycemia (fasting glucose < 126 mg/dL), within the previous 3 months.

Participants were excluded at baseline if they had clinically symptomatic pancreatitis within the previous 12 months, history of lung or liver transplant, significant kidney or liver dysfunction, were pregnant or nursing females, or had a history of anaphylaxis, angioedema or Stevens-Johnson syndrome. Baseline and 6-month follow-up studies were delayed for acute illness necessitating change of antibiotics or treatment with oral or intravenous glucocorticoids within the prior 4 weeks. Additional protocol details are available at ClinicalTrials.gov (NCT01879228). The study was approved by the institutional review boards of the University of Pennsylvania and the Children’s Hospital of Philadelphia. All participants provided written informed consent to participate.

Eligible subjects were randomized equally to 2 groups (sitagliptin 100 mg daily in the morning vs placebo for 6 months). Randomization was performed centrally by the University of Pennsylvania Investigational Drug Services to ensure allocation concealment; the investigative team and study participants were blinded to allocation. Balance between treatment arms was maintained by use of randomly permutated blocks with unequal block sizes and with stratification by glucose tolerance (early glucose intolerance, impaired glucose tolerance, CFRD). Sitagliptin and matching placebo tablets were provided by Merck Sharp and Dohme Corp and distributed by the University of Pennsylvania Investigational Drug Services.

Glycemic and Safety Assessments

All subjects received a OneTouch® Ultra® glucometer to monitor blood glucose and report any hypoglycemia (<70 mg/dL). Participants had in-person visits at baseline, and during follow-up at 1 month, 3 months, and 6 months for physical examination; glycosylated hemoglobin A1c (HbA1c), complete blood count, amylase, lipase, and a comprehensive metabolic panel were obtained. Additional phone visits occurred at 2 months, 4 months, and 5 months. During in-person or by-phone monthly visits, semistructured interviews, medications/intercurrent illnesses, glucose monitoring, insulin dosing (if applicable), and side effects were reviewed. The primary safety outcomes were the occurrence of any severe hypoglycemia episode requiring assistance of another person for correction (21), the occurrence of any episode of pancreatitis, and the occurrence of any hypersensitivity reaction. Additional safety outcomes included gastrointestinal complaints and weight loss.

Mixed-Meal Tolerance Test

A standardized mixed-meal tolerance test (MMTT) was used to evaluate postprandial glucose tolerance and incretin and islet hormone secretion (4,5) at baseline and 6 months. After a 12-h overnight fast, an antecubital or forearm vein catheter was placed for blood sampling. After approximately 20 min of acclimatization to the catheter, baseline blood samples were taken at t = −10 and −1 min before consumption of an 820-kcal meal over 15 min starting at t = 0. Meal composition was 47% carbohydrate, 40% fat, and 13% protein of the total energy content (22). Participants took their regularly prescribed dose of pancreatic enzyme replacement with the test meal. Additional blood samples were collected at t = 10, 15, 20, 30, 60, 90, 120, 150, 180, 210, and 240 min from the start of the meal. Rapid-acting insulin or repaglinide (23) was held for 12 h (if applicable) prior to testing. At the 6-month assessment, blinded sitagliptin or placebo was administered the morning of the MMTT.

Glucose-Potentiated Arginine Test

The glucose-potentiated arginine (GPA) test was performed at baseline and 6 months according to established methodology for evaluation of β-cell sensitivity to glucose and secretory capacity and α-cell suppression by glucose (4,5,24-26). After an overnight fast, an intravenous catheter was placed for infusions and a second catheter was placed for blood sampling. Baseline blood samples were taken at t = −5 and −1 min relative to injection of 10% arginine (5 g) over 1 min starting at t = 0. Additional blood samples were collected at t = 2, 3, 4, and 5 min. Beginning at t = 10 min, a hyperglycemic clamp technique (27) using a variable rate infusion of 20% dextrose was performed to achieve a plasma glucose (PG) concentration of ~230 mg/dL. Blood samples were taken every 5 min to adjust the infusion rate and achieve the desired PG concentration. After 45 min of glucose infusion (at t = 55 min), a second arginine pulse was injected with identical blood sampling. The first administration of arginine has no effect on the subsequent response to arginine using this protocol (28). A subsequent 2-h period without glucose infusion allowed PG to return to baseline. A second hyperglycemic clamp was then performed to achieve a PG concentration of ~340 mg/dL. After 45 min of glucose infusion, a third arginine pulse was injected with identical sampling. If applicable, rapid-acting insulin or repaglinide (23) was held for 12 h prior to testing; given the NPO status during the GPA, sitagliptin or placebo was held the morning of the 6-month assessment.

Biochemical Analyses

PG was measured in duplicate by the glucose oxidase method using an automated glucose analyzer (YSI 2300; Yellow Springs Instruments, Yellow Springs, OH, USA). Additional blood samples were collected into tubes on ice containing ethylenediamine tetra-acetate and protease inhibitor cocktail and, for the MMTT test, DPP-4 inhibitor (Sigma-Aldrich, St. Louis, MO, USA). Samples were centrifuged at 4°C, separated, and frozen at −80°C for subsequent analysis. Plasma free fatty acids (FFAs) were measured in duplicate using enzymatic colorimetrics (Wako Chemicals, Richmond, VA, USA). Plasma insulin (29), C-peptide (30), glucagon (31), and proinsulin (32) were assayed in duplicate by double-antibody radioimmunoassay (Millipore, Billerica, MA, USA). Intact GLP-1 (Millipore) (4) and intact GIP (Crystal Chem, Downers Grove, IL, USA) (33,34) were measured in duplicate by enzyme-linked immunosorbent assay. All immunoassays for baseline and 6-month assessments were run simultaneously.

