Application of testosterone to epitestosterone ratio to horse urine - a complimentary approach to detect the administrations of testosterone and its pro-drugs in Thoroughbred geldings
Marjaana Viljantoa,e, James Scartha, Pamela Hincksa, Lynn Hillyerb, Adam Cawleyc, Craig Suannc, Glenys Nobled, Christopher J. Walkere, Andrew T. Kicmane and Mark C. Parkine
aLGC, Newmarket Road, Fordham, Cambridgeshire, CB7 5WW, UK
bBritish Horseracing Authority, 75 High Holborn, London, WC1V 6LS, UK
cRacing NSW, Level 11, 51 Druitt Street, Sydney, NSW, Australia, 2000
d School of Animal and Veterinary Sciences, Charles Sturt University, Locked Bag 588, WaggaWagga, NSW, Australia, 2678
e Drug Control Centre, Analytical and Environmental Sciences Research Divisions, King’s College London, 150 Stamford Street, London, SE1 9NH, UK
ABSTRACT
Detection of testosterone and/or its pro-drugs in the gelding is currently regulated by the application of an internationalthreshold for urinary testosterone of 20 ng/mL. The use of steroid ratios may provide a useful supplementary approach to aid in differentiating between the administration of these steroids and unusual physiological conditions that may result in atypically high testosterone concentrations. In the current study, an ultra high performance liquid chromatography tandem mass spectrometry (UHPLC-MS/MS) method was developed to quantify testosterone (T) and epitestosterone (E).The method was used to analyse 200post-race urine samples from geldingsin orderto generate the ratios for the reference population. Following statistical analysis of the data, an upper limit of 5 for T:E ratio in geldings is proposed. Samples collected from 15 geldings with atypical urinary testosterone concentrations (> 15 ng/mL) but otherwise normal steroid profile, had T:E ratios within those observed for the reference population.The applicability of an upper T:E ratio to detect an administrationwas demonstrated by the analysis of a selection of incurred samples from testosterone propionate, dehydroepiandrosterone (DHEA) and a mixture of DHEA and pregnenolone (Equi-Bolic®) administrations. These produced testosterone concentrations above the threshold of 20 ng/mL, but also T:E ratios above the proposed limit of 5. In conclusion, consideration of the T:E ratio appears to be a valuable complimentaryaid to evaluate whether an atypical testosterone concentration could be caused by a natural biological outlier as opposed to the administration of these steroids.
Keywords
Endogenous steroids, testosterone:epitestosterone ratio, gelding, urine, threshold.
1. INTRODUCTION
Detection of the administration of anabolic-androgenic steroids, which are identical in structure to those produced endogenously in the horse, can be challenging. Currently, the detection of doping with testosterone and its pro-drugs is regulated through the use of international thresholds in urine and plasma [1] and by the detection of synthetic steroid esters in plasma and hair [2]–[5].
The current international urinary threshold of 20 ng/mL for free and conjugated testosterone in geldings was based on a population study conducted in Hong Kong for post-race (n=110) and training (n=382) urine samples [6]. The proposed threshold was supported byinternational post-race population data from the UK (n=105) and France (n=117) [7] and a further population study by Hong Kong (n=2012) [8]. This threshold has been supported to be appropriate for regulatory purposes by the results of studies involving the intramuscular (IM) administrations of 500 mg of testosterone [9],a mix of testosterone esters (100 mg of testosterone propionate, 200 mg of testosterone isocaproate and 200 mg of testosterone phenylpropionate)[6]and oral administrations of 1 mg/kg of testosteronepro-drugs; dehydroepiandrosterone (DHEA) and androstenedione [10].
In recent years, an apparent increase in the incidence of atypically high testosterone concentrations in Thoroughbred gelding post-race urine samples has been observed in the UK and Australia. A number of these samples breached the testosterone threshold of 20 ng/mL, but the overall steroid profile was not indicative of testosterone administration since it should not cause increases in epitestosterone and DHEA concentrations. Consideration of two other causative factors that could have elevated urinary testosterone concentrations were also ruled out,these being dehydration[11]or cryptorchidism, where the detection of estr-5(10)-ene-3β,17α-diol would be expected [12]. As a result these samples were categorised as ‘atypical’ and other explanations for their presence were explored. A reasonable hypothesis is that the atypical results have arisen due to considerable adrenal stimulation markedly raising the urinary testosterone production rate, with these geldings having experienced more physiological ‘stress’ compared to others.
The adrenalcortex is likely to be the main contributor to testosterone and epitestosterone production in the gelding.However, the proportion of testosterone and/or epitestosterone that arise fromperipheral conversion of precursor steroids secreted as opposed todirect secretion is currently unknown.Artificial stimulation with synthetic adrenocorticotrophic hormone (tetracosactide) results in an increase in the urinary concentration of testosterone, and also DHEA and theisomers of androst-5-ene-3,17-diols [13].Natural causes may also disturb the pituitary-adrenal axis, such as physiological stress caused by transportation, exercise and disease. Increased plasma/saliva hydrocortisone (cortisol) concentrations have been observed in horses transported for short (< 300 km) [14] and long (< 1370 km) distances [15], [16], and in horses affected by diseases, such as laminitis and acute abdominal syndrome [17]. The increase in adrenal steroidogenesiscould also result in increased testosterone and epitestosterone concentrations in horses, especially since exercise induced stress has been shown to have an augmentative effect on urinary/plasma hydrocortisone concentrations [18]–[20], as well as on plasma testosterone concentrations [21]. Even though this evidence points tothe aetiology of atypical urinary concentrations to be due to unusually raised adrenal steroidogenesis, the possibility of a rare polymorphism(s) in the metabolism of testosterone cannot be discounted.
Steroid ratios have been used historically to monitor the use of testosterone and its pro-drugs in geldings and mares/fillies. Prior to the current concentration threshold of 20 ng/mL in geldings, a testosterone to DHEA (T:DHEA) ratio greater than 5 was proposed for both geldings and mares/fillies[22]. This was established using a population of 153 geldings and 65 mares/fillies, and shown to be exceeded by the IM administration of 50 mg of testosteronephenylpropionate. Furthermore, a T:E ratio of 12wasadopted as a threshold for the detection of testosterone in mares/fillies[7]. Thisthreshold was based on data collected from the analysis of 294 post-race urine samples and also following the IM dose of 1,000 mg of testosterone hexahydrobenzoate[23].Thecurrent international concentration based threshold of 55 ng/mLin filliesreplaced the T:E ratio of 12 in 2000following a study of 1531 animals in six international laboratories[24].
Historically, the detection and quantification of low concentrations of endogenous anabolic steroids (low ng/mL to pg/mL amounts) has been analytically challenging, such as that for testosterone and epitestosterone in the urine from the gelding. However, more sensitive liquid chromatography tandem mass spectrometry (LC-MS/MS) methods have been developed since the aforementioned studieswere conducted using gas chromatography mass spectrometry (GC-MS). Chemical derivatisation can also be used to further improve the electrospray ionisation of neutral steroids, pertinent to the analysis of testosterone and epitestosterone in equine plasma and urine[2], [25]–[27].
The aim of the current study was to determine whether a urinary T:E ratio could be used as a supplementary approach to help differentiateatypical testosterone concentrations in geldings from scenarios involving administration of testosterone or its prodrugs. This required a development of a sensitive and robust analytical method incorporating LC-MS/MS and chemical derivatisation for the quantitative analysis of testosterone and epitestosterone in equine urine. Extraction of a total fraction (free and conjugated) was required since testosterone is mainly excreted as a sulphate conjugate and epitestosterone as a glucuronide conjugate in the horse [28]. Furthermore, the current internationally approvedthreshold of 20 ng/mL testosterone in gelding urine is based on determination of the total fraction [1],so defining the reference range of T:E values using the same approach would then allow for meaningful use of the resulting data in relation to routine anti-doping processes. The resulting population data would be used to statistically determine an upper limit for a T:E ratio using a normal post-race sample size of 200 geldings. The proposed upper T:Eratio limit was applied and evaluated for itsabilityto detect the administrations of testosterone propionate, DHEA and a mixture of DHEA and pregnenolone (Equi-Bolic®) and to determine whether the observed atypical samples could be biological outliers.
2. EXPERIMENTAL
2.1 Reagents and standards
Methoxylamine hydrochloride, sodium phosphate, acetyl chloride and pancreatin (8 x USP specifications) were obtained from Sigma-Aldrich (UK) and methanol, ethyl acetate, sodium hydroxide and potassium dihydrogen orthophosphate were from Fisher Scientific Ltd (UK). Laboratory water was purified using a Triple Red Duo Water system (Triple Red Laboratory Technology, Bucks, UK).
Testosterone sulphate, epitestosterone sulphate, 16,16,17α-d3-testosterone sulphateand 2,2,4,4-d4-androsterone glucuronide were purchased from the Australian Government National Measurement Institute (Australia).
Stock solutions containing individual compounds at a concentration of 1 mg/mL were prepared in methanol, and stored at – 20 °C. Mixed stock solutions for analytes and deuterated internal markers were prepared in methanol at a concentration of 10 µg/mL, and they were subsequently diluted with methanol in order to obtain separate spiking solutions at appropriate concentrations.
2.2 Equine urine samples
2.2.1 Population samples
The referencepopulationconsisted of 200anonymised urine samples collected post-race from Thoroughbred geldings at different race meetings in the UK during January (n=63), April (n=65) and September (n=72) 2015as part of the anti-doping programmecarried out by the British Horseracing Authority (BHA). Samples were stored at – 20 ºC prior to analysis.
A further 15 urine samples collected post-race over a longer time framefrom Thoroughbred geldings known to haveatypically high concentrations of testosterone (≥15 ng/mL), but notably otherwise having an unremarkable steroid profile, were also analysed and compared with the reference population.
2.2.2 Administration samples
A selection of pre- and post-administration urine samples were analysed following the administrations of testosterone propionate, DHEA and a mix of DHEA and pregnenolone to Thoroughbred geldings. The UK studies were conducted under Licence from the Home Office according to the framework of the Animal Scientific Procedures Act (1986) with statutory ethical review, whilst the Australian study was conducted following the approval of Charles Sturt University (CSU) Animal Care and Ethics Committee.
Testosterone propionate (Testoprop®, 50 mg/mL by Jurox) was administered intramuscularly in five consecutive weekly doses of 50 mg to two Thoroughbred geldings at the BHA’s Centre for Racehorse Studies (CRS), UK. The horseswere 7 and 10 years old, and their bodyweights were 482 and 469 kg, respectively. Urine samples were collected pre-administration and at 6 and 173 hours following the final dose for horse 1 and horse 2, respectively. The samples were stored at -20 ºC prior to analysis.
DHEA (by BOVA Compounding Pharmacy) was administered at a dose of 1 mg/kg in a capsule via naso-gastric tube to six Thoroughbred geldings atCSU School of Animal and Veterinary Sciences, Australia. The ages of the animals ranged from 2 to 15 years old, and the mean bodyweight was 575.3 ± 60.1 kg. Urine samples were collected pre-administration and up to 96 hourspost-administration for each horse. The samples were stored at -20 ºC at Racing NSW, Australia and they were delivered in dry ice to LGC, UK for further analysis.
A supplement containing a mix of DHEA and pregnenolone (Equi-Bolic, purportedly 500 mg/15mL each) was administered orally in five consecutive daily doses of 500 mg to two Thoroughbred geldings at the BHA’s CRS. The animals were 8 and 9 years old, and they weighed 510 and 528 kg, respectively.Urine samples were collectedpre-administration and 7 and 4 hours following one dose for horse 1 and horse 2, respectively. The samples were stored at -20 ºC prior to analysis.
2.3Sample extraction
1.5 mLaliquots of equine urine were diluted with 3.6 mL of phosphate buffer (1 M, pH 6.3) containing d3-testosterone sulphate at a concentration of 28.2 ng/mL and 100 µL of pancreatin solution. Aliquots were also spiked with 75 µL of d4-androsterone glucuronide at a concentration of 1 µg/mL. Solid phase extraction (SPE) was performed using Waters Oasis® WCX cartridges (3 mL, 60 mg), a mixed mode sorbent consisting of a polymeric reversed-phase incorporating a weak ion exchanger. This phase was used in the clean-up of neutral steroids to remove potential cationic interferences, whilst the alkaline washes were used to remove potential anionic interferences. The phase was conditioned with 2 mL of methanol followed by 2 mL of purified water prior to sample loading. Cartridges were washed with 2 mL of sodium phosphate buffer (0.25 M, pH 8.3) followed by 2 mL of purified water prior to eluting with 3 mL of methanol. Aliquots were dried and subsequently hydrolysed by adding 0.5 mL of anhydrous methanolic hydrochloride (1 M) and incubating at 60 ºC for 15 minutes[29]. The reaction was quenched with 3 mL of sodium phosphate buffer (0.25 M, pH 8.3) prior to another SPE step. Again, WCX cartridges were conditioned with 2 mL of methanol followed by 2 mL of purified water prior to sample loading and they were washed with 2 mL of sodium hydroxide (0.1 M) and 2 mL of purified water. Elution was carried out with 3 mL of ethyl acetate and aliquots were dried prior to reconstitution with 200 µL of anhydrous ethyl acetate. 25 µL portions of samples were transferred to glass HPLC vials and they were dried. To enhance analytical sensitivity, a derivatisation step was then performed, converting the oxo function on ketonic steroids, such as testosterone and epitestosterone to their methyloxime derivatives[2], [25]. 200 µL of methoxylamine hydrochloride solution in 80 % methanol in purified water (0.1 M) was added to each vial and then incubated at 80 ºC for an hour. 2 µL of each extract was injected into the LC-MS/MS system.
2.4LC-MS/MS analysis
Sample analysis was performed on a LC-MS/MS system consisting of a Waters Acquity I-Class UPLC coupled with a Waters Xevo TQ-S triple quadrupole mass spectrometer in positive electrospray ionisations mode at a capillary voltage of 2.0 kV, a source temperature of 150 °C and a desolvation gas temperature at 550 °C. Collision gas was argon at a flow rate of 0.15 mL/min. Analysis was carried out in selected reaction mode (SRM), and selected transitions, cone voltages and collision energies are shown in Table 1.Data was processed using TargetLynx software.The peak areas of the analytes and internal marker were recorded and used to calculate testosterone and epitestosterone concentrations and their ratios.
