Effect of sirolimus on insulin dynamics in horses

Abstract

Background

Sirolimus, a mechanistic target of rapamycin inhibitor, suppresses insulin production in other species and has therapeutic potential for hyperinsulinemia in horses.

Hypothesis/Objective

Determine the pharmacokinetics (PKs) of sirolimus and evaluate its effect on insulin dynamics in healthy and insulin dysregulation (ID) horses.

Animals

Eight Standardbred geldings.

Methods

A PK study was performed followed by a placebo‐controlled, randomized, crossover study. Blood sirolimus concentrations were measured by liquid chromatography‐mass‐spectrometry. PK indices were estimated by fitting a 2‐compartment model using nonlinear least squares regression. An oral glucose test (OGT) was conducted before and 4, 24, 72, and 144 hours after administration of sirolimus or placebo. Effects of time, treatment and animal on blood glucose and insulin concentrations were analyzed using mixed‐effects linear regression. Sirolimus was then administered to 4 horses with dexamethasone‐induced ID and an OGT was performed at baseline, after ID induction and after 7 days of treatment.

Results

Median (range) maximum sirolimus concentration was 277.0 (247.5‐316.06) ng/mL at 5 (5‐10) min and half‐life was 3552 (3248‐4767) min. Mean (range) oral bioavailability was 9.5 (6.8‐12.4)%. Sirolimus had a significant effect on insulin concentration 24 hours after a single dose: median (interquartile range) insulin at 60 min (5.0 [3.7‐7.0] μIU/mL) was 37 (−5 to 54)% less than placebo (8.7 [5.8‐13.7] μIU/mL, P = .03); and at 120 min (10.2 [8.4‐12.2] μIU/mL) was 28 (−15 to 53)% less than placebo (14.9 [8.4‐24.8] μIU/mL, P = .02). There was minimal effect on glucose concentration. Insulin responses decreased toward baseline in ID horses after 7 days of treatment.

Conclusion and Clinical Importance

Sirolimus decreased the insulinemic response to glucose and warrants further investigation.

Abbreviations

AUC

area under the curve

C _max

maximum plasma concentration

EDTA

ethylenediaminetetraacetic acid

EMS

equine metabolic syndrome

FKBP

FK506 binding protein

GIP

glucose‐dependent insulinotropic peptide

GLP‐1

glucagon‐like peptide 1

HAL

hyperinsulinemia associated laminitis

HLPC

high‐performance liquid chromatography

ID

insulin dysregulation

IQR

interquartile range

ISTD

internal standard

LC‐MS

liquid chromatography‐mass‐spectrometry

LLOQ

Lower Limit of Quantitation

mTOR

mechanistic target of rapamycin

mTORC1

mechanistic target of rapamycin complex 1

mTORC2

mechanistic target of rapamycin complex 2

NCA

non‐compartmental analysis

OGT

oral glucose test

PK

pharmacokinetic

PPID

pituitary pars intermedia dysfunction

SGLT‐2

sodium‐glucose‐linked cotransport‐2

T _Cmax

time to reach the maximum plasma concentration

1. INTRODUCTION

Laminitis most commonly affects horses with an underlying endocrinopathy: either pituitary pars intermedia dysfunction or equine metabolic syndrome. ¹ , ² , ³ The central role of insulin dysregulation (ID), specifically hyperinsulinemia, in the development of endocrinopathic laminitis has been established through multiple studies using experimental models, ⁴ , ⁵ , ⁶ , ⁷ prompting introduction of the term “hyperinsulinemia‐associated laminitis” (HAL) to describe this form of laminitis. ¹ , ² Although restriction of dietary non‐structural carbohydrates remains a mainstay of ID management in horses, ⁸ , ⁹ pharmaceutical control of the exaggerated post‐prandial insulinemic response would assist in the prevention of HAL. ⁶ , ⁸ There are currently limited pharmacological options for controlling insulinemic responses in horses. Metformin is an oral biguanide with anti‐hyperglycemic and insulin sensitizing effects, but variable clinical efficacy is reported in horses attributed to its low oral bioavailability. ¹⁰ , ¹¹ , ¹² , ¹³ Sodium‐glucose‐linked cotransport‐2 inhibitors decrease insulin secretion by promoting the renal elimination of glucose and can prevent or ameliorate HAL under experimental conditions ¹⁴ , ¹⁵ ; however, these drugs are still under investigation and are not registered for veterinary use. There is a clear need for the development of pharmaceutical therapies for the management of ID and HAL in horses.