Calculations

Mixed-Meal Tolerance Test

Insulin secretory rates (ISRs) were calculated from C-peptide values and derived by parametric deconvolution of C-peptide kinetics using a 2-compartment model (35) in WinSAAM software 3.0.8 (University of Pennsylvania, New Bolton Center, Kennett Square, PA, US). Thirty- and 180-min incremental areas under the curve (AUCs) for glucose, ISRs, glucagon, FFAs, GLP-1, and GIP were calculated with baseline values subtracted using the trapezoidal method in STATA 15 software (StataCorp LP, College Station, TX, USA).

GPA test

Acute insulin, C-peptide, proinsulin, and glucagon responses to arginine (AIR_arg, ACR_arg, APR_arg, and AGR_arg, respectively) were calculated as mean of 2-, 3-, 4-, and 5-min values minus mean of baseline values (27). Acute responses during the 230 mg/dL clamp enable determination of glucose potentiation of arginine-induced insulin (AIR_pot), C-peptide (ACR_pot), and proinsulin (APR_pot) release and glucose inhibition of arginine-induced glucagon release. Acute responses during the 340 mg/dL clamp allow for determination of maximum arginine-induced insulin (AIR_max), C-peptide (ACR_max), and proinsulin (APR_max) release (ie, β-cell secretory capacity) and minimum arginine-induced glucagon (AGR₃₄₀) release (24).

The fasting proinsulin-to-C-peptide ratio was calculated as the molar concentration of proinsulin divided by the molar concentration of C-peptide × 100 (36). Estimation of proinsulin-to-C-peptide ratio within β-cell secretory granules is most reliable after acute stimulation of secretion (37); therefore, we examined the proinsulin secretory ratio (PISR) in response to each injection of arginine from respective acute proinsulin:C-peptide responses (4,5).

As a measure of second-phase insulin response and β-cell sensitivity to glucose, the difference between insulin levels during the 230 mg/dL glucose clamp and fasting divided by the difference in corresponding PG concentrations was calculated. This change represents the slope over time in 2 dimensions of insulin (y) and glucose (x): (insulin at glucose 230 mg/dL – fasting insulin)/ΔPG. The glucose-potentiation slope (GPS) was calculated as the difference in acute insulin responses at the 230 mg/dL and fasting glucose concentrations divided by the difference in PG at the AIR_pot and the fasting glucose at the AIR_arg: (AIR_pot − AIR_arg)/ΔPG. β-cell sensitivity to glucose was further assessed as the PG₅₀, the PG at which half-maximal insulin secretion is achieved, using the y-intercept (b) of the GPS to solve the equation ½(AIR_max) = (GPS ∙ PG₅₀) + b (4,5,24-26). Similar calculations substituting C-peptide for insulin were performed. Insulin sensitivity (M/I) was determined by dividing mean glucose infusion rate required during the 230 mg/dL glucose clamp (M) by mean prestimulus insulin level (I) between 40 and 45 min of glucose infusion.

Subject Compliance Monitoring

Patients were instructed to bring study medication to each visit. Pill counts; monthly calendars for charting; and completion of Adherence Questionnaire, a semistructured interview, were conducted monthly by study staff by phone or in-person to assess adherence. If a patient was found to have reduced compliance (defined as <80% expected), a member of the study staff phoned the patient on a regular basis to remind the patient to take their study medication.

Spirometry (Pulmonary Function Tests)

Pulmonary function data were obtained from the pulmonary outpatient chart or PortCF database from the 6 months prior to enrollment through the intervention period. Measured forced vital capacity and forced expiratory volume in 1 s (FEV₁) were converted to percentage predicted values (FEV₁%-predicted) using the Global Lung Initiative predicted equations (38).

Statistical Analyses

Graphs were first generated to permit visual inspection of data by allocation group and to inform analyses surrounding MMTT-related glucose and FFA excursion as well as hormonal responses. Baseline and follow-up continuous data are reported as median (min; max) and categorial data as percentages, except where indicated. Safety data are described. Initial comparisons of (1) baseline data between allocation groups and (2) baseline and 6-month data within allocation group were performed using (i) unpaired t-test or Wilcoxon rank-sum test for continuous data and chi² for categorical data and (ii) paired t-test or Wilcoxon matched-pairs signed-rank test, respectively. Significance was considered at P < 0.05 (2-tailed).

The primary endpoint designated was the change in second-phase insulin response derived from the GPA test; this second-phase response was chosen as the primary endpoint based upon findings of significant increases in second phase insulin secretion in response to GLP-1 infusion in islet transplant recipients who are known to have reduced β-cell secretory capacity similar to that found in individuals with pancreatic insufficient CF and abnormal glucose tolerance (39). To assess for within subject treatment effectiveness, defined as the relative influence of sitagliptin on slope of insulin to glucose, we used longitudinal mixed effects models with the slope as the outcome, time (after intervention vs baseline) and treatment (sitagliptin vs placebo) as main effects, and time-by-treatment interaction as the estimate of interest (40). This method (1) allows for robust variance estimates that account for possible nonnormality of slope measures, (2) takes advantage of improved efficiency in a statistical model that adjusts for subject-level covariates, (3) benefits from the strong correlation of slope within a subject over time, and (4) models variance-covariance of the outcome with subject-specific slope terms with unstructured correlation structure through the specification of random effect terms. Model convergence and goodness of fit were assessed for all models. For these subject-specific models, convergence issues will be seen for measures where the variance decreases over time, referred to as negative curvature of the variance [ie, (41), pp. 52-54]. In this situation, to account for the within-subject correlation, our focus was modeling the variance-covariance matrix of the repeated measures rather than subject-specific effects, which Verbeke and Molenbergh refer to as the marginal model structure (41). Similar models were performed to assess for differences by allocation group in changes in all assessed MMTT and GPA parameters. With more timepoints, more flexible patterns of change over time can be fit such as multiple phases of change connected through the longitudinal period with separate slopes estimated per phase (42). Subject-specific slopes per phase are specified with an unstructured correlation structure for the random slope terms. Graphical inspection of MMTT-related ISR reflected 3 phases of change: 0- to 60-min, 60- to 120-min, and 120- to 180-min. To test for between treatment group changes in the ISR trajectory during these 3 phases, slopes were compared between group over time using mixed effects models (group * visit * phase) adjusted for ISR at the start of the phase. Analyses were performed using Stata v15 (StataCorp LP, College Station, TX, USA).

