Methods
This study was performed in accordance with Project Licence PPL30/3156
issued under the Animal (Scientific) Procedures Act 2013 (EU Directive
2010/63/EU) and local ethics committee as part of a larger study
investigating nociceptive withdrawal reflexes and diffuse noxious
inhibitory controls. This study is reported in accordance with the
ARRIVE 2.0 guidelines (du Sert et al., 2020).
Study design
The NC3Rs experimental design assistant (EDA) was used to design the
study. The population PK model experiment was conducted in tandem with a
larger study investigating nociceptive withdrawal reflexes (NWR) and the
influence of anaesthesia on diffuse noxious inhibitory controls (DNIC)
in the rat. This in vivo experimental system in rats replicates
aspects of the human pain pathway (Le Bars et al., 1992, 2001; White et
al., 2018). This study design facilitated the population PK model
generating matched groups of male and female SD and Lewis rats
undergoing nociceptive testing and DNIC studies. The motivation for
undertaking the study was to interrogate the influence of the
anaesthesia per se , which is essential prerequisite although
oftentimes overlooked or ignored. This is particularly important in
electrophysiology experiments and nociception paradigms conducted under
anaesthesia, where the anaesthesia delivered can have a major influence
on the reflexes being studied.
Animals
Sixteen adult (8-12 weeks) Lewis rats, 9 males (308 ± 49 g) and 7
females (222 ± 9 g) and 24 adult (9-12 weeks) SD rats, 8 males (422 ± 41
g) and 16 female (304 ± 15 g) (Charles River Laboratories, Margate, UK)
were used.
Animals were housed in single sex groups of 4, in double layer
ventilated cages, given access to food (Teklad 2018, Harlan) and tap
water ad libitum and maintained on a 12-hour light/dark cycle.
All cages had play tubes, bedding material and chew blocks for
enrichment. All experiments started at 10:00 h each day.
General anaesthesia
The methods for instrumentation of animals used were identical to those
previously described by (White et al., 2017). Rats were anaesthetised
with isoflurane (3% for induction of anaesthesia, 1-1.5% during
surgery) in nitrous oxide/oxygen (2:1) mixture. Lidocaine 2% (Lignol,
Dechra, Shrewsbury, UK) 3 mg kg-1 was infiltrated
subcutaneously prior to skin incision. Using aseptic techniques the left
jugular vein was surgically cannulated using 0.63 mm O.D. polyethylene
tubing (Fisher Scientific, Loughborough, UK) for administration of drugs
and isotonic fluids. The left carotid artery was surgically cannulated
using 1mm O.D. polyethylene tubing (Fisher Scientific, Loughborough, UK)
to monitor arterial blood pressure and for sampling.
Monitoring anaesthesia
The hypnotic characteristics of the anaesthetic were evaluated by
monitoring the paw withdrawal reflex in response to pinch, corneal
reflex in response to light brushing, spontaneous blinking and gross
purposeful movement and cardiopulmonary parameters. Arterial blood
pressure was monitored by an arterial pressure transducer (SensoNor 840;
SensoNor, Horten, Norway) and recorded using a PC running Spike2
software (CED Ltd, Cambridge, UK). Heart rate was recorded via two 25g
needles inserted subcutaneously on the lateral sides of the thoracic
wall. The ECG signal was amplified and used to trigger an instant rate
meter (Neurolog NL253, Digitimer, Welwyn Garden City, UK) and again
recorded using Spike2 software. Respiratory rate and effort were
assessed by observing chest excursion and measuring end tidal carbon
dioxide (CapStar 100, Linton, Diss, UK). Intermittent positive pressure
ventilation (IPPV) was instigated (SAV04, Vetronic, Abbotskerwell, UK)
in the face of hypoventilation to maintain normocapnia.
Infusion of alfaxalone
Infusion regimens of alfaxalone (Alfaxan®, Jurox, Malvern, UK) were
administered to rats using a calibrated syringe driver (SP100iz, WPI,
Hitchin, UK). All animals received a loading dose (1.67 mg
kg-1 for 2.5 minutes) followed by a constant rate
infusion (CRI). For all animals isoflurane and nitrous oxide were
stopped 2.5 minutes after starting the alfaxalone infusion, but oxygen
was supplied throughout the experiment. The male and female Lewis rats
(n=16) were administered a 60-minute CRI (0.75 mg kg-1min-1) followed by a reduced CRI (0.57 mg
kg-1 min-1) for the remainder of the
experiment. The Sprague Dawley males (n=8) received a 0.75 mg
kg-1 min-1 CRI throughout the
experiment. The Sprague Dawley females received a 60-minute CRI (0.57 mg
kg-1 min-1) followed by a reduced
dose (0.42 mg kg-1 min-1) for the
remainder of the experiment.
Sampling
Arterial blood was withdrawn from the carotid cannula into lithium
heparin tubes and placed on ice. Blood samples (200 μl) were collected
at baseline (prior to alfaxalone administration) and at standardised
time points across the alfaxalone infusion period. Arterial blood gases,
biochemistry and haematology parameters (pH, pCO2,
pO2, bicarbonate, sodium, potassium, chloride, calcium,
glucose, lactate and creatinine concentrations) were also measured
(EPOC, Woodley Instrumentation, Bolton, Lancashire, UK). All rats
received an equal volume of balanced electrolyte solution after blood
sampling (Vetivex 11 (Hartmann’s), Dechra, Shrewsbury, UK). Samples were
centrifuged (4000g for 10 minutes) within 30 minutes of
collection. Plasma was harvested and stored at -20°C until determination
of plasma alfaxalone concentration.