Sirolimus (rapamycin) is used in human medicine for its anti‐proliferative, immunosuppressive, and neuroprotective effects. ¹⁶ Sirolimus inhibits mechanistic target of rapamycin (mTOR), a highly conserved serine‐threonine protein kinase which integrates numerous signals from growth factors, hormones, nutrients, and cellular energy levels to regulate protein translation and cell growth, proliferation, and survival. ¹⁶ The mTOR signaling pathway is activated within the lamellae during laminitis development in experimental models of HAL. ¹⁷ Furthermore, mTOR is also a key positive regulator of pancreatic β‐cell function, ¹⁸ and its inhibition suppresses insulin production and secretion, suggesting that it could be an attractive target for the management of ID and prevention of HAL. Published information on the effects of sirolimus in horses is lacking, therefore the objectives of this study were (a) to determine the pharmacokinetics (PKs) of sirolimus, (b) to evaluate the effect of a single dose of sirolimus on insulin dynamics in healthy horses, and (c) to obtain preliminary data on the effects of oral sirolimus in horses with experimentally induced ID. We hypothesized that sirolimus would decrease the insulinemic response to a glucose challenge in horses.

2. MATERIALS AND METHODS

2.1. Study design

Eight healthy Standardbred geldings (median age: 9 years; range: 6‐11 years; mean bodyweight: 458.9 ± 33.3 kg) were acclimated in a paddock for 1‐week before commencing the study. Horses were considered clinically healthy based on results of a physical examination and plasma biochemistry and serum hematology analyses. The horses were confined to individual stalls with free‐choice Rhodes grass/alfalfa hay and water during all experiments, unless otherwise specified. A PK study was performed first (experiment 1), followed (after a 6‐week washout period) by a placebo‐controlled crossover study evaluating the effect of sirolimus on insulin dynamics in healthy adult horses (experiment 2). After a further 6‐week washout period, the effect of oral sirolimus was evaluated in 4 of the horses with dexamethasone‐induced ID. No anxiolytic or sedative medications were used during the experiments.

In experiment 1, all 8 horses were administered an intravenous dose of sirolimus (0.06 mg/kg bodyweight; 0.2% suspended in propylene glycol) injected over 5 min, with blood samples collected before dosing (T0), and then at 5, 10, 20, 30, 45, 60, 90 min and 2, 3, 4, 6, 8, 12, 24, 36, 48, and 72 hours after sirolimus administration to measure sirolimus concentration. After a 1‐week washout, all 8 horses were administered an enteral dose of sirolimus (0.06 mg/kg bodyweight; 0.2% suspended in propylene glycol) via nasogastric tube, with serial blood sampling performed as for the intravenous dosing. The horses were then returned to the paddock for a 6‐week washout period before commencing experiment 2.

Experiment 2 was conducted with the same 8 horses in a placebo‐controlled, non‐blinded, randomized, crossover design. After a 24‐hour period of individual stall acclimation, an oral glucose test (OGT) was performed 24 hours before, and at 4, 24, 72, and 144 hours after, the intravenous administration of sirolimus (0.06 mg/kg) or equivalent volume of placebo (propylene glycol). Treatment order (sirolimus vs placebo) was randomly assigned for each horse by coin toss. The OGT timing was intended to facilitate testing of insulin dynamics at specific blood sirolimus concentrations based on data from experiment 1 (the predicted blood concentrations at 4, 24, 72, and 144 hours after a single 0.06 mg/kg intravenous dose were approximately 60, 20, 10, and 5 ng/mL, respectively). Sirolimus (or placebo) was administered slowly (over 5 min) via a 14‐gauge intravenous catheter which had been placed in the right jugular vein at least 2 hours before dosing and was then removed. Blood sampling was performed via a 14‐gauge catheter placed in the left jugular vein the evening before each OGT and maintained during the experiment. The OGT procedure was performed as described ¹⁹ : 1 g/kg of dextrose powder dissolved in 2 L of warm water was administered via a nasogastric tube after a 6‐hour fast (all feed was removed from the stall 6 hours before the OGT). Blood samples were collected at 0, 15, 30, 60, 90, 120, 150, 180, 210, and 240 min after dextrose administration. Hay was withheld during the testing period. Blood glucose concentration was measured stall‐side, and samples for serum insulin concentration (0, 60, 90, and 120 min) were stored for later analysis. Samples for measurement of sirolimus concentration in plasma and whole blood were also obtained at the commencement of each OGT. After the 144 hours OGT, the horses were returned to a paddock for a 3‐week washout period. The experiment was then repeated with each horse receiving the opposite treatment.

