Introduction

Trenbolone (Tren) belongs to the class of synthetic anabolic-androgenic steroids (AAS) and is structurally characterized by a 4,9,11-triene-3-one structure composing a highly conjugated π-electron system (Figure 1). The significant anabolic properties of Tren resulting in increased muscle size and strength have generated an incentive for illicit applications including doping and livestock breeding. In sports, the use of trenbolone has been prohibited by the World Anti-Doping Agency (WADA) at all times, categorized under S1 1. (anabolic androgenic steroids) in the Prohibited List (WADA, 2020). According to WADA’s annual statistics, anabolic agents are the most frequently misused substance group in sports with a total of 1,823 adverse analytical findings (AAFs) in 2018. Within this group, Tren occurrences account for 6% (WADA, 2019). The statistics however can only reflect cases of detectable Tren and does not conclusively address the question whether Tren is less favored by users of AAS or if available detection strategies do not offer the required analytical retrospectivity.

For that reason the objective of this project was to re-investigate the metabolism of Tren in order to probe for metabolic products potentially supporting the extension of the detection window. The number of previously reported Tren metabolites is scarce, and for doping control purpose the analysis focuses at present on the main human urinary metabolites epitrenbolone (EpiTREN), epitrenbolone glucuronide (EpiTREN Glu), and trenbolone glucuronide (TREN Glu) (De Boer et al., 1991; Schänzer, 1996). Regarding the detection windows, two data sets have been published spanning from approximately 3 days (Spranger and Metzler, 1991) to 32 days (Sobolevsky and Rodchenkov, 2015). Besides glucuronides, sulfates (Rzeppa et al., 2015), and cysteine conjugates (Sobolevsky and Rodchenkov, 2015) of Tren and Epitren were reported. Cysteine conjugates are produced by phase-II metabolism, where the tripeptide glutathione is covalently bound via the sulfur atom by glutathione transferase and, subsequently, glutamate and glycine are eliminated. In general, cysteine conjugates are hydrolyzed employing alkaline conditions (Blair, 2006; Fabregat et al., 2010, 2011, 2013; Pozo et al., 2010), but the cysteine conjugate of trenbolone described by Sobolevsky and Rodchenkov (2015). was found to be stable during alkaline hydrolysis and was analyzed by high performance liquid chromatography electrospray ionization tandem mass spectrometry (HPLC-ESI-MS/MS) as in-source fragment in ESI negative mode and as intact conjugate in ESI positive mode (Sobolevsky and Rodchenkov, 2015). During in-vitro studies, several monohydroxylated metabolites, and trenbolone-diketone were generated (Metzler and Pfeiffer, 2001; Kuuranne et al., 2008).

Nowadays, LC-MS-based methods are commonly used for the analysis of Tren and its metabolites (Thevis et al., 2005a, 2009; Tudela et al., 2015) as GC-MS-based methods were found to be of limited utility due to derivatization artifacts and low thermal stability of the target analytes (De Boer et al., 1991; Ayotte et al., 1996; Casademont et al., 1996; Marques et al., 2007; Brun et al., 2011).

For systematic metabolism studies, a method for metabolite identification using hydrogen isotope ratio mass spectrometry was developed and successfully applied for the first time in 2013 (Thevis et al., 2013). The fundamental principle is analogous to metabolism studies using radioactively labeled compounds (Sano et al., 1976). The compounds can be detected selectively because of their isotopic labeling by measuring the radioactivity in case of tritium or ¹⁴C labeled compounds or by measuring the hydrogen isotope ratios in case of deuterium labeled compounds.

Hydrogen isotope ratios are determined by gas chromatography/thermal conversion/isotope ratio mass spectrometry (GC-TC-IRMS). The organic compounds are converted under reducing conditions to CO and N₂ as well as molecular hydrogen (H₂), respectively the deuterated isotopologe HD. After ionization, detection is accomplished using m/z 2 for H+2H2+ and m/z 3 for HD⁺ by Faraday cups with different amplification factors (factor 1000 difference). Since the natural hydrogen abundance amounts on average to 99.985% for H and 0.015% for D (Dunn and Carter, 2018), comparable signals for H+2H2+ and HD⁺ are obtained for samples at natural abundance, while deuterated compounds lead to a significant increase of signals at m/z 3. Compounds resulting in diagnostic HD⁺ signals are subsequently comprehensively characterized by gas chromatography/electron ionization/high accuracy/high resolution mass spectrometry (GC-EI-HRMS). This concept has been proven in several studies (Thevis et al., 2013; Piper et al., 2016a,b, 2018, 2019).

Within this project, liquid chromatography/electrospray ionization/high accuracy/high resolution mass spectrometry (LC-ESI-HRMS) was applied for further characterization of trenbolone and its metabolites after GC-TC-IRMS analysis. Twenty metabolites were identified with a detectability of up to 6 days. Four metabolites exhibiting the longest detection windows were characterized by parallel reaction monitoring (PRM) experiments and comparison to reference material.

Materials and Methods

Chemicals and Steroids

Trenbolone reference material and the internal standard 2,2,4,6,6-d₅-trenbolone were purchased from Toronto Research chemicals (Toronto, Canada), and epitrenbolone from the National Measurement Institute (Sydney, Australia). Steroid reference material for HPLC separation including ETIO (etiocholanolone), A (androsterone), T (testosterone), and PD (pregnanediol) was supplied by Sigma-Aldrich, and 11 K (11-ketoetiocholanolone), 5a (5α-androstanediol), and 5b (5β-androstanediol) were obtained from Steraloids (Newport, RI). Chromabond C18 solid-phase extraction (SPE) cartridges (500 mg, 6 mL) were purchased from Macherey & Nagel (Düren, Germany) and β-glucuronidase from Escherichia coli (140 U/mL) from Roche Diagnostics (Mannheim, Germany). Ultrapure water was prepared by a Barnstead™ GenPure™ xCAD Plus system (Thermo, Germany). All reagents and solvents were of analytical grade. Acetonitrile (ACN), formic acid (FA), methanol (MeOH), tert-butyl methyl ether (TBME), cyclohexane, pyridine, sodium hydroxide (NaOH), sulfuric acid (H₂SO₄), glacial acetic acid, and potassium tri-sec-butylborohydride (1 M in THF) were provided by Merck (Darmstadt, Germany). Acetic anhydride was supplied by Sigma Aldrich (Taufkirchen, Germany) and tetrahydrofuran (THF) by VWR (Darmstadt, Germany). N-methyl-N-trimethylsilyltrifluoroacetamide (MSTFA) was purchased from Chemische Fabrik Karl Bucher (Waldstetten, Germany).

Excretion Study

An excretion study was conducted in order to re-investigate the trenbolone metabolism. Following written informed consent, 10 mg of 5-fold deuterated trenbolone (Figure 1) dissolved in ultra-pure water/EtOH (80:20, v/v) was orally administered to one healthy male volunteer (43 years, 84 kg) who declared not to have used any medication or nutritional supplements during this study and for a wash-out period of at minimum 3 month (any compounds), respectively 6 month (deuterated compounds) before the study. Three blank samples were collected pre-administration, and post administration samples were collected for up to 30 days. During the first 48 h after trenbolone intake, every urine was collected. From day three until day five, two to three urine samples per day were collected, and afterwards only the first morning urine was sampled until the end of the study. The administration study was approved by the Ethics Committee of the National Institute of Sports of Romania (Bucharest, Romania, #2283, 2016).

Analysis of Hydrolyzed Steroids

Sample Preparation

An extensive sample preparation was required in order to reach adequate purity and undecomposed volatility of the metabolites for GC-TC-IMRS analysis. Urine samples were prepared according to established protocols for isotope ratio analysis of steroids. Here, every sample is separated into four main fractions: unconjugated steroids, glucuronic acid conjugates (Piper et al., 2008), sulfo-conjugates (Piper et al., 2010), and cysteine conjugates (Fabregat et al., 2010, 2011, 2013; Pozo et al., 2010). Subsequently, the fractions of glucuronides and sulfates are further divided by HPLC into seven sub-fractions (Thevis et al., 2013). Sample preparation and analysis is summarized in Figure 2.

For each sample, a volume of 20 mL urine was required. First, two C-18 SPE cartridges per sample were pre-conditioned with 2 mL of MeOH and subsequently washed with 2 mL of water. Then, every sample was split and 10 mL of urine were applied to each cartridge, which was subsequently washed with 2 mL of water, and finally eluted with 2 mL of MeOH. Both eluates of each sample were combined and evaporated to dryness under a gentle stream of compressed air at 50°C. Following evaporation, samples were reconstituted with 1 mL of an aqueous 0.2 M sodium phosphate buffer at pH 7, and a liquid-liquid-extraction (LLE) step with 5 mL of TBME was performed. Therefore, the mixture was shaken for 5 min and subsequently centrifuged for 5 min at 600 g before separating both layers. The organic layer yielded the fraction of the unconjugated steroids (fraction f; free).

The remaining aqueous layer was incubated with 100 μL of β-glucuronidase at 50°C for 1 h. To terminate hydrolysis, 500 μL of an aqueous 20% potassium carbonate buffer (pH 10) were added. In order to extract the liberated steroids, a second LLE step with TBME was performed and the resulting organic layer contained the former glucuronic acid conjugates (fraction Gluc). Then, the pH of the aqueous layer was adjusted to 5 with glacial acetic acid, and purified by SPE as described above. After evaporation, samples were incubated with 2.5 mL of EtOAc/MeOH (70/30, v/v) and 1 mL of EtOAc/H₂SO₄ (100 mL/200 ng, v/w) at 50°C for 1 h. Subsequently, 0.5 mL of methanolic NaOH (1 M) were added, the mixture was evaporated as described above, and reconstituted with 5 mL of water. A third LLE with TBME was conducted to extract the formerly sulfo-conjugated steroids (fraction Sulf). This was followed by alkalization of the aqueous layer with 300 μL of 6 M KOH and an incubation step at 60°C for 15 min. Finally, alkaline-labile steroids comprising potential cysteine conjugates were extracted by the last LLE (fraction Cys).

The four resulting TBME extracts per sample were evaporated to dryness and the fractions of glucuronides and sulfates further purified by HPLC. For that purpose, samples were reconstituted in 100 μL ACN/H₂O (50/50, v:v), and the entire volume was injected into an Agilent 1100 HPLC-UV system (Waldbronn, Germany) equipped with a X Bridge Shield RP18 column (4.6 × 250 mm) with 5 μm particle size (Waters, Eschborn, Germany). UV signals were acquired at 195 and 360 nm. Gradient elution was conducted as follows with a flow rate of 1 mL/min: Starting at 20% ACN/80% water, the gradient increased to 100% ACN within 25 min, was held for 10 min, and re-equilibrated for 5 min. With support of a Foxy R1 automatic fraction collector (Axel Semrau, Sprockhövel, Germany), the following HPLC sub-fractions were produced using the retention time markers shown in brackets. I: 3.00–10.00 min, II: 10.01–13.50 min (Tren, Epitren), III: 13.51–14.80 min (T), IV: 14.81–17.00 min (EpiT, DHEA, 5b, 5a, ETIO), V: 17.01–19.50 min (PD), VI: 19.51–24.50, and VII 24.51–33.00 min (16 EN).

The eluted HPLC sub-fractions were evaporated to dryness. Derivatization of all fractions derived from HPLC clean-up, as well as the fractions containing the cysteine adducts and free steroids was performed in accordance with the applied chromatographic system as described below.

Derivatization Techniques

Acetylation

For acetylation, samples were reconstituted in 75 μL of pyridine and 75 μL of acetic anhydride and derivatized for 1 h at 70°C. Subsequently, the derivatization mixture was evaporated. Samples were subjected to GC-TC-IRMS/MS (section GC-TC-IRMS/MS Setup), GC-EI-HRMS (QTOF) (section GC-EI-HRMS (QTOF) Setup) and LC-ESI-HRMS (section LC-ESI-HRMS (LC Orbitrap) Setup) analysis.

Trimethylsilylation

For trimethylsilylation, samples were reconstituted in 80 μL of MSTFA:NH₄:ethanethiol (1000:2:3, v:w:v), incubated at 60°C for 45 min (Mareck et al., 2004, 2008), and measured as described in section GC-EI-HRMS (Orbitrap) Setup.

Instrument Methods

GC-TC-IRMS/MS setup

After evaporation of the derivatization mixture, the acetylated samples were reconstituted in an appropriate volume of cyclohexane (typically 20 μL) for GC-TC-IRMS/MS analysis. Analysis was performed on a Delta V Plus IRMS coupled via a GC Isolink CNH for thermal conversion at 1,450°C with a ceramic reduction reactor and ConFlow IV to a Trace 1310 GC (Thermo, Bremen, Germany). Chromatographic separation was accomplished on a DB-17 MS column (30 m × 0.25 mm) with a film thickness of 0.25 mm. The temperature gradient was as follows: The temperature remained constant at 100°C for 1.5 min and increased with 40°C/min to 240°C and subsequently with 5°C/min to 320°C with a hold time of 2 min. Samples were injected in splitless mode at 300°C with an injection volume of 5 μL. A single taper inlet liner (900 μL volume, 4 mm inner diameter, 6.47 mm outer diameter, 78.5 mm length) with glass wool from Agilent (part number: 5190-2293) was used. After passing the GC column, the flow was split by a ratio of approximately 1:10 to an ISQ single quadrupole mass spectrometer (Thermo, Bremen, Germany). Data acquisition and processing was accomplished using Isodat 3.0 and Xcalibur 2.2 software (Thermo, Bremen, Germany).