Chromatographic separation was achieved on an Acquity BEH C18 (100 mm x 2.1 mm, 1.7 µm) reversed phase UPLC column using 0.1 % formic acid in methanol (A) and 0.1 % formic acid in water (B) as mobile phases. Gradient was operated at 60 ºC and at a flow rate of 0.4 mL/min. It was started at 20 % organic for 0.5 minutes followed by increases to 60 % A at 1 minute, 80 % A at 6 minutes and 99.9 % at 6.3 minutes. 99.9 % A was held for1 minute before resuming the initial conditions and re-equilibrating for 1 minute. A total run time was 8.5 minutes.
2.5Method validation
A matrix ‘standard addition’ approach was used to prepare calibration lines and quality control (QC) samples for each precision and accuracy batches, since both testosterone and epitestosterone are endogenous to all genders in horses[11], [30]. Pooled gelding urine was spiked at known concentrations of the analytes and the slope and intercept of the generated calibration line were used to calculate their endogenous concentrations in pooled urine. Subsequently, the calibration line and QC samples were adjusted to account for that endogenous concentration.
The method was quantitatively validated using measures of linearity, intra- and inter-batch precision and accuracy, specificity, selectivity and sensitivity (adhering to unpublished European Horserace Scientific Liaison Committee quantitative method validation guidelines). Three separate precision and accuracy batches were extracted and analysed and each of them contained a calibration line in duplicateat concentrations of endogenous (E) only, E+2, E+5, E+10, E+20, E+50, E+100, E+200 and E+500 ng/mL and six replicates of QC samples at concentrations of E, E+2, E+20, E+200 and E+400 ng/mL. Sample dilution was also investigated by spiking pooled gelding urine in six replicates at a concentration of 2 µg/mL and performing 1-in-10 dilution with pooled gelding urine to enable its quantification in the validated range.
Cross-talk between testosterone, epitestosterone and d3-testosterone were investigated by spiking pooled urine in duplicate with each standard and monitoring the changes in peak abundancesat the respective retention times of the other analytes. Furthermore, the selectivity of the method was assured for each analyte through the monitoring of quantifier and qualifier ion ratios for both endogenous and augmented analyte concentrations.
2.6 Sample analysis
Following the validation, the method was used to analyse 200 post-race gelding urine samples, 15 gelding urine samples with atypicaltestosterone concentrations and a selection of administration samples. These samples were analysed alongside a calibration line and QC samples in duplicate at concentrations of E+3 (low QC), E+250 (medium QC) and E+400 ng/mL (high QC). A matrix ‘standard addition’ approach was again used to calculate the endogenous concentrations for the calibrators, whilst the endogenous concentrations generated in the validation were used for the QC samples to ensureacceptableinter-batch performance.
2.7Statistical analysis of the population data
Statistical analysis of the data was performed using SPSS Statistics software (version 22). The Kolmogorov-Smirnov test for normality was used to determine whether the population data were Gaussian.
3. RESULTS AND DISCUSSION
3.1 Method validation
Chromatographic separation was achieved for testosterone and epitestosterone (Figure 1). Production of methoxylamine derivatives resulted in formation of two geometric isomers for both testosterone and epitestosterone due to their 3-keto-Δ4 structure [31].Ideally from a computation point of view, either no separation or alternatively complete separation of the syn and anti isomers of the methoxylamine derivatives of testosterone and epitestosterone is desirable, even though the latter is unlikely to be ever achievable given the number of theoretical plates associated with the UPLC column employed. Whereas the two forms of derivatisedepitestosterone were not resolved by liquid chromatography using the selected gradient, there was partial resolution for that of derivatised testosterone. Even so, satisfactory results for method validation were achieved for both steroids, as described below, based on the use of peak area data.
Method selectivity was assuredby monitoring the ion ratios of quantifier (m/z 318>126) and qualifier (m/z 318>138) MS/MS transitions for both testosterone and epitestosterone. These product ions (Figure 2) are in agreement to the fragmentation patterns previously observed for oxime derivatives [25].No cross-talkwas observed for the urine samples spiked at a concentration of 400 ng/mL between testosterone, epitestosteroneand d3-testosterone.
Method selectivitywas further demonstrated by the analysis of 18 non-spiked urine samples from different genders (colt, gelding and mare/filly) and no interfering matrix peaks were observed at the respective retention times of the analytesusing the quantifying transition (m/z 318>126).Both testosterone and epitestosterone were detected in all samples analysed due to their endogenous nature.