The pharmacodynamics, safety and pharmacokinetics of single doses of the motilin agonist, camicinal, in type 1 diabetes mellitus with slow gastric emptying

Abstract

Background and Purpose

Here we have investigated the pharmacokinetics, pharmacodynamics and safety of single doses of camicinal in type 1 diabetes mellitus (T1DM) patients with a history of slow gastric emptying with symptoms consistent with gastroparesis.

Experimental Approach

In a randomized, double‐blind, placebo‐controlled, incomplete block, three‐period, two‐centre crossover study, patients received oral administration of placebo and two of the three possible doses of camicinal (25, 50 or 125 mg). Gastric emptying (¹³C‐octanoic acid breath test), pharmacokinetics and safety were primary outcomes.

Key Results

Nine of the 10 patients enrolled completed the study. Gastric half‐emptying time decreased by −95 min (95% CI: −156.8, −34.2) after a single dose of camicinal 125 mg compared with placebo (52 vs. 147 min, P < 0.05), representing a 65% improvement. A decrease of the gastric half‐emptying time compared with placebo (approximately 39 min) was observed with camicinal 25 and 50 mg, representing a 27% reduction for both doses (not statistically significant). A positive exposure–response relationship was demonstrated across all doses. The effects of camicinal on gastric half‐emptying time were not influenced by fasting glucose levels. Single doses up to 125 mg were well tolerated. Camicinal was well absorbed, exhibiting linear and approximately dose‐proportional pharmacokinetic characteristics and a clear exposure–response relationship with gastric emptying.

Conclusions and Implications

Camicinal significantly accelerated gastric emptying of solids in T1DM patients following administration of a single oral dose. Camicinal was well tolerated and exhibited similar pharmacokinetic characteristics in diabetic patients to those previously reported in healthy volunteers.

Abbreviations

GLP‐1

glucagon‐like peptide‐1

HbA1c

haemoglobin A1c

T1DM

type 1 diabetes mellitus

Tables of Links

TARGETS
GPCRs
Motilin receptor

LIGANDS
Alemcinal	GLP‐1
Azithromycin	Mitemcinal
Camicinal	Motilin
Erythromycin	Pancreatic polypeptide
Ghrelin

These Tables list key protein targets and ligands in this article which are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Pawson et al., 2014) and are permanently archived in the Concise Guide to PHARMACOLOGY 2015/16 (Alexander et al., 2015).

Introduction

Gastroparesis is defined as a delay in gastric emptying in the absence of mechanical obstruction (Camilleri et al., 2011). Gastroparesis may be primary (idiopathic) or secondary (diabetes, connective tissue disorders) and results in poor delivery of nutrition and/or drugs to the small bowel (Hasler, 2012). The long‐term complications of untreated gastroparesis can lead to gastric failure (i.e. severe nutritional compromise necessitating parenteral feeding (Krishnan et al., 2013). Abdominal symptoms (nausea, vomiting, postprandial fullness, early satiety, bloating and pain/discomfort) often occur in gastroparesis, although the severity of symptoms and the degree of delayed gastroparesis are poorly correlated (Janssen et al., 2013). Although a number of gastric prokinetic agents with different mechanisms of action have been developed, most are not currently available for clinical use because of poor efficacy or the potential for the development of significant adverse effects (Sanger and Alpers, 2008). Therefore, there is a need for safe and efficacious gastric prokinetic agents to serve this unmet medical need.

One approach to the identification of a novel gastroprokinetic drug is to develop compounds that act as motilin receptor agonists. Motilin is an endogenous peptide hormone, produced mainly in specific endocrine cells in the epithelium of the duodenum (Polak et al., 1975). When released, motilin exerts activity via the motilin receptor, a G protein‐coupled, 7‐transmembrane receptor. In humans, motilin receptors are expressed by cholinergic neurons within the enteric nervous system and also by smooth muscle cells of the gastrointestinal tract (Broad et al., 2012), with additional expression by vagal afferent neurons in certain animals (Javid et al., 2013) and only limited expression outside the gastrointestinal tract (Sanger et al., 2013b). During fasting, the release of motilin is believed to mediate the short‐lived bursts of contractions within the stomach, which occur at regular intervals and migrate along the intestine (known as phase III of the migrating motor complex) (Nakajima et al., 2010). These migrating contractions perform a ‘housekeeping role’ by clearing the stomach and small intestine of luminal secretions, undigested material and bacteria in preparation for the next meal (Vantrappen et al., 1979). Recent evidence suggests that this increase in motilin concentration and subsequent phase III activity is also associated with an increase in hunger (Tack et al., 2014).