Sample size

Based upon a generalized estimating equation model, power calculations were performed to account for the within subject correlation of measures over time and thus to take advantage of the efficiency of a pre-post design using the function “compare_slope2” prepared by Shults and Hilbe for Stata [(43), chapter 9] and confirmed by (44). Work from our group has found a second-phase insulin response of 0.063 with an SD of 0.08 in adults with pancreatic insufficient CF and abnormal glucose tolerance. In response to an acute infusion of GLP-1, the slope increased by 0.25 (µU/mL)/(mg/dL) ~ effect size = 3.1. This response was less robust in CFRD [0.1 (µU/mL)/(mg/dL)] ~ effect size = 1 to 1.25; we expected smaller effect sizes with chronic sitagliptin therapy. Assuming strong intraclass correlations (based upon our longitudinal work in CF) of 0.7 to 0.85, effect sizes of 0.63 to 0.92 can be found with 80% power with 12 participants per group. Previous work from our group suggests that among normal subjects, 1 SD in this response equates to 0.06 (µU/mL)/(mg/dL) (45).

Results

Participant Characteristics

All 26 screened individuals were randomized to either sitagliptin (n = 13) or placebo (n = 13), and 24 [sitagliptin (n = 12) and placebo (n = 12)] completed the study (Fig. 1). One participant randomized to sitagliptin developed a rash within a few days of starting study drug and was withdrawn from the study; 1 participant in the placebo group was withdrawn due to history of angioedema that was identified after the randomization visit.

Figure 1.

CONSORT flow diagram. Eligibility was assessed at a screening visit, and double-blind randomization occurred following completion of baseline mixed-meal tolerance and glucose-potentiated arginine test visits. One subject randomized to placebo provided additional medical history at the randomization visit that met an exclusion criterion (history of angioedema) and so was removed from the study prior to initiating intervention. One subject randomized to sitagliptin developed a pruritic maculopapular rash affecting both arms after 1 week of intervention and was withdrawn from the study. All remaining subjects completed the 6-month intervention and follow-up mixed-meal tolerance and glucose-potentiated arginine test visits and are included in the analysis.

Baseline characteristics were similar between groups, except for body weight and body mass index (BMI), which were higher in the sitagliptin group (P = 0.014 and P = 0.011, respectively) (Table 1) and translated into 2 obese participants in the sitagliptin group but none in the placebo group despite randomization. As expected by the stratified randomization scheme, glucose tolerance categories were similarly represented between the 2 groups (Table 1). CFRD without fasting hyperglycemia was present in 2 participants in the sitagliptin group (only 1 of whom was treated with insulin: 1 unit of rapid acting insulin for 13 grams of carbohydrates; total daily dose of 36 units) and 3 participants in the placebo group (only 1 of whom was treated with insulin: 14 units of 70/30 insulin prior to overnight continuous enteral feed). Four subjects in each group entered the study on CFTR modulator therapy [sitagliptin group: ivacaftor (n = 2), lumacaftor/ivacaftor (n = 1), tezacaftor/ivacaftor (n = 1); placebo group: ivacaftor (n = 1), lumacaftor/ivacaftor (n = 3)]. Two subjects in the sitagliptin group, both homozygous for ΔF508 mutations in CFTR, started lumacaftor/ivacaftor after their 4-month visit. Adherence to sitagliptin or placebo was similar in both groups and was generally >85% (Fig. 2A). Following 3 month of intervention, HbA1c was slightly lower in the sitagliptin group (−0.05; 95% CI: −0.17 to 0.08), a difference that largely reflected an increase in the placebo group (0.16; 95% CI: 0.04 to 0.30; P = 0.023); these differences were no longer present at 6 months (P = 0.89) (Fig. 2B).

Table 1.

Participant characteristics at baseline

	Sitagliptin (n = 12)	Placebo (n = 12)
Age, years	24 (18-43)	28.5 (21-40)
Female, n (%)	5 (42)	5 (42)
Weight, kg	71.6 (57-102.5)	59.8 (35.8-78)a
BMI, kg/m2	24.5 (18.8-33.9)	21.6 (17.2-24.4)*
PERT dose per meal, lipase IUb	135 000 (840 000-200 000)	120 000 (48 000-214 000)
FEV1, %-predicted	84 (50-109)	68 (33-97)
CFTR mutation status
delF508 homozygous	7	6
delF508 heterozygous	5	5
other	0	1
Glucose Tolerance
EGI, n	4	4
IGT, n	6	5
CFRD w/o FH, n	2	3
OGTT profile
fasting glucose, mg/dL	95 (77-103)	92 (77-108)
1-h glucose, mg/dL	220 (131-253)	233 (163-303)
2-h glucose, mg/dL	172 (89-309)	166 (34-269)
HbA1c, %	5.6 (4.8-6.1)	5.7 (5.4-5.9)

Data are medians and ranges (min-max) for continuous variables and number and percentage (%) for categorical variables.

Abbreviations: BMI, body mass index; CFRD w/o FH, cystic fibrosis-related diabetes without fasting hyperglycemia; CFTR, cystic fibrosis transmembrane conductance regulator, EGI, early glucose intolerance; FEV₁, forced expiratory volume in 1-second; IGT, impaired glucose tolerance; OGTT, oral glucose tolerance test; PERT, pancreatic enzyme replacement therapy.

^a P < 0.05 for comparison between sitagliptin and placebo groups.

^b PERT dose given is that administered for the mixed-meal tolerance test at baseline and 6 months.