GC-EI-HRMS (QTOF) setup

Following GC-TC-IRMS/MS analysis, the acetylated samples were diluted to a final volume of 200 μL cyclohexane and subjected to GC-EI-HRMS measurements on an Agilent 7200 QTOF system hyphenated to an Agilent 7890A gas chromatograph (Santa Clara, CA). The chromatography setup was equivalent to GC-TC-IRMS/MS described above (section GC-TC-IRMS/MS Setup), including the same analytical column and the same temperature program. The injection volume was reduced to 4 μL. Data were acquired within a range of m/z 50–800 with an acquisition rate of 5 spectra/s and evaluated with MassHunter software (version B.06, Agilent). Mass calibration was performed before and during each analytical batch.

GC-EI-HRMS (Orbitrap) setup

The trimethylsilylated samples were injected (2 μL injection volume) into a Q Exactive GC Orbitrap (Thermo, Bremen, Germany). Due to the different derivatization technique, the system was operated with modified chromatographic conditions adapted from routine protocols (Thevis et al., 2011). The GC was equipped with a HP-Ultra 1 column (17 m × 0.2 mm) of 0.11 mm film thickness. The temperature program started at 180°C and raised with 3°C/min to 240°C and with 40°C/min to 320°C, where it remained constant for 2 min. Samples were injected in split mode with a split flow of 5 mL/min. The mass range for full MS experiments was m/z 50–700 and a resolution of 60,000 FWHM was applied. Data were evaluated with Xcalibur software.

LC-ESI-HRMS (LC Orbitrap) setup

For LC-ESI-HRMS measurements, the acetylated samples were evaporated and reconstituted with 100 μL of ACN/H₂O (50:50, v/v) acidified with 0.1% FA. A Vanquish UHPLC (Thermo, Bremen, Germany) equipped with a Poroshell 120 EC-C8 (2.7 μm, 3 × 50 mm) (Agilent, Santa Clara, CA) was hyphenated to a Q Exactive HF-X (Thermo, Bremen, Germany). ACN and ultrapure water both containing 0.1% FA were employed as solvents, and the flow rate was set to 400 μL/min. A volume of 5 μL was injected per sample. The LC gradient run was as follows: Starting at 40% ACN, it was increased to 99% within 9 min, and hold for further 3 min until re-equilibration, yielding a total analysis time of 15 min. MS experiments comprised a full scan (m/z 200–800), AIF (all ions fragmentation), and PRM in positive ionization mode at a resolution of 60,000 FWHM. For PRM experiments, the isolation window was set to m/z 0.4 and in the higher-energy collisional dissociation (HCD) cell, collision energies of 20, 30, 35 or 40 eV were applied. For pseudo MS³ experiments, the source induced dissociation (SID) energy was set to 20 eV.

Analysis of Conjugated Steroids

For the analysis of intact conjugated steroids, samples were diluted 1:1 with ultrapure water. In order to perform PRM experiments, selected samples were 5-fold pre-concentrated by SPE as described in section 2.3.1, and reconstituted in 50 μL 10% aqueous ACN. The setup for the LC-ESI-HRMS (Orbitrap) system was similar to the settings described in chapter LC-ESI-HRMS (LC Orbitrap) Setup, but employing a modified gradient. Starting at 1% ACN (containing 0.1% FA), the gradient increased to 40% ACN within 9 min, to 99% until 10.9 min, and to 1% until 11 min. The system was re-equilibrated for 3 min. Full MS, AIF, and PRM experiments were conducted and ESI with positive polarity was applied. In selected experiments, also negative ionization was used, which is explicitly indicated in the corresponding data sets.

Synthesis of Trenbolone-diol

Trenbolone-diol was synthesized by reduction of Tren under argon atmosphere. To 10 mg of Tren dissolved in 10 mL of anhydrous THF, 100 μL of potassium tri-sec-butylborohydride (1 M in THF) were added under constant stirring. After 15 min, the reaction was stopped with 10 mL of ultrapure water, and subsequently, a LLE with 20 mL of TBME was performed. An aliquot of the yielded products was acetylated as described in section Acetylation.

Results and Discussion

Metabolite Identification With GC-TC-IRMS

The use of GC-TC-IRMS allows for comprehensive metabolite studies, particularly when investigating biotransformation products of deuterated compounds.

A significant increase of the ratio m/z 3 to m/z 2 indicates the presence metabolites of the administered substance, in this case trenbolone. Exemplary GC-TC-IRMS chromatograms of both a pre- and post-administration sample are displayed in Figure 3. The upper part (A) shows the chromatogram of the fraction containing the hydrolyzed glucuronides (fraction Gluc) of a pre-administration sample with deuterium levels at natural abundance. By contrast, the lower chromatogram (B) shows a sample collected 45 h following drug administration-, and several peaks of deuterated molecules corresponding to metabolites of Tren are visible. In the inset, details of the metabolites eluting at retention times between 1108 and 1110 s are shown.