Experiment 3 was conducted using 4 of the healthy Standardbred geldings used in the previous experiments (median age: 8 years; range: 6‐11 years; mean bodyweight: 486.0 ± 50.3 kg). After a 24‐hour period of individual stall acclimation, a baseline OGT (day 0) was performed and treatment with dexamethasone (0.08 mg/kg IM) every other day was then commenced to induce ID. ¹³ , ²⁰ On day 8, a second OGT was performed, then sirolimus was added (0.06 mg/kg orally once a day) concurrently with dexamethasone for a further 7 days. A final OGT was performed on day 15. Whole blood samples for estimation of sirolimus peak and trough concentrations were obtained at 2 and 24 hours (just before repeat administration) after each sirolimus dose, respectively. A daily “resting” blood glucose measurement was performed at the time of dexamethasone administration.

2.2. Insulin and glucose measurement

Blood glucose concentration was measured immediately at all sample time points using a portable glucometer validated for use in horses (AlphaTRAK2, Zoetis Inc, Michigan, USA) with a recorded intra‐assay variability of 5.3%. ²¹ , ²² Blood for measurement of insulin concentration (0, 60, 90, and 120 min) was immediately placed into a plain serum tube, allowed to clot at room temperature, and then centrifuged for 10 min at 450g at 4°C with the separated serum subsequently frozen at −80°C and analyzed within 6 weeks. Insulin concentration was measured using a validated benchtop chemiluminescent assay (Immulite 1000, Siemens, Bayswater, Australia) with a recorded intra‐assay variability of 6.5%. ²³ , ²⁴

2.3. Whole blood and plasma sirolimus measurement

Whole blood (EDTA) and plasma (EDTA, centrifuged and separated) were stored at −20°C until analysis of sirolimus concentration within 4 weeks by liquid chromatography‐mass‐spectrometry (LC‐MS). Sirolimus was extracted by protein precipitation with 0.1 M zinc sulfate, followed by solid‐phase extraction (full details in Supplementary Item 1). The eluant was then transferred to high‐performance liquid chromatography vials and analyzed using mass spectrometry (MS; MS/MS mode). Sirolimus, the internal standard (ISTD; 1000 ng/mL tacrolimus) and normal equine whole blood were used for calibration and quality controls. A matrix‐matched calibration curve spanning the concentration range (0.5‐500 ng/mL) was prepared. The detector response based on peak area ratio to the internal standard was linear (r > .99) in the concentration range with weighting 1/x applied. Sciex OS software was used for data processing. Quantitation was based on the peak area ratio of sirolimus detected in each sample to the ISTD (ie, tacrolimus) using the calibration curve. The Lower Limit of Quantitation (LLOQ) for sirolimus in equine whole blood or plasma was 0.5 ng/mL as determined by standard analytical method validation guidelines. ²⁵

2.4. PK analysis

Sirolimus whole‐blood concentration vs time data were fit to a compartmental model using nonlinear least squares regression via a mathematical modeling software package (WinSAAM, University of Pennsylvania, Kennett Square, Pennsylvania, USA) as performed previously. ²⁶ , ²⁷ , ²⁸ Visual inspection of the IV‐time profile of blood sirolimus suggested that the data fit a 2‐compartment model. The best fit was obtained by minimizing fractional SD of the micro‐rate constants and by convergence of the predicted and observed plasma concentration curves. Maximum plasma concentration (C _max), and the time to reach the maximum plasma concentration (T _Cmax) were determined directly from the data. Terminal half‐life (IV) was calculated using a non‐compartmental analysis.