In human stomach, motilin receptor agonists increase enteric cholinergic activity, a function not shared by the structurally related hormone ghrelin (Broad et al., 2014). However, it is increasingly recognized that when compared with motilin itself, certain non‐peptide motilin receptor agonists can act in a different manner, evoking a relatively prolonged increase in cholinergic activity within the stomach, particularly in the antrum, to increase gastric emptying of meals after single or repeated dosing. This difference has been observed for erythromycin and azithromycin – antibiotic drugs with complex structures and additional abilities to activate the motilin receptor – and with camicinal, a low MW, selective motilin receptor agonist (Dass et al., 2003; Broad et al., 2012; Broad and Sanger, 2013). The reason for this difference is not clear, but circumstantial evidence, including the existence of different binding sites for motilin agonists at the motilin receptor, has led to the proposal that the functions of different agonists are ‘biased’ towards the generation of one type of activity over another, a phenomenon often referred to as ‘biased agonism’ (Sanger, 2014). Regardless of mechanism, erythromycin has been used to stimulate gastric emptying in patients with diabetic, idiopathic and post‐vagotomy gastroparesis (Janssens et al., 1990; Mozwecz et al., 1990; Camilleri, 1993; Peeters, 1993; Richards et al., 1993; Sturm et al., 1999). These gastroprokinetic effects occur at doses lower than those required for antibacterial activity (Ritz et al., 2005), whereas at higher doses, erythromycin causes direct stimulation of the muscle (Coulie et al., 1998) and is associated with increased gastric‐related adverse events, such as nausea and diarrhoea. Erythromycin is also associated with QTc interval prolongation and drug interactions with drugs such as statins (e.g. rhabdomyolysis) (Guo et al., 2010; Patel et al., 2013). A link between the use of azithromycin and sudden cardiac arrhythmia and death has also been reported (Rao et al., 2014). Nevertheless, the beneficial effects of low doses of erythromycin have also been established after repeat dosing. For example, in pilot studies, low doses of erythromycin or repeated intravenous erythromycin, dose‐titrated to achieve efficacy and tolerability in each patient, and improved symptoms associated with gastroparesis (Richards et al., 1993; DiBaise and Quigley, 1999; Sturm et al., 1999; Dhir and Richter, 2004; Ritz et al., 2005). In addition, erythromycin (600 mg⋅day⁻¹ for 4 weeks) improved glucose control with lowering of haemoglobin A1c (HbA1c) in a population of type 2 diabetes mellitus patients with slowed gastric emptying; the gastric emptying effect was reversible and returned to pretreatment levels after washout (Ueno et al., 2000).

Other motilin receptor agonists derived from the same macrolide structure as erythromycin (known as the motilides) including GM611 (mitemcinal), ABT‐229 (alemcinal), EM523L, EM574, and KC 11458, and the peptide motilin receptor agonist atilmotin also increased the rate of gastric emptying in humans (Okano et al., 1996; Verhagen et al., 1997; Park et al., 2006). Mitemcinal, in a randomized withdrawal designed study, progressively improved gastroparesis symptoms in a subset of diabetic patients over 3 months of therapy, and symptoms returned to pretreatment levels upon withdrawal (McCallum and Cynshi, 2007).

Camicinal is a first‐in‐class low MW (non‐macrolide) motilin receptor agonist with potential use in conditions where slow gastric emptying is a primary aetiology for the clinical presentation (Sanger et al., 2009; Westaway et al., 2009). Camicinal was designed using the recombinant human motilin receptor to enhance the specificity for the receptor and potentially decrease the off‐target characteristics encountered with compounds that have complex and non‐specific motilide structures (Sanger et al., 2013a; Sanger et al., 2013b), followed by evaluation of its ability to induce prolonged enhancement of rabbit and then human gastric cholinergic activity (Broad et al., 2012). Camicinal (GlaxoSmithKline, Stevenage, UK) has since been shown to accelerate gastric emptying by 30–40% in healthy volunteers as single (50–150 mg) or 14 day repeat oral doses (50–125 mg) (Dukes et al., 2010).

