A mass spectrometric study on meloxicam metabolism in horses and the fungus Cunninghamella elegans, and the relevance of this microbial system as a model of drug metabolism in the horse

Article abstract of DOI:10.1002/jms.1575

This paper describes a study where the metabolism of the non-steroidal anti-inflammatory drug meloxicam was investigated in six horses and in the filamentous fungus Cunninghamella elegans. The metabolites identified were compared between the species, and then the fungus was used to produce larger amounts of the metabolites for future use as reference material. C. elegans proved to be a good model of phase Imeloxicam metabolism in horses since all four metabolites found were the same in both species. A part from the two main metabolites, 5′-hydroxymethylmeloxicam and 5′-carboxymeloxicam, a second isomer of hydroxymeloxicam and dihydroxylated meloxicam were detected for the first time in horse urine and themicrobial incubations. Phase II metabolites were not discovered in the C. elegans samples but hydroxymeloxicam glucuronide was detected intact in horse urine for the first time in this study. Urine from six horses was further analyzed in a semi-quantitative sense and 5′-hydroxymethylmeloxicam gave peaks with much higher intensity compared to the parent drug and the other metabolites, and was detected for at least 14 days after the last given dose in some of the horses. From the results presented in this article, we suggest that analytical methods developed for the detection of meloxicam in horse urine after prohibited use should focus on the 5′-hydroxymethyl metabolite and that C. elegans can be used to produce large amounts of this metabolite for potential future use as a reference compound. Copyright

Full text of DOI:10.1002/jms.1575

Research Article
Received: 7 January 2009
Accepted: 11 February 2009
Published online in Wiley Interscience: 16 March 2009
(www.interscience.com) DOI 10.1002/jms.1575
A mass spectrometric study on meloxicam
metabolism in horses and the fungus
Cunninghamella elegans, and the relevance of
this microbial system as a model of drug
metabolism in the horse
Annica Tevell Åberga, Charlotte Olssonb†, Ulf Bondessona,b and
Mikael Hedelanda,b∗
This paper describes a study where the metabolism of the non-steroidal anti-inflammatory drug meloxicam was investigated
in six horses and in the filamentous fungus Cunninghamella elegans. The metabolites identified were compared between the
species, and then the fungus was used to produce larger amounts of the metabolites for future use as reference material. C.
elegans proved to be a good model of phase I meloxicam metabolism in horses since all four metabolites found were the same in
both species. Apart from the two main metabolites, 5ꢀ-hydroxymethylmeloxicam and 5ꢀ-carboxymeloxicