Pramlintide but Not Liraglutide Suppresses Meal-Stimulated Glucagon Responses in Type 1 Diabetes

Abstract

Context

Postprandial hyperglycemia remains a challenge in type 1 diabetes (T1D) due, in part, to dysregulated increases in plasma glucagon levels after meals.

Objective

This study was undertaken to examine whether 3 to 4 weeks of therapy with pramlintide or liraglutide might help to blunt postprandial hyperglycemia in T1D by suppressing plasma glucagon responses to mixed-meal feedings.

Design

Two parallel studies were conducted in which participants underwent mixed-meal tolerance tests (MMTTs) without premeal bolus insulin administration before and after 3 to 4 weeks of treatment with either pramlintide (8 participants aged 20 ± 3 years, hemoglobin A_1c 6.9 ± 0.5%) or liraglutide (10 participants aged 22 ± 3 years, hemoglobin A_1c 7.6 ± 0.9%).

Results

Compared with pretreatment responses to the MMTT, treatment with pramlintide reduced the peak increment in glucagon from 32 ± 16 to 23 ± 12 pg/mL (P < 0.02). In addition, the incremental area under the plasma glucagon curve from 0 to 120 minutes dropped from 1988 ± 590 to 737 ± 577 pg/mL/min (P < 0.001), which was accompanied by a similar reduction in the meal-stimulated increase in the plasma glucose curve from 11,963 ± 1424 mg/dL/min pretreatment vs 2493 ± 1854 mg/dL/min after treatment (P < 0.01). In contrast, treatment with liraglutide had no effect on plasma glucagon and glucose responses during the MMTT.

Conclusions

Adjunctive treatment with pramlintide may provide an effective means to blunt postmeal hyperglycemia in T1D by suppressing dysregulated plasma glucagon responses. In contrast, plasma glucose and glucagon responses were unchanged after 3 to 4 weeks of treatment with liraglutide.

We examined the effect of pramlintide or liraglutide as adjunctive therapy on postprandial glucagon and glucose response to mixed-meal feedings in youths with type 1 diabetes.

Postprandial hyperglycemia remains a challenge in type 1 diabetes (T1D) due to a number of factors that include delays in the absorption and action of premeal boluses of insulin from the subcutaneous space and dysregulated glucagon secretion in response to mixed-meal feedings (1–5). In individuals without diabetes, plasma glucagon levels change very little after eating a mixed meal that includes protein and carbohydrates because the stimulation of glucagon secretion by increases in plasma amino acids is offset by the suppression of glucagon secretion by increases in plasma glucose levels. In contrast, it has been demonstrated that children with T1D have higher plasma glucagon responses after mixed-meal feedings compared with healthy peers (3). Moreover, plasma glucagon responses to mixed-meal feeding increase over time, presumably due to the progressive loss of residual β-cell function (1, 6).

Due to the adverse impact of postprandial hyperglycemia on overall glycemic control and the risk of complications (7, 8), it has been suggested that treatment with agents approved for use in type 2 diabetes may be effective in lowering postprandial glucose peaks in T1D by mechanisms independent of stimulation of insulin secretion (9, 10). Specifically, sodium-glucose cotransporter 2 inhibitors reduce plasma glucose by lowering the renal threshold for glucose excretion (11), whereas it has been suggested that the glucose-lowering effects of both pramlintide (an analog of amylin) and liraglutide (a glucagon-like peptide-1 agonist) are due, in part, to slowing of gastric emptying and suppression of exaggerated postmeal increases in plasma glucagon (12, 13). In two parallel studies, we used a full closed-loop (CL) insulin delivery system to control postmeal glucose excursions before and after 3 to 4 weeks of treatment with pramlintide and liraglutide at maximally recommended doses. Those studies showed durable slowing of gastric emptying with pramlintide but not with liraglutide (14). In those studies, we also performed mixed-meal tolerance tests (MMTTs) in the morning following 24 hours of CL control to examine and compare whether 3 to 4 weeks of treatment with pramlintide or liraglutide suppressed meal-stimulated increases in plasma glucagon in T1D (14). The results of these MMTTs are reported in this study.

Materials and Methods

Participants

Participants were eligible to enroll in the pramlintide and liraglutide studies if they had a clinical diagnosis of T1D for at least 1 year, hemoglobin A_1c ≤9% (≤75 mmol/mol), and a normal hematocrit and serum creatinine level. Participants were excluded if they had a history of an eating disorder, celiac disease, gastroparesis, another disorder of intestinal absorption or motility, a history of a hypoglycemic seizure in the past 3 months, another chronic medical condition (except treated hypothyroidism), current use of medications (other than insulin) known to affect blood glucose level or gastrointestinal motility, and prior adverse reactions to the drug under study. Female participants could not be pregnant or lactating. The studies were reviewed and approved by the Yale University Human Investigation Committee, and written informed consent was obtained by adult participants. Parental consent with participant assent was obtained for participants <18 years.

Procedures

Dose-titration phase

Participants in the studies underwent two 24-hour periods of CL glucose control before and after 3 to 4 weeks of treatment with pramlintide or liraglutide (14). During outpatient treatment, the dose of pramlintide was uptitrated from 30 to 60 μg given 15 minutes prior to each meal, and the once-daily dosing of liraglutide before breakfast was uptitrated from 0.6 to 1.8 mg/d. In both studies, insulin doses were adjusted, as needed, by frequent telephone contacts with the study participants.

MMTTs

In each study, the participants underwent two MMTTs, the first performed before therapy with pramlintide or liraglutide and the second performed after 3 to 4 weeks of treatment with one of the drugs. All MMTTs were performed at ∼8 am in the morning after an 8- to 12-hour overnight fast, during which glucose levels were regulated with a Medtronic CL system (Medtronic, Northridge, CA) (14). The CL system used in both of these studies consisted of four components: a Medtronic Paradigm 715 insulin pump, a Medtronic MiniLink REAL-Time transmitter (MMT-7703) adapted for 1-min transmission, a Medtronic continuous glucose sensor (Sof-Sensor in the pramlintide study and an Enlite sensor in the liraglutide study), and the Medtronic external Physiological Insulin Delivery algorithm modified to include insulin feedback, which was on a laptop computer (all from Medtronic).

An intravenous catheter was used for frequent blood sampling during the MMTTs. At the start of the 4-hour MMTTs, baseline samples for measurement of plasma glucose and plasma glucagon were obtained, the CL system was shut off, and participants were placed back on their usual open-loop basal rate settings. Participants then consumed 6 mL/kg of Boost High Protein 6 cc/kg (Nestlé HealthCare Nutrition, Fremont, MI) to a maximum dose of 360 mL. Additional blood samples for measurements of plasma glucose and plasma glucagon were obtained every 15 to 30 minutes for 240 minutes following ingestion of Boost High Protein (3); the macronutrient content per 100 mL of Boost High Protein is protein 6.3 g, carbohydrate 13.9 g, and fat 2.5g. During the second MMTT for each participant, pramlintide (60 μg) or liraglutide (1.8 mg) was injected just prior to meal ingestion.

The primary outcome of the two parallel studies reported in this study was the difference in the incremental area under the curve (iAUC) in plasma glucagon levels from baseline to 120 minutes (Glucagon iAUC_{0–120 min}). Secondary outcomes included the iAUC in plasma glucagon levels from 120 to 240 minutes (Glucagon iAUC_{120–240 min}), incremental area under the meal-stimulated increase in the plasma glucose curve from 0 to 120 minutes (Glucose iAUC_{0–120 min}), incremental area under the meal-stimulated increase in the plasma glucose curve from 120 to 240 minutes (Glucose iAUC_{120–240 min}), changes in peak plasma glucagon and peak plasma glucose levels, and the time-to-peak for glucagon and glucose over 240 minutes.

Breakfast during CL

Participants in the liraglutide group had their glucagon and glucose response to breakfast assessed during the 24-hour CL admissions. Meals were self-selected and not limited by calorie or carbohydrate content. Samples were obtained every 15 to 30 minutes for 180 minutes after the breakfast to assess both glucose and glucagon levels. CL insulin delivery was maintained during the meal test.

Laboratory measurements

Plasma glucose was analyzed using the YSI 2300 STAT Plus glucose analyzer (YSI Life Sciences, Yellow Springs, OH). Glucagon was measured by a double-antibody radioimmunoassay (GL-32K; EMD Millipore, Burlington, MA). The lower limit of detection of plasma glucagon was 20 pg/mL, and the higher limit of the standard assay curve was 400 pg/mL. The accuracy of the assay was 97 ± 0.8%.