Figure 2.

(A) Adherence as assessed by pill count at 1, 3, and 6 months following randomization to sitagliptin 100 mg daily or matched placebo. (B) Glycemic control as assessed by HbA1c at baseline and at 3 and 6 months following randomization to sitagliptin 100 mg daily or matched placebo. HbA1c was slightly lower in the sitagliptin group at 3 months (−0.05; 95% CI: −0.17 to 0.08), a difference that largely reflected a nonsustained increase in the placebo group (0.16; 95% CI: 0.04 to 0.30; P = 0.023); no differences were present at 6 months (P = 0.89).

Mixed-Meal Tolerance Test

Following 6 months of intervention, postprandial responses for intact GLP-1 and intact GIP increased with sitagliptin vs no change with placebo over both 30- and 180-min AUCs (P < 0.01 for all comparisons) (Table 2, Fig. 3), consistent with the anticipated DPP-4 inhibition activity of sitagliptin. Glucose peaked on average at 60 min, and the peak glucose was not different between groups at baseline [sitagliptin: 192 (115; 272) mg/dL vs placebo: 207 (168; 280) mg/dL; P = 0.19], or following 6 months of intervention (Fig. 4A and 4B). No differences were present after 6 months in fasting, 60-min, or 120-min peak glucose; timing of peak glucose; or AUC_glc following meal ingestion between groups (Table 2, Fig. 4).

Table 2.

Mixed-meal tolerance test responses during the 30- and 180-min postingestion prior to and following 6 months of sitagliptin or placebo

	Sitagliptin (n = 12)			Placebo (n = 12)			Δ sitagliptin vs Δ placebo, P-value
	Baseline	6 months	P-value	Baseline	6 months	P-value
30-min AUC
Intact GLP-1, pmol·min/L	34 (−74 to 159)	223 (68.3 to 358)	0.002	58.2 (−17.3 to 156)	95.4 (−20 to 236)	0.43	<0.001
Intact GIP, pmol·min/La	159 (53.5 to 953)	555 (152 to 1584)	0.028	260 (39.3 to 783)	315 (−25 to 617)	0.95	0.009
Glucose, mg·min/dL	544 (220 to 698)	609 (101 to 1259)	0.21	683 (236 to 1108)	566 (95 to 1021)	0.69	0.18
ISR, mU/L	650 (112 to 2120)	981 (163 to 1311)	0.58	482 (78 to 1226)	502 (91 to 1692)	0.58	0.45
ISR/Glucose mU·min/mg	0.022 (0.004 to 0.106)	0.021 (0.006 to 0.056)	0.75	0.016 (0.002 to 0.032)	0.016 (0.004 to 0.032)	0.48	0.99
Glucagon, pg·min/mL	150 (0 to 320)	73.8 (−175 to 438)	0.2	52.5 (−167 to 380)	134 (−35 to 498)	0.17	0.1
Free Fatty acids, mmol·min/L	0.07 (−3.0 to 2.2)	−1.5 (−8.7 to 5.1)	0.29	−0.7 (−17.9 to 2.0)	−1.0 (−5.9 to 0.9)	0.67	0.3
180-min AUC
Intact GLP-1, pmol·min/L	460 (−801 to 1107)	1334 (902 to 4111)	0.022	730 (−29 to 1839)	638 (−368 to 1323)	0.31	<0.001
Intact GIP, pmol·min/La	3608 (2020 to 7303)	9709 (3608 to 29 524)	0.005	6070 (1488 to 10 789)	5632 (2203 to 8175)	0.78	0.001
Glucose, mg·min/dL	5128 (2606 to 13 105)	6063 (1676 to 11 518)	0.43	9915 (4581 to 19 079)	7159 (4567 to 19 198)	0.64	0.96
ISR, mU/L	7759 (4175 to 30 581)	9309 (3972 to 14 954)	0.58	10 836 (4586 to 13 641)	9184 (4485 to 13 626)	0.82	0.6
ISR/Glucose mU·min/mg	0.021 (0.009 to 0.156)	0.024 (0.009 to 0.071)	0.27	0.017 (0.009 to 0.039)	0.02 (0.006 to 0.036)	0.48	0.6
Glucagon, pg·min/mL	541 (−1378 to 3693)	473 (−2118 to 2153)	0.18	143 (−1000 to 3065)	910 (−500 to 2583)	0.038	0.025
Free Fatty acids, mmol·min/L	−32.4 (−62.8 to −16.5)	−42.4 (−109 to −13.8)	0.39	−50.8 (−185 to −22.4)	−44.4 (−75.2 to −24.8)	0.75	0.15

Data are medians and ranges (min-max).

Abbreviations: AUC, incremental area-under-the-curve; GLP-1, glucagon-like peptide-1; GIP, glucose-dependent insulinotropic polypeptide; ISR, insulin secretory rate.

^an = 10 in the sitagliptin group and n = 9 in placebo group.

Figure 3.

Plasma levels of intact GLP-1 and intact GIP prior to (-●-) and 6 months following (-o-) intervention with sitagliptin 100 mg daily (A and C) or matched-placebo (B and D) in response to the mixed-meal tolerance test in subjects with pancreatic insufficient CF. Responses for intact GLP-1 (AUC_GLP-1) and intact GIP (AUC_GIP) increased after 6 months of sitagliptin vs placebo over the first 30 and 180 min (P < 0.01 for all comparisons). Data are given as mean ± SE.

Figure 4.

Plasma glucose and insulin secretory rates prior to (-●-) and 6 months following (-o-) intervention with sitagliptin 100 mg daily (A and C) or matched placebo (B and D) in response to the mixed-meal tolerance test in subjects with pancreatic insufficient CF. No differences in glucose excursion were found between sitagliptin (A) and placebo (B). While there was no difference in the ISR response assessed by 30- or 180-min AUC_ISR, ISR during the first 60 min was faster (P = 0.049) and declined more rapidly in the subsequent 60 min (P < 0.0001) following 6 months of sitagliptin (C) vs placebo (D). Data are given as mean ± SE.

