In vitro metabolism studies of desoxy-methyltestosterone (DMT) and its five analogues, and in vivo metabolism of desoxy-vinyltestosterone (DVT) in horses

Article abstract of DOI:10.1002/jms.3613

The positive findings of norbolethone in 2002 and tetrahydrogestrinone in 2003 in human athlete samples confirmed that designer steroids were indeed being abused in human sports. In 2005, an addition to the family of designer steroids called 'Madol' [also known as desoxy-methyltestosterone (DMT)] was seized by government officials at the US-Canadian border. Two years later, a positive finding of DMT was reported in a mixed martial arts athlete's sample. It is not uncommon that doping agents used in human sports would likewise be abused in equine sports. Designer steroids would, therefore, pose a similar threat to the horseracing and equestrian communities. This paper describes the in vitro metabolism studies of DMT and five of its structural analogues with different substituents at the 17α position (R￡H, ethyl, vinyl, ethynyl and ²H₃-methyl). In addition, the in vivo metabolism of desoxy-vinyltestosterone (DVT) in horses will be presented. The in vitro studies revealed that the metabolic pathways of DMT and its analogues occurred predominantly in the A-ring by way of a combination of enone formation, hydroxylation and reduction. Additional biotransformation involving hydroxylation of the 17α-alkyl group was also observed for DMT and some of its analogues. The oral administration experiment revealed that DVT was extensively metabolised and the parent drug was not detected in urine. Two in vivo metabolites, derived respectively from (1) hydroxylation of the A-ring and (2) di-hydroxylation together with A-ring double-bond reduction, could be detected in urine up to a maximum of 46 h after administration. Another in vivo metabolite, derived from hydroxylation of the A-ring with additional double-bond reduction and di-hydroxylation of the 17α-vinyl group, could be detected in urine up to a maximum of 70 h post-administration. All in vivo metabolites were excreted mainly as glucuronides and were also detected in the in vitro studies.

Full text of DOI:10.1002/jms.3613

Journal of
MASS
SPECTROMETRY
Research article
Received: 9 March 2015
Revised: 13 April 2015
Accepted: 4 May 2015
Published online in Wiley Online Library
(wileyonlinelibrary.com) DOI 10.1002/jms.3613
In vitro metabolism studies of desoxy-
methyltestosterone (DMT) and its five
analogues, and in vivo metabolism of desoxy-
vinyltestosterone (DVT) in horses
Wai Him Kwok,* Karen Y. Kwok, David K. K. Leung, Gary N. W. Leung,
Colton H. F. Wong, Jenny K. Y. Wong and Terence S. M. Wan*
The positive findings of norbolethone in 2002 and tetrahydrogestrinone in 2003 in human athlete samples confirmed that de-
signer steroids were indeed being abused in human sports. In 2005, an addition to the family of designer steroids called ‘Madol’
[also known as desoxy-methyltestosterone (DMT)] was seized by government officials at the US–Canadian border. Two years later,
a positive finding of DMT was reported in a mixed martial arts athlete’s sample. It is not uncommon that doping agents used in
human sports would likewise be abused in equine sports. Designer steroids would, therefore, pose a similar threat to the
horseracing and equestrian communities. This paper describes the in vitro metabolism studies of DMT and five of its structural
analogues with different substituents at the 17α position (R¼H, ethyl, vinyl, ethynyl and 2H3-methyl). In addition, the in vivo me-
tabolism of desoxy-vinyltestosterone (DVT) in horses will be presented.
The in vitro studies revealed that the metabolic pathways of DMT and its analogues occurred predominantly in the A-ring by
way of a combination of enone formation, hydroxylation and reduction. Additional biotransformation involving hydroxylation
of the 17α-alkyl group was also observed for DMT and some of its analogues. The oral administration experiment revealed that
DVT was extensively metabolised and the parent drug was not detected in urine. Two in vivo metabolites, derived respectively
from (1) hydroxylation of the A-ring and (2) di-hydroxylation together with A-ring double-bond reduction, could be detected in
urine up to a maximum of 46h after administration. Another in vivo metabolite, derived from hydroxylation of the A-ring with
additional double-bond reduction and di-hydroxylation of the 17α-vinyl group, could be detected in urine up to a maximum of
70h post-administration. All in vivo metabolites were excreted mainly as glucuronides and were also detected in the in vitro
studies. Copyright © 2015 John Wiley & Sons, Ltd.
Keywords: designer steroids; doping control; equine urine; in vitro and in vivo metabolism
human sports would likewise be abused in equine sports. Therefore,
designer steroids would pose a similar threat to the horseracing
Introduction
Designer steroids can be derived from a simple chemical modifica-
tion of known anabolic androgenic steroids. They have been synthe-
sised by unscrupulous chemists for the purpose of performance
enhancement and escaping detection. As long as these steroids
remain unknown, it is difficult to detect their presence in biologi-
cal samples by either target analysis or mass spectral library
search, as both approaches would require first establishing the
analytical characteristics with suitable reference materials. In re-
cent years, abuses of designer steroids, such as norbolethone[1]
and tetrahydrogestrinone,[2] have been reported in human sports.
Strictly speaking, norbolethone is not a designer steroid because
it is not novel but was originally synthesised in 1966 with
intended pharmacological applications. It was, however, never
marketed. As norbolethone was intended to be used in human
sports to avoid detection, it can be considered to fall within the
same category of designer steroids. ‘Madol’ (17α-methyl-5α-
androst-2-ene-17β-ol), also known as desoxy-methyltestosterone
(DMT),[3] was an addition to the family of designer steroids and
was seized by government officials at the US–Canadian border
in 2005. Two years later, DMT was detected in a mixed martial arts
athlete’s sample.[4] It is not uncommon that doping agents used in
and equestrian communities.
The detection of unknown designer steroids has been proposed
by Thevis et al.[5,6] and Nielen et al.[7] by using respectively precursor
ion scans in triple quadrupole mass spectrometers and a combina-
tion of androgen bioactivity testing and quadrupole time-of-flight
mass spectrometry. Pozo et al. later proposed a more comprehen-
sive method to detect unknown anabolic steroids and their metab-
olites based on precursor ion scanning of three ions (m/z 105, m/z
91 and m/z 77) using liquid chromatography–tandem mass
spectrometry.[8] The detection of unknown anabolic androgenic
steroids in doping control was recently reviewed.[9] While these
generic approaches to detect unknown steroids are valid and excel-
lent concepts, there are limitations (such as uncertainty in detecting
truly unknown substances), and these should only be used as
*
Correspondence to: Wai Him Kwok and Terence S. M. Wan, Racing Laboratory,
The Hong Kong Jockey Club, Sha Tin Racecourse, Sha Tin, NT, Hong Kong, China.
E-mail: wh.kwok@hkjc.org.hk; terence.sm.wan@hkjc.org.hk
Racing Laboratory, The Hong Kong Jockey Club, Sha Tin Racecourse, Sha Tin, NT,
Hong Kong, China
J. Mass Spectrom. 2015, 50, 994–1005
Copyright © 2015 John Wiley & Sons, Ltd.

