Content uploaded by Dayamin Martinez Brito
Author content
All content in this area was uploaded by Dayamin Martinez Brito on May 19, 2021
Content may be subject to copyright.
SPECIAL ISSUE - RESEARCH ARTICLE
7-keto-DHEAmetabolism in humans. Pitfalls in interpreting the
analytical results in the antidoping field
Dayamin Martinez-Brito
1
| Xavier de la Torre
1
| Cristiana Colamonici
1
|
Davide Curcio
1
| Francesco Botrè
1,2
1
Laboratorio Antidoping FMSI, Rome, Italy
2
Department of Experimental Medicine,
“Sapienza”University of Rome, Rome, Italy
Correspondence
Xavier de la Torre, Scientific Deputy-Director,
Laboratorio Antidoping FMSI, Largo Giulio
Onesti, 1, 00197 Rome RM, Italy.
Email: xavier.delatorre@gmail.com; xavier.
delatorre@fmsi.it
Abstract
7-keto-DHEA (3β-hydroxy-androst-5-ene-7,17-dione) is included in section S1 of the
World Antidoping Agency (WADA) List of Prohibited Substances. The detection of
its misuse in sports needs special attention, since it is naturally present in urine sam-
ples. The main goal of this study is to investigate the in vivo metabolism of 7-keto-
DHEA after a single administration to healthy volunteers and to better describe the
relationship between arimistane (androst-5-ene-7,17-dione) and 7-keto-DHEA after
the application of the common routine procedures to detect anabolic steroids in
WADA accredited antidoping laboratories. Free, glucuro-, and sulpho-conjugated ste-
roids extracted from urine samples obtained before and after the administration of
7-keto-DHEA were analyzed by different gas chromatographic (GC)–mass spectro-
metric (MS) techniques. Gas chromatography coupled to tandem MS to study the
effect on the endogenous steroid profile, coupled to isotope ratio mass spectrometry
(IRMS) to investigate the potential formation of androgens derived from DHEA and
coupled to high resolution accurate mass spectrometry (HRMS) to investigate new
diagnostic metabolites. The analysis by IRMS confirmed that there is no formation of
DHEA from 7-keto-DHEA. Ten proposed metabolites, not previously reported, were
described. These include reduced and hydroxylated structures that are not consid-
ered part of the steroid profile in antidoping analyses. They showed considerable
responses in all fractions analyzed. Some deoxidation reactions (including arimistane
formation) were found and most probably can be linked to the sample preparation or
instrumental analysis. This is important when interpreting the results after the appli-
cation of procedures to detect steroids in urine currently used in antidoping laborato-
ries. 7-keto-DHEA metabolism in humans for antidoping purposes was studied and
unexpected results were found. This could lead to a misinterpretation of the data,
depending on the procedure applied and the analytical instrumentation used.
KEYWORDS
7-keto-DHEA, arimistane, IRMS, mass spectrometry, metabolism
Received: 27 August 2019 Revised: 31 October 2019 Accepted: 31 October 2019
DOI: 10.1002/dta.2734
Drug Test Anal. 2019; :1629–1643. wileyonlinelibrary.com/journal/dta © 2019 John Wiley & Sons, Ltd. 1629
11
1|INTRODUCTION
7-keto-DHEA (3β-hydroxy-androst-5-ene-7,17-dione or 7-oxo-
DHEA) is included in the section S1. Anabolic Androgenic Steroid
(AAS), Subgroup B. Endogenous AAS and their Metabolites and Iso-
mers, when Administered Exogenously, of the World Antidoping
Agency (WADA) List of Prohibited Substances.
1
7-keto-DHEA is
included in this group since it can be produced endogenously and it
could also be a metabolite of an endogenous anabolic androgenic ste-
roid. In addition, it is also available as a dietary supplement. No criteria
for performing an isotope ratio mass spectrometric (IRMS) confirma-
tion exist, as is instead required by WADA for other pseudo-
endogenous steroids.
2
In human metabolism, 7-keto-DHEA is produced from DHEA via
its 7α-hydroxylated derivative followed by its oxidation. It does not
activate the androgen receptors, is not converted to androgen and
because of its carbonyl function on the 7 position is not a substrate
for aromatase to form estrogens. It is approximately 2.5 times more
active than DHEA as an inducer of thermogenic enzymes such as
mitochondrial glycerophosphate dehydrogenase and cytosolic malic
enzyme. It enhances interleukin-2 production by human mononuclear
leukocytes.
3,4
The available scientific literature on human metabolism of this
compound is mainly focused on the reduction of the ketone at the C7
position leading to the formation of 7β- and 7α-hydroxylated isomeric
metabolites. It has been reported that the 7β-isomer is formed in
greater proportion and that its activity is at least five times higher
than its 7α-isomer and its 7-keto precursor.
5,6
Due to its steroidal structure, presumably 7-keto-DHEA
undergoes an extensive phase I metabolism characterized by reduc-
tions and/or oxidations and the formation of hydroxylated metabo-
lites in all possible combinations. In the context of doping analysis, 7α-
and 7β-hydroxylated metabolites and 7ξ-hydroxy-androstenedione
have been described as the main in vivo metabolites,
5-7
while other
authors have described hydroxylations in C9α,C1β, and C12βunder
in vitro conditions.
8
Reviewing the anecdotal information available on some websites
marketing this supplement, arimistane (androst-3,5-diene-7,17-dione)
is postulated to be the main metabolite of 7-keto-DHEA. To the best
of our knowledge, no peer-reviewed reference is available to date.
Arimistane is the 3-desoxy reduced derivative of 7-keto-DHEA and it
is included in section S4. Hormone and Metabolic Modulators. Sub-
group 1. Aromatase Inhibitors of the WADA Prohibited List.
1
On the other hand, several authors have described the synthesis
of androst-3,5-diene-7-one related structures from androst-
5-ene,3–7-dione or androst-5-ene-3β-ol-7-one structures under
acidic conditions.
9,10
At the same time, another group demonstrated
the competitive inhibition of the mitochondrial aromatase activity of
androstan-3,5-diene-7,17-dione; androstan-3,5-diene-7-one-17-ol;
androstan-5-ene-7,17-dione, and androstan-5-ene-7-one-17-ol struc-
tures. The structures with an hydroxyl group in C17 showed a less
inhibitory effect than 17-one structures, but in any case the
configuration androst-5-ene-7-one is present,
11-13
confirming the
adequate inclusion of arimistane in the section of the WADA
prohibited list.
The analysis of the urinary steroids sulfate fraction has been
described widely and has been gaining more attention in recent years
in the antidoping field.
14-18
The use of either enzymatic and/or chemi-
cal hydrolysis has its specific advantages and disadvantages. Shackle-
ton reported that most steroid sulfates can be hydrolyzed
enzymatically, although this is probably not the universal method of
choice for all the compounds of interest. Solvolysis may in some
instances be preferable as it only hydrolyzes sulfates and does not dis-
criminate the steroid disulfates.
19
Several authors have published results after solvolysis by the use
of ethyl acetate, methanol, and sulfuric acid in different proportions,
also considering the effect of the temperature and time of incubation.
Most publications describe no pH adjustment before the solid phase
extraction (SPE) previous to the hydrolysis; if so, an acid buffer was
used.
14-17,20,21
Finally, others described the use of glacial acetic acid
before the solvolysis, but in some cases, acetylated artifacts were
reported to have been produced.
18
At this point, a careful revision of
the data must be done in order to differentiate the proposed real
metabolites from potential artifacts formed during the sample
preparation.
Figure 1 shows the relationship among the chemical structures of
DHEA, 7-keto-DHEA, and arimistane as well as the main metabolic
routes. The main goal of this paper was to investigate the in vivo
metabolism of 7-keto-DHEA after a single oral administration and
additionally to better describe the relationship between arimistane
and 7-keto-DHEA after the application of the common routine proce-
dure to detect anabolic androgenic steroids in WADA accredited anti-
doping laboratories.
2|EXPERIMENTAL
2.1 |Reference material and reagents
Androst-5-ene-3β-ol-7,17-dione (7-keto-dehydroepiandrosterone,
7-keto-DHEA), androst-5-ene-3β,7α-ol-17-one (7α-OH-DHEA),
androst-5-ene-3β,7β-ol-17-one (7β-OH-DHEA), 5α-androstane-
3α,17β-diol (5α-Adiol), and 5β-androstane-3α,17β-diol (5β-Adiol) were
supplied by Steraloids Inc (Newport, USA). Androst-5-ene-3β-ol-
17-one (dehydroepiandrosterone, DHEA), androst-4-ene-3-one-17β-
ol (testosterone, T), testosterone-D3 (T-D3), androsterone (A),
etiocholanolone (Etio), and 17α-methyltestosterone (MT) were pur-
chased from Sigma-Aldrich (Milan, Italy). Androst-4-ene-3-one-17α-ol
(epitestosterone, E), etiocholanolone-D5 (Etio-D5), androsterone-D4
(A-D4), and androsterone-D4 glucuronide were purchased from the
National Measurement Institute (North Ryde, Australia) and androst-
3,5-diene-7,17-dione (arimistane) was purchased from MuseChem
(CA, USA). Androst-3,5-diene-7β-ol-17-one (arimistane metabolite)
was supplied by WAADS (World Association of Antidoping Scientists).