Postprandial responses for ISR did not change following 6 months of intervention with either sitagliptin or placebo as assessed by 30- and 180-min AUC, including when adjusted for the corresponding AUC_glu (Table 2). However, prompted by the ISR graphs in Figure 4C, we tested for between treatment group changes in the ISR trajectory (slope) during 3 phases: insulin secretion increased more quickly through the first 60 min (P = 0.049) and decreased more rapidly over the subsequent 60 to 120 min (P < 0.0001) of the MMTT following 6 months of sitagliptin vs placebo (Fig. 4D). Not unexpectedly, with this shift to earlier insulin secretion, the ISR slope decreased during the 120 to 180 min interval in the sitagliptin group vs no change with placebo (P < 0.0001). Postprandial glucagon suppression was modestly more robust over 180 min after 6 months of sitagliptin vs placebo intervention; this difference likely reflects the unexpected finding of less suppression in the placebo group in whom the AUC_gln was higher at 6 months combined with a modest enhancement in glucagon suppression with sitagliptin (P = 0.025). Table 2 presents glucagon incremental AUC while Figure 4 shows the dynamic response during the MMTT. No differences in suppression of FFAs were observed between groups over the 6-month interval (Table 2, Fig. 5).

Figure 5.

Plasma glucagon and free fatty acids prior to (-●-) and 6 months following (-o-) intervention with sitagliptin 100 mg daily (A and C) or matched-placebo (B and D) in response to the mixed-meal tolerance test in subjects with pancreatic insufficient CF. Postprandial glucagon suppression was greater over 180 min (AUC_gln; P = 0.025) following 6 months of sitagliptin (A) vs placebo (B). There were no differences in free fatty acid suppression between sitagliptin (C) and placebo (D). Data are given as mean ± SE.

Glucose-Potentiated Arginine

Despite randomization, baseline acute insulin responses were greater in the sitagliptin cohort (all P < 0.05) and were overall higher in the 2 obese participants. Following 6 months of intervention, no significant differences in fasting insulin or C-peptide or the acute insulin or C-peptide responses to arginine under fasting, 230-mg/dL hyperglycemic clamp, or 340-mg/dL hyperglycemic clamp conditions were found between the sitagliptin and placebo groups (Table 3, Figs. 6A and 6B and 7C and 7D). Fasting glucagon, AGR_arg and AGR_min were lower after 6 months of sitagliptin compared to placebo (P < 0.05 for all comparisons) (Table 3, Fig. 6C and 6D), with a weak trend for reduction in AGR_inh (P = 0.12). Neither fasting proinsulin nor the acute proinsulin response under fasting conditions (APR_arg) differed over 6 months, but a trend for lower acute proinsulin responses under glucose-potentiated conditions (APR_pot and APR_max) after 6 months of sitagliptin was found when compared to placebo (P ≤ 0.1 for both comparisons) (Table 3, Fig. 7A and 7B). However, no differences in PISRs occurred following 6 months of intervention (Table 3).

Table 3.

Glucose-potentiated arginine test parameters prior to and following 6 months of sitagliptin or placebo

	Sitagliptin (n = 12)			Placebo (n = 12)
	Baseline	6 months	P-value	Baseline	6 months	P-value	Δ sitagliptin vs Δ placebo P-value
Insulin, µU/mL
Fasting	9.5 (4.6-19.2)	8.0 (2.6-21.1)	0.86	5.9 (3.4-13.3)	6.5 (2.9-8.4)	0.35	0.5
AIRarg	25.5 (9.7-36.9)	15.6 (3.7-51.6)	0.14	13.1 (4.4-26.1)	9.2 (4.8-22)	0.24	0.36
AIRpot	53.9 (6.1-154)	46.8 (15-103)	0.14	27.6 (16.7-54.1)	32.8 (9.9-62.7)a	0.48	0.075
AIRmax	87.1 (36.2-144.5)b	71.5 (17-146.7)	0.14	36.7 (15.6-83)	35.4 (15.6-78)	0.58	0.068
C-peptide, ng/mL
Fasting	1.1 (0.32-3.19)	1.4 (0.36-2.41)	0.58	0.9 (0.38-1.85)	0.91 (0.41-1.64)	0.98	0.6
ACRarg	0.92 (0.32-2.01)	0.88 (0.2-1.7)	0.39	0.72 (0.33-1.06)	0.67 (0.34-0.99)	0.35	0.89
ACRpot	2.6 (0.79-5.59)	2.89 (0.65-6.39)	0.75	1.99 (0.80-3.88)	1.91 (0.9-3.66)	0.94	0.98
ACRmax	3.66 (1.46-6.34)	2.98 (0.93-5.51)b	0.44	1.93 (1-5.15)	1.78 (1-6.65)	0.88	0.34
Glucagon, pg/mL
Fasting	58.8 (27-84)	46.3 (33.5-78)	0.13	51.5 (40-83.5)	54.5 (41-79.5)	0.48	0.037
AGRarg	50.5 (1.25-117)	37.8 (8.5-54.8)	0.063	35 (17.5-108)	35.4 (17.8-70)	0.94	0.016
AGRinh	34.1 (11-79.8)	34 (4-45.3)	0.1	28.1 (11-79.8)	30.3 (16.5-50.5)	0.94	0.12
AGRmin	34.8 (4-63)	23.4 (0-43)b	0.07	27.3 (6.8-75.8)	28.4 (12.5-57.1)	0.72	0.049
Proinsulin, pmol/L
Fasting	7.73 (2-15.25)	5.93 (2-31.65)	0.18	6.28 (2-10.45)	6.45 (2-13.4)	0.81	0.87
APRarg	6 (1.53-8.93)	3.55 (0-12.95)	0.48	3.78 (0-8)	1.88 (0-10.78)	0.08	0.78
APRpot	13.59 (5.7-29.1)	11.3 (4-21.63)	0.06	8.59 (2.24-15.45)	8.33 (0-18.15)	0.33	0.16
APRmax	17.2 (5.38-24.2)	11.84 (4.78-24.2)b	0.07	12.94 (4.28-16.75)	11.63 (4.63-17.33)	0.58	0.14
Proinsulin secretory ratios (PISR)
Fasting proinsulin:C-peptide ratio, %	1.94 (1.02-4.83)	1.4 (0.45-5.09)	0.31	1.63 (0.73-3.62)	1.58 (0.69-3.87)	0.64	0.5
PISRarg, %	1.62 (0.76-4.13)	1.72 (0-5.04)	0.94	1.34 (0-2.88)	0.96 (0-3.28)	0.07	0.60
PISRpot, %	1.82 (0.75-5.88)	1.7 (0.54-3.97)	0.14	1.37 (0.53-4.4)	1.3 (0-2.61)	0.31	0.79
PISRmax, %	1.55 (0.3-3.97)a	1.3 (0.67-2.97)	0.28	1.73 (0.081-2.85)	1.44 (0.72-2.84)	0.53	0.44
β-cell sensitivity to glucose
Insulin230− Insulin0 PG230-PG0	0.11 (0.05-0.906)	0.085 (0.024-0.459)	0.39	0.055 (0.022-0.13)	0.053 (0.025-0.128)	0.93	0.32
GPS, from insulin	0.23 (0.12-1.3)	0.29 (0.1-0.74)	0.54	0.15 (0.03-0.29)	0.17 (0.06-0.39)	0.06	0.11
GPS, from C-peptide	0.011 (0.002-0.04)	0.013 (0.004-0.045)	0.48	0.01 (0.002-0.024)	0.01 (0.004-0.023)	0.75	0.67
PG50, mg/dL, from insulin	155 (101-214)	162 (133-174)a	0.58	139 (70.3-198)	142 (80.9-188)a	0.96	0.83
PG50, mg/dL from C-peptide	156 (110-251)b	152 (106-172)b	0.80	121 (95-182)	140 (87-214)	0.75	0.24
Insulin sensitivity
M/I, mg/kg per min/μU/mL	0.41 (0.06-0.65)	0.44 (0.11-0.95)	0.21	0.61 (0.32-1.06)	0.62 (0.41-1.21)	0.53	0.88

Data are medians and ranges (min-max).

Abbreviations: GPS, glucose potentiation slope; M/I, insulin sensitivity calculated as the mean glucose infusion rate during 230 mg/dL glucose clamp (M) divided by the mean pre-stimulus insulin level (I); PG₅₀, plasma glucose at half-maximal insulin secretion.

^a n = 11.

^b n = 10.

Figure 6.

Acute insulin responses to arginine under fasting, 230-mg/dL hyperglycemic clamp, and 340-mg/dL hyperglycemic clamp conditions prior to (-●-) and 6 months following (-o-) intervention with sitagliptin 100 mg daily (A) or matched-placebo (B) identify no augmentation of insulin secretion. Acute glucagon responses to arginine were lower under fasting (AGR_arg; P = 0.016), trended lower under 230-mg/dL hyperglycemic clamp (AGR_inh; P = 0.12), and were lower under 340-mg/dL hyperglycemic clamp (AGR_min; P = 0.049) conditions following 6 months of sitagliptin (C) vs placebo (D). Data are given as mean ± SE.

Figure 7.

Acute proinsulin and C-peptide responses to arginine prior to (-●-) and 6 months following (-o-) intervention with sitagliptin 100 mg daily or matched placebo. Acute proinsulin responses were not different under fasting but trended lower under 230-mg/dL hyperglycemic clamp, and 340-mg/dL hyperglycemic clamp conditions prior to (-●-) and 6 months following (-o-) intervention with sitagliptin 100 mg daily (A) or matched-placebo (B) identify no augmentation of insulin secretion. Acute C-peptide responses to arginine under fasting, 230-mg/dL hyperglycemic clamp, and 340-mg/dL hyperglycemic clamp conditions prior to (-●-) and 6 months following (-o-) intervention were not different after sitagliptin 100 mg PO daily (A) or matched-placebo (B). Data are given as mean ± SE.

After 6 months of intervention, the primary outcome, second-phase insulin response, was unchanged, as were the glucose potentiation slope and PG₅₀ in both sitagliptin and placebo groups (Table 3). While at baseline, fasting insulin was higher (P = 0.023) and insulin sensitivity as measured by M/I (P = 0.015) was lower in the sitagliptin group due to the 2 obese participants in this cohort (data not shown), no change in fasting insulin (data not shown) or M/I occurred in either group with the 6-month intervention (Table 3).

Safety of Sitagliptin

One participant in the sitagliptin group was withdrawn due to rash, presumably from hypersensitivity, a rare but known potential allergic reaction with DPP-4 inhibition (46). No patients experienced severe hypoglycemia. The participant with CFRD randomized to the sitagliptin group had 4 episodes of glucose <54 mg/dL within month 1 and required a decrease in the insulin to carbohydrate ratio from 1:13 to 1:15 grams; 3 additional episodes occurred in the subsequent 5 months. Two participants in the placebo group each had 1 episode of glucose <54 mg/dL within the first month of intervention.

No episodes of pancreatitis and no elevations in either amylase or lipase occurred. Two weeks after starting study sitagliptin, 1 participant reported decreased appetite, nausea, and vomiting occurring over approximately 10 days; symptoms resolved without recurrence or interruption of therapy.

While BMI was higher in the sitagliptin group at baseline and follow-up, the 6-month change in BMI was not different between groups (P = 0.25). Maximum weight loss in the sitagliptin group was 2.6 kg in a female with baseline BMI = 33.9 kg/m² and in the placebo group was 2.5 kg in a male with baseline BMI = 22.8 kg/m².