2.5. Statistical analysis

All analyses were conducted using Stata (16MP), with 2‐sided tests of hypotheses and a P < .05 as the criterion for statistical significance. Tests of normal distribution (Shapiro‐Wilk) were performed to determine the extent of skewness. For data with normal distribution, mean and SD were reported. Significantly skewed data were reported as medians and interquartile ranges (IQR). For the outcome variables (glucose and insulin concentrations) inferential statistical analysis was based on a mixed‐effects linear regression model with the fixed effects of the interaction between categorical time (0, 60, 90, and 120 min) treatment (sirolimus vs placebo) and interval after treatment (−24, 4, 24, 72, and 144 hours) and the fixed effect of bodyweight. Random effects were set on the level of the animal. To permit for possible departures from normality for some of the outcomes, robust estimation of the variance was used. Post‐hoc analysis was used for calculating the marginal (model adjusted) means and to calculate the pairwise effects. Fisher protected least significant difference method was used to adjust the significance values for multiple comparisons. All marginal means and effects are reported with their respective 95% CI. Results of experiment 3 were reported descriptively.

3. RESULTS

A single intravenous and oral dose of sirolimus was well tolerated in this group of 8 healthy adult horses, with heart rate, respiratory rate, and rectal temperature remaining within normal limits in all animals in experiment 1 (Table 1). Transient hemolysis with concurrent pigmenturia occurred after intravenous administration of sirolimus in 4 horses during experiment 1, but not after oral administration. Transient pigmenturia was also observed in 7 horses during experiment 2 after intravenous infusion of either sirolimus or the placebo; voided urine appeared grossly normal by 6 hours in these horses. All blood samples were hemolyzed at the 4‐hour timepoint in experiment 2 but were grossly normal by 24 hours, and at all other timepoints. Three horses developed a transient increase in rectal temperature (up to 41°C) during experiment 2, but physical examination and additional diagnostics (routine hematology and serum biochemistry) were otherwise unremarkable. Two of these horses had received sirolimus and 1 had received the placebo 4 to 7 days earlier.

3.1. Pharmacokinetics

Whole blood concentrations (mean ± SD) of sirolimus after a single intravenous dose in experiment 1 are shown in Figure 1. The corresponding PK parameters are shown in Table 1. Volume of distribution was 371.2 mL/kg (range, 306.2‐386.3) and clearance was 0.36 mL/kg/min (range, 0.33‐0.40). Terminal half‐life was calculated to be 3552 min (range, 3248‐4767). Mean (range) oral bioavailability was 9.5 (range, 6.8‐12.4)%.

FIGURE 1.

Mean (±SD) whole blood concentrations of sirolimus over time in 8 horses administered a single intravenous dose (0.06 mg/kg). Timing of each oral glucose test in experiment 2 was based on predicted whole blood concentrations of sirolimus from the pharmacokinetic data: concentrations at 4, 24, 72, and 144 hours after a single 0.06 mg/kg dose were approximately 60, 20, 10, and 5 ng/mL, respectively (dotted lines, 144 hours not shown).

TABLE 1.

Pharmacokinetic parameter estimates (median, range) for sirolimus following a single dose (0.06 mg/kg) in 8 healthy adult horses.

Parameters	Sirolimus, 0.06 mg/kg (n = 8)
C max (ng/mL)	277.0 (247.5‐316.06)
T Cmax (min)	5 (5‐10)
Vd (mL/kg)	371.2 (306.2‐386.3)
CL (mL/kg/min)	0.36 (0.33‐0.40)
k10 (L/min)	0.0010 (0.0009‐0.0012)
k12 (L/min)	0.0036 (0.0026‐0.0042)
k21 (L/min)	0.0011 (0.0008‐0.0013)
T 1/2 (min)	677.1 (594.73‐809.15)
Terminal half‐life (min)	3552 (3248‐4767)
F	0.0954 (0.0680‐0.1246)

Abbreviations: CL, clearance; C _max, peak blood concentration; F, extravascular bioavailability; T _1/2, half‐life; T _Cmax, time until maximum concentration; Vd, volume of distribution.

Plasma concentrations of sirolimus were also measured in 4 horses at 3 timepoints during experiment 2 and remained at (0.63 ± 0.3 ng/mL at 4 hours after dosing) or below (<0.5 ng/mL at 24 and 144 hours after dosing) the LLOQ. In experiment 3, the median (range) peak and trough whole blood sirolimus concentrations (Table 3) were highest on the final day of treatment: 2.3 ng/mL (1.8‐3.6) and 1.5 ng/mL (1.1‐1.7), respectively.