MARTINEZ-BRITO ET AL.
1630
2.1.1 |Sulfate reference material used to optimize
the solvolysis incubation time
Etiocholanolone D5 (Etio-D5) sulfate, DHEA sulfate, etiocholanolone
sulfate, testosterone sulfate, and epitestosterone sulfate were pro-
vided by NMI (North Ryde, Australia) and 11β-hydroxy-
etiocholanolone sulfate, epiandrosterone sulfate, and 5-androstene-
3β,17β-diol disulfate by Steraloids (Newport, USA).
All reagents and solvents (sodium bicarbonate, potassium carbon-
ate, sodium phosphate, sodium hydrogen phosphate, tert-butylmethyl
ether, and ammonium hydroxide) were of analytical or HPLC grade
and provided by Carlo Erba (Milan, Italy). Enzyme β-glucuronidase
from Escherichia coli K12 was from Roche Diagnostic (Mannheim, Ger-
many). Water was from a Milli-Q water purification system (Millipore
S.p.A, Milan, Italy). Mercaptoethanol (2-ME), ammonium iodide, and
N-methyl-N-trimethylsilyl-trifluoracetamide (MSTFA) were purchased
from Sigma-Aldrich (Milan, Italy), Macherey-Nagel (Germany), and
Sigma-Aldrich (MO, USA), respectively.
2.2 |Description of the urinary excretion samples
Urinary excretion studies were carried out with the participation of
two healthy volunteers (male and female, 46 and 51 years old, 69 and
88 kg body weight respectively, normal BMI in both cases). A single
oral dose of 100 mg of 3-acetyl-7-keto-DHEA (Now Foods,
Bloomingdale, IL, USA) was administered. Samples were collected
before administration and up to 48 hours (male) and 77 hours (female)
post-administration. The samples were kept frozen until analysis. This
observational study fulfills the recommendations for research involv-
ing human subjects described in the Declaration of Helsinki.
22
Signed
informed consent was obtained from the volunteers after approval of
the Ethical Committee (Lazio 1) of the Lazio Region (Rome, Italy).
2.3 |Evaluation of the capsule content
The content of the capsule was dissolved in methanol to reach a con-
centration of 10 μg/mL. Then 5 μL of the solution was injected on an
Agilent 1290 Infinity II LC System coupled to an Agilent Ultivo LC/TQ;
analytical column Eclipse Plus C18 (50 mm length ×2.1 mm internal
diameter ×1.8 μm particle size); the mobile phase flow was
0.4 mL/min; mobile phase A: water (0.1% formic acid) and B: methanol
(0.1% formic acid): A(90%) –B(10%) for 1 min, A(5%) –B(95%) in
5 min total run 7 min. The ion source was AJS ESI, fragmentor 135 V,
acquisition in positive ionization and scan mode (m/z 55–500); scan
time 500 ms. The nebulizer nitrogen gas pressure was set at 40 psi.
Simultaneously reference material arimistane (retention time
4.20 min) and 7-keto-DHEA (retention time 3.9 min) were analyzed.
No presence of arimistane nor any other impurity was observed in the
capsule content. The results agree absolutely with the observation of
Lorenz et al.
23
FIGURE 1 Relationship among
DHEA, 7-keto-DHEA, and arimistane
chemical structures and probable
metabolic routes. Inside the dashed line
are the most described target metabolites
for 7-keto-DHEA (7α- and 7β-hydroxy-
DHEA)
MARTINEZ-BRITO ET AL.1631
2.4 |Optimization of the solvolysis incubation
time
In order to evaluate a suitable incubation time, a mix of 1 μg of DHEA
sulfate, Etio sulfate, T sulfate, E sulfate, 11β-hydroxy-Etio sulfate,
EpiA sulfate, and 5-androstene-3β,17β-diol sulfate was incubated for
3 hours at 55C in 1 mL of the solvolysis mixture (ethyl
acetate/methanol/sulfuric acid, 80: 20: 0.12, v:v:v). Measurements
were done after 1, 1.5, 2, 2.5, and 3 hours of hydrolysis (n = 3) and
after neutralization with ammonium hydroxide, liquid–liquid extrac-
tion (LLE) with tert-butylmethyl ether (TBME) of the deconjugated
steroids. The TMS derivatives were analyzed by GC-TOF and the
absolute areas were used to evaluate the responses after 1 and
3 hours by the application of Student's t-test (α= 0.05).
The internal standard for sulfate analysis (ISTD) was prepared
from methanolic solutions of MT, T-D3, A-D4, and Etio-D5 sulfate.
The recovery of these compounds was evaluated by comparison of
aliquots (n = 3) after the solvolysis (at 1, 1.5, 2, 2.5, and 3 hours) and
LLE, with the same amount analyzed without solvolysis nor LLE
(n = 3).
2.5 |Experiment to verify the formation of
arimistane from 7-keto-DHEA
Reference material of 7-keto was derivatized with decreasing propor-
tions of MSTFA: ammonium iodine: 2-mercaptoethanol (1000:4:6;
1000:2:3; 1000:1:1.5; 1000:0.5:0.75; 1000:0.25:0.38; v:w:v) and
MSTFA (n = 3 each). The replicates of the derivatized 7-keto-DHEA
were immediately injected onto GC-QTOF.
2.6 |Sample treatment
The analysis of the free plus glucuronated (F + G) fraction of the uri-
nary steroids for antidoping analysis has been described earlier.
24
Briefly, to 2 mL of urine, ISTD containing among others MT, T-D3, A-
D4-glucuronide, and Etio-D5 was added. Then the pH was adjusted
using phosphate buffer (750 μL, 0.8 M, pH 7) to carry out the hydroly-
sis of glucuronates by β-glucuronidase from E. coli (50 μL). The sam-
ples were incubated for 1 h at 55C and then the pH was elevated to
9–10 using carbonate/bicarbonate buffer (0.5 mL, 20%). Double
liquid–liquid extraction was carried out with TBME (2 ×4 mL) and the
remaining aqueous phase was conserved. The dry residue was
derivatized with MSTFA/NH
4
I/2-ME (1000:2:6, v:w:v) at 70C for
30 min to form the per-trimethylsilyl (enol-TMS) derivatives. The
hydrolysis efficiency was monitored by evaluating the A-D4 (from A-
D4 glucuronide) to Etio-D5 ratio in all samples.
The remaining aqueous phase from the previous step was used
for the analysis of the sulfate (S) fraction by solvolysis in an organic
solvent. The residual TBME was evaporated under a stream of nitro-
gen and then solid phase extraction was conducted on C18 columns
(Sep-Pak
®
C18, Waters SpA, Milan, Italy) after adding the ISTD
mixture for the sulfate fraction. The solvolysis using a mixture of ethyl
acetate/methanol/sulfuric acid (1 mL, 80: 20: 0.12, v:v:v) was done by
incubating the samples 1.5 hours at 55C. The pH was neutralized
using ammonium hydroxide and a double liquid–liquid extraction was
done using TBME (2 ×4 mL). The evaporated residue was derivatized
using the same derivatization mixture described before at 70C for
30 min to form the pertrimethylsilyl (enol-TMS) derivatives. The
hydrolysis efficiency was monitored by evaluating the A-D4 to Etio-
D5 (from Etio-D5 sulfate) ratio in all samples.
25
3|INSTRUMENTAL
3.1 |Endogenous steroid profile
The endogenous steroid profile was measured using an Agilent 7890A
gas chromatographer coupled to an Agilent 7000 triple quadrupole
mass spectrometer (Agilent Technologies SpA, Cernusco sul Naviglio,
Milan, Italy). The chromatography was performed using a HP1MS
(Agilent J&W, CPS Analitica, Milano, Italy) methyl fused-silica capillary
column (17 m ×0.2 mm, 0.11 μm film thickness). The initial tempera-
ture was 188C for 2.5 min, then increased at 3C/min to 211C
maintained 2 min, ramped at 10C/min to 238C, then ramped at
40C/min to 320C, and kept for 3.5 min at the final temperature.