Statistical considerations

Comparisons between the pretreatment and during treatment measurements were calculated using the paired Student t test for continuous variables. Fisher exact test was adopted for categorical variables. Changes in plasma glucose and glucagon during the MMTTs were expressed as incremental values from baseline (0 minutes) to the specified time points. The iAUCs and peak value for both plasma glucose and glucagon were calculated as difference from the baseline measure (0 minutes). Data are expressed as mean ± standard deviation (SD). Data were analyzed using Prism 7 software (GraphPad Software, Inc., La Jolla, CA).

Results

Participants

Ten out of 11 participants who enrolled in the CL study (14) with liraglutide agreed to undergo the two MMTTs, as did 8 out of 10 participants from the pramlintide study. The clinical characteristics of the 10 liraglutide and 8 pramlintide participants are shown in Table 1.

Table 1.

Baseline Characteristics

Clinical Characteristics	Liraglutide (n = 10)	Pramlintide (n = 8)	P
Sex (female/male)	6/4	5/3	>0.99
Age, y (age range)	21.9 ± 3.5 (18–27)	19.6 ± 2.8 (16–23)	0.151
BMI, kg/m2	23.5 ± 2.9	23.0 ± 1.5	0.665
Hemoglobin A1c at enrollment, % (mmol/mol)	7.5 ± 1.0 (58.0 ± 10.9)	6.9 ± 0.5 (52.0 ± 5.5)	0.142
Diabetes’ duration, y	9.9 ± 6.5	9.4 ± 4.6	0.857
Weight, kg	67.1 ± 9.6	70.7 ± 14.6	0.537
Total daily insulin dose, U/kg/d	0.8 ± 0.1	0.9 ± 0.3	0.355

Data are mean (SD) unless otherwise indicated.

Abbreviation: BMI, body mass index.

MMTT results

Baseline plasma glucagon and glucose levels

As shown in Table 2, in each of the experiments, overnight CL insulin delivery resulted in baseline fasting plasma glucagon and glucose levels that were similar in both groups of subjects both before and during treatment with pramlintide and liraglutide.

Table 2.

Baseline Glucose and Glucagon Values Prior to and Posttreatment With Adjunctive Therapy

	Pretreatment	Posttreatment	P
Pramlintide
Glucose, mg/dL	121 ± 22	115 ± 21	0.586
Glucagon, pg/mL	42 ± 22	45 ± 19	0.775
Liraglutide
Glucose, mg/dL	100 ± 16	116 ± 27	0.124
Glucagon, pg/mL	52 ± 19	47 ± 19	0.563

Data are mean (SD) unless otherwise indicated.

Increments in plasma glucagon and glucose before and during treatment with pramlintide

The patterns of incremental changes in plasma glucagon and glucose before and during treatment with pramlintide are shown in Fig. 1A and 1B, respectively. During the first 2 hours of the MMTTs, treatment with pramlintide markedly reduced the rise in plasma glucagon levels following meal ingestion (Fig. 1A), and increases in plasma glucose levels were also blunted (Fig. 1B). Moreover, between 2 and 4 hours, glucagon levels remained suppressed during treatment with pramlintide (Fig. 1A), even in the face of a delayed rise in plasma glucose levels (Fig. 1B). As shown in Table 3, adjunctive therapy with pramlintide markedly suppressed Glucagon iAUC_{0–120 min} and the peak increment in plasma glucagon, as well as Glucose iAUC_{0–120 min,} Glucose iAUC_{120–240 min}, and the peak increment in glucose. Pramlintide treatment also delayed the time-to-peak glucagon and glucose levels (Table 3).

Figure 1.

Glucagon and glucose profile during MMTT. Glucagon profile before and after the treatment with (A) pramlintide and (C) liraglutide during MMTT. Glucose profile before and after the treatment with (B) pramlintide and (D) liraglutide during MMTT. Glucagon and glucose are expressed as incremental value from the baseline.

Table 3.

Outcome Measures

	Liraglutide			Pramlintide
	Pretreatment	Posttreatment	P	Pretreatment	Posttreatment	P
Glucagon
Glucagon iAUC0–120, pg/min/mL	1904 ± 651	1801 ± 906	0.774	1988 ± 590	737 ± 577	0.0007
Glucagon iAUC120–240, pg/min/mL	311 ± 564	701 ± 860	0.246	560 ± 807	933 ± 789	0.366
Glucagon incremental peak, pg/mL	29 ± 16	35 ± 20	0.309	32 ± 16	23 ± 12	0.026
Glucagon time-to-peak, min	47 ± 30	64 ± 41	0.281	49 ± 22	173 ± 66	0.0097
Glucose
Glucose iAUC0–120, mg/min/dL	13,001 ± 1207	12,029 ± 1500	0.619	11,963 ± 1424	2493 ± 1854	<0.0001
Glucose iAUC120–240, mg/min/dL	20,241 ± 1794	18,135 ± 2580	0.138	17,505 ± 2721	13,397 ± 2841	0.051
Glucose incremental peak, mg/dL	200 ± 29	171 ± 47	0.070	181 ± 46	150 ± 63	0.011
Glucose time-to-peak, min	132 ± 41	135 ± 29	0.85	128 ± 31	221 ± 32	<0.001

