Stanozolol-N-glucuronide metabolites in human urine samples as suitable targets in terms of routine anti-doping analysis

The exogenous anabolic-androgenic steroid (AAS) stanozolol stays one of the most detected substances in professional sports. Its detection is a fundamental part of doping analysis, and the analysis of this steroid has been intensively investigated for a long time. This contribution to the detection of stanozolol doping describes for the first time the unambiguous proof for the existence of 17-epistanozolol-1 0 N-glucuronide and 17-epistanozolol-2 0 N-glucuronide in stanozolol-positive human urine samples due to the access to high-quality reference standards. Examination of excretion study samples shows large detection windows for the phase-II metabolites stanozolol-1 0 N-glucuronide and 17-epistanozolol-1 0 N-glucuronide up to 12 days and respectively up to almost 28 days. In addition, we present appropriate validation parameters for the analysis of these metabolites using a fully automatic method online solid-phase extraction (SPE) method already published before. Limits of identification (LOIs) as low as 100 pg/ml and other validation parameters like accuracy, precision, sensitivity, robustness, and linearity are given.


| INTRODUCTION
The family of anabolic-androgenic steroids (AAS) belongs to one of the most common illicitly used substance class in the world of professional sports. Within this large group of different drugs, the synthetic steroid stanozolol (17α-methyl-5α-androst-2-eno [3,2-c]pyrazol-17βol) attributes to the highest number of positive cases according to World Anti-Doping Agencies (WADA) statistics. 1,2 This exogenous steroid is well known analytically and various strategies for its detection are described in the literature. Because this steroid was synthesized in the late 1950s, there was plenty of time to develop many different approaches to analyze stanozolol and its metabolites. 3 In 1986, the team around Donike and Schänzer developed the first method for the analysis of the metabolite 3 0 -OH-stanozolol applying gas chromatography-mass spectrometry (GC-MS). 4 In the following 35 years, many other techniques, primarily based on mass spectrometric techniques coupled to on either gas (GC-MS) or liquid chromatography (LC-MS), for analyzing a large number of different stanozolol metabolites, were published. [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20] In general, the traditional approach for the simultaneous analysis of several different steroids is to perform enzymatic hydrolysis to cleave highly polar phase-II conjugates, like glucuronic acids and sulfates, followed by liquid-liquid extraction and the analysis of remaining phase-I metabolites and parent molecules with GC-or LC-MS. 21,22 For the measurement with GC-MS, the analytes are additionally derivatized with N-methyl-N-trimethylsilyltrifluoroacetamide (MSTFA) to reduce their polarity. This kind of approach is the gold standard nowadays and is commonly performed by anti-doping laboratories worldwide for the routine initial testing procedure (ITP), often including the detection of stanozolol parent or phase-I metabolites.
However, with the emergence of more powerful LC-MS devices, a new, modern way of steroid analysis was developed. With this approach, time-and resource-consuming steps of enzymatic hydrolysis, extraction and derivatization are omitted. Phase-II conjugates of steroids are analyzed directly without further extraction or concentration steps. [15][16][17][18][19][20][23][24][25][26][27][28][29] In 2015, the team around G. Balcells already proposed the analysis of a high number of relevant phase-II metabolites for anti-doping screening purposes. 16 Nowadays, high-resolution (HR) LC-MS devices are frequently used in order to increase sensitivity and selectivity of the measurement. In 2013, Van Eenoo et al.
showed the promising potential of this approach for the detection of stanozolol abuse for the first time. 17 The team developed an approach for the direct analysis of 3 0 -OH-stanozolol glucuronide in human urine.
This idea was adopted by developing a simple but powerful method for the detection of phase-II metabolites of steroids, as previously published. This approach was optimized by placing a fully automated online solid-phase extraction (SPE) procedure upstream of the analytical measurement with LC-HRMS. 18 Next to the aspect of saving time and resources by direct analysis of phase-II conjugates, no enzymatic hydrolysis step using, for example, β-glucuronidase from Escherichia coli is required. Consequently, issues like incomplete or inhibited hydrolysis to yield phase-I metabolites, as necessary for GC-MS methods, are no longer relevant. Literature and own experience demonstrates that, for example, stanozolol-N-glucuronides are hardly hydrolyzed with enzymes commonly used in anti-doping laboratories. 19 As a consequence, these metabolites are usually not detected in routine ITP at all.
We have observed that the excretion profile for stanozolol-Nglucuronides is consistent in most positive samples, depending on the drug's application time.  Stock solutions with a concentration of 1 μg/ml for IS and standard substances were prepared by dissolving 1 μg of standard substance in 1-ml MeOH. Standard working solutions were prepared by diluting stock solutions with MeOH. Until use, solutions were stored at À20 C. Reference samples were prepared by adding working solutions directly to blank urine.