3.2. Insulin and glucose

Serum insulin concentrations during the OGT at each interval after intravenous administration of sirolimus (vs placebo) for experiment 2 are shown in Figure 2. There was a significant effect of treatment on serum insulin concentrations only during the OGT performed 24 hours after treatment (Table 2). At 24 hours after administration of sirolimus, median (IQR) serum insulin concentration was 37% (−5% to 54%) lower than placebo at the 60 min OGT time point (P = .03) and 28% (−15% to 53%) lower than placebo at the 120 min OGT time point (P = .02; Figure 3). There was no concurrent change in glucose concentration at 24 hours after administration of sirolimus. At 4 hours after administration, blood glucose was lower at the 0 min OGT time point with sirolimus (103.5 [98.1‐107.6] mg/dL) compared to placebo (108.9 [104.4‐114.3] mg/dL); P = .006, and blood glucose was also lower at the 60 min OGT time point with sirolimus (143.1 [142.2‐149.4] mg/dL) compared to placebo (163.8 [146.7‐168.8] mg/dL); P = .009. There was no effect of administration of sirolimus on blood glucose concentration at any other OGT time point (Table 4). There was no change in AUC for insulin or glucose for any OGT, apart from a significant decrease (P = .02) in the AUC_glucose at the 4 hours OGT with administration of sirolimus (996.9 ± 194.1 mg/dL × min) compared to placebo (1058 ± 195.3 mg/dL × min; Table 5).

FIGURE 2.

Serum insulin concentration in healthy adult horses (n = 8; mean ± SD) during sequential oral glucose test (60, 90, and 120 min timepoints) performed before (baseline; 0 hour) and 4, 24, 72, and 144 hours after a single dose of sirolimus (0.06 mg/kg IV) or placebo. * = significant difference between sirolimus and placebo; P < .05.

TABLE 2.

Serum insulin concentrations (median ± IQR) obtained during successive oral glucose test (OGT) performed before (ie, baseline) and after the administration of a single dose of sirolimus (0.06 mg/kg IV) or placebo in 8 healthy Standardbred horses.

Time	OGT timepoint	Insulin concentration (μIU/mL)
		Sirolimus	Placebo	P‐value
Before treatment	0 min Blood sirolimus conc.: <0.5 ng/mL	2.0 (2.0‐2.0)	2.0 (2.0‐2.0)	.61
	60 min	6.7 (4.2‐7.7)	6.3 (3.3‐8.2)	.52
	90 min	6.0 (4.8‐10.6)	9.0 (5.4‐16.4)	.32
	120 min	7.6 (6.0‐14.6)	10.1 (3.9‐14.6)	.66
4 hours	0 min Blood sirolimus conc.: 60.3 ± 9.4 ng/mL	2.0 (2.0‐2.0)	2.0 (2.0‐2.0)	.74
	60 min	3.3 (2.0‐3.8)	4.0 (2.0‐7.7)	.14
	90 min	4.7 (3.3‐6.6)	6.4 (3.5‐10.4)	.13
	120 min	6.2 (4.2‐8.0)	8.5 (4.0‐12.0)	.94
24 hours	0 min Blood sirolimus conc.: 19.2 ± 3.3 ng/mL	2.0 (2.0‐2.0)	2.0 (2.0‐2.0)	.73
	60 min	5.0 (3.7‐7.0)	8.7 (5.8‐13.7)	.03
	90 min	8.1 (6.8‐8.7)	12.6 (7.2‐16.4)	.07
	120 min	10.2 (8.4‐12.2)	14.9 (8.4‐24.8)	.02
72 hours	0 min Blood sirolimus conc.: 9.5 ± 2.0 ng/mL	2.0 (2.0‐2.0)	2.0 (2.0‐2.0)	.99
	60 min	7.0 (5.3‐8.8)	7.3 (5.5‐11.9)	.28
	90 min	12.1 (9.2‐18.4)	10.4 (8.0‐16.3)	.58
	120 min	15.7 (10.3‐21.3)	13.1 (5.4‐20.2)	.47
144 hours	0 min Blood sirolimus conc.: 4.5 ± 1.3 ng/mL	2.0 (2.0‐2.5)	2.0 (2.0‐2.1)	.77
	60 min	11.7 (7.4‐12.3)	8.1 (7.1‐17.4)	.83
	90 min	15.8 (10.7‐22.5)	15.4 (11.5‐19.9)	.98
	120 min	18.6 (12‐36.9)	16.5 (16.0‐24.3)	.75

Note: The measured concentration of sirolimus (mean ± SD) in whole blood at the start of each OGT is also provided.