The primary objective of this study was to assess the pharmacodynamic effects of a single dose of camicinal on gastric emptying, as well as safety, tolerability and pharmacokinetics of single oral doses of camicinal in type 1 diabetes mellitus (T1DM) patients with slow gastric emptying. Exploratory objectives included assessment of the effect of camicinal on blood concentrations of glucose and gut hormones [glucagon‐like peptide 1 (GLP‐1), ghrelin and pancreatic polypeptide], together with changes in sensation of hunger, food intake and symptoms of gastroparesis. These data were presented, in part, at the Digestive Disease Week Conference 2011 (Hellström et al., 2011),

Methods

Patients

This study was conducted in accordance with good clinical practice guidelines and the Declaration of Helsinki after obtaining a written informed consent from each patient. The study was approved by the Institutional Review Boards at each study site.

Patients for this study were male and female T1DM patients between 18 and 70 years of age, HbA1c less than 86 mmol⋅mol⁻¹ with onset of disease at less than 30 years of age and previously documented slow gastric emptying (greater than 30% retention at 2 h as determined by scintigraphy or gastric half‐emptying time greater than 109 min as determined by the ¹³C‐octanoate breath test) with symptoms. The patients were required to have a minimum of 3 months' history of relevant symptoms for gastroparesis (i.e. chronic postprandial fullness, postprandial nausea, vomiting, loss of appetite and early satiety). Although patients may not have had gastroparesis by strict scintigraphic criteria, all had slow emptying associated with symptoms typical of delayed gastric emptying. This criterion was operationalized to allow demonstration of pharmacological effect across a range of emptying times. Exclusions included acute severe gastroenteritis, gastric pacemaker, chronic parenteral feeding, persistent severe vomiting, pronounced dehydration, diabetic ketoacidosis (within the prior 4 weeks), history of eating disorders and thyroid dysfunction. Entry criteria required that prokinetic and antiemetic agents could not be used within at least 1 week of dosing.

Study design

This two‐centre study was conducted as a randomized, double‐blind, placebo‐controlled, incomplete block, three‐period crossover design consisting of a screening period (within 28 days of study initiation), three treatment periods and a post‐treatment follow‐up visit. The patients remained in the study unit for each treatment period. Each patient participated in three single oral dose treatment periods with at least 7 days between study periods. This washout period was greater than five times the plasma half‐life of the drug (25–30 h). For each dosing session, patients were admitted to the clinic on day 1 prior to dosing for baseline assessments. Following completion of these assessments, patients were randomly assigned to receive a single oral dose of camicinal 25, 50 and 125 mg or placebo in one of the nine sequences, in equal ratios. All patients received placebo in one of these treatment periods, and two of the three camicinal doses in the other two periods. Drug dosing took place between 7:30 and 9:00 am, in a fasted state. Patients remained fasted until 5.5 h post‐dose.

Study outcomes

Gastric half‐emptying time was measured using the ¹³C‐octanoic acid breath test. Eighty minutes after dosing, patients were given a ¹³C‐octanoate test meal (consumed within 10 min). The test meal consisted of a pancake (8.4 g protein, 11.2 g fat, and 26.7 g carbohydrates; 243.5 kcal) including 91 mg ¹³C‐octanoic acid (Euriso‐TOP, Saint‐Aubain Cédex, France). After consumption of the meal, breath samples were collected at 15‐min intervals for 4 h (approximately 5.5 h post‐dose). Patients received a lunch following gastric emptying testing and dinner. Meals were identical in content for each meal in all study periods.

Patients rated the severity of pain, bloating, early fullness, post‐meal fullness, nausea, vomiting, belching, and heartburn on a 4‐point scale from 0 to 3 with lower scores representing less symptom severity. Patients completed this daily gastrointestinal symptom diary at baseline and after each meal for a 24 h period. A hunger questionnaire was administered every 30 min throughout the gastric emptying test period. Total energy intake was calculated by a dietician from the calories consumed at ad libitum meal times for the 24 h post‐dose period. Following each dosing, stools were monitored for 24 h post‐dose by capturing the following parameters: (i) time to first bowel movement after dosing; (ii) number of bowel movements within 24 h post‐dose; and (iii) stool consistency as an overall rating of all stools evacuated within 24 h of dosing according to the Bristol Stool Form scale (Lewis and Heaton, 1997). Safety and tolerability were assessed by monitoring patients for adverse events, vital signs, ECGs and clinical laboratory parameters (the day before dosing and 24 h after dosing).