| Urine samples
According to WADA's collection guidelines, all positive urine samples used in this project were collected by accredited sample collection authorities. 30 The samples have previously been analyzed by the accredited anti-doping laboratory Seibersdorf Labor GmbH.
All samples are unanimously confirmed positive for stanozolol. The samples were subsequently anonymized and approved for research.
Previously, the athletes gave permission to use the urine samples for research purposes, according to the International Standard for Laboratories (ISL). 31 Samples used for the excretion study were provided by the accredited anti-doping laboratory Cologne, Institute of Biochemistry-German Sport University Cologne, Germany. For these samples, a male healthy volunteer received a single oral dose of 5 mg of stanozolol (Winstrol ® ). Urine samples were then collected up to 28 days after administration of the substance. A written agreement was received from the participant and the project was accepted by the local ethical committee. 19 The anonymized blank urine samples were provided from healthy female and male volunteers. Until analysis, all urine samples were stored at À20 C. An online-SPE-LC-HRMS approach was chosen as analytical method.

| Sample preparation
The method is described in detail in a previous publication. 18 Analytes extraction is carried out fully automatically upstream the injection into the Vanquish Horizon UHPLC+ system (Thermo Fisher, Austin, Texas, USA). An Accucore Phenyl-Hexyl, 10 Â 3-mm column with 2.6-μm particle and 80-Å pore size (Fischer Scientific, Loughborough, UK) was used as extraction column. As analytical column, a Kinetex EVO C-18, 100 Â 2.1-mm column with 2.6-μm particle-and 100-Å pore size  Table 1 with an ion extraction range of 5 ppm were used. Isolation windows were set to ±1 m/z. Collision energies (CEs) were optimized by injection of methanolic working solutions of reference substances.
Diagnostic ions and corresponding CEs are also shown in Table 1. The software Thermo Xcalibur Qual Browser 4.1.45 was used for data procession and calculation of monoisotopic masses. All systems were supervised with Xcalibur 4.0 (Thermo Fischer).

| Method validation
Method validation parameters for qualitative and semi-quantitative purposes were used according to the ISL. The following parameters were acquired: specificity, precision, robustness, linearity, accuracy, matrix effects, carryover and limit of identification (LOI). Detailed descriptions of all parameters are given below. Method validation was carried out by using the above described PRM method. Peak areas gained from product ion 1 were used for all semi-quantitative parameters. Concentrations were corrected with the IS and calculated with an internal calibration curve measured in each sequence. Data processing used the software Thermo Xcalibur Quan Browser 4.1.45 and parameters were calculated with Microsoft Excel 2010. The minimum required performance level (MRPL) for free stanozolol is 2 ng/ml, as defined in the WADA Technical Document TD2019MRPL. 32 Therefore, 50% of MRPL, 1 ng/ml, were used for most validation parameters. According to the WADA identification criteria, comparison of retention times and ratios of relative abundances of two ion transitions were used to evaluate the specificity, robustness and LOI. 33 For comparisons, matrix-free (MQ) samples were spiked with reference substances at the respective concentrations.

| Specificity
Five different female and five different male blank urine samples from healthy volunteers were analyzed (n = 10). Furthermore, a second set of these 10 samples were spiked with 1-ng/ml standard working solution. Relative abundances (peak area) of two ion transitions and retention times were compared in order to verify the absence of interferences for both diagnostic ions.

| Precision
Three sets of 10 replicates of blank urine samples were spiked with standard working solution at three different concentrations, low 1 ng/ml, medium 10 ng/ml, and high 50 ng/ml (n = 3 Â 10) and were analyzed. Coefficient of variation (CV) of areas (normalized with IS) for intra-and inter-day precision for three concentration levels was calculated by measuring samples on three consecutive days.