FIGURE 3.

Spaghetti plots comparing serum insulin and blood glucose concentrations in healthy adult horses (n = 8) taken during an oral glucose test (60, 90, and 120 min timepoints) performed 24 hours after a single dose of sirolimus (0.06 mg/kg IV) or placebo. Serum insulin was lower with sirolimus treatment compared with placebo at the 60 and 90 min timepoints, but there was no significant effect on blood glucose. * = significant difference between sirolimus and placebo; P < .05.

In experiment 3, the OGT performed after 7 days of dexamethasone administration showed an exaggerated serum insulin response to glucose challenge relative to the respective baseline OGT in each horse (Figure 4). After sirolimus was added to the administration of dexamethasone for an additional 7 days, insulin responses in the final OGT appeared similar to that of the baseline OGT (Figure 4). The median (range) peak insulin concentration was 13.6 (5.9‐15.8) μIU/mL at the baseline OGT; 81.6 (41‐254) μIU/mL after 7 days of dexamethasone administration and 28.2 (13.5‐54.4) μIU/mL after a further 7 days of dexamethasone with added administration of sirolimus. Peak blood glucose concentrations were 186.3 (156.6‐189) mg/dL at baseline, 261.9 (226.8‐282.6) mg/dL after dexamethasone and 237.6 (181.8‐257.4) after sirolimus. The resting blood glucose concentrations were 115.7 (99.5‐132.8) mg/dl during the first 7 days of dexamethasone administration and 111.6 (98.1‐121.5) mg/dl during administration of sirolimus and dexamethasone.

FIGURE 4.

Individual horse blood insulin responses to an oral glucose test at baseline, after a week of dexamethasone treatment to induce insulin dysregulation (ID), and then after a further week of dexamethasone treatment combined with 0.06 mg/kg sid oral sirolimus (ID + SIROLIMUS). In each case, treatment with sirolimus returned post‐prandial insulin responses to near baseline values.

4. DISCUSSION

This study establishes the PKs of sirolimus and its effects on insulinemic responses to OGT in healthy adult horses. A significant decrease in the insulinemic response (vs placebo) was observed only during the OGT performed 24 hours after administration of the single sirolimus dose, with no concurrent significant effects on blood glucose. Although the absolute decrease in blood insulin concentration was small in healthy horses treated with a single intravenous dose, the results of the preliminary multi‐dose oral study demonstrate that sirolimus has the potential to suppress exaggerated insulin responses to OGT in experimentally induced ID. Since hyperinsulinemia is central to the development of HAL, ⁷ sirolimus warrants further investigation using placebo‐controlled studies in horses with ID in order to determine if its effects could be clinically useful. Although there were no observed adverse effects attributable to sirolimus, the safety of longer‐term administration also requires further investigation.

The insulinemic responses at the 4 and 72 hours OGTs were not different than placebo, suggesting that sirolimus affects insulin dynamics in healthy horses for a period beginning >4 hours but ending <72 hours after administration of a single dose. Furthermore, the lack of an apparent effect when blood sirolimus concentrations were highest (at the 4‐hour OGT) suggests that the effects of sirolimus on insulin dynamics in horses are not solely dependent on blood concentration. The reported effects of sirolimus on insulin and glucose dynamics vary with species, dose rate, and duration of treatment. ²⁹ , ³⁰ , ³¹ , ³² Evidence primarily from in vitro studies demonstrates that treatment with sirolimus leads to reduced insulin secretion and affects β‐cell mass by decreasing β‐cell survival and proliferation in human and murine pancreatic islets. ³³ These effects are thought to contribute to the increased risk of post‐transplant diabetes mellitus in human patients treated with sirolimus. ³⁴ However, a recent in vivo study of human pancreatic islets transplanted into mice treated with sirolimus for 4 weeks (maintaining typical therapeutic concentrations of 5‐20 ng/mL) showed that there was no effect of sirolimus on β‐cell proliferation or survival. ²⁹ Rather, sirolimus had multiple functional effects on human islets, including impaired insulin secretion, insulin processing and β‐cell granule formation, which resolved within 4 weeks of cessation of treatment. ²⁹