Data are mean (SD) unless otherwise indicated. P values in boldface are significant.

Increments in plasma glucagon and glucose before and during treatment with liraglutide

The incremental changes in plasma glucagon and glucose before and during treatment with liraglutide are shown in Fig. 1C and 1D, respectively. As seen in these figures, after 3 to 4 weeks of liraglutide treatment, there were no noteworthy differences in the plasma glucagon and glucose responses during the first 120 minutes and second 120 minutes following Boost ingestion. As shown in Table 3, the peak increment in plasma glucagon, time-to-peak glucagon, and iAUC for glucagon were not significantly different before and during treatment with liraglutide. Furthermore, there were no marked changes in the peak increment in plasma glucose, time-to-peak glucose, and iAUC for glucose after treatment with liraglutide.

Plasma glucagon and glucose responses during CL insulin delivery before and during treatment with liraglutide

To validate that a liquid meal response would be reflective of the physiologic changes in glucagon and glucose following a standard meal under controlled insulin delivery conditions, a self-selected breakfast was provided to participants during both CL admissions. The average macronutrient content of the standardized breakfast was 75 ± 49 g of carbohydrates, 24 ± 16 g of protein, and 17 ± 14 g of fat. Corroborating the findings demonstrated during the MMTT, no difference in the glucagon or glucose response was appreciated in the 3 hours following the standardized breakfast meal (Fig. 2 and Supplemental Table 1).

Figure 2.

Change in glucagon and glucose levels during full CL insulin delivery with a standardized breakfast prior to and 3 to 4 weeks posttreatment with liraglutide.

Discussion

Pramlintide and liraglutide have been widely investigated as adjunctive therapies aimed at limiting postmeal hyperglycemia in T1D (9, 12, 14–25) due to putative modes of action that include the ability to suppress dysregulated glucagon responses to mixed-meal feedings, slowing of gastric motility, and earlier satiety (9, 26, 27). It should be noted, however, that clinical studies have demonstrated differences between the two drugs on glucose control, glucagon secretion, and gastric emptying (14, 16, 18, 19, 22, 28–30), with liraglutide surprisingly increasing the glucagon responses to mixed-meal feedings after chronic treatment in type 2 diabetes and pramlintide being highly effective in delaying gastric emptying but with conflicting effects on glucagon secretion in T1D (14, 24).

Our parallel studies of the use of pramlintide and liraglutide as adjunctive agents to improve control of postprandial glucose excursions during CL insulin delivery provided a unique opportunity to examine and compare the effects of these agents on dysregulated increases in plasma glucagon levels after meals (14). Consequently, the most important findings of the current study are that we were unable to observe any suppressive effects of liraglutide on plasma glucagon responses to mixed-meal feedings or any suggestion of a delay in gastric emptying after only 3 to 4 weeks of treatment. In contrast, marked suppression of 2-hour plasma glucagon, as well as reduced and delayed increases in plasma glucose levels, were sustained after the same duration of treatment with pramlintide.

These findings are consistent with previous studies that supported approval of pramlintide for use as an adjunctive agent in T1D (12, 22, 23, 30, 31) and more recent phase 3 studies of liraglutide that indicated little improvement in metabolic control in patients with T1D (16, 17, 20). Our results are also consistent with previous reports indicating that liraglutide is a less effective drug of its class in modulating the gastric motility, with more pronounced action from short-term glucagon-like peptide-1 analogs, like exenatide or lixisenatide (26, 32). As noted in our previous publication in this group of patients (24) and by others (19, 33), liraglutide may be of benefit to overweight or obese patients with T1D due to its suppression of appetite, which may support weight loss and reductions in insulin doses (17, 27).