Journal of
MASS
SPECTROMETRY
Metabolism studies of DMT and its five analogues
complementary approaches. The identification of screening targets
via in vitro and in vivo metabolism studies,[10–14] although requiring
many resources, remains to be the most certain way to control the
abuse of selected designer steroids. This paper describes a series of
in vitro metabolism studies of DMT and five of its structural ana-
logues with different substituents at the 17α position (R¼H, ethyl,
vinyl, ethynyl and 2H3-methyl). In addition, the in vivo metabolism
of one of the DMT analogues, desoxy-vinyltestosterone (DVT), in
horses will be presented.
was extracted twice with ethyl acetate (5 ml), and the combined
organic extract was evaporated to dryness. The residue was
analysed by GC–MS as trimethylsilyl (TMS) derivatives. Control
experiments in the absence of either (1) the steroid under study
or (2) microsomes were performed in parallel.
Drug administration studies
Two thoroughbred geldings were each administered orally with a
single dose of 500 mg of DVT. Urine samples were collected before
administration and then at least twice daily for up to 8 days post-
administration. Urine samples were extracted with and without
prior hydrolysis in order to identify respectively conjugated and free
DVT and its metabolites. The extracts were analysed by GC–MS after
trimethylsilylation.
Materials and methods
DMT, desoxy-testosterone (DT), desoxy-ethyltestosterone (DET),
DVT, desoxy-ethynyltestosterone (DENT) and d3-desoxy-
methyltestosterone (d3-DMT) were custom synthesised by the
Hong Kong Polytechnic University, and their identities were con-
firmed by 1H NMR and 13C NMR. 5α-Androst-1-ene-3α,17β-diol and
5α-androst-1-ene-3β,17β-diol were custom synthesised by The
Chinese University of Hong Kong.[13] [16,16,17]-d3-5α-Androstane-
3α,17β-diol (d3-androstanediol) was synthesised in house according
to the literature.[15] 17α-Methyl-1-testosterone and 4-androstene-
3β,17β-diol were obtained from Steraloids (Rhode Island, USA).
Epiandrosterone, testosterone, β-nicotine adenine dinucleotide
(β-NAD), glucose 6-phosphate, glucose-6-phosphate dehydroge-
nase (G6PDH), magnesium chloride (MgCl2), N-methyl-N-
trimethylsilylfluoroacetamide (MSTFA), ammonium iodide (NH4I),
1,4-dithioerythritol (DTE), protease (from bovine pancreases, type
I, 6.9 U/mg solid) and β-glucuronidase (from Patella vulgata,
lyophilised powder) were purchased from Sigma (St. Louis, MO,
USA). N,O-Bis(trimethylsilyl)trifluroacetamide (BSTFA) with 5%
trimethylchlorosilane (TMCS) was prepared by mixing BSTFA from
Pierce (Illinois, USA) with TMCS from Aldrich (St. Louis, MO, USA).
β-Glucuronidase (Escherichia coli) was from Roche (Indianapolis,
USA). Diisopropyl ether (GR grade), ethyl acetate (GR grade), meth-
anol (GR grade), sulfuric acid (Suprapur®, 96%), sodium chloride
(GR grade) and sodium dihydrogen phosphate (GR grade) were
obtained from Merck (Darmstadt, Germany). Sodium hydroxide
(pro analysis grade) was from RDH (Seelze, Germany). Sodium
sulfate was purchased from Farco Chemical Supplies (Beijing,
China). ABS Elut NEXUS cartridge (60 mg, 3 ml) was obtained
from Varian (Harbor City, CA, USA).
Extraction procedures for phase I metabolism study on DVT
post-administration urine samples
Urine (3ml) was fortified with d3-androstanediol (30 ng) as the in-
ternal standard and diluted with potassium phosphate buffer
(pH 6.0, 0.1 M, 1 ml). A solution of protease (5 mg/ml, 60 μl) and
β-glucuronidase (P. vulgata, 18 000 U/ml, 360 μl) was added, and
the sample incubated at 65 °C for 3.5 h. The enzyme-treated urine
was then extracted twice with diisopropyl ether (5ml). The com-
bined extract was washed with base (2.0ml containing 1 M sodium
hydroxide and 0.15 M sodium chloride), passed through an anhy-
drous sodium sulfate drying tube and evaporated to dryness under
nitrogen at 60°C.
Extraction procedures for phase II metabolism studies
Unconjugated steroids
Urine (5ml) was mixed with sodium sulfite solution (0.25 ml, 25%
w/v), adjusted to pH 6.8 and extracted with diisopropyl ether
(5ml). The organic layer was base washed (4.0ml containing 1 M
sodium hydroxide and 0.15 M sodium chloride) and centrifuged
(2100 g for 1 min). The supernatant was passed through an anhy-
drous sodium sulfate drying tube and evaporated to dryness under
nitrogen at 60°C.
Glucuronide-conjugated steroids
Preparation of horse liver microsomes
The remaining aqueous layer was then adjusted to pH 6.4 and incu-
bated at 45°C for 1 h with β-glucuronidase from E. coli (100 μl). The
enzyme-treated steroids were extracted from the mixture accord-
ing to the aforementioned procedures for unconjugated steroids.
Horse liver microsomes were isolated in house from fresh horse
liver, which was supplied by the Equine Hospital of The Hong Kong
Jockey Club. Small pieces of horse liver were homogenised in
Tris/KCl buffer (0.05 M, pH 7.4). The homogenate was centrifuged
at 10 000 g for 25 min. Microsomes were isolated from the superna-
tant by centrifuging at 105 000 g for 1 h. The pellet of microsomes
was then washed twice with Tris/KCl buffer (0.05 M, pH 7.4). All
the preparation steps and reagent storage were carried out at 4 °C.
Sulfate and sulfate glucuronide-conjugated steroids
The sulfate and sulfate-glucuronide conjugates remaining in the
aqueous layer after enzyme hydrolysis were extracted using solid-
phase extraction (SPE) (ABS Elut Nexus cartridge). The loaded SPE
cartridge was washed with deionised water (3ml) and eluted with
a solvolysis reagent (3ml; ethyl acetate: methanol: conc. H2SO4,
100:20:0.2, v/v/v). The eluate was incubated at 55 °C for 2 h.
NaOH/NaCl (1M/0.15M, 1.2 ml) was then added. The mixture was
vortexed for 0.5min. The separated organic layer was evaporated
to dryness under nitrogen at 60 °C. The cooled residue was re-
dissolved in diisopropyl ether (3ml) and washed with deionised
water (1ml). The organic layer was passed through an anhydrous
sodium sulfate drying tube and evaporated to dryness under nitro-
gen at 60 °C.
Microsomal incubation
For the in vitro metabolic studies, a mixture of freshly prepared horse
liver microsomes (30 μl), β-NAD (1.5 mM), glucose-6-phosphate
(7.5 mM), magnesium chloride (4.5 mM), EDTA (1 mM) and G6PDH
(1 U/ml) in phosphate buffer (5 ml, 50 mM, pH 7.4) was incubated
with the steroid under study (200 μg) at 37 °C overnight with shaking.
The reaction was terminated by boiling at 100 °C for 10 min. The
mixture was centrifuged at 2100 g for 10 min. The supernatant
J. Mass Spectrom. 2015, 50, 994–1005
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms

Journal of
MASS
SPECTROMETRY
W. H. Kwok et al.
Figure 1. Synthetic route for DMT and its five analogues.
Table 1. Structures of DVT, DMT, DT, DET, DENT and d3-DMT, and major metabolites identified after incubation with horse liver microsomes
Structure In vitro metabolites
Enone formation
in the A-ring
Mono-hydroxylation
Reduction + di-hydroxylation
in the A-ring
Additional
metabolites
in the A-ring
DVT-M1a
DVT-M1b
DVT-M1c
DVT-M1d
DVT-M2a
DVT-M2b
DVT-M2c
DVT-M2d
DVT-M3a
DVT-M3b
DVT-M3c
DVT-M3d
DVT-M4a
DVT-M4b
DVT-M5a
DVT-M5b
DVT-M6
DVT-M7a
DVT-M7b
DVT-M8
DMT-M1a
DMT-M2a
DMT-M2b
DMT-M3a
DMT-M3b
DMT-M4a
DMT-M4b
DMT-M5
DMT-M1b*
DT-M1a
DT-M2a
DT-M3
Not detected
DT-M1b*
DT-M2b*
DT-M2c
DT-M2d
DT-M2e*
DT-M2f*
Not detected
DET-M1a
DET-M1b
DET-M3a
DET-M3b
DET-M3c
DET-M4
DET-M5
DENT-M1a
DENT-M1b
DENT-M2a
DENT-M2b
DENT-M3a
DENT-M3b
Not detected
d3-DMT-M1a
d3-DMT-M1b
d3-DMT-M1c
d3-DMT-M2a
d3-DMT-M2b
d3-DMT-M3a
d3-DMT-M3b
d3-DMT-M4a
d3-DMT-M4b
d3-DMT-M5
* Confirmed with reference standards.
wileyonlinelibrary.com/journal/jms
Copyright © 2015 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2015, 50, 994–1005

Journal of
MASS
SPECTROMETRY
Metabolism studies of DMT and its five analogues
Derivatisation for GC–MS analysis
to 300 °C at 15 °C/min and finally held at 300 °C for 6 min. Samples
(1 μl) were injected at 250 °C in the splitless mode. All GC–MS analyses
were performed in the EI mode with full-scan acquisition.
TMS derivatives were prepared by adding either BSTFA (5% TMCS,
30 μl) or MSTFA/NH4I/DTE (MND; 1000:2:4, v/w/w, 30μl) to the dried
residue. The mixture was incubated at 60 °C for 30 min or 80 °C for
30 min for respectively BSTFA or MND derivatisation. The resulting
mixture was injected directly for GC–MS analysis.
Results and discussion
Synthesis of DMT and its five analogues
Instrumentation
DMT and its five analogues were custom synthesised by the Hong
Kong Polytechnic University. The synthetic route is shown in Fig. 1.
Epiandrosterone was used as the starting material. The hydroxyl
group at C3 was tosylated. Elimination of toluenesulfonic acid gave
a carbon double bond in the A-ring. The mixture was then alkylated
at C17 using the appropriate Grignard reagents to give the final
products. In the case of DT, the 17β-hydroxy functional group
An Agilent 6890N Network GC system coupled to an Agilent 5973
Network Mass Selective Detector (Agilent Technologies, CA, USA) was
used. Separation was performed on an HP-1MS (~30m×0.25mm,
0.25 μm film thickness; Agilent Technologies, CA, USA) column with
a constant helium flow of 1.2 ml/min. The oven temperature was set
initially at 60 °C for 1.0 min, increased to 120 °C at 30 °C/min and then
Figure 2. EI mass spectra of trimethylsilylated (a) DVT-M1a, (b) DVT-M2d, (c) DVT-M3a, (d) DVT-M4b, (e) DVT-M5a, (f) DVT-M6, (g) DVT-M7a and (h) DVT-
M8 obtained from the in vitro incubation mixture of DVT with horse liver microsomes. All TMS derivatives were prepared using MND except for DVT-M1a
where BSTFA (5% TMCS) was used.
J. Mass Spectrom. 2015, 50, 994–1005
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms

Journal of
MASS
SPECTROMETRY
W. H. Kwok et al.
was formed by reduction of the 17-keto group with sodium borohy-
dride. As the elimination step was not regiospecific, both the Δ2 and
Δ3 isomers were formed in the final products (in a ratio of approx-
imately 3:1). This was consistent with the sample of DMT seized
by government officials at the US–Canadian border reportedly con-
taining around 25% of the Δ3 isomer.[3] The resulting DMT and its
analogues were used for the subsequent in vitro and in vivo studies
without further purification.
spectral interpretations for in vitro metabolites of the other five
analogues will be discussed using DVT as a model compound
and divided into four groups as follows.
(i) Enone metabolites
In order to facilitate the interpretation of the TMS derivatives of
the M1-enone metabolites, a mild TMS derivatising reagent (BSTFA
with 5% TMCS) was used to suppress the formation of enol-TMS de-
rivatives in the A-ring. The TMS derivatives of DVT-M1 (four iso-
mers) had molecular ions at m/z 386, which was consistent with
the addition of a keto function to DVT (Fig. 2(a)). The characteristic
pair of D-ring fragments at m/z 155 and m/z 142 confirmed that the
D-ring structure of DVT remained intact. The mechanism for the
formation of such D-ring fragments was first investigated by
Middleditch et al.[16] The in vitro studies of the other five analogues
DMT, DT, DET, DENT and d3-DMT gave similar enone metabolites,
DMT-M1a and DMT-M1b, DT-M1a and DT-M1b, DET-M1a
and DET-M1b, DENT-M1a and DENT-M1b, and d3-DMT-M1a to
d3-DMT-M1c, respectively. The mass spectra of the TMS derivatives
In vitro biotransformation studies
Results of the in vitro studies of DMT and its five analogues are
summarised in Table 1. All proposed metabolites are 5α-steroids
unless they have a carbon double bond at the C5 position. The ma-
jor biotransformations of DMT and its five analogues can be
categorised into a combination of enone formation, hydroxylation
and reduction in the A-ring. In the case of DVT, DMT, DET and d3-
DMT, additional metabolites derived from other biotransformation
pathways have been identified. For simplicity, results and mass
Figure 2. (Continued).
wileyonlinelibrary.com/journal/jms
Copyright © 2015 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2015, 50, 994–1005

Journal of
MASS
SPECTROMETRY
Metabolism studies of DMT and its five analogues
of representative M1 metabolites, DMT-M1b, DT-M1b, DET-M1a,
DENT-M1a and d3-DMT-M1c, are shown in Figs 3(a)–7(a), respectively.
All these spectra showed peaks at their corresponding molecular ions
and characteristic pairs of D-ring fragments. Of these metabolites, the
identities of DMT-M1b and DT-M1b were confirmed with authentic
reference standards of 17α-methyl-1-testosterone and testosterone,
respectively, suggesting that enone formation was in the A-ring
bearing either the 1-ene-3-one or 4-ene-3-one structure. As the re-
tention time of trimethylsilylated d3-DMT-M1c matched well with that
of trimethylsilylated 17α-methyl-1-testosterone, and its mass spectrum
(Fig. 7(a)) also closely resembled that of DMT-M1b (Fig. 3(a)) with a
shift of 3 m/z units at their molecular ions and some fragment ions,
Figure 3. EI mass spectra of trimethylsilylated (a) DMT-M1b (17α-methyl-1-testosterone), (b) DMT-M2b, (c) DMT-M3b, (d) DMT-M4b and (e) DMT-M5
obtained from the in vitro incubation mixture of DMT with horse liver microsomes. All TMS derivatives were prepared using MND except for DMT-M1b
where BSTFA (5% TMCS) was used.
J. Mass Spectrom. 2015, 50, 994–1005
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms

Journal of
MASS
SPECTROMETRY
W. H. Kwok et al.
Figure 4. EI mass spectra of trimethylsilylated (a) DT-M1b (testosterone), (b) DT-M2b (5α-androst-1-ene-3α,17β-diol), (c) DT-M2e (4-androstene-3β,17β-
diol), (d) DT-M2f (5α-androst-1-ene-3β,17β-diol) and (e) DT-M3 obtained from the in vitro incubation mixture of DT with horse liver microsomes. All TMS
derivatives were prepared using MND except for DT-M1b where BSTFA (5% TMCS) was used.
d3-DMT-M1c could be assigned as 17α-2H3-methyl-1-testosterone.
However, the mass spectra did not allow categorical assignment of
the enone position in the A-ring of other M1 metabolites. The exis-
tence of approximately 25% of the Δ3-isomer in each parent substance
(Fig. 1) further complicated the structural interpretation of the in vitro
metabolites. A similar metabolic pathway has been reported for
ethylestrenol in horses,[17] where the corresponding enone
(norethandrolone) was the major urinary metabolite. In addition, the
formation of the enone metabolite from DMT has also been reported
in human urine.[18]
(ii) Hydroxylated metabolites
The TMS derivatives of DVT-M2 (four isomers) had molecular
ions at m/z 460, which was consistent with the addition of a
hydroxyl group to DVT (Fig. 2(b)). Diagnostic ions at m/z 155 and
m/z 142 confirmed that the D-ring was intact. The lack of a prominent
m/z 147 (Me3SiO+¼SiMe2) indicated that the two trimethylsiloxyl
groups were not in close proximity. The in vitro studies of the
other analogues except for DET gave similar mono-hydroxylated
M2 metabolites (DMT-M2a and DMT-M2b, DT-M2a to DT-M2f,
wileyonlinelibrary.com/journal/jms
Copyright © 2015 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2015, 50, 994–1005