The direct suppressive effects of sirolimus on insulin secretion are likely due to inhibition of the mTOR pathway through interaction with its 2 discrete complexes: mTOR complex 1 (mTORC1; primarily responsible for regulation of cell growth); and mTOR complex 2 (mTORC2; responsible for cytoskeleton regulation and cell survival). ³¹ , ³⁵ Treatment with sirolimus mimics starvation and activates autophagy, exerting its effects mostly via direct binding and inhibition of mTORC1. Although direct binding to mTORC2 by sirolimus does not occur, mTORC2 inhibition has been observed in specific tissues and with longer exposure to the drug. ³¹ , ³⁶ Both mTORC1 and mTORC2 might have roles in regulation of insulin production and secretion in response to glucose, ³⁷ however the effects of sirolimus treatment on insulin dynamics are generally attributed to mTORC2 inhibition. ³⁸ Treatment of cells in vitro with sirolimus caused mTORC2 inhibition only after at least 24 hours of exposure to the drug, and the effect varied with cell type. ³⁶ It is therefore plausible that this delayed inhibition of mTORC2 was responsible for suppression of the insulinemic response in the current study, which was not observed until 24 hours after administration.

Reported adverse effects in humans include hyperglycemia and hyperlipidemia due to sirolimus‐induced changes in glucose and lipid metabolism, as well as insulin sensitivity. ²⁹ , ³⁰ , ³⁵ , ³⁹ The effect of mTOR inhibitors on glucose regulation is not fully understood but has been utilized therapeutically in the management of hyperinsulinemic hypoglycemia resulting in regulation of blood glucose in cases referrable to abnormal activation of the mTOR pathway. ⁴⁰ The exact mechanisms by which mTOR inhibitors normalize blood glucose in these patients are not known, ⁴¹ but are associated with a decrease in the secretion and production of insulin ⁴² , ⁴³ and perhaps also increased hepatic gluconeogenesis ³⁵ and reduced hepatic insulin sensitivity. ³⁸

In the current study, the only significant effect of sirolimus on blood glucose was observed at the 4‐hour OGT (a small decrease); however, blood glucose concentrations remained within reference range and were of limited clinical relevance after a single sirolimus dose. Tissue insulin resistance was not directly assessed in this study. Although insulin resistance is a component of ID in many horses, blood insulin concentration has been demonstrated to be directly associated with the development, recurrence, and recovery from laminitis ⁴ , ⁵ , ⁷ , ⁴⁴ ; therefore the effect of sirolimus on blood insulin concentrations after carbohydrate challenge was the focus of the current study.

Sirolimus has been investigated in other species ⁴⁵ and the PKs appear to be non‐linear, varying between species and with different dosing regimens. ⁴⁶ Pharmacokinetics of a single intravenous dose of sirolimus (0.5 and 0.05 mg/kg) in the rabbit demonstrated dose‐dependent, nonlinear PKs and delayed clearance (terminal half‐life: 768.6 ± 128.4 min; clearance: 0.977 ± 0.279 mL/min/kg at 0.05 mg/kg). ⁴⁶

As in our study, the data fit a 2‐compartment model. Using a similar dose (0.06 mg/kg), a much longer terminal half‐life and a slower clearance rate was observed in the horses of the current study. There is a dearth of other published PK studies investigating a single dose of intravenous sirolimus in other species. In our study, oral bioavailability of sirolimus was low and the maximal trough concentrations after 7 days of oral administration were also low. Despite this, there still appeared to be an effect on insulin dynamics in each of the 4 horses. The PK properties of sirolimus change significantly after consecutive dosing, and so further studies investigating long‐term oral dosing regimens (using a larger group of horses) are indicated. A preliminary investigation into the PKs of oral low‐dose (0.1 mg/kg) sirolimus in healthy dogs showed that blood concentrations peaked at 2‐ and 6‐hours after administration, but this effect could not be explained and was not repeatable with consecutive dosing. ⁴⁷ The current study shows that in horses, as in other species, the plasma concentration of sirolimus is much less than that of whole blood. In human blood, sirolimus is predominantly sequestered in erythrocytes (mean % distribution ± SD: 94.5% ± 4.9%). ⁴⁵ , ⁴⁸ Drug sequestration in red blood cells likely occurs due to proteins such as FK506 binding protein, ⁴⁵ and might also contribute to its pharmacologic activity due to cell‐to‐cell transfer. ⁴⁵