Helium was used as a carrier gas at constant pressure (1 bar) and 2 μL
of extract was injected in split mode (1:20). Both the injector and
transfer line were operated at 280C. The data acquisition was per-
formed in multiple reaction monitoring (MRM). This method was used
for the determination of the steroid profile as defined in the WADA
technical document TD2018EAAS which is routinely used by the
WADA accredited laboratories. The steroid profile included quantita-
tion of T, E, A, Etio, 5α-Adiol, 5β-Adiol, and the ratios T/E, A/Etio,
A/T, 5α/5β-Adiol, and 5α-Adiol/E. The method is approved by the
National Body of Accreditation of Italy as part of the laboratory
17025 scope of accreditation.
3.2 |Metabolism study
The metabolism study was carried out on an Agilent 7890B gas chro-
matograph coupled to a 7200 Accurate Mass Q-TOF GC/MS system.
The chromatographic conditions were as the same as described
before. The acquisition was in SCAN mode in the range from 50 to
750 Da, scan time 500 ms/spectrum. The molecular formulae of the
proposed structures were based on the measurement of the accurate
mass and the difference (Δ) between the experimental and theoretical
mass expressed as ppm. Errors below 5 ppm were considered accept-
able. Characteristic mass fragments were used as additional evidence
of the proposed structures. The combined information derived from
enol-TMS and MSTFA (only hydroxyl group derivatized) allowed for a
preliminary structural elucidation.
In order to compare the relative responses on the excretion of
the proposed metabolites detected, and since no reference materials
of most of the compounds are available, the results are presented as
MARTINEZ-BRITO ET AL.
1632
the relative response of the metabolite base peak to the response of
MT added in the samples at a concentration of 250 ng/mL.
3.3 |HPLC sample purification
For the evaluation of the isotopic ratio content of the parameters of
the steroid profile a method published earlier was used.
26
Briefly, uri-
nary extracts from samples at 0, 6, and 13 hours after the administra-
tion, were purified on an Agilent 1100 Series liquid chromatograph
(Agilent Technologies SpA, Cernusco sul Naviglio, Milan, Italy). The
selected fractions were collected by an Agilent 1100 fraction collec-
tor. The separation conditions were established by monitoring the sig-
nal of a UV lamp at 192 nm (Agilent 1100 UV DAD detector). Sample
purification to obtain adequate extracts of the target compounds
(TCs) and endogenous reference compounds (ERCs) was performed
using a C18 column + guard column from ACE (CPS Analitica, Milan,
Italy) (25 cm, 4.6 mm, 5 μm and 2 cm, 4.6 mm, 5 μm) at 38C. Separa-
tion was programmed with a mobile phase composed of water (sol-
vent A) and acetonitrile (solvent B). The flow rate was set at
1 mL/min. The collected fractions were taken to dryness under a
nitrogen stream at 70C. Before analysis by GC-C-IRMS, the fractions
were dissolved with an adequate volume of a mixture cyclohexane:
isopropanol (4:1, v/v) to reach a response on the linear range of the
procedure.
Additionally, for better elucidation of the proposed structures
during the metabolic study, a clean-up step by HPLC was applied to
two samples extracted in duplicate: blank urine and sample collected
3 hours after 7-keto-DHEA administration. These four aliquots were
extracted as described previously for F + G and S conjugates. After
injection into the HPLC with the same conditions described previ-
ously, 1 min fractions were collected for a total of 38 fractions (F1 to
F38). One set of fractions for each sample was derivatized using the
same protocol described earlier (MSTFA/NH
4
I/2-ME) and the other
one with only MSTFA (50 μL for 30 min at 70C) with the aim of dis-
tinguishing the hydroxyl and keto groups present in the proposed
metabolites. All extracts were analyzed by GC-TOF.
3.4 |GC-C-IRMS analyses
GC-C-IRMS analyses were performed on an HP6890 gas chromato-
graph connected to a combustion furnace linked to a Thermo Delta
Advantage isotope ratio mass spectrometer through an Isolink-
Conflow IV interface (ThermoElectron, Bremen, Germany). The chro-
matography was performed in a HP5MS (J&W Scientific, CPS
Analitica, Milan, Italy) 5% phenylmethyl fused-silica capillary column
(30 m ×0.25 mm i.d. ×0.25 μm film thickness). The oven temperature
program was the following: initial temperature 150C for 1 min,
increased at 25C/min to 260C maintained 4 min, increased at
25C/min to 310C, and kept 2.7 min at the final temperature. Helium
was used as the carrier gas at 2.1 mL/min and injection, 2 μLof
extract, was performed in splitless mode at 280C.
26
4|RESULTS AND DISCUSSION
4.1 |Endogenous steroid profile and
measurements of δ
13
C
It has been reported that 7-keto-DHEA is not a biologically active pre-
cursor of endogenous androgens or estrogens, nevertheless, it is con-
sidered to be a marker of DHEA administration under the current
WADA criteria. The collection of the samples covered a period of
48 hours (volunteer 1) and 77 hours (volunteer 2), therefore at least
two circadian cycles were included in the study by which the influ-
ence of normal variations can be ruled out. After the steroid profile
analysis by GC–MS/MS, fluctuations in the urinary concentrations
could not be correlated either with the oral administration of 7-keto-
DHEA or exogenous pseudoendogenous steroids, but with the normal
fluctuation of the circadian cycle. Ratios between T/E, A/Etio, and
5α/5β-Adiols also showed no alteration.
The concentrations of DHEA and its main 7-hydroxylated metab-
olites were monitored as part of the alternative hydroxylated profile
for endogenous steroids analyzed in GC–MS/MS. Figure 2 shows the
time-course of DHEA, 7α-hydroxy-DHEA, 7β-hydroxy-DHEA, and
7-keto-DHEA in the F + G fraction of urine. Differences in the excre-
tion profile of DHEA were observed for the two volunteers. While the
male showed no alterations, the female volunteer showed increased
concentrations up to 450 ng/mL in the F + G fractions that returned
to basal at 10 hours after the administration.
Both subjects showed an apparent alteration in the urinary con-
centrations of 7α-hydroxy-DHEA compatible with an oral administra-
tion, but by comparing the data obtained after the analysis of the
steroid profile by GC–MS/MS and the analysis by GC-TOF, interest-
ing information emerged. The RT of 7α-OH-DHEA (13.54 min, vali-
dated qualifier transition m/z 415 > 325) is very close to androst-
4,6-diene-3,17-dione (13.56 min). The molecular mass for 4,6-diene-
TMS derivative is m/z 428, and the main fragments in its MS corre-
spond to M
+
−CH
3
(m/z 413) and M
+
−[OTMS+CH
3
] (m/z 323). The
isotopic contribution of the ions m/z 413 to m/z 415 and m/z 323 to
m/z 325 is evident when a positive sample is analyzed by GC–
MS/MS. Therefore, it is easy to misinterpret the results because there
is an apparent alteration of the 7α-OH-DHEA when, actually, the
response belongs to a different analyte. The presence of androst-4,-
6-diene-3,17-dione (instead of the 7α-OH-DHEA), in samples col-
lected after the administration of 7-keto-DHEA, was confirmed by
GC-TOF with the simultaneous analysis of the suitable reference
material (Figure 3). The origin of androst-4,6-diene-3,17-dione will be
presented below.
The fact that the 7β-hydroxy-isomer showed no visible alteration
in the concentration (for both volunteers) does not agree with previ-
ous publications about 7-keto-DHEA metabolism.
27-29
The results
obtained by GC–MS/MS are supported by the measurements in GC-
TOF, and no explanation as was observed for the 7α-isomer could be
found.
An increment of the 7-keto-DHEA was observed as expected.
In order to investigate the potential formation of DHEA from
MARTINEZ-BRITO ET AL.1633
7-keto-DHEA, IRMS measurements (δ
13
C‰) of A, Etio, T, 5α-Adiol,
5β-Adiol, and DHEA were done. Their values, as well the differences
with two ERC showed “no positivity to any exogenous administra-
tion”(Table 1 shows the results using only pregnanediol as endoge-
nous reference compound). The scientific literature reported that no
androgenic (or estrogenic) effects were observed after the adminis-
tration of 7-keto-DHEA, so the results based on GC/C/IRMS in the
present study seem to be in agreement with these observations
3,4
although additional studies with a larger number of individuals are
needed.
Once the formation of DHEA from 7-keto-DHEA was discarded
based on the IRMS results, the difference in the urinary DHEA con-
centrations between the volunteers could only have a pharmacologi-
cal explanation based on the differences between males and females.
FIGURE 2 Urinary concentration time course for DHEA, apparent 7α-hydroxy-DHEA, 7β-hydroxy-DHEA, and 7-keto-DHEA in free +
glucuronated fraction for volunteer 1 (M, male) and volunteer 2 (F, female) measured in the routine GC–MS/MS procedure for establishing the
steroid profile
MARTINEZ-BRITO ET AL.
1634
It is known that the variety of metabolites and their concentrations
depends on the enzymatic and genetic endowment of each individual.