Similarly, while FEV₁%-predicted tended to be higher in the sitagliptin group at baseline and follow-up, the 6-month change in FEV₁%-predicted was not different between groups (P = 0.48); the maximum FEV₁%-predicted decrease was 9.3 in the sitagliptin group and 13.5 in the placebo group.

Discussion

In this preliminary study of chronic incretin-based therapy in pancreatic insufficient, glucose intolerant CF, postprandial concentrations of intact incretin hormones were more robust, glucagon suppression improved modestly, and insulin secretory phases during a meal partly restored following 6 months of sitagliptin therapy. Despite these findings, no improvements were seen in meal-related glucose excursion or in second phase, or arginine-stimulated insulin, responses. There was a reduction in fasting and glucose-inhibited acute glucagon responses to arginine after 6 months of sitagliptin, findings consistent with the greater meal-related suppression of glucagon and supporting an effect of DPP-4 inhibition to enhance glucose-dependent regulation of islet α-cell function. An initial, limited, but not sustained difference in HbA1c was found after 3 months of sitagliptin compared to placebo, a difference likely attributable to the concomitant slight HbA1c increase in the placebo group. While these findings do not provide robust evidence for glucose-lowering effects of chronic DPP-4 inhibitor treatment in adults with pancreatic insufficient CF, the earlier phased secretion of insulin following nutrient ingestion and greater glucose-dependent suppression of glucagon secretion suggest an improvement of islet β- and α-cell function in pancreatic insufficient CF.

As a DPP-4 inhibitor, sitagliptin interferes with catabolism of both GLP-1 and GIP, which are otherwise rapidly degraded following stimulation of secretion. Indeed, enhanced postprandial responses for intact GLP-1 and GIP accompanied 6 months of sitagliptin therapy in our CF cohort. Pancreatic enzyme replacement therapy is recognized to improve total GLP-1 and total GIP concentrations following a meal (11,12). The enhanced intact GLP-1 and intact GIP secretion during the meal test cannot be attributed to pancreatic enzyme replacement therapy in the sitagliptin group; the experimental condition for the mixed meal test included administration of the same, individualized clinically prescribed pancreatic enzyme replacement therapy dose at baseline and 6 months. The increase in intact GLP-1 and GIP responses were comparable to those found following a single sitagliptin dose of 100 mg given before a meal test in adults with early T2D and healthy controls in whom glucose excursion was reduced by 14% and insulin secretion rates were left-shifted but, similarly to our study, had no effect on insulin or C-peptide AUC or measures of β-cell sensitivity to glucose (33). This intact GLP-1 and GIP profile, however, did not translate into dampened glucose excursion in our CF cohort with abnormal glucose tolerance. We previously reported a similar lack of glucose-lowering effect with sitagliptin 100 mg daily for 6 months in a randomized clinical trial involving individuals with early T2D (26). Recognizing that in pancreatic insufficient CF (1) loss of early-phase insulin secretion is present even with subtle glucose abnormalities (oral glucose tolerance test 1-h glucose > 155 mg/dL but 2-h glucose < 200 mg/dL) (5), (2) this loss of early-phase insulin secretion is accompanied by delayed, compensatory increased insulin secretion at 60 to 90 minutes, and (3) this compensatory late-phase insulin gradually declines with worsening glucose tolerance (4,5), the meal-related glucose and insulin secretion profile was further examined. Our data suggest some restoration of this earlier insulin secretion and subsequent reduction of compensatory late-phase insulin secretion with sitagliptin.

Alterations in glucagon secretion in the CF pancreas may provide 1 mechanism for the absent improvement in glucose excursion during MMTT with sitagliptin. Hyperglucagonemia is well-recognized as a major contributor to hyperglycemia in T2D. Thus, GLP-1 may be a particularly effective therapeutic for T2D because it suppresses glucagon secretion (17). Indeed, the intact GLP-1 responses following a single sitagliptin dose in healthy controls and adults with well-controlled T2D were associated with overall enhanced insulin secretion but a trend toward reduced glucagon secretion only in the T2D group, the latter of which could contribute to improved glucose tolerance (17). In contrast, frank hyperglucagonemia is not a feature of abnormal glucose tolerance in CF. In fact, nondiscriminatory islet cell destruction contributes to an overall glucagon reduction in pancreatic insufficient CF (4). Thus, in the absence of frank hyperglucagonemia, sitagliptin-mediated enhancements in GLP-1 may not impact glucose excursion, while still affecting glucose-dependent regulation of remaining islet cell function.

Our group previously observed no difference in acute glucagon responses during GPA testing in individuals with pancreatic insufficient CF by groups categorized by worsening glucose tolerance (5). Because graded reductions in acute insulin and C-peptide responses during GPA testing and in early-phase insulin secretion during MMTT were associated with progressive impairment in oral and postprandial glucose tolerance, we concluded that defects in insulin secretion rather than enhanced glucagon secretion was responsible for the progression of glucose intolerance in pancreatic insufficient CF. We subsequently observed that 16 weeks of CFTR modulatory treatment (ivacaftor in individuals with at least 1 CFTR gating or conductance mutation) increased insulin secretion in response to GPA that was associated with a proportionate reduction in the glucose-inhibited acute glucagon response to arginine (9). Thus, glucagon secretion in CF may become relatively elevated for the declining insulin secretory capacity during the progression of islet dysfunction and still represent a target of intervention for improving islet health and delaying worsening of abnormal glucose tolerance. Reciprocal impairment of glucose-potentiation and inhibition of arginine-induced release of insulin and glucagon, respectively, has been described in non-CF individuals with impaired glucose tolerance (47). Further study of the interaction of CFTR modulators and incretin-based therapeutic approaches on islet function in pancreatic insufficient CF are warranted.