Sirolimus was generally well tolerated in this group of horses over the short treatment periods. Adverse effects observed after intravenous administration included transient pigmenturia with hemolysis, and a transient increase in rectal temperature in some horses, both of which were also seen in horses receiving placebo. Propylene glycol was used as a solvent for sirolimus and administered as the placebo compound. Intravenous administration of propylene glycol can result in hypotension, pain on intravenous administration (due to its osmolarity), lactic acidosis, and pulmonary hypertension. ⁴⁹ Intravascular hemolysis has also been reported after a single intravenous infusion in humans due to its high osmolarity. ⁴⁹ Considering the long terminal half‐life of sirolimus, there is a possibility of drug accumulation with repeated administration which is important to establish over longer treatment periods. The results of our preliminary oral study indicate that the effects of sirolimus on insulin dynamics after oral administration could be of sufficient magnitude to be clinically useful in horses with ID; however, interpretation of these results is limited by the lack of a placebo group and the small number of horses. Further placebo‐controlled studies in horses with either naturally occurring or experimentally induced ID are needed to evaluate the effects on insulin and glucose dynamics. The anti‐proliferative effects of sirolimus cause dose‐dependent immunosuppression primarily through inhibition of T and B cell expansion, ⁵⁰ a desirable clinical effect in transplant patients, but a potentially serious adverse effect if the drug is to be used for insulin control in horses. We did not specifically evaluate immune function in the current study, however maximum trough blood sirolimus concentrations in the oral study (1.5 ng/mL) were a fraction of the target trough concentrations used for immunosuppressive therapy in humans (12‐30 ng/mL). ⁵¹ , ⁵² Further studies evaluating the effects of sirolimus on immune function in horses, particularly with longer term administration, are required.

Although administration of sirolimus decreased the insulinemic response to the OGT 24 hours after dosing, we did not evaluate the mechanism for this.

Sirolimus had a significant effect on insulin concentration 24 hours after dosing, with a decrease in insulin compared to baseline. However, the intra‐assay variability of the Immulite 1000 is 6.5% ²³ , ²⁴ which might have resulted in a type‐1 error. Although a decrease in blood insulin concentrations was evident at 24 hours after dosing, changes in the corresponding AUC of insulin were not significant. Only 4 time points were used to calculate the AUC of insulin at each OGT, increasing the risk of a type‐2 error. In future studies evaluating sirolimus in horses with ID, a demonstrable reduction in not only the peak but also the AUC of blood insulin concentrations after carbohydrate challenge will be important to establish a therapeutic effect relevant to HAL prevention. An effect of sirolimus on insulin dynamics was not evident at 4 hours after administration but was evident by 24 hours; suggesting that daily treatment might be required for a clinical effect. It is possible that this effect on insulin response was present earlier, but missed, due to the timing of our OGT. Similarly, an effect was no longer evident by 72 hours; but it is not known exactly when the effect on insulin dynamics was no longer evident.

In conclusion, the results of this study suggest that therapeutic mTOR inhibition using sirolimus has potential for controlling blood insulin in horses. With growing evidence of active mTORC1 signaling in experimental models of HAL, ¹⁷ sirolimus warrants further investigation as a potential therapeutic for horses with ID and HAL.

CONFLICT OF INTEREST DECLARATION

Authors declare no conflict of interest.

OFF‐LABEL ANTIMICROBIAL DECLARATION

Authors declare no off‐label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION

Approved by Animal Ethics of the University of Queensland, AE63763.

HUMAN ETHICS APPROVAL DECLARATION

Authors declare human ethics approval was not needed for this study.

Supporting information

Data S1: Supporting Information

de Tonnerre DJ, Medina Torres CE, Stefanovski D, et al. Effect of sirolimus on insulin dynamics in horses. J Vet Intern Med. 2023;37(2):703‐712. doi: 10.1111/jvim.16650