| Accuracy
Three sets of 10 replicates of blank urine samples were spiked with standard working solution at three different concentrations, low 1 ng/ml, medium 10 ng/ml, and high 50 ng/ml (n = 3 Â 10) and were measured. Accuracy (determined concentration/nominal concentration*100%) was calculated.

| Matrix effects
Six different blank urine samples and one matrix-free sample (MQ) were spiked with 1-ng/ml standard working solution and measured. Average matrix effects (ion suppression or enhancement) were calculated by comparing signal area (normalized with IS) of urine samples to the matrix-free sample.

| Carryover
Blank urine sample was spiked with 400-ng/ml standard working solution and measured directly prior to a blank urine sample. The intensity of signal area (normalized with IS) in the blank sample was calculated (%).

| Limit of identification
Three sets of three different blank urine samples were spiked with standard working solution at three concentrations (0.05, 0.075, and 0.1 ng/ml, n = 3 Â 3), close to an estimated LOI and were analyzed.
According to WADA specifications, LOI was defined as the lowest concentration level at which the analytical signal meets the regulations for relative abundance and retention times. The acronym LOI, used by WADA, is coequal with the more known term limit of detection (LOD).

| Method validation
The method validation parameters of the 17-epistanozolol-N-glucuronides are quite similar to the values observed for stanozolol-N-glucuronides in our previous work. 18 In Table 2  Abbreviations: c, concentration; CV, coefficient of variation; LOI, limit of identification; ME, matrix effects; n, number of samples; pr., precision; RSD, relative standard deviation. product ion 2 (m/z 81) showed a disproportionately increased abundance compared to product ion 1 (m/z 329), leading to a bigger area ratio than a reference sample without matrix and with smaller injection volume. No carryover effect at all was observed after injection of a high concentration sample. Probably due to the lack of comprehensive sample preparation, high matrix effects (177% and 184%) were observed, which, however, do not seem to have a negative influence on precision and accuracy of the method. Nevertheless, for pure quantitative measurements a matching deuterated IS is recommended. Fulfilling WADAs identification criteria, we could detect both 17-epistanozolol-1 0 N-glucuronide and 17-epistanozolol-2 0 Nglucuronide at the lowest concentration of 100 pg/ml. By applying alternative criteria for the calculation of the LOI, for example, a signal/ noise ratio of >3, the LOIs would be even lower (50 pg/ml). These suitable validation parameters promise a reliable use of this method for the confirmation of stanozolol doping in routine anti-doping analysis.

| Identification of 17-epistanozolol-Nglucuronides
In order to identify the two metabolites in question, 17-epistanozolol-  Comparing retention times and at least two MS/MS transitions of the targeted analyte in a positive sample and a reference sample is required to fulfill WADA identification criteria. The relative abundance of diagnostic ions can be determined from peak areas or heights. In this work, peak areas were used. Table 3 shows the comparative calculations of retention times and abundances, as well as the criteria to be met.
With 0.2% difference for 17-epistanozolol-1 0 N-glucuronide and 0.1% for 17-epistanozolol-2 0 N-glucuronide, for both metabolites, the relative differences of retention times were significantly below the maximum tolerance of 1%. Furthermore, the relative area abundances' differences were 0.6% and 0.8%, which is also far below the tolerated 5% aberrance. These data provide the unequivocal proof of the existence of 17-epistanozolol-1 0 N-glucuronide and 17-epistanozolol-2 0 N-glucuronide in human urine after ingestion of the exogenous steroid stanozolol.  of stanozolol. 15,16 However, due to the lack of proper reference substances, in all cases, metabolite elimination data were presented based on relative signal intensities rather than metabolite concentrations.  Furthermore, the direct analysis of glucuronide metabolites delivers promising results for many other substances, too.

| Excretion study
Therefore, consideration should be given to complementing the usual ITP with an approach involving the direct analysis of glucuronide metabolites of doping substances without the use of glucuronidase.
Direct analysis of steroid phase-II metabolites is deemed to bring many advantages to the field of anti-doping analysis. Therefore, the characterization of new unknown metabolites and the subsequent production of reference substances should stay in focus of current research.