Pharmacokinetics

The pharmacokinetic profiles of camicinal were determined by blood sampling over a 24 h period after dosing. Blood samples were collected into EDTA tubes for preparation of plasma. Camicinal concentrations were analysed using a validated HPLC–MS/MS method. The lower and upper limits of quantitation were 1 and 2000 ng⋅mL⁻¹, respectively.

Physiological markers

The effects of camicinal on the gut hormones ghrelin, pancreatic polypeptide, and GLP‐1 were assessed. An ELISA was used for analysing ghrelin and pancreatic polypeptide levels. The ghrelin assay (Millipore UK Ltd., Billerica, MA) had a reportable range of 25 to 2000 pg⋅mL⁻¹. The pancreatic polypeptide assay (Peninsula Laboratories, San Carlos, CA, USA) had a reportable range of 20 to 25 000 pg⋅mL⁻¹. Prior to analysis, a protein precipitation step was performed using absolute ethanol. GLP‐1 was analysed by an electrochemiluminescence immunoassay (Meso Scale Discovery, Rockville, MD, USA) with a reportable range of 2 to 10 000 pg⋅mL⁻¹. All gut hormones were analysed in duplicate. In addition, plasma glucose levels were monitored by Glucometer every 15 min in the first hour after dosing and then every 30 min until 6 h post‐dose.

Statistical methods

It was planned to enrol approximately 18 patients into the study so that data would be available for 12 patients per dose level (incomplete block design). Based on previous placebo data from Tack et al. (mean t1/2b = 144 min) and a within‐patients standard deviation (28.55) from previous FTIH study MOT107043, there was 90% power (for N = 12) to detect a decrease of 101 min (30% decrease), assuming a two‐sided α = 0.05 level (Tack et al., 2014). After a planned interim analysis (N = 8), the results provided sufficient evidence to progress the compound into a larger repeat dose study, and the study was stopped with N = 10. Data from all patients were analysed (i.e. the intent‐to‐treat population) using sas version 9.1.

For the ¹³C‐octanoic acid breath test, the percent ¹³CO₂ cumulative values versus time curves were fitted to a modified power exponential model, and the following parameters were derived: gastric half‐emptying time (i.e. the time at which half of the ¹³CO₂ is excreted relative to the cumulative excretion when time is infinite), gastric lag phase (time to onset of gastric emptying) and gastric emptying coefficient (i.e. an index of the gastric emptying rate) (Ghoos et al., 1993). The gastric emptying parameters were statistically analysed by a mixed model, fitting period and treatment as fixed effects, and subject as a random effect. The point estimate and corresponding 95% confidence interval for the difference ‘camicinal – placebo’ were constructed for each dose level, using the residual error from the model. Exploratory endpoints, including physiological markers, were summarized descriptively (Senn, 2002).

A population inhibitory sigmoid maximum response (E_max) model was fitted to gastric half‐emptying time versus camicinal AUC_0−∞ for all patients at all doses using nonlinear mixed effects modelling in NONMEM version 7 as follows:

$${\mathrm{GE}=\mathrm{E}}_0‐\left(\left({\mathrm{E}}_{max}*\left({\mathrm{AUC}}^{\mathrm{\gamma }}\right)\right)/\left(\left({\mathrm{AUC}}^{\mathrm{\gamma }}\right)+\left({{\text{EAUC}}_{50}}^{\mathrm{\gamma }}\right)\right)\right.$$

where E_max is the maximal reduction in gastric half‐emptying time versus placebo, EAUC₅₀ represents the camicinal exposure (AUC_0−∞) at which 50% of maximal reduction in gastric half‐emptying time is observed, E₀ represents the mean gastric half‐emptying time following placebo, and γ is a shape parameter. An additive residual error model was used to describe the intra‐individual variability on gastric half‐emptying time. The final model was used to simulate 1000 subjects' exposures and gastric half‐emptying time data to calculate the average and 90% empirical prediction interval from the resulting simulation. These predicted gastric half‐emptying time and AUC values were overlaid with the observed data for a visual predictive check, to confirm that the model describes the central tendency and extent of variability, including random effects in parameters and those due to residual error of the observed data.