Recent data suggest GIP plays a greater role in meal-related insulin secretion (48), and loss of GIP insulinotropic effect in T2D is well-recognized (49). Unlike GLP-1, GIP augments glucagon secretion (50). However, despite augmented, meal-related intact GIP concentrations in response to sitagliptin, glucagon secretion is not enhanced by DPP-4 inhibition in healthy controls or early T2D (33). Similarly, sitagliptin resulted in enhanced glucagon suppression in response to arginine stimulation in our CF cohort; consistent with this finding but less robust was the modest enhancement in glucagon suppression during a meal. Thus, as in non-CF individuals with glucose intolerance, the glucagonostatic effects of increased active GLP-1 action appear to dominate over the glucagon stimulating effects of increased active GIP during treatment with DPP-4 inhibitors. An additional consideration is inclusion of intact GIP and intact GLP-1 in our study of DPP-4 inhibition. Reduced degradation of active incretin concentrations is expected with sitagliptin therapy and (51) was confirmed here. A feedback mechanism is suspected to limit GLP-1 and GIP secretion and thus total GLP-1 and total GIP. Indeed, no change in total GIP was found after 6 months of sitagliptin (data not shown), a finding that highlights the importance of measuring active incretin hormones when considering effects of DPP-4 inhibition on the enteroinsular axis. Whether the increased potency of GLP-1 receptor agonists for augmenting glucose-dependent insulin secretion may lead to improvement in glucose tolerance for individuals with pancreatic insufficient CF also requires further study (52).

Preservation of insulin secretion and interruption of progression to diabetes are important goals in CF, but such efficacious interventions have yet to be identified. Incretin-based therapies are attractive because they may interfere with β-cell apoptosis. The extent to which sitagliptin could preserve β-cell function and delay or prevent emergence of worsening glucose tolerance and CFRD would require a longer intervention with a larger sample size and cannot be addressed by this study. Progressively higher PISR in response to arginine during maximal glucose-potentiation have been demonstrated with worsening glucose tolerance in CF and may indicate increased insulin secretory demand, which could exacerbate underlying insulin secretion defects (5). Acute infusion of GLP-1 during GPA testing has been shown to increase the APR_pot and PISR_pot in islet and pancreas transplant recipients (39), suggesting that incretin therapy may lead to depletion of mature β-cell secretory granules in certain populations. In addition to the earlier shift in the insulin secretory response to nutrient ingestion, another reassuring finding in support of a positive β-cell effect from 6-month sitagliptin intervention in pancreatic insufficient CF is the trend toward lower proinsulin secretion under glucose-potentiated conditions and absence of higher PISR during hyperglycemia with sitagliptin therapy. The preliminary data on the safety and tolerability of 6 months of sitagliptin therapy are also reassuring.

A number of limitations should be acknowledged. This study has a number of strengths including its randomized, double-blind comparative design, near-complete participant retention, well-tolerated intervention with high rates of adherence, and use of gold-standard methods for assessing islet β- and α-cell function. ISRs were calculated using deconvolution of peripherally measured C-peptide, a method that addresses hepatic insulin extraction, which can be altered in settings of impaired glucose tolerance and limit the use of peripheral insulin concentrations as measures as insulin secretion. While individuals across a spectrum of abnormal glucose tolerance were studied only a limited number had the diagnosis of CFRD. Nonetheless, the lack of any improvement in glucose tolerance when reductions in glucose excursions have been demonstrated in otherwise healthy, glucose tolerant individuals (33) suggests sitagliptin is not an effective glucose-lowering agent in this population with abnormal glucose tolerance. Given the emerging overweight and obesity in the CF population [reviewed by Litvin et al (53)], important considerations are the contributions of excess weight and insulin resistance to unmasking underlying insulin secretion defects and the use of T2D therapeutics in this new landscape. Two participants in the sitagliptin group were obese, were less insulin sensitive, and tended to have the greatest acute insulin responses, likely reflecting β-cell compensation for the underlying insulin resistance. The smaller sample size and limited representation of obese individuals do not permit subgroup analysis, but graphical inspection did not suggest baseline obesity status impacted sitagliptin responsiveness (data not shown). While we measured HbA1c, improvements in HbA1c were not expected as this surrogate tends to be normal even in the setting of early CFRD. Continuous glucose monitoring would have provided more extensive glucose excursion data over days rather than a single meal and, accordingly, additional evidence for the presence or absence of a glycemic effect of sitagliptin but was not performed as part of this study. Finally, this study is unable to directly assess for insulinotropic effects of GLP-1 and GIP; resistance to the latter is well-recognized in T2D (54,55). A separate, ongoing study is addressing islet β- and α-cell responses to acute GLP-1 and GIP infusions in people with pancreatic insufficient CF across the spectrum of glucose tolerance (ClinicalTrials.gov Identifier: NCT01851694).

We conclude that 6 months of treatment with the DPP-4 inhibitor, sitagliptin, evokes some improvement of meal-related insulin secretion dynamics without overall improvement in glucose excursion. Augmentation of glucagon suppression was present with sitagliptin therapy, but in the absence of the frank hyperglucagonemia that features in T2D, the difference may have limited impact upon hyperglycemia in CF. Whether these limited effects of incretin hormones upon β-cell insulin and α-cell glucagon secretion indicate improved islet function in pancreatic insufficient CF remains to be determined, but further investigation of these findings may inform strategies aimed at preservation of islet function toward the ultimate goal of diabetes prevention in CF.

Additional Information

Disclosures: No potential conflicts of interest relevant to this article were reported. Sitagliptin and placebo were provided by Merck Sharp & Dohme Corp. through Merck Investigator Initiated Study Program #50512. The opinions expressed in this paper are those of the authors and do not necessarily represent those of Merck Sharp & Dohme Corp. DDDL reports stock options with Merck & Co. through spouse employment.

Data Availability

Data sets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.