Moreover, the depletion of cofactors and enzymes (generally pro-
voked by xenobiotics) favor the use of alternative metabolic routes. In
the specific case of DHEA, numerous references describe the differ-
ences in its metabolism in women and men. For example, it has been
reported that women excreted nearly twice as much non-glucuronide
and non-sulfate conjugates in the urine compared with men. Several
authors have also described the sex difference in the plasma DHEA to
DHEA-S ratio (higher in women) that may well be explained on the
basis of greater in vivo cleavage of DHEA-S to DHEA in women than
in men.
30-34
FIGURE 3 (A) Chromatogram of the transition m/z 415 > 325 at 13.54 min. Screened for 7α-OH-DHEA for a positive control and reference
material of androst-4,6-diene-3,17-dione and 7α-OH-DHEA after the analysis by GC–MS/MS, (B.1) mass spectrum of 7α-OH-DHEA reference
material molecular ion at m/z 520.3228, (B.2) mass spectrum of androst-4,6-diene-3,17-dione reference material, (B.3) mass spectrum of in a
sample collected after the administration of 7-keto-DHEA analyzed by GC-TOF, and (B.4) extracted chromatogram of theoretical mass m/z
520.3224 ± 20 ppm for the same sample in B.3 showing no presence of 7α-OH-DHEA after the administration of 7-keto-DHEA analyzed by GC-
TOF
MARTINEZ-BRITO ET AL.1635
In summary, in a real doping scenario and by interpretation of the
List of Prohibited Substances (section S1.b), the concentrations found
for 7-keto-DHEA after the analysis of routine GC–MS/MS, should be
considered as a “normal individual physiological level”since the results
obtained by GC/C/IRMS excluded any exogenous administration of
DHEA or other pseudoendogenous compounds.
On the other hand, the use of 7-keto-DHEA as a target metabo-
lite probably requires a study to establish the population reference
intervals in order to set a concentration threshold in the WADA TD-
EAAS, as is presently established for T and E (200 ng/mL) or A and
Etio (10 000 ng/mL). This is, however, not free of difficulties since
many parameters, such as sex, age, ethnicity, genetic polymorphism,
among others, should be considered; and reference values for 7-keto-
DHEA are described for a Caucasian population only.
35
In order to
avoid this work, it is necessary to search outside the “common”
endogenous steroid profile to differentiate the administration of
DHEA from 7-keto-DHEA. This search included the free + glucuronide
(F + G) and sulfate (S) fractions of the urine.
4.2 |Optimization of the solvolysis incubation
time and stability under solvolysis conditions
A set of experiments was conducted in order to assess the opti-
mum incubation time of hydrolysis. The results obtained could be
used to evaluate the stability of the compounds as well. Theoreti-
cally, a decreasing response of the metabolites with time could
indicate a degradation process. On the contrary, an increment of
the response could mean incomplete hydrolysis of the sulfates and
then a modification of the experimental protocol would probably
be needed.
Despite the same initial concentration of the analytes, the differ-
ences observed are due to the response factors on the chromato-
graphic capillary column and the mass spectra and because the ratio
to MT calculation was done using the base peak. The application of
Student's t-test (α= 0.05) showed no significant differences between
the absolute areas of the free analytes after 1 and 3 hours of hydroly-
sis. This result confirms also the stability of the compounds over
3 hours of hydrolysis, liquid–liquid extraction (LLE), and derivatization
steps.
The responses of MT, T-D3, and And-D4 (ISTD) after 1, 1.5,
2, 2.5, and 3 hours of hydrolysis (n = 3) was compared with the
response of the same analytes without solvolysis. Assuming the
responses of the analytes without solvolysis (nor LLE) as 100%, the
recovery could be evaluated preliminarily. Figure 4B shows the results
where stability is evident. The Student's t-test (α= 0.05) showed no
significant differences between the responses after 1 and 3 hours of
hydrolysis. Recovery was observed to be between 98–112% for MT,
89–98% for T-D3, and 80–93% for A-D4.
The hydrolysis incubation time was set to 1 hour in the study of
the S fraction of urine collected after the 7-keto-DHEA administra-
tion. Notwithstanding the satisfactory statistical analysis, low amounts
(lower than 4%) of acetylated by-products were found after the first
hour. This issue was considered not significant, since the aim of the
study was not quantitative.
4.3 |Metabolism of 7-keto-DHEA
Considering the known metabolic routes for steroid metabolism,
redox reactions, hydroxylations, and isomerization reactions are
expected. In addition to glucuronated metabolites, the 3β-
TABLE 1 Results obtained after the application of isotope ratio mass spectrometry (IRMS) assay to the parameters of the endogenous steroid
profile to two volunteers after the administration of the single oral dose of 7-keto-DHEA
Time
(h)
A
(δ
13
C)
Etio
(δ
13
C)
T
(δ
13
C)
5α-Adiol
(δ
13
C)
5β-Adiol
(δ
13
C)
DHEA
(δ
13
C)
PD*
(δ
13
C)
11-keto-et*
(δ
13
C)
Volunteer 1 (male) 0 −22.7 −23.3 −22.7 −24.2 −23.5 −21.6 −23.6 −23.3
6−22.6 −23.4 −22.9 −23.8 −23.5 - −22.9 −23.2
13 −22.5 −23.3 −22.8 −23.5 −23.2 −21.9 −22.6 −22.8
Δδ-PD
0 0.9 0.3 −0.9 −0.6 0.2 2.0 Negative sample
6 0.4 −0.4 0.0 −0.9 −0.5 -
13 0.2 −0.6 0.2 −0.9 −0.6 0.8
Volunteer 2
(female)
0−21.6 −21.9 - −23.2 −23.9 −20.4 −21.3 −21.7
6−21.6 −22.5 - −23.0 −24.3 −21.0 −21.8 −22.5
13 −21.4 −22.1 - −22.7 −23.8 −21.0 −21.4 −21.7
Δδ-PD
0−0.3 −0.6 - −1.8 −2.7 0.9 Negative sample
6 0.2 −0.7 - −1.2 −2.5 0.8
13 −0.02 −0.7 - −1.3 −2.3 0.4
*
Endogenous reference compound (ERC).
MARTINEZ-BRITO ET AL.
1636
configuration of 7-keto-DHEA suggests a considerable formation of
sulfoconjugated metabolites. The analysis of the S fraction revealed
several interesting compounds after the administration of 7-keto-
DHEA. However, the signals observed in the first part (ca 7 min) of
the chromatogram showing m/z 340, could be degradation products
from 7-keto-DHEA and/or from its metabolites, probably produced
by the acidic conditions of the solvolysis.
4.4 |Free plus glucuronated urinary fraction
In the free + glucuronated (F + G) urinary fraction, some metabolites
eluted at a RT close to a very complex zone of hydroxylated endoge-
nous steroids. In addition, structures of some poly-hydroxylated
metabolites at higher RT were observed as well. Table 2 shows the
most significant metabolites found in urine after a single oral adminis-
tration of 7-keto-DHEA, which were not present in negative urine col-
lected before the administration or they could be present below the
detection capacity of the procedure. Most of the metabolites pro-
posed here have not been described in previous investigations on
DHEA,
4,36-39
so they could be interpreted as specific metabolites of
7-keto-DHEA.
4.5 |Reduced metabolites of 7-keto-DHEA
The most interesting compound observed was the probable 3α-isomer
of 7-keto-DHEA itself (I). Its presence is possible by the formation of
an intermediate with 3-keto configuration as described for the iso-
meric analytes after the activity of 3βand 3α-hydroxysteroid dehydro-
genases. The intermediate structure of androsten-3,7,17-trione was
observed at 15.8 min showing a time-course consistent with an oral
administration (C
28
H
48
O
3
Si
3
, m/z 516.2893, Δ. 2.9 ppm). The maxi-
mum response of the 3α-isomer was estimated at 3–5μg/mL in the
F + G fraction, and it was not observed in the S fraction, corroborating
its 3α- configuration. The last sample collected by volunteer 1 (77 h)
showed a relative response close to 60 ng/mL, while 7-keto-DHEA
was not detected after 23 h after the administration. The responses in
the last collected sample for volunteer 2 were estimated close to
150 and 10 ng/mL for 3α-isomer and 7-keto-DHEA, respectively.
Having a 3β- configuration, it would be expected that higher concen-
trations of 7-keto-DHEA would be found in the S fraction, but the
levels were about 85% of those found in the F + G fraction (see expla-
nation below). Figure 4 shows the mass spectra of 7-keto-DHEA and
its proposed 3α-isomer in the F + G fraction. After the purification of
the samples by HPLC, prior and 7 hours after the administration, this
3α-isomer could be observed in fraction 2 (F2). The trimethylsilylation
(without enolization) of the compound supported the postulated
structure (Table 2).