Adverse events were collected from the administration of study drug until the final follow‐up contact. Other safety analyses were summarized descriptively [i.e. change from baseline and number of patients outside the normal range for blood pressure, heart rate and electrocardiography parameters (i.e. 12‐lead electrocardiogram) and change from baseline in clinical chemistry and haematology parameters].

Results

Patient demographics

A total of 10 patients were enrolled in this study (nine patients at one site and one patient at another site). The majority of patients (90%) were white women of Caucasian heritage, mean age was 45 years (SD = 13), and median age was 50 years (range: 29–64 years). The mean body mass index, height and weight were 24.9 years (range: 19.8 to 31.8), 72.5 kg (range: 70.8 to 81.6) and 137.7 cm (range: 137.2 to 208.3), respectively.

One patient was withdrawn in the first period after receiving a single dose of placebo due to a serious adverse event of moderately abnormal blood glucose and QT prolongation (associated with hypokalaemia resulting from vomiting). Therefore, nine of the 10 patients received camicinal during the study and completed all dosing periods. The number of patients exposed to each treatment was as follows: placebo, n = 10; camicinal 25 mg, n = 7; camicinal 50 mg, n = 7; and camicinal 125 mg, n = 6.

Gastric emptying

Camicinal significantly reduced (P < 0.05) gastric half‐emptying time by 95 min after a single dose of 125 mg compared with placebo (52 min versus 147 min, respectively) (95% CI: −156.8, −34.2). This represents a 65% improvement, with all patients exhibiting decreases in gastric half‐emptying time (Table 1). A decrease of the gastric half‐emptying time from placebo (−39 min) was also observed at both the 25 and 50 mg dose levels, representing a 27% reduction for both doses (not statistically significant). As exposures at these doses overlapped, a PK/PD (inhibitory E_max) model was developed using camicinal AUC versus gastric half‐emptying time; a positive exposure–response relationship was observed with this model. The model output and predictions showed similar results as shown in Table 1 with a placebo (E₀) value of 151 min and maximum effect, a reduction, of 41 min at the highest dose. Camicinal showed a positive exposure–response relationship in gastric half‐emptying time versus AUC_0−∞ as shown in Figure 1. As exposures increased, gastric half‐emptying time was reduced, indicating more rapid gastric emptying.

Table 1.

Summary of statistical analysis of gastric emptying parameters: camicinal versus placebo comparisons

Parameter	Comparison (mg)	Camicinal mean estimate	Placebo mean estimate	Estimate of difference	95% CI of difference
Gastric half‐emptying Time (min)	25	107.7	147.1	−39.4	(−101.6, 22.7)
	50	108.6	147.1	−38.5	(−100.7, 23.6)
	125	52.1	147.1	−95.0*	(−155.8, −34.2)*
Tlag (min)	25	51.9	98.9	−47.0	(−95.3, 1.3)
	50	68.8	98.9	−30.2	(−78.4, 18.1)
	125	23.7	98.9	−75.2*	(−122.5, −27.9)*
GEC	25	2.92	2.59	0.33	(−0.2, 0.8)
	50	2.99	2.59	0.40	(−0.1, 0.9)
	125	3.60	2.59	1.00*	(0.5, 1.5)*

T_lag, gastric lag phase/time to onset of gastric emptying; GEC, gastric emptying coefficient.

^* P < 0.05, significant difference from placebo.

Figure 1.

Exposure–response relationship of gastric half‐emptying time by AUC (0–∞); median prediction (solid line) and 90% prediction interval (dashed lines) from population PK/PD model shown.

In addition, the effects of camicinal on gastric emptying were confirmed with observations in gastric lag phase and gastric emptying coefficient (Table 1). The percentage of shortening of gastric lag phase was 47%, 30% and 76% with camicinal 25, 50 and 125 mg respectively. Similarly, the percentage of improvement in gastric emptying coefficient was 13%, 15% and 39% with camicinal 25 mg, 50 mg, and 125 mg, respectively. As part of a post hoc analysis, the effect of camicinal on gastric half‐emptying time was unaffected by the fasting glucose concentration (mean range at predose: 7.61 to 9.48 mmol⋅L⁻¹) as illustrated in Figure 2. Camicinal 125 mg had an effect on gastric emptying regardless of whether or not the patient's gastric emptying rate on placebo was above or within the normal range (upper limit of normal = 109 min). Compared with placebo, the mean increase in caloric intake over the 24 h period following drug dosing was 92 calories (10%) with camicinal 25 mg, 292 calories (33%) with camicinal 50 mg and 215 calories (24%) with camicinal 125 mg (post hoc analysis).

Figure 2.

Scatter plot of gastric half‐emptying time by fasting plasma glucose concentration.

Pharmacokinetics

Camicinal was well absorbed and demonstrated a linear and approximately dose‐proportional behaviour in gastric emptying. Selected pharmacokinetic parameters are summarized in Table 2. A full characterization of the elimination phase was not possible because pharmacokinetic sampling was up to 24 h, similar to the half‐life of the compound (approximately 32 h).

Table 2.

Summary of selected plasma camicinal pharmacokinetic parameters

Parameter		Camicinal doses
	N	25 mg	50 mg	125 mg
AUC (0–∞)(h * ng⋅mL−1)	6
Geom. mean		4103.1	6193.8	16774.2
CVb%		(46.8)	(61.9)	(59.6)
Cmax (ng⋅mL−1)	6
Geom. mean		288.3	488.0	1124.9
CVb%		(65.2)	(64.7)	(58.7)
Tmax (h)	6
Median		1.2	1.0	1.5
Range		(0.5–3.0)	(0.5–3.0)	(0.5–3.0)

CVb, coefficient of variation.

Physiological markers

The concentrations of GLP‐1 or ghrelin during the 2 h after dosing with camicinal or placebo did not appear to differ from pre‐dose levels (Table 3). There was evidence to suggest that pancreatic polypeptide increased following the administration of the highest dose (125 mg) of camicinal (Figure 3). The maximum increase of pancreatic polypeptide coincided with the approximate time of maximum concentration of the drug (T_max) (Table 3).

Table 3.

Biomarker levels

Parameter	Placebo	Camicinal 25 mg	Camicinal 50 mg	Camicinal 125 mg
Ghrelin (pg⋅mL−1)
Pre‐dose
Mean	205.1	306.8	263.0	344.6
SD	130.8	155.4	73.8	84.1
Median	194.0	284.0	230.5	372.0
95% CI	95.8, 314.5	113.9, 499.7	185.6, 340.4	240.2, 449.0
2 h post‐dose
Mean	240.5	265.5	227.2	185.0
SD	119.5	155.32	95.7	85.27
Median	267.0	198.5	166.0	211.0
95% CI	140.6, 340.4	102.5, 428.5	108.4, 346.0	79.1, 290.9
GLP‐1 (pg⋅mL−1)
Pre‐dose
Mean	12.4	6.6	13.5	6.8
SD	13.2	4.0	17.7	1.3
Median	8.0	6.0	7.5	7.0
95% CI	2.3, 22.6	1.7, 11.5	−5.0, 32.0	5.2, 8.4
2 h post‐dose
Mean	12.9	7.2	17.8	9.4
SD	13.4	3.5	22.6	1.52
Median	9.5	7.0	8.5	10.0
95% CI	1.7, 24.1	3.5, 10.8	−5.9, 41.5	7.5, 11.3
Pancreatic polypeptide (pg⋅mL−1)
Pre‐dose
Mean	78.0	75.5	75.7	75.5
SD	48.6	39.9	49.2	39.9
Median	66.0	61.5	54.0	61.5
95% CI	33.1, 122.9	12.1, 138.9	−46.6, 197.9	12.1, 138.9
2 h post‐dose
Mean	94.5	54.5	222.0	324.5
SD	39.7	23.9	178.3	249.8
Median	105.5	45.0	155.0	310.0
95% CI	52.9, 136.1	16.5, 92.5	34.9, 409.1	−72.9, 721.9

SD, standard deviation.

Figure 3.

Pancreatic polypeptide 30 min post‐meal.

The glucose concentration‐time profiles for the 6 h period post‐dose did not appear to be substantially different between camicinal doses or placebo. The 1 h postprandial glucose concentration was up to 15% higher following camicinal compared with placebo; however, at later time points (5–6 h post‐dose), the concentrations following camicinal 125 mg were lower than those following placebo, neither of which was considered to be clinically significant. Figure 4 illustrates that the between‐patient variability in plasma glucose concentrations following camicinal 125 mg, which appeared to be lower with more predictable glucose absorption than the between‐patient variability following placebo. No excursions into the hypoglycaemic or hyperglycaemic ranges were observed following any dose of camicinal.

Figure 4.

Plot of mean (±SD) plasma glucose profiles following placebo or camicinal 125 mg. Black line is the mean placebo glucose profile with SD as gray shaded area; red line is the camicinal 125 mg mean glucose profile with SD as pink shaded area.

Other exploratory endpoints

As expected, given the single‐dose design of the study, there appeared to be minimal differences or changes between placebo and any dose of camicinal based on the following exploratory parameters: time to first bowel movement post‐dose, bowel movement count, stool consistency (Bristol Stool Form Score), hunger sensation and gastroparesis symptoms.

Safety and tolerability

Within the constraints of the small sample size, the overall frequency and types of reported adverse events was low and similar when patients were administered camicinal (all doses) or placebo. Adverse events reported by more than one patient were headache, flatulence, vomiting, constipation and decreased blood glucose, which were adverse events expected in this population of T1DM patients with gastroparesis. Adverse events greater than or equal to two patients are reported in Table 4. The majority of adverse events considered by the investigator to be related to administration of camicinal were headache, flatulence, vomiting, and other gastrointestinal disorders. One patient receiving placebo experienced a serious adverse event of moderate abnormal blood glucose and QT interval prolongation, which was not considered by the investigator to be related to treatment. No clinically significant trends in electrocardiograms, clinical laboratory tests (including plasma glucose) and vital signs were detected when dosed with camicinal.

Table 4.

Summary of adverse events in two or more patients

Adverse event	Placebo N = 10	Camicinal 25 mg N = 6	Camicinal 50 mg N = 6	Camicinal 125 mg N = 6
Headache	1	1	3	1
Flatulence	2	1	1	1
Vomiting	1	0	1	1
Constipation	0	1	0	1
Decreased blood glucose	1	0	1	1

Discussion and conclusions

The results of this single‐dose, crossover study show that the selective motilin receptor agonist camicinal stimulates gastric emptying in patients with T1DM. Thus, as assessed by ¹³C‐octanoic acid breath test, the gastric half‐emptying time was decreased in a dose‐dependent fashion by approximately 27% with camicinal 25 and 50 mg, and approximately 65% after 125 mg compared with placebo. The exposures achieved following 25 and 50 mg dosing did overlap. In addition, an exposure–response model was developed and showed a positive relationship with camicinal exposure (AUC) and decreasing gastric emptying half‐time. These effects of camicinal on gastric emptying half‐time were also supported by a reduction in gastric emptying lag phase and increased gastric emptying coefficient. Hence, the data indicate that camicinal exerts stimulatory action on all phases of the gastric emptying process. Notably, administration of camicinal at single doses of 25, 50 and 125 mg was well tolerated in T1DM patients experiencing delayed gastric emptying with symptoms. Few side effects related to camicinal were reported, and these effects occurred at a similar frequency as those reported with placebo. Of particular interest was the absence of nausea, vomiting and diarrhoea. This is an adverse event profile commonly reported with erythromycin when used at doses required to treat infections (Catnach and Fairclough, 1992; Boivin et al., 2003), but which are greater than necessary to produce maximal effect on gastric emptying (Desautels et al., 1995). The present data are therefore consistent with the suggestion that different, dose‐dependent mechanisms underpin the abilities of motilin receptor agonists to either increase gastric emptying or induce nausea (Desautels et al., 1995; Javid et al., 2013). Finally, the paucity of effects on lower gut functions such as first bowel movement post‐dose, bowel movement count and stool consistency are consistent with the absence of functional motilin receptors affecting enteric neuromuscular functions in the lower gut (Sanger et al., 2009; Broad et al., 2012).