At the end of the experiments animals were euthanised by intravenous
injection of pentobarbitone (pentobarbital, Ayrton Saunders Ltd,
Runcorn, UK) followed by cervical dislocation (by a trained individual
as required by UK Home Office regulations).
Sample analyses
Standard quantification (STD) curves for in vivo plasma samples were
generated using authentic alfaxalone standard samples giving
concentrations from 200 ng ml-1 to 40 μg
ml-1 in addition to the use of quality controls (QC).
Spiking solutions for standards and QCs were made from separate accurate
weighing of drug compounds. The methanol standard curve and QCs were
prepared by spiking 10 µL of a known concentration spike solution into a
solution of 40 µL methanol + 100 µL methanol containing 3 µM of
lansoprazole as internal standard + 50 µL of either male or female blank
plasma (Charles River, Margate, UK). Alfaxalone in vivo plasma
samples were prepared by adding 50 µL of the plasma samples + 50 µL
methanol + 100 µL methanol containing 3 µM of lansoprazole as internal
standard.
Samples, standards, and QCs were then vortexed, stored in a freezer at
-20 °C overnight prior to centrifugation at 4000g for 20 minutes at 4
°C. The supernatant was then transferred into LC-MS vials for analysis
and concentration determination. Finally, the STD curves were analysed
at the beginning and end of the run to determine any variation or
deterioration of LC/MS performance. The analytical methods were
validated to ensure suitable precision and accuracy, lower limit of
quantification (LLOQ), linearity, calibration range and selectivity.
The samples were analysed using a Micromass Quattro Premier mass
spectrometer incorporating an Agilent 1100 HPLC. An Ascentis® C18 column
(2.1 × 50 mm, 3 μm) (Sigma, UK) protected by a Phenomenex C18 guard
cartridge (Phenomenex, UK) was used with the following LC conditions:
Solvent A = 10% methanol, 90% water and 0.02% formic acid, Solvent B
= 100% methanol and 0.02% formic acid, flow rate = 0.4 ml/min, column
temperature = 60°C. LC gradient went from 70 % solvent A:30 % solvent
B to 1 % solvent A:99 % solvent B over a 3 minute interval. The MS/MS
method used electrospray positive mode with a 333.2 ˃315.2 and 297.2
transitions for the detection of alfaxalone. The lower limit of
quantification (LLOQ) was 200 ng ml-1. Two separate
LC/MS/MS runs were performed for the male and female samples,
respectively.
Data and Statistical analysis
Pharmacokinetic analyses
Pharmacokinetic analyses were conducted using an IV infusion
compartmental model for (a) individual Lewis rat data using dose per kg
and (b) Lewis and Sprague Dawley population data using total dose. Both
analyses used Phoenix® WinNonlin® version 8.3 software (Certara USA,
Inc., Princeton, NJ). A 2-stage approach was applied to (a) which
firstly involved the estimation of clearance (CL), half-life
(T1/2), mean residence time (MRT) and steady-state
volume of distribution (Vdss) for alfaxalone in each rat. Secondly,
statistical tests were performed on pharmacokinetic parameters to
determine any differences between male and female rats.
Compartmental non-linear mixed effects methods (NLME) models were
applied to (b) using total dose given to each rat to determine whether
body weight influenced PK parameters. Two populations were analysed:
population 1 (28 rats) comprised the Lewis rat data with Sprague Dawley
rat data described by White et al. (2017); population 2 (52 rats)
comprised population 1 plus the additional Sprague Dawley rat data
(Figure 1). Model residual error was based on a mixed ratio error model.
An exponential random effect model was chosen to describe
inter-individual variability i.e. parameter = typical parameter x
exp(eta). Categorical covariates were implemented for sex (male = 0,
female = 1) and strain (Lewis = 0, SD = 1) on the model parameters in a
multiplicative exponential way. A continuous covariate for log of
centralised body weight (LCBW) was applied in a multiplicative way. The
model analysis started from the basic compartmental models without the
covariates. Next, the contribution of the covariates on fixed effects
and correlation on random effects to the PK parameters were assessed by
a reduction in the objective function using stepwise forward inclusion.
Selection of the best model was based on the lowest value of the Akaike
and Bayesian Information Criteria (AIC and BIC), chi-square p-value
based on the likelihood ratio test, visual inspection of the population
predicted concentration versus the observed concentrations and the
resulting conditional weighted residual errors. Finally, the best model
was checked for robustness using a bootstrap resampling method. Monte
Carlo simulations were used to determine a 95% confidence tolerance
interval for the 5th and 95thpercentile of the population.
Statistical Analyses
Statistical tests were performed using GraphPad Prism (GraphPad
Software, La Jolla, CA, USA) version 9. The male and female log
transformed Lewis rat pharmacokinetic parameters were compared using an
unpaired, two tailed Student’s t-test (α = 0.05) and a p value of
< 0.05 was considered significant. Data are reported as mean ±
standard deviation (SD) unless stated otherwise. The male and female
arterial blood pressure or plasma concentration data were compared at
different time points using 2-way ANOVA with post hoc Sidak multiple
comparison test.