Two reduced structures from 7-keto-DHEA (II and III) were
observed and it can be assumed as well that the configuration is 3α-
hydroxy because their excretion was only in the F + G fraction. The
similarity among the MS spectra of these metabolites, 7-keto-DHEA,
and the 3α-isomer is remarkable, because of the delocalization of one
electron in the conjugated system 5-ene-7-one (Figure 5). In the F1
fraction, where only hydroxyl groups are derivatized (MSTFA), two
FIGURE 4 Extracted chromatogram for ion at m/z 517.3 (± 20 ppm) and mass spectra of the 3α-isomer of 7-keto-DHEA (I, 15.52 min) and
7-keto-DHEA (15.6 min) in free + glucuronated fraction of urine
MARTINEZ-BRITO ET AL.1637
compounds at m/z 448 were detected, supporting the proposed
structures as androst-5-ene-3α,17ξ-diol-7-one. The accurate mass is
compatible with the presence of one keto group and two hydroxyl
groups. The effect of the conjugated π-electron system resulting in
charge delocalization is additional evidence of the presence of
androst-5-ene-7-keto configuration.
40,41
A similar observation was
made in 7a-hydroxy-testosterone tris-O-TMS resulting in a
3,7-dihydroxy-3,5-diene structure.
The presence of the ion at m/z 169 is generally linked to a
trimethylsilylated 17-keto group. The mass spectra of the proposed
metabolites II and III showed the non-relevant presence of this ion
because of the proximity of two analytes showing an intense ion at
m/z 169. Figure 5 shows the extracted mass chromatogram at m/z
519 and m/z 169 as well as the mass spectra of the metabolites.
Based on this, we could hypothesize the presence of one hydroxy
group in C17 instead of a 17-keto group. These “nearby metabolites”
showed higher responses but the presence of common ions in their
mass spectra with endogenous compounds (m/z 520) diminish the
specificity of the results, at least under our assay conditions.
Being two isomers with similar mass spectra, the 17β-hydroxyl (II)
configuration was assigned to the peak showing a higher response
since excretion as glucuronate is favored. In any case, additional ana-
lytical techniques are required to establish the correct configuration
of the structures.
The analyte showing the highest response relative to MT was not
observed in the blank urine samples collected before the administra-
tion and it was estimated at 17.4 μg/mL (volunteer 1, t
max
5.5 h) and
4.4 μg/mL (volunteer 2, t
max
3 h). After the HPLC purification, this
compound (IV) could be observed in F1 and F2. The strong intensity
of the ion at m/z 169 indicates the presence of the C17-oxo group.
Besides, the accurate mass corresponded to a structure showing two
hydroxyl groups and two unsaturations, one of them is the C17-keto
group and the other is supposed to be at Δ5–6. Because it was found
only in the F + G fraction, the 3α- configuration can be proposed.
TABLE 2 Proposed metabolites found in urine after the administration of a single oral dose of 7-keto-DHEA
ID Description Derivative/fraction Formula
RT
(min)
Theor. Mass
(Da)
Exp. mass
(Da)
Error
(ppm)
I7-keto-dehydroepiandrosterone (3α-isomer of
7-keto-DHEA)
TMS-enol C
28
H
49
O
3
Si
3
15.49 517.2989 517.2978 2.2
TMS/F2 C
22
H
34
O
3
Si 15.16 374.2277 374.2286 −2.3
II Androst-5-ene-3α,17[β]-diol-7-one TMS-enol C
28
H
51
O
3
Si
3
15.67 519.3146 519.3146 0.0
TMS/F1 C
25
H
44
O
3
Si
2
15.90 448.2829 448.2842 −2.9
III Androst-5-ene-3α,17[α]-diol-7-one TMS-enol C
28
H
51
O
3
Si
3
15.77 519.3146 519.3140 1.16
TMS/F1 C
25
H
44
O
3
Si
2
15.78 448.2829 448.2833 −0.9
IV Androst-5-ene-3α,7ξ-diol-17-one TMS-enol C
28
H
52
O
3
Si
3
15.20 520.3224 520.3224 0.05
TMS/F1-F2 C
25
H
44
O
3
Si
2
14.33 448.2828 448.2826 0.67
V5α-androstane-3α,7ξ-diol-17-one (7ξ-OH-A) TMS-enol C
28
H
54
O
3
Si
3
14.50 522.3380 522.3383 −0.43
TMS/F2 C
25
H
46
O
3
Si
2
13.85 450.2985 450.2982 0.77
VI 5β-androstane-3α,7ξ-diol-17-one (7ξ-OH-Etio) TMS-enol C
28
H
54
O
3
Si
3
15.50 522.3380 522.3386 −1.0
TMS/F2 C
25
H
46
O
3
Si
2
15.10 450.2985 450.2995 −2.11
VII Androst-3,5-diene-3,7ξ,17ξ,16ξ-tetraol TMS-enol C
31
H
60
O
4
Si
4
16.10 608.3568 608.3582 −2.19
C
30
H
57
O
4
Si
4
593.3333 593.3348 −2.37
TMS/F1 C
31
H
60
O
4
Si
4
16.10 608.3568 608.3574 −0.88
C
30
H
57
O
4
Si
4
593.3333 593.3326 1.33
VIII Androst-5-ene, 7-keto, 3α,17ξ,16ξ-triol TMS-enol C
31
H
60
O
4
Si
4
15.90 608.3568 608.3547 3.5
C
30
H
57
O
4
Si
4
593.3333 593.3327 1.17
TMS/F1-F2 C
28
H
52
O
4
Si
3
16.18 536.3173 536.3163 1.94
IX Androst-5-ene,3α,7ξ,17ξ,16ξ-tetraol TMS-enol C
31
H
62
O
4
Si
4
15.98 610.3725 610.3716 1.5
C
30
H
59
O
4
Si
4
595.3490 595.3509 −3.12
TMS/F1-F2 C
28
H
54
O
4
Si
3
15.27 538.3329 538.3301 5.3*
C
27
H
51
O
4
Si
3
523.3095 523.3104 −1.69
C
22
H
42
O
2
Si
2
394.2723 394.2727 −0.93
XAndrost-5-ene,3α,7ξ,17ξ,16ξ-tetraol TMS-enol C
31
H
62
O
4
Si
4
16.22 610.3725 610.3724 0.19
C
30
H
59
O
4
Si
4
595.3490 595.3493 −0.43
TMS/F1-F2 C
31
H
62
O
4
Si
4
16.22 610.3725 610.3721 0.68
C
30
H
59
O
4
Si
4
595.3490 595.3497 −1.11
*
The molecular ion m/z 538.3301 is present at 2% of the base peak in the mass spectrum.
MARTINEZ-BRITO ET AL.
1638
Additionally, in the F + G fraction, two apparently isomeric com-
pounds with an intense ion at m/z 169 were detected. The RT of
these two analytes were, respectively, 14.4 min (V) and 15.5 min (VI).
Based on the accurate mass of the enol-TMS and TMS derivatives of
the hydroxyl groups as well as their presence in the F + G fraction, the
structures proposed could be 7ξ-hydroxy-A and 7ξ-hydroxy-Etio.
Only 8% of the 7ξ-hydroxy-Etio was found in the S fraction. Their
responses relative to MT were estimated at 2 μg/mL (volunteer 1, t
max
5.5 h, volunteer 2, t
max
3 h) for 7ξ-hydroxy-A, and 3 μg/mL (volunteer
1, t
max
5.5 h) and 15 μg/mL (volunteer 2, t
max
3 h) for 7ξ-hydroxy-Etio.
Of the reduced metabolites, those identified as IV and VI showed
a longer excretion time and higher responses. The last sample col-
lected showed a relative response of metabolite IV of 200 ng/mL (vol-
unteer 1, 48 h) and 80 ng/mL (volunteer 2, 77 h); and metabolite VI
showed a relative response of 500 ng/mL for both volunteers (volun-
teer 1, 48 h and volunteer 2, 77 h).
4.6 |Hydroxylated metabolites of 7-keto-DHEA
In addition to the previously presented reduced metabolites, the pres-
ence of hydroxylated metabolites related to 7-keto-DHEA
administration could also be detected. Two signals showing a molecu-
lar ion at m/z 608.35 whose molecular formula corresponds to struc-
tures with two unsaturations and four oxygens could be corroborated
by their accurate mass as enol-TMS and TMS derivatives. None of
them was found in the urinary S fraction.
After purification by HPLC and subsequent derivatization, one of
them showed the same RT and mass spectrum found in F1 (VII), lead-
ing us to propose the presence of four hydroxyl groups and two C-C
unsaturations most probably in positions Δ5–6 and Δ3–4. Meanwhile,
the analysis of the other one showed a molecular formula consistent
with the presence of a keto group (VIII) because of the difference of
72 Da with respect to the enol-TMS derivative. The fact that both
mass spectra show an intense ion at m/z 503 (M
+
-CH
3
-OTMS) could
be an indication that the position of the additional hydroxyl group is
in C16, which is expected by comparison with the related structure to
DHEA. Taking into account that they were observed only in the F + G
fraction and the steric hindrance of the D ring for conjugation, it could
be hypothesized that the configuration is 3α-hydroxy-7-keto. The
response of these hydroxylated analytes relative to MT were
0.3 μg/mL (volunteer 1, t
max
5.5 h) and 1.5 μg/mL (volunteer 2, t
max
3 h) for compound VII, while for compound VIII it was 0.03 μg/mL
(volunteer 1, t
max
5.5 h) and 0.35 μg/mL (volunteer 2, t
max
3 h).
FIGURE 5 Chromatogram and mass spectra of the trimethylsilyl derivatives of the reduced metabolites of 7-keto-DHEA (II and III) in free
+glucuronated fraction of the urine as well as the interference showing an intense ion at m/z 169
MARTINEZ-BRITO ET AL.1639
Finally, two compounds equally hydroxylated but reduced were
observed at 15.9 min (IX) and 16.22 min (X). The MS showed a molec-
ular ion at m/z 610.37 Da as enol-TMS as well as an intense fragment
for the loss of CH
3
+ OTMS. After the purification by HPLC and pos-
terior hydroxyl derivatization, a signal was observed at the same RT
and mass spectra leading us to propose the presence of four hydroxyl
groups and only one C-C unsaturation. In the case of Δ5–6 reduction,
two isomers 5α- and 5β- should be formed but instead, only com-
pound X, was found. So far, it can be postulated that the Δ5–6
unsaturation remains, the Δ3–4 is reduced, and no isomers are
formed.
The accurate mass and the molecular formula of the hydroxyl-
TMS derivative of compound IX, showed the presence of four oxy-
gens including one keto-group (underivatized). After analysis of the
HPLC fractions, we observed one signal for both enol-TMS (16.1 min)
and TMS (16.2 min) showing a much lower intensity. The molecular
mass presented an ion at m/z 610.3705 Da (error 0.68 ppm) as enol-
TMS and m/z 538.3312 Da (error 3.32 ppm) as TMS. Considering the
similarity and the intensities observed, it is probable that the structure
for compound Xis 16β-hydroxy-17-one, while the lower intensity sig-
nal could be its isomer 17β-hydroxy-16-one. This hypothesis agrees
with the results described by Numazawa et al.
42
Using deuterated
compounds and making measurements of the activation and
enolization energies, he concluded that the mechanism of equilibrium
of hydration-dehydration in the D ring favors the configuration 16β-
hydroxy-17-one over the isomer 17β-hydroxy-16-one.
The maximum response of hydroxylated-reduced analytes relative
to MT were 0.5 μg/mL (volunteer 1, t
max
5.5 h) and 0.9 μg/mL (volun-
teer 2, t
max
3 h) for compound IX, while for compound Xthey were
0.6 μg/mL (volunteer 1, t
max
5.5 h) and 1.6 μg/mL (volunteer 2, t
max
3 h).
At the end of the study, of the four reduced metabolites, those
identified as IX and Xwere those excreted for longer and showed a
higher response. The last sample collected showed a relative response
of metabolite IX of 30 ng/mL (volunteer 1, 48 h) and 1 ng/mL (volun-
teer 2, 77 h); and the metabolite Xshowed a relative response of
about 50 ng/mL for both volunteers (volunteer 1, 48 h and volunteer
2, 77 h).
4.7 |Degradation products or in vivo metabolites?
The solvolysis conditions for the cleavage of the sulfate conjugates
are quite aggressive for some steroidal structures. Specifically, de-
oxidation of the 3β-hydroxy group in a 3,5-diene-7-keto configuration
has been described widely under acidic conditions.
9,10
The usual pro-
cedure to detect steroids in antidoping laboratories, include a derivati-
zation step, which occurs in acidic pH medium. This condition and the
high temperature of the GC injector are enough to degrade the
7-keto-DHEA present in the urine sample to produce androst-3,-
5-diene-7,17-dione, also known as arimistane. In fact, a rapid experi-
ment was carried out to demonstrate the formation of arimistane in
the presence of 7-keto-DHEA under the common conditions of
trimethylsilylation to detect steroids and to evaluate the endogenous
steroid profile. To obtain the TMS derivative of the 7-keto-DHEA ref-
erence material, different derivatization mixtures with different pro-
portions of MSTFA, ammonium iodide, and 2-mercaptoethanol were
used. After the analysis by GC-TOF a signal at the RT with an MS
corresponding to arimistane was observed. Figure 6 shows the per-
centage of arimistane that was formed from 7-keto-DHEA in each
condition.
The results obtained after the instrumental analysis of the enol-
TMS of a reference material of 7-keto-DHEA showed arimistane (m/z
428 at 11.2 min) the response being 7 fold the response of 7-keto-
DHEA. This in situ formation of arimistane was shown to be depen-
dent on the quantity of 7-keto-DHEA available in the sample. It
means that faced with an adverse analytical finding to a 7-keto-DHEA
sample, reasonable amounts of arimistane and its metabolite (androst-
3,5-diene-7β-ol-17-one) should be detected. The last one probably
originated from one of the reduced metabolites described earlier,
leading to a results misinterpretation.
Reviewing the results of the urinary excretion of 7-keto-DHEA
in the F + G fraction, it was observed that the highest response of
the 7β- metabolite of arimistane was at 6 ng/mL for volunteer 1 (t
max
7.5 h) and was not detected in volunteer 2. Meanwhile, arimistane
showed a maximum response for volunteers 1 (t
max
5.5 h) and 2 (t
max
3 h), corresponding to a relative response of 63 and 84 ng/mL,
respectively. On the other hand, the S fraction showed the highest
value of these two compounds in agreement with the previous com-
ment regarding the acidic conditions. The maximum value of the 7β-
metabolite was ca 50 and 10 ng/mL for volunteers 1 (t
max
7.5 h) and
2(t
max
6 h), respectively, while arimistane itself was observed at
9 and 2.5 μg/mL for volunteers 1 (t
max
3h)and2(t
max
3 h),
respectively.
The analysis of the HPLC purified fractions of the F + G fraction,
showed the arimistane response was 37 times higher than 7-keto-
DHEA itself in F3. Considering the lower polarity of arimistane struc-
ture (two keto groups and two double C-C bonds) it is not expected
to find the presence of this compound in this early fraction in which
the poly-hydroxyl structures were collected. In fact, the elution of the
arimistane reference material under the HPLC conditions was much
later. Consequently, its presence here can be attributed to the deriva-
tization step. While analysis of the S fraction showed arimistane and
7-keto-DHEA with similar responses observed in F3, for F32 the
arimistane response was 78 fold higher than that found in F3. Based
on that, two different origins for arimistane can be suggested: one as
a degradation product of 7-keto-DHEA because of the acidic condi-
tions during sample preparation and/or as an in vivo metabolite of
7-keto-DHEA.
The analysis of the HPLC F2 (F + G fraction), that contains the
3α- isomer of 7-keto-DHEA, did not show the presence of
arimistane (m/z 428 at 11.2 min). On the contrary, it showed the
presence of androst-4,6-diene-3,17-dione (13.5 min, identified by
MS and RT compared with the pure reference material). The
response of this diene-dione was only 10% of the 3α-isomer proba-
bly because of the higher stability of the 3α- configuration
MARTINEZ-BRITO ET AL.
1640
compared with the 3β- configuration. Additionally, the F17 for
F + G and S fractions showed the presence of this diene-dione
compound in a similar response but close to 10% of the response
of arimistane in F32. Considering them as degradation products, the
origin is quite easy to hypothesize because all the metabolites with
androst-3,5-diene-7-one or androst-5-ene-7-one configurations will
yield the corresponding 3-deoxidated product. Once the presence
of androst-4,6-diene-3,17-dione is confirmed, it is probable that its
formation will have its origin in the proposed metabolite identified
as VII (androst-3,5-diene-3,7ξ,17ξ,16ξ-tetraol). In any case, these
biomarkers (degradation products or in vivo metabolites) should be
considered to be of minor interest, providing additional information
to better understand a 7-keto-DHEA finding.
Figure 7 shows the time course of the urinary elimination of these
compounds in the F + G and S fractions. All the metabolites described
showed an elimination time higher than 7-keto-DHEA itself observed
in the F + G fraction. At 15 hours after the administration, 7-keto-
DHEA started to return to basal values, while in the last sample col-
lected by volunteer 1 (48 h) the relative responses of seven proposed
metabolites were over 100 ng/mL. The same behavior was observed
for volunteer 2. The S fraction showed that the main compounds
observed were arimistane (androst-3,5-diene-7,17-dione) and
androst-4,6-diene-3,17-dione considered as degradation metabolites,
so far.
4.8 |Preliminary IRMS measurements of δ
13
Cof
the metabolites
Direct δ
13
C(‰) measurement of the 3-acetyl-7-keto-DHEA from the
capsule administered was done. The detection of arimistane and
7-keto-DHEA was confirmed by the simultaneous analysis of the ref-
erence material. The signal for arimistane was observed to be around
30% of 3-acetyl-7-keto-DHEA, and a very low signal (below 5%) of
7-keto-DHEA could be observed. This is in agreement with a recent
publication on the analysis of 5-en-7-keto structures.
23
The three
compounds were analyzed by IRMS without any additional derivatiza-
tion and the δ
13
C values were −32.2‰,−31.6‰, and −32.0‰for
3-acetyl-7-keto-DHEA, arimistane, and 7-keto-DHEA, respectively.
The homogeneous delta values obtained for the three compounds is
mainly related to the steroid core carbon structure.
According to the HPLC fractions obtained during the clean-up
and avoiding isotopic fractionation, only three metabolites were mea-
sured (the whole peak was included in a unique fraction). The 3α-
isomer of 7-keto-DHEA (I) showed a δ
13
Cof−31.6‰(non-derivative
mass 302.1874 Da, Δ2.6 ppm), the reduced metabolite II showed a
δ
13
Cof−31.2‰(non-derivative mass 304.2033 Da, Δ−0.01 ppm),
and finally, the reduced metabolite Vshowed a δ
13
Cof−31.6‰(non-
derivative mass 306.2186 Da, Δ−0.2 ppm). The other metabolites
were not measured since they appeared in more than one fraction or
FIGURE 6 Formation of the arimistane-
TMS observed after the analysis of reference
material of trimethylsilyl 7-keto-DHEA by
GC-TOF, related to the different
derivatization composition used (MSTFA:
NH
4
I: 2-mercaptoethanol)
FIGURE 7 Time course of proposed metabolites described in the study in the free + glucuronidate fraction (F + G) and sulfate fraction. MT:
Methyltestosterone used as internal standard
MARTINEZ-BRITO ET AL.1641
were interference. These additional IRMS data support the fact that
there is no back formation of DHEA from 7-keto-DHEA.
5|CONCLUSION
The detection of 7-keto-DHEA administration based on the endoge-
nous steroid profile of ABP and the consequent analysis by GC-C-
IRMS failed. The additional IRMS data on specific proposed metabo-
lites support the fact that there is no back formation of DHEA from
7-keto-DHEA. At the same time, the monitoring of 7α- and 7β-
hydroxylated metabolites to monitor DHEA administration has the old
and well-known disadvantage that both are substances that are pro-
duced physiologically and are found, albeit in traces, in urine samples.
The identification of specific metabolites is required.
After a single 100 mg oral dose of 7-keto-DHEA, ten proposed
metabolites, in addition to 7-keto-DHEA itself, were newly described
(not present in the blank urine pre-administration samples above the
LOD of the method). These include reduced and hydroxylated struc-
tures that are independent of the steroid profile. The described
metabolites showed considerable responses in both the F + G and S
fractions. Among them, there were four metabolites excreted in the
F + G for a longer time and at higher responses, two reduced (IV and
VI) and two tetrahydroxylated metabolites (IX and X). The proposed
structures are based on the accurate mass measurement and compari-
son with similar structure fragmentation patterns. Additional spectro-
metric techniques or the syntheses of the proposed structures are
needed for a definitive confirmation of the configurations.
Other compounds that can be linked to a degradation reaction by
deoxidation in positions 3 or 7 were found. In this sense, extreme care
should be taken when interpreting the results after the application of
procedures to detect steroids in urine currently in place in many
laboratories.
The presence of arimistane in concentrations higher than its 7β-
hydroxy metabolite could be an adequate marker of 7-keto-DHEA
administration because after administration of arimistane, arimistane
itself is almost undetectable in urine.
43
To disclose whether arimistane is or is not an in vivo metabolite
of 7-keto-DHEA requires the use of methodologies avoiding derivati-
zation and/or the removal of the sulfate moiety as liquid chromatogra-
phy coupled to mass spectrometry.
ORCID
Dayamin Martinez-Brito https://orcid.org/0000-0002-7066-4916
Xavier de la Torre https://orcid.org/0000-0001-8037-6750
Francesco Botrè https://orcid.org/0000-0001-5296-8126
REFERENCES
1. WADA. The World Antidoping Code –The 2019 Prohibited List
International Standard. https://www.wada-ama.org/sites/default/
files/wada_2019_english_prohibited_list.pdf. Published 2019. (Last
accessed 10 June 2019)
2. WADA. TD2019IRMS - Detection of Synthetic Forms of Endogenous
Anabolic Androgenic Steroids by GC/C/IRMS. 2019. https://www.
wada-ama.org/sites/default/files/td2019irms_final_eng_clean.pdf.
Published 2019. (Last accessed 10 June 2019)
3. Davidson M. Safety and pharmacokinetic study with escalating doses
of 3-acetyl-7-oxo-dehydroepiandrosterone in healthy male volun-
teers. Clin Invest Med. 2000;23(5):300-310.
4. Lardy H, Kneer N, Wei Y, Partridge B, Marwah P. Ergosteroids II: bio-
logically active metabolites and synthetic derivatives of dehydroepi-
androsterone. Steroids. 1998;63(3):158-165. https://doi.org/10.
1016/S0039-128X(97)00159-1
5. Šulcová J, Hampl R, Hill M, Stárka L, Novácˇek A. Delayed effects of
short-term transdermal application of 7-oxo-dehydroepiandrosterone
on its metabolites, some hormonal steroids and relevant
proteohormones in healthy male volunteers. Clin Chem Lab Med.
2005;43(2):221-227. https://doi.org/10.1515/CCLM.2005.038
6. Van Eenoo P, Delbeke FT. Metabolism and excretion of anabolic ste-
roids in doping control –new steroids and new insights. J Steroid Bio-
chem Mol Biol. 2006;101(4–5):161-178. http://www.sciencedirect.
com/science/article/pii/S096007600600210X
7. Okada M, Fukushima DK, Gallagher TF. Isolation and characterization
of 3β-hydroxy-Δ5-steroids in adrenal carcinoma. J Biol Chem. 1959;
234(7):1688-1692. http://www.jbc.org/content/234/7/1688.short
8. Świzdor A, Panek A, Milecka-Tronina N. Biohydroxylation of 7-oxo-
DHEA, a natural metabolite of DHEA, resulting in formation of new
metabolites of potential pharmaceutical interest. Chem Biol Drug Des.
2016;88(6):844-849. https://doi.org/10.1111/cbdd.12813
9. Nickon A, Bagli JF. Reactivity and geometry in allylic systems.
I. Stereochemistry of photosensitized oxygenation of monoolefins.
J Am Chem Soc. 1961;83(6):1498-1508. https://doi.org/10.1021/
ja01467a049
10. Marwah A, Marwah P, Lardy H. Ergosteroids VII: perchloric acid-
induced transformations of 7-oxygenated steroids and their bio-
analytical applications –a liquid chromatographic-mass spectrometric
study. Bioorg Chem. 2002;30(4):233-248. https://doi.org/10.1016/
S0045-2068(02)00010-X
11. Numazawa M, Mutsumi A, Hoshi K, Tanaka Y. Androst-5-ene-7,-
17-dione: a novel class of suicide substrate of aromatase. Biochem
Biophys Res Commun. 1992;186(1):32-39. https://doi.org/10.1016/
S0006-291X(05)80771-5
12. Numazawa M, Tachibana M. Studies directed toward a mechanistic
evaluation of aromatase inhibition by androst-5-ene-7,17-dione:
time-dependent inactivation by the 19-nor and 5β,6β-epoxy deriva-
tives. Steroids. 1997;62(7):516-522. https://doi.org/10.1016/S0039-
128X(97)00002-0
13. Numazawa M, Mutsumi A, Tachibana M, Hoshi K. Synthesis of
androst-5-en-7-ones and androsta-3,5-dien-7-ones and their related
7-deoxy analogs as conformational and catalytic probes for the active
site of aromatase. J Med Chem. 1994;37(14):2198-2205. https://doi.
org/10.1021/jm00040a012
14. Torrado S, Roig M, Farré M, Segura J, Ventura R. Urinary metabolic
profile of 19-norsteroids in humans: glucuronide and sulphate conju-
gates after oral administration of 19-nor-4-androstenediol. Rapid
Commun Mass Spectrom. 2008;22(19):3035-3042. https://doi.org/10.