Journal of
MASS
SPECTROMETRY
Metabolism studies of DMT and its five analogues
Figure 5. EI mass spectra of trimethylsilylated (a) DET-M1a, (b) DET-M3a, (c) DET-M4 and (d) DET-M5 obtained from the in vitro incubation mixture of DET
with horse liver microsomes. All TMS derivatives were prepared using MND except for DET-M1a where BSTFA (5% TMCS) was used.
DENT-M2a and DENT-M2b, d3-DMT-M2a and d3-DMT-M2b).
The mass spectra of TMS derivatives of representative M2 metab-
olites, DMT-M2b (Fig. 3(b)), DT-M2b (Fig. 4(b)), DT-M2e (Fig. 4(c)),
DT-M2f (Fig. 4(d)), DENT-M2a (Fig. 6(b)) and d3-DMT-M2b (Fig. 7(b)),
are shown in the corresponding figures. All these spectra show
peaks at their corresponding molecular ions, [M-90]+• ions (neutral
loss of HOSiMe3 from the molecular ions), and characteristic pairs
of D-ring fragments. Of these metabolites, DT-M2b, DT-M2e
and DT-M2f were confirmed with authentic reference standards
to be 5α-androst-1-ene-3α,17β-diol, 4-androstene-3β,17β-diol
and 5α-androst-1-ene-3β,17β-diol, respectively. Based on these
findings, one of the hydroxylation sites of M2 metabolites should
be at the C3 position in the A-ring. In the case of DMT-M2b, the
mass spectrum of its TMS derivative resembled that reported by
Gauthier et al.[18] and Rodchenkov et al.,[19] although the hydrox-
ylation site of this human metabolite proposed by Gauthier et al.
was at C4. It was noteworthy that both the mass spectrum and
retention time of trimethylsilylated d3-DMT-M2b (Fig. 7(b))
matched well with those of trimethylsilylated DMT-M2b (Fig. 3(b))
with shifts of 3 m/z units at their molecular ions and some frag-
ment ions. Therefore, the sites of hydroxylation in d3-DMT-M2b
and DMT-M2b should be the same.
(iii) Reduced and hydroxylated metabolites
The TMS derivatives of DVT-M3 (four isomers) had molecular
ions at m/z 550, which was consistent with the addition of two hy-
droxyl groups together with reduction of the carbon double bond
in the A-ring (Fig. 2(c)). Again, diagnostic ions at m/z 155 and m/z
142 confirmed that the D-ring was intact. The presence of a prom-
inent m/z 147 indicated that the two trimethylsiloxyl groups were in
close proximity, possibly formed by reduction and di-hydroxylation
of the carbon double bond in the A-ring. The in vitro studies of the
other five analogues all gave similar M3 metabolites (DMT-M3a
and DMT-M3b, DT-M3, DET-M3a to DET-M3c, DENT-M3a and
DENT-M3b, d3-DMT-M3a and d3-DMT-M3b). The mass spectra of
TMS derivatives of representative M3 metabolites, DMT-M3b
(Fig. 3(c)), DT-M3 (Fig. 4(e)), DET-M3a (Fig. 5(b)), DENT-M3a
J. Mass Spectrom. 2015, 50, 994–1005
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms

Journal of
MASS
SPECTROMETRY
W. H. Kwok et al.
Figure 6. EI mass spectra of trimethylsilylated (a) DENT-M1a, (b) DENT-M2a and (c) DENT-M3a obtained from the in vitro incubation mixture of DENT with
horse liver microsomes. All TMS derivatives were prepared using MND except for DENT-M1a where BSTFA (5% TMCS) was used.
(Fig. 6(c)) and d3-DMT-M3b (Fig. 7(c)), are shown in the correspond-
ing figures. All these spectra showed peaks at their corresponding
molecular ions, m/z 147 (Me3SiO+¼SiMe2), and characteristic pairs
of D-ring fragment ions. It was reported that DMT-M3 was one of
the major urinary metabolites of DMT in human, and the mass spec-
trum of trimethylsilylated DMT-M3b (Fig. 3(c)) matches well with
those reported.[18–20] Moreover, d3-DMT-M3b should have the
same configuration as DMT-M3b because their retention times
and mass spectra matched well with each other except for a shift
of 3 m/z units at their molecular ions and some fragment ions.
modification of the 17α-vinyl group to a 17α-(1,2-dihydroxyethyl)
group. Similarly, DVT-M6 (Fig. 2(f)) and DVT-M7a (Fig. 2(g); two iso-
mers) were assigned to be the 17α-(1,2-dihydroxyethyl) analogues
of DVT-M2 and DVT-M3, respectively. The trimethylsilylated
DVT-M8 had a molecular ion at m/z 726 (Fig. 2(h)). The prominent
ion of [M-205]+ indicated that this metabolite also contained a
17α-(1,2-dihydroxyethyl) group. Based on these data, DVT-M8
was assigned to be a dihydroxylated DVT with a 17α-(1,2-
dihydroxyethyl) group. However, the exact locations of the two hy-
droxyl groups could not be determined. In summary, the proposed
structures of the 20 DVT in vitro metabolites are shown in Fig. 8.
In the case of DMT, three more minor in vitro metabolites DMT-
M4a, DMT-M4b (Fig. 3(d)) and DMT-M5 (Fig. 3(e)) were observed.
DMT-M4a and DTM-M4b were two stereoisomers having similar
mass spectra of their TMS derivatives. The TMS derivatives of
DMT-M4 and DMT-M5 had molecular ions at m/z 626, which was
consistent with the addition of three hydroxyl groups together with
reduction of the double bond in the A-ring (Fig. 3(d) and (e)). The
diagnostic D-ring fragment ions m/z 231 and m/z 218 indicated that
one hydroxylation occurred in the D-ring, and the characteristic ion
m/z 147 indicated that at least two trimethylsiloxyl groups were in
close proximity with each other. Therefore, DMT-M4a and DMT-
M4b were possibly formed from DMT-M3 with additional hydrox-
ylation at C16 in the D-ring. The structures of DMT-M4 were
confirmed by matching their mass spectra with those reported by
Rodchenkov et al.[19] as one of the urinary metabolites in
human.[18,20] A prominent characteristic ion [M-103]+, representing
the loss of a Me3SiOCH2 radical from the molecular ion (m/z 626),
(iv) Additional metabolites
Eight additional in vitro metabolites (DVT-M4 to DVT-M8) were
detected for DVT (Table 1). These metabolites were believed to
arise from the reduction and di-hydroxylation of the 17α-vinyl
group in addition to the three metabolic pathways described earlier.
This metabolic pathway was unique to DVT and was not observed
in the other five analogues studied. The mass spectra of TMS deriv-
atives of representative metabolites DVT-M4 to DVT-M8 (Fig. 2(d)
to (h)) all contained a prominent characteristic [M-205]+ ion,
representing the loss of a Me3SiOCHCH2OSiMe3 radical from the cor-
responding molecular ion. A similar loss of the trimethylsilylated
17α-hydroxyalkyl moiety had been observed for the metabolites
of norethandrolone,[17] 17α-methyltestosterone[21] and turinabol.[22]
Based on their trimethylsilylated mass spectra, DVT-M4b (Fig. 2(d);
two isomers) and DVT-M5a (Fig. 2(e); two isomers) were assigned
to be the enone metabolites of DVT (DVT-M1) with additional
wileyonlinelibrary.com/journal/jms
Copyright © 2015 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2015, 50, 994–1005

Journal of
MASS
SPECTROMETRY
Metabolism studies of DMT and its five analogues
Figure 7. EI mass spectra of trimethylsilylated (a) d3-DMT-M1c (17α-2H3-methyl-1-testosterone), (b) d3-DMT-M2b, (c) d3-DMT-M3b, (d) d3-DMT-M4b and
(e) d3-DMT-M5 obtained from the in vitro incubation mixture of d3-DMT with horse liver microsomes. All TMS derivatives were prepared using MND except
for d3-DMT-M1c where BSTFA (5% TMCS) was used.
was observed in the mass spectrum of trimethylsilylated DMT-M5
(Fig. 3(e)), indicating that the hydroxylation occurred at the
17α-methyl group, similar to the additional metabolites of DVT
(DVT-M4 to DVT-M8). To our knowledge, this novel 17-hydroxymethyl
metabolite DMT-M5 has not been reported previously.
In the case of d3-DMT, three similar minor metabolites d3-DMT-M4a,
d3-DMT-M4b and d3-DMT-M5 were observed. The presence of the di-
agnostic ions m/z 147 (Me3SiO+¼SiMe2), m/z 234 and m/z 221 (charac-
teristic D-ring fragments) in the mass spectrum of trimethylsilylated
d3-DMT-M4b (Fig. 7(d)) further supported its proposed structure.
J. Mass Spectrom. 2015, 50, 994–1005
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms

Journal of
MASS
SPECTROMETRY
W. H. Kwok et al.
Figure 8. A proposed scheme for the phase I metabolism of DVT in horse.
The mass spectrum of tetrakis-trimethylsilylated d3-DMT-M4a was
highly similar to that of d3-DMT-M4b (Fig. 7(d)). In addition, the re-
tention times of trimethylsilylated d3-DMT-M4a, d3-DMT-M4b and
d3-DMT-M5 matched well with those of trimethylsilylated DMT-
M4a, DMT-M4b and DMT-M5, respectively; and their mass spectra
(Fig. 7(d) and (e)) closely resembled those of trimethylsilylated DMT-
M4b (Fig. 3(d)) and DMT-M5 (Fig. 3(e)), respectively, with the corre-
sponding deuterium mass shift at their molecular ions and some
fragment ions. Therefore, the additional hydroxylation in d3-DMT
should also occur at the C16 and 17α-methyl positions in the D-ring.
In contrast, similar 16-hydroxy metabolites were not observed in
the in vitro study of DET, probably owing to steric hindrance from
the 17α-ethyl group. Instead, DET-M4 was detected, with a compa-
rable abundance with those of DET-M1 and DET-M3. Based on the
mass spectrum of trimethylsilylated DET-M4 (Fig. 5(c)), with a
prominent [M-29]+ ion representing the loss of ethyl radical from
the molecular ion (m/z 640), and the characteristic D-ring frag-
ments m/z 157 and m/z 144 showing the intact D-ring, the structure
of DET-M4 was proposed to be tri-hydroxy DET, possibly formed
from DET-M3 with additional hydroxylation at a site other than
the D-ring. In addition, a minor metabolite DET-M5 was observed,
with hydroxylation possibly occurring at the α-carbon of the 17α-
ethyl group, similar to the metabolic pathway of DMT to DMT-
M5. The mass spectrum of trimethylsilylated DET-M5 (Fig. 5(d))
showed a prominent characteristic ion [M-117]+ representing the
loss of a Me3SiOCHCH3 radical from the molecular ion m/z 640, frag-
ment ions m/z 433, m/z 343 and m/z 253 (sequential losses of
HOSiMe3 from [M-117]+), a pair of diagnostic D-ring fragments m/z
245 and m/z 232, and a prominent m/z 147 (Me3SiO+¼SiMe2),
which further supported its proposed structure.
di-hydroxylation with double-bond reduction in the A-ring and (3)
mono-hydroxylation in the A-ring with di-hydroxylation of the re-
duced 17α-vinyl group, were observed in post-administration urine
samples. All three in vivo metabolites were confirmed to be the
same as those identified in the in vitro study. The other 17 in vitro
metabolites identified could not be detected from the post-
administration urine samples using the GC–MS methods in this
study. This was possibly due to their low in vivo concentrations, es-
pecially after extensive distribution and hepatic as well as non-
hepatic metabolism, and interferences from the horse urine matrix.
A phase I metabolic pathway for DVT in horses is proposed in
Fig. 8. The major phase I metabolic processes involved a combina-
tion of enone formation, hydroxylation and reduction in the A-ring
to give DVT-M1, DVT-M2 and DVT-M3, respectively. Upon further
di-hydroxylation of the reduced 17α-vinyl group, DVT-M4 and
DVT-M5, DVT-M6 and DVT-M7 could be formed from respectively
DVT-M1, DVT-M2 and DVT-M3. In addition, DVT-M8 could be
formed by di-hydroxylation of the reduced 17α-vinyl group to-
gether with di-hydroxylation elsewhere in the molecule.
Oral administration study with DVT: phase II metabolism
To study phase II metabolism, unconjugated, glucuronide-conjugated,
sulfate-conjugated and sulfate glucuronide-conjugated metabolites
were fractionated and deconjugated. Enzyme hydrolysis and sol-
volysis were used to deconjugate the urinary metabolites because
DVT was found to be unstable under methanolysis condition.[23]
Metabolites DVT-M2d and DVT-M3a were found exclusively in
the glucuronide fraction, indicating that they were excreted
Table 2. Detection periods for the urinary metabolites of DVT by GC/MS
Oral administration study with DVT: phase I metabolism
Metabolites of DVT
Horse A (h)
Horse B (h)
Urine samples from oral administration of DVT were collected up to
8 days post-administration. The parent drug was not detected.
Three metabolites (DVT-M2d, DVT-M3a and DVT-M6), derived
respectively from (1) mono-hydroxylation in the A-ring, (2)
DVT-M2d
DVT-M3a
DVT-M6
4–22
4–46
4–70
4–22
4–46
4–46
wileyonlinelibrary.com/journal/jms
Copyright © 2015 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2015, 50, 994–1005

Journal of
MASS
SPECTROMETRY
Metabolism studies of DMT and its five analogues
[3] M. H. Sekera, B. D. Ahrens, Y. C. Chang, B. Starcevic, C. Georgakopoulos,
D. H. Catlin. Another designer steroid: discovery, synthesis, and detection
of ‘Madol’ in urine. Rapid Commun. Mass Spectrom. 2005, 19, 781–784.
[4] News. Available at: http://mmajunkie.com/2007/11/strikeforces-dennis-
mainly as glucuronide conjugates. Metabolite DVT-M6 was ex-
creted predominantly in the glucuronide fraction and to a smaller
extent in the sulfate and/or sulfate-glucuronide fraction.
hallman-and-alexander-crispim-test-positive-for-steroids (accessed
9
March 2015)
[5] M. Thevis, U. Bommerich, G. Opfermann, W. Schänzer. Characterization
of chemically modified steroids for doping control purposes by
electrospray ionization tandem mass spectrometry. J. Mass Spectrom.
2005, 40, 494–502.
[6] M. Thevis, H. Geyer, U. Mareck, W. Schänzer. Screening for unknown
synthetic steroids in human urine by liquid chromatography–tandem
mass spectrometry. J. Mass Spectrom. 2005, 40, 955–962.
[7] M. W. F. Nielen, T. F. H. Bovee, M. C. van Engelen, P. Rutgers,
A. R. M. Hamers, J. A. van Rhijn, L. A. P. Hoogenboom. Urine testing
for designer steroids by liquid chromatography with androgen
bioassay detection and electrospray quadrupole time-of-flight mass
spectrometry identification. Anal. Chem. 2006, 78, 424–431.
[8] O. J. Pozo, K. Deventer, P. van Eenoo, F. T. Delbeke. Efficient approach
for the comprehensive detection of unknown anabolic steroids and
metabolites in human urine by liquid chromatography-electrospray-
tandem mass spectrometry. Anal. Chem. 2008, 80, 1709–1720.
[9] O. J. Pozo, N. de Brabanter, A. Fabregat, J. Segura, R. Ventura,
P. van Eenoo, K. Deventer. Current status and bioanalytical challenges
in the detection of unknown anabolic androgenic steroids in doping
control analysis. Bioanalysis 2013, 5, 2661–2677.
[10] G. N. W. Leung, F. P. W. Tang, T. S. M. Wan, C. H. F. Wong, K. K. H. Lam,
B. D. Stewart. In vitro and in vivo studies of androst-4-ene-3,6,17-trione
in horses by gas chromatography–mass spectrometry. Biomed.
Chromatogr. 2010, 24, 744–751.
[11] J. Scarth, A. Clarke, P. Teale, C. Pearce. Comparative in vitro metabolism
of the ‘designer’ steroid estra-4,9-diene-3,17-dione between the
equine, canine and human: identification of target metabolites for
use in sports doping control. Steroids 2010, 75, 643–652.
[12] A. Clarke, J. Scarth, P. Teale, C. Pearce, L. Hillyer. The use of in vitro
technologies and high-resolution/accurate-mass LC–MS to screen for
metabolites of ‘designer’ steroids in the equine. Drug Test. Anal.
2011, 3, 74–87.
[13] W. H. Kwok, E. N. M. Ho, G. N. W. Leung, F. P. W. Tang, T. S. M. Wan,
H. N. C. Wong, J. H. K. Yeung. Metabolic studies of 1-testosterone in
horses. Drug Test. Anal. 2013, 5, 81–88.
[14] G. N. W. Leung, W. H. Kwok, T. S. M. Wan, K. K. H. Lam, P. J. Schiff. Metabolic
studies of formestane in horses. Drug Test. Anal. 2013, 5, 412–419.
[15] E. Houghton, L. Grainger, M. C. Dumasia, P. Teale. Application of gas
chromatography/mass spectrometry to steroid analysis in equine
sports: problems with enzyme hydrolysis. Oreg. Mass Spectrom. 1992,
27, 1061–1070.
[16] B. S. Middleditch, P. Vouros, C. J. W. Brooks. Mass spectrometry in the
analysis of steroid drugs and their metabolites: electron-impact-
induced fragmentation of ring D. J. Pharm. Pharmacol. 1973, 25, 143–149.
[17] A. R. McKinney, D. D. Ridley, C. J. Suann. The metabolism of
norethandrolone in the horses: characterization of 16-, 20- and 21-
oxygenated metabolites by gas chromatography/mass spectrometry.
J. Mass Spectrom. 2001, 36, 145–150.
[18] J. Gauthier, D. Poirier, C. Ayotte. Characterisation of
desoxymethyltestosterone main urinary metabolite produced from
cultures of human fresh hepatocytes. Steroids 2012, 77, 635–643.
[19] G. Rodchenkov, T. Sobolevsky, V. Sizoi. New designer anabolic steroids
from Internet. In Recent Advances in Doping Analysis, W. Schänzer, H.
Geyer, A. Gotzmann, U. Mareck (Eds). SPORTVERLAG Strauss: Koln,
2006, 141–150.
[20] T. Sobolevsky, G. Rodchenkov. Mass spectrometric description of novel
oxymetholone and desoxymethyltestosterone metabolites identified
in human urine and their importance for doping control. Drug Test.
Anal. 2012, 4, 682–691.
Target analytes for detecting DVT administration
Table 2 gives the detection times for the urinary metabolites of DVT
by GC/MS. Both DVT-M3a and DVT-M6 could be the target
analytes to detect DVT administration as they could be detected
consistently in urine from both horses for at least 46 h after admin-
istration. In view of the relatively short detection times for these
metabolites, other alternatives, such as hair analysis and the use
of steroidomics approach to establish a statistical model of bio-
markers profile, are being explored in order to detect the abuse
of DVT and its analogues more effectively in performance horses.
Conclusion
In vitro biotransformation studies of DMT and its analogues (DVT,
DT, DET, DENT and d3-DMT) using horse liver microsomes showed
that the biotransformations occurred predominantly in the A-ring by
way of a combination of enone formation, hydroxylation and reduc-
tion. In the case of DVT alone, additional biotransformations via re-
duction and di-hydroxylation of the 17α-vinyl group were observed;
in the case of DMT, d3-DMT and DET, additional biotransformations
via hydroxylation of the 17α-alkyl group were observed; and in the
case of DMT and d3-DMT, additional biotransformations were ob-
served via hydroxylation at C16. For the oral administration of DVT
to horses, no parent drug was detected. Three urinary metabolites,
namely DVT-M2d, DVT-M3a and DVT-M6, were observed in post-
administration urine, mainly in the glucuronide fraction. The meta-
bolic pathway for DVT was postulated. Metabolites DVT-M3a and
DVT-M6 are the analytes of choice for detecting DVT administra-
tion, as they gave the longest detection time in urine (up to a max-
imum of 70 h post-administration observed). The three urinary
metabolites of DVT could all be produced from in vitro microsomal
incubation. Although in vivo metabolism studies of DMT and its four
analogues (DT, DET, DENT and d3-DMT) in horses have not been
carried out, inclusion of the mass spectra of their in vitro metabolites
in a GC/MS library could permit to detect the use of these designer
steroids in racehorses through library search. Furthermore, an isolate
of in vitro metabolites from a microsomal incubation can also be
used as an acceptable reference material.[24]
Acknowledgements
The authors would like to thank the veterinary surgeons of The
Hong Kong Jockey Club for their assistance in performing drug ad-
ministration experiments to horses and for arranging sample collec-
tion. They also wish to thank Pauly Chan, L. H. Yau, Alan Wan, CoCo
Ng, Christina Tang, Sandra Chan, Stephen Cheung, Celia Wong and
Richard Chan for their technical assistance.
[21] M. C. Dumasia. In vivo biotransformation of 17α-methyltestosterone in
the horse revisited: identification of 17-hydroxymethyl metabolites in
equine urine by capillary gas chromatography/mass spectrometry.
Rapid Commun. Mass Spectrom. 2003, 17, 320–329.
[22] E. N. M. Ho, W. H. Kwok, D. K. K. Leung, T. S. M. Wan, S. Y. Wong. Metabolic
studies of turinabol in horses. Anal. Chim. Acta 2007, 586, 208–216.
[23] P. W. Tang, D. L. Crone. A new method for hydrolyzing sulfate and
glucuronyl conjugates of steroids. Anal. Biochem. 1989, 182, 289–294.
[24] ILAC-G7: 06/2009, Accreditation Requirements and Operating Criteria
for Horseracing Laboratories. Available at: http://accab.org/images/
pdf/ILAC_G7_06_2009.pdf (accessed 9 March 2015).
References
[1] D. H. Catlin, B. D. Ahrens, T. Kucherova. Detection of norbolethone, an
anabolic steroid never marketed, in athletics’ urine. Rapid Commun.
Mass Spectrom. 2002, 16, 1273–1275.
[2] D. H. Catlin, M. H. Sekera, B. D. Ahrens, B. Starcevic, Y. C. Chang,
C. K. Hatton. Tetrahydrogestrinone: discovery, synthesis, and detection
in urine. Rapid Commun. Mass Spectrom. 2004, 18, 1245–1249.
J. Mass Spectrom. 2015, 50, 994–1005
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms