Detection of the selective androgen receptor modulator GSK2881078 and metabolites in urine and hair after single oral administration.

Hair and urine concentrations of the non-steroidal selective androgen receptor modulator GSK2881078 were examined following single oral administration to investigate its hair incorporation and estimate the general suitability of hair testing for selected androgen-receptor modulators. Hair segments were collected following a single dose of 1.5 mg GSK2881078 by repeated shaving of scalp hair at week 0 (blank), week 1 (representing the pre-application period), week 3 (ideally focusing the time of incorporation), weeks 5 and 9 (post-administration period). The intact compound and various (at least 4) hydroxy-metabolites exhibited similar elimination profiles. The peak urinary concentration (approx. 920 pg/mL) was observed after 8 hours, and is reduced to the detection limit (2 pg/mL) on day 42 following administration of 760 μg GSK2881078. Correspondingly, hair concentrations of GSK2881078 (intact compound only) following a single oral dose of 1.5 mg GSK2881078 reached a peak concentration of 1.7 pg/mg in the segments collected 3 weeks post administration, representing the time of ingestion. The concentration rapidly declined to trace amounts of 0.7 pg/mg (week 5) and 0.2 pg/mg (week 9), respectively. In conclusion, measurement of the intact compound GSK2881078 is feasible for both urine and hair analysis. However, concentrations in hair after single oral administration are in the low pg/mg-range and can only be detected, if the segments cover the administration period.

tion), and Weeks 5 and 9 (post-administration period). The intact compound and various (at least 4) hydroxy-metabolites exhibited similar elimination profiles. The peak urinary concentration (approximately 920 pg/ml) was observed after 8 h and is reduced to the detection limit (2 pg/ml) on Day 42 following administration of 760 μg GSK2881078. Correspondingly, hair concentrations of GSK2881078 (intact compound only) following a single oral dose of 1.5 mg GSK2881078 reached a peak concentration of 1.7 pg/mg in the segments collected 3 weeks post administration, representing the time of ingestion. The concentration rapidly declined to trace amounts of 0.7 (Week 5) and 0.2 pg/mg (Week 9), respectively. In conclusion, measurement of the intact compound GSK2881078 is feasible for both urine and hair analysis. However, concentrations in hair after single oral administration are in the low pg/mg range and can only be detected, if the segments cover the administration period.  [5][6][7] Those trials showed safety and anabolic effects of daily doses of 0.75 to 4 mg GSK2881078 in healthy males and females as well as in cachexic males and females. A recent study described the mass spectrometric analyses of GSK2881078. 8 Hair as matrix may provide additional information compared with urine samples in doping control. According to WADA's International Standard for Laboratories, analysis of hair is approved but shall not be used to counter analytical findings from urine or blood. 9 Due to possible long-term detection in hair, valuable methods are established to differentiate food contamination from clenbuterol abuse 10 and to identify administered intact compounds, for example, testosterone esters in connection to elevated T/E ratio in urine. 11 With regard to SARMs, only few studies investigated their incorporation in the hair matrix, among them LGD-4033, SARMs S4 and S22. [12][13][14] Aim of the present study was the elucidation of the in vivo metabolism of GSK2881078 to identify metabolites as potential targets for doping analysis. After oral administration of GSK2881078 by one healthy volunteer, urine samples were collected up to 42 days post administration. Beside the excretion study, the study investigated the incorporation of GSK2881078 and its metabolites detected in urine in hair as an alternative matrix.

| MATERIALS AND METHODS
For additional experimental details, refer to the Supporting Information.
GSK2881078 with a purity of ≥98% was purchased from Cayman Chemical (Ann Arbour, Michigan, USA). LGD-4033 was received from WAADS without any certificate or purity statement. All solvents used were purchased from Chemsolute (Th Geyer, Renningen, Germany).

| Sample preparation
Urine sample extraction with and without hydrolysis was carried out as described previously. 15

| HPLC-MS measurements
High-performance liquid chromatography-mass spectrometry (HPLC-MS) measurements were performed using a HPLC (1290 Infinity II

| Analytical
GSK2881078 is well ionized in both polarities. Product ions observed in positive ESI are identical to those already described by Thevis and Volmer in a recent publication. 8 The unusual temperature dependence in positive ESI mode-which was confirmed using different instruments and is likely to be influenced by gradient composition and ion source design-needs to be taken into consideration ( Figure S1). There was a steep intensity decline of the protonated molecule observed at m/z 331 showing a minimum intensity at 500 C (i.e., typically the optimum temperature in most screening methods). This behavior is not well understood, because no formation of apparent fragments or adducts was observed at this temperature. Moreover, the conventional temperature dependence in negative ionization mode suggests a sufficient thermal stability of the molecule ( Figure S1).
Collision-induced dissociation was in both polarities characterized by side-chain cleavage, losses of HF, and combinations thereof ( Figure S2). Acetate adducts of GSK2881078 were formed in negative ionization at lower temperatures, but maximum intensities could be achieved by using the in-source fragment (i.e., trifluoromethyl-indolecarbonitrile) as precursor of subsequent -HF fragmentation reactions ( Figures 1B and S3). This was found to be the most sensitive approach for screening of intact GSK2881078 and its hydroxy-metabolites. To include structural information of the whole molecule and to differentiate hydroxylation position, an alternative QTrap MS3 procedure was chosen for confirmation purposes in positive ionization mode (Table 1 and Figure 2A).
Identical methods were used for urine and hair extracts, respectively. Corresponding detection limits of the negative MRM procedure (defined as signal-to-noise ratios of the qualifier ion better than three) were found to be 2 pg/ml (urine) or 0.1 pg/mg (hair). There was a good linearity of calibration over at least three decimal powers. The intraday repeatability of urinary concentrations was below 10% variation. Carry-over was not relevant in both matrices (estimated carryover 0.03%).

| Urine
The examination of GSK2881078 revealed immediate renal excretion of the unmodified compound. Moreover, four hydroxy-metabolites were identified based on their corresponding fragmentation pattern. windows up to 21 days, following administration of 10 mg ligandrol, were observed. 18 The dihydroxy-metabolite was assumed to be the best long-term marker for ligandrol in doping analysis.

| Hair
Hair segments were collected-following the administration of 1.5 mg GSK2881078 by repeated shaving of scalp hair at Weeks 0 (blank), 1 (representing the pre-application period), 3 (ideally focusing the time of incorporation), and 5 and 9 (post administration of 1.5 mg GSK2881078).
The assignment of hair segments to corresponding elimination times is based on the assumptions of hair growth speed as well as the temporal offset before the first collection, that is, approximately 1 week between hair formation and entering the skin surface plus approximately 4.5 mm residue length of remaining hair segments.
The individual hair growth rate was examined by comparison of segment lengths to be 1.0 cm/month (sd 13%). In contrast, retrospective segmentation of long hair strands is less accurate due to misalignment of hair within the bundle and the increasing proportion of non-growing (catagen) hair. Methanol extraction was found to be the preferential sample preparation procedure, whereas sodium hydroxide disintegration of hair failed due to substance instability and acidic hydrolysis (using HCl) did not markedly improve recovery (data not shown).
Intact GSK2881078 was detected at a maximum concentration of 1.7 pg/mg in the 5 mm segments of hair collected 3 weeks post administration (i.e., perfectly covering administration period, Figure 4).
None of the hydroxy-metabolites could be identified. Concentrations dropped to approximately 0.7 pg/mg in adjacent segments (collected after 1 and 5 weeks, respectively) and were above the detection limit (0.1 pg/mg) in samples collected after 9 weeks (Figure 4).
In accordance to the physicochemical similarity between neutral lipophilic steroids and GSK2881078, their incorporation rate into hair appears rather poor, as compared with basic drugs of abuse or most pharmaceuticals. A single high therapeutic oral administration of GSK2881078 was found to be detectable in hair at low pg/mg levels, if segmentation matches perfectly with administration time and superior detection limits below 1 pg/mg hair are achieved.
F I G U R E 3 Urinary concentration profile of GSK2881078 and abundance of the two mono-hydroxylated metabolites relative to GSK2881078 after oral administration of 760 μg GSK2881078 [Colour figure can be viewed at wileyonlinelibrary.com] F I G U R E 4 Hair concentration of GSK2881078. Samples were collected 1, 3, 5, and 9 weeks after oral administration of 1.5 mg GSK2881078. The "3-week segment" represents the period since recent cutting (at 1 week post administration). Due to the delay between hair root formation and first cutting of approximately 2 weeks, it is assumed to represent best the administration time 4 | CONCLUSION GSK2881078 was well detectable in urine samples after oral administration of 760 μg, which was assumed to represent a single therapeutic dose. Negative MRM experiments were found to be most sensitive and suitable for screening purposes (in urine or hair). Owing to its reduced structural information (following in-source fragmentation), the confirmation should be preferentially carried out in positive ionization mode.
Secondary fragmentations following the initial loss of HF (MS3 mode) was found to be sufficiently sensitive and comprehensive (e.g., to differ-