1002/rcm.3689
15. Fragkaki AG, Angelis YS, Kiousi P, Georgakopoulos CG, Lyris E. Com-
parison of sulfo-conjugated and gluco-conjugated urinary metabolites
for detection of methenolone misuse in doping control by LC-HRMS,
GC-MS and GC-HRMS. J Mass Spectrom. 2015;50(5):740-748.
https://doi.org/10.1002/jms.3583
16. Rzeppa S, Heinrich G, Hemmersbach P. Analysis of anabolic
androgenic steroids as sulfate conjugates using high performance liq-
uid chromatography coupled to tandem mass spectrometry. Drug
Test Anal. 2015;7(11–12):1030-1039. https://doi.org/10.1002/dta.
1895
17. Rzeppa S, Viet L. Analysis of sulfate metabolites of the doping agents
oxandrolone and danazol using high performance liquid
MARTINEZ-BRITO ET AL.
1642
chromatography coupled to tandem mass spectrometry. J Chromatogr
B. 2016;1029–1030:1-9. https://doi.org/10.1016/j.jchromb.2016.
06.028
18. Piper T, Schänzer W, Thevis M. Revisiting the metabolism of
19-nortestosterone using isotope ratio and high resolution/high accu-
racy mass spectrometry. J Steroid Biochem Mol Biol. 2016;162:80-91.
https://doi.org/10.1016/J.JSBMB.2015.12.013
19. Shackleton CHL. Profiling steroid hormones and urinary steroids.
J Chromatogr B Biomed Sci Appl. 1986;379:91-156. https://doi.org/
10.1016/S0378-4347(00)80683-0
20. Stanley SM, Kent S, Rodgers J. Biotransformation of 17-alkyl steroids
in the equine: high-performance liquid chromatography–mass spec-
trometric and gas chromatography–mass spectrometric analysis of
fluoxymesterone metabolites in urine samples. J Chromatogr B Biomed
Sci Appl. 1997;704(1–2):119-128. https://doi.org/10.1016/S0378-
4347(97)00440-4
21. Hauser B, Deschner T, Boesch C. Development of a liquid
chromatography–tandem mass spectrometry method for the determi-
nation of 23 endogenous steroids in small quantities of primate urine.
J Chromatogr B. 2008;862(1–2):100-112. https://doi.org/10.1016/j.
jchromb.2007.11.009
22. Declaration of Helsinki (1964). Recommendations guiding physicians
in biomedical research involving human subjects. BMJ. 1996;313
(7070):1448–1449. https://doi.org/10.1136/bmj.313.7070.1448a
23. Lorenz LM, Toomey VM, Lanzarotta AC, Flurer RA, Falconer TM.
Identification of the designer steroid androsta-3,5-diene-7,17-dione
in a dietary supplement. Drug Test Anal. 2019;11(7):1109-1115,
dta.2589. https://doi.org/10.1002/dta.2589
24. de la Torre X, Colamonici C, Curcio D, Botrè F. Fast IRMS screening
of pseudoendogenous steroids in doping analyses. Drug Test Anal.
2017;9(11–12):1804-1812. https://doi.org/10.1002/dta.2321
25. Martinez Brito D, Notarianni M, Iannone M, Botrè F, de La Torre X.
Validation of sulphates deconjugation for metabolic studies. XXXVII
Manfred Donike Work Dope Anal. 2019.
26. de la Torre X, Colamonici C, Curcio D, Molaioni F, Botrè F. A compre-
hensive procedure based on gas chromatography-isotope ratio mass
spectrometry following high performance liquid chromatography
purification for the analysis of underivatized testosterone and its ana-
logues in human urine. Anal Chim Acta. 2012;756(0):23-29. https://
doi.org/10.1016/j.aca.2012.10.013
27. Hampl R, Lapcˇík O, Hill M, et al. 7-Hydroxydehydroepiandrosterone –
a natural antiglucocorticoid and a candidate for steroid replacement
therapy? Physiol Res. 2000;49(S1):S107-S112.
28. Delbeke FT, Van Eenoo P, Van Thuyne W, Desmet N. Prohormones
and sport. J Steroid Biochem Mol Biol. 2002;83(1–5):245-251. https://
doi.org/10.1016/S0960-0760(02)00274-1
29. Cawley AT, George AV. Complementary stable carbon isotope ratio
and amount of substance measurements in sports anti-doping. Drug
Test Anal. 2012;4(12):897-911. https://doi.org/10.1002/dta.1378
30. Flórez J, Armijo JA, Mediavilla A. Farmacologia Humana
31. Goodman G. In: Brunton LL, ed. The Pharmacological Basis of Thera-
peutics. Eleventh v. New York, USA: McGraw-Hill; 2006.
32. Zumoff B, Bradlow HL. Sex difference in the metabolism of
dehydroisoandrosterone sulfate*.J Clin Endocrinol Metab. 1980;51(2):
334-336. https://doi.org/10.1210/jcem-51-2-334
33. Zumoff B, Rosenfeld RS, Strain GW, Levin J, Fukushima DK. Sex dif-
ferences in the twenty-four-hour mean plasma concentrations of
dehydroisoandrosterone (DHA) and dehydroisoandrosterone sulfate
(DHAS) and the DHA to DHAS ratio in normal adults*.J Clin
Endocrinol Metab. 1980;51(2):330-333. https://doi.org/10.1210/
jcem-51-2-330
34. Frye RF, Kroboth PD, Kroboth FJ, et al. Sex differences in the phar-
macokinetics of dehydroepiandrosterone (DHEA) after single- and
multiple-dose administration in healthy older adults. J Clin Pharmacol.
2000;40(6):596-605. http://www.ncbi.nlm.nih.gov/pubmed/
10868310. Accessed 8 August 2019
35. Van Renterghem P, Van Eenoo P, Geyer H, Schänzer W, Delbeke FT.
Reference ranges for urinary concentrations and ratios of endoge-
nous steroids, which can be used as markers for steroid misuse, in a
Caucasian population of athletes. Steroids. 2010;75(2):154-163.
http://www.sciencedirect.com/science/article/B6TC9-4XVC4S1-1/
2/5db55c7691f51abf3e864856ca5e848c
36. Dehennin L, Ferry M, Lafarge P, Pérès G, Lafarge J-P. Oral administra-
tion of dehydroepiandrosterone to healthy men: alteration of the uri-
nary androgen profile and consequences for the detection of abuse in
sport by gas chromatography-mass spectrometry. Steroids. 1998;63
(2):80-87. https://doi.org/10.1016/S0039-128X(97)00138-4
37. Shelby MK, Crouch DJ, Black DL, Robert TA, Heltsley R. Screening
indicators of dehydroepiandosterone, androstenedione, and dihydro-
testosterone use: a literature review. J Anal Toxicol. 2011;35(9):638-
655. https://doi.org/10.1093/anatox/35.9.638
38. Van Renterghem P, Van Eenoo P, Sottas P-E, Saugy M, Delbeke F.
Subject-based steroid profiling and the determination of novel bio-
markers for DHT and DHEA misuse in sports. Drug Test Anal. 2010;2
(11–12):582-588. https://doi.org/10.1002/dta.206
39. Cawley A, Hine E, Trout G, George A, Kazlauskas R. Searching for
new markers of endogenous steroid administration in athletes:
“looking outside the metabolic box.”.Forensic Sci Int. 2004;143(2–3):
103-114. https://doi.org/10.1016/J.FORSCIINT.2004.02.031
40. Brooks CJW, Johnson AW, Beynon JH. Some aspects of mass spec-
trometry in research on steroids. Philos Trans R Soc London Ser a, Math
Phys Sci. 1979;293(1400):53-67. https://doi.org/10.1098/rsta.1979.
0079
41. Harvey DJ, Vouros P. Influence of the 6-trimethylsilyl group on the
fragmentation of the trimethylsilyl derivatives of some 6-hydroxy-
and 3,6-dihydroxy-steroids and related compounds. Biomed Mass
Spectrom. 1979;6(4):135-143. https://doi.org/10.1002/bms.
1200060402
42. Numazawa M, Nagaoka M, Matsuzaki H, Yamashita K, Komatsu S,
Osawa Y. Mechanistic aspects of rearrangement of 16α-hydroxy-
17-keto steroids to the 17β-hydroxy-16-keto isomers. Steroids. 2008;
73(8):798-805. https://doi.org/10.1016/J.STEROIDS.2008.03.001
43. Martinez Brito D, Torre X, Botrè F. Detection of urinary metabolites
of arimistane in humans by gas chromatography coupled to high accu-
racy mass spectrometry for antidoping analyses. Rapid Commun Mass
Spectrom. 2019;33(24):1894-1905. https://doi.org/10.1002/rcm.
8529
How to cite this article: Martinez-Brito D, de la Torre X,
Colamonici C, Curcio D, Botrè F. 7-keto-DHEAmetabolism in
humans. Pitfalls in interpreting the analytical results in the
antidoping field. Drug Test Anal. 2019;11:1629–1643. https://
doi.org/10.1002/dta.2734
MARTINEZ-BRITO ET AL.1643