A strength of the study is that the MMTTs were performed after completion of 24 hours of CL insulin delivery, including overnight control just prior to the start of the MMTTs, with the last meal being consumed >12 hours earlier. Most MMTT protocols mandate that fasting glucose levels between 70 and 200 mg/dL be achieved prior to meal ingestion. Our use of the CL system ensured that participants had even tighter glycemic control, thus minimizing the potential confounding effects of differences in fasting plasma glucose prior to performance of the procedure. Even plasma glucagon levels were similar prior to the conduct of the pre- and posttreatment MMTTs performed. Use of the CL system also ensured precision in regard to the insulin delivery prior to the start of the MMTTs, eliminating a potential confounding factor of overinsulinization prior to meal ingestion. Thus, in both sets of experiments, the only difference between the two MMTTs in each participant was the injection of the study drug prior to the second MMTT.

Compared with the sharp increases in plasma glucagon and plasma glucose during the pretreatment MMTT, only a slight increase in plasma glucagon and glucose levels was observed during the first 60 minutes of the MMTT during treatment with pramlintide (Fig. 1D). However, it is possible that diminished increases in plasma glucagon and plasma glucose during the first 2 hours of the MMTT were both due to delays in gastric emptying rather than by suppression of glucagon secretion by pramlintide. Arguing against this conclusion is the observation that the relatively flat glucagon response following meal ingestion with pramlintide was present for the full 4 hours of the MMTT, despite delayed absorption of carbohydrate and amino acids and corresponding increases in plasma glucose after meal ingestion. These data suggest that the ability of pramlintide to mitigate postmeal hyperglycemia is related to both its ability to delay gastric emptying and suppress α-cell secretion.

A limitation of the current study was that it was not designed to allow for formalized comparison between the two agents. However, comparison of the individual treatments prior to and posttreatment in the two separate cohorts studied allows extrapolation of how the two adjunctive therapies may differ. It should also be noted that studies examining physiologic changes induced by pharmacologic agents often have larger sample sizes. However, the present analysis is a substudy nested within inpatient CL studies that were designed to examine the feasibility and potential efficacy of adjunctive therapies in conjunction with CL insulin delivery. Although our sample size is relatively small, it was sufficient to show a change in hormonal response during treatment with pramlintide. Although we have no direct means of assessing participants’ compliance in taking the study drugs during the outpatient phase of the studies, participants were contacted frequently by telephone during this time. These phone calls allowed investigators to encourage compliance and assess for adverse effects of the study medications. Importantly, all participants in both studies tolerated the full therapeutic doses of the drugs during the inpatient studies, which would have been unlikely if they had not been compliant in the outpatient dose-titration phase. Finally, as a surrogate marker for compliance, it is notable that during both studies, participants on average had a lowering of their total daily insulin dose (24).

Finally, it is possible that use of a standardized meal instead of a liquid mixed meal would have provided better approximation of how these therapies impact day-to-day life. Although not performed in the pramlintide study, plasma glucagon responses to a standardized breakfast during CL insulin delivery were assessed in the liraglutide study. Furthermore, the standard breakfast meal study conducted during the inpatient CL admissions provided the opportunity to see if dynamic insulin delivery affected the results of meal-stimulated glucagon and glucose responses. As demonstrated in Fig. 2, no difference was appreciated with the standard meal, thus providing justification for assessment of the MMTT in the present analysis (Supplemental Table 1) and confirming the reliability of the use of MMTT, instead of a real meal, to assess the effect of the adjunctive therapy on glucagon and glucose response.

In conclusion, this study has highlighted that liraglutide did not suppress dysregulated increases in plasma glucagon responses to meals even after a relatively short period of treatment. We have also confirmed the effect of pramlintide in limiting the early meal-stimulated increases in plasma glucagon and glucose levels. However, the requirement for subcutaneous injections of pramlintide before each meal has limited the use of this agent in patients with T1D in clinical practice.

Abbreviations:

CL

closed-loop

Glucagon iAUC_{0–120 min}

incremental area under the plasma glucagon curve from 0 to 120 minutes

Glucagon iAUC_120–240min

incremental area under the curve in plasma glucagon levels from 120 to 240 minutes

Glucose iAUC_{0–120 min}

incremental area under the meal-stimulated increase in the plasma glucose curve from 0 to 120 minutes

Glucose iAUC_{120–240 min}

incremental area under the meal-stimulated increase in the plasma glucose curve from 120 to 240 minutes

iAUC

incremental area under the curve

MMTT

mixed-meal tolerance test

SD

standard deviation

T1D

type 1 diabetes.