Camicinal was well absorbed in T1DM patients with gastroparesis and had similar mean exposures to those observed in healthy volunteers, but with higher variability (Dukes et al., 2010). Pharmacokinetic data in both studies showed a linear and approximately dose‐proportional behaviour as expected for a low MW compound. Its estimated half‐life of about 26 h confirmed that once daily dosage of camicinal would be appropriate in forthcoming use. In the current study, there was a clear exposure–response relationship with gastric emptying in diabetic patients with gastroparesis. We have previously demonstrated an exposure–response relationship to camicinal upon gastric emptying in healthy volunteers (Dukes et al., 2010). The mean maximum concentrations of camicinal achieved with the doses employed in this study (680–2649 nmol⋅L⁻¹) were greater than the EC₅₀ value of camicinal (13 nmol⋅L⁻¹ (pEC₅₀ 7.9) reported previously (Sanger et al., 2009), which indicates an appropriate receptor stimulation should have been achieved to account for effect on gastric emptying noted in the current study. As plasma concentrations and drug exposures increased, gastric emptying was enhanced supporting the dose‐related increased effect on gastric emptying.

In line with its effects on gastric emptying, camicinal 125 mg increased glucose uptake with the 1 h postprandial glucose concentration being up to 15% higher compared with placebo. At later time points (5–6 h post‐dose), the glucose concentrations following camicinal were lower than those following placebo. These data may indicate that camicinal, by normalizing gastric emptying, improves the absorption profile of glucose following a meal. This may translate into fewer excursions into the extreme glucose range and potentially lower HbA1c, which may provide better metabolic control.

Symptomatic improvement was not observed in this study. This is not surprising given some of the limitations of the study including (i) only a small number of patients were enrolled in the study; (ii) patients only received a single dose per study period; and (iii) each period consisted of intense study procedures, which may have contributed to abnormal reporting of symptoms. Furthermore, subjective gastric sensations are prone to variability and therefore constitute a less reliable readout than direct or indirect measures of changes in gastric emptying. One indirect assessment of gastric emptying in this study was the amount of food ingested at each meal, which was increased in the 24 h following camicinal administration. At present, it is not clear if this observation is related to the proposed link between the release of motilin and increased feeling of hunger (Tack et al., 2014) or if the volunteers simply ate more because the increased emptying of food from the stomach facilitated the return of hunger (Janssen et al., 2011). Further studies are needed to answer this question. Additional limitations of the study, which may have confounded interpretation of overall results, included short duration of the study and the inherently variable nature of gastric emptying measurements, which are not likely to address all of the complex processes involved in gastroparesis symptoms (Yarandi and Srinivasan, 2014).

In conclusion, single doses of the novel, selective motilin receptor agonist, camicinal, accelerated the gastric emptying of solids in T1DM patients with gastroparesis in an exposure‐related manner. Camicinal was well tolerated and exhibited similar pharmacokinetic characteristics in diabetic patients as in healthy volunteers. Further studies are now needed to investigate the ability of camicinal to improve symptoms in this group of patients, bearing in mind the different ways in which this drug could potentially affect gastric functions. Camicinal seems to be a promising agent in the treatment of diabetic gastroparesis with a possible additional capability of rectifying metabolic control.

Author contributions

P.M.H. performed the research and contributed to writing the paper. J.T. performed the research and contributed to writing the paper. L.V.J., K.H. and G.E.D. contributed to study design, data analysis and to writing the paper. M.E.B., D.B.R. and D.H.A. contributed to study design and to writing the paper. G.J.S. contributed to the data analysis and to writing the paper.

Conflict of interest

L.V.J., K.H., M.E.B., D.B.R. and G.E.D. are all employees of GlaxoSmithKline, Inc. P.M.H. has no conflict of interest to declare. J.T. has acted as a consultant to GlaxoSmithKline. D.H.A. was a paid consultant for GlaxoSmithKline R&D. G.J.S. has previously received research funding for research into the mechanisms of action of the motilin receptor agonists camicinal (GSK) and RQ‐00201894 (RaQualia).

Declaration of transparency and scientific rigour

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research recommended by funding agencies, publishers and other organizations engaged with supporting research.

Hellström, P. M. , Tack, J. , Johnson, L. V. , Hacquoil, K. , Barton, M. E. , Richards, D. B. , Alpers, D. H. , Sanger, G. J. , and Dukes, G. E. (2016) The pharmacodynamics, safety and pharmacokinetics of single doses of the motilin agonist, camicinal, in type 1 diabetes mellitus with slow gastric emptying. British Journal of Pharmacology, 173: 1768–1777. doi: 10.1111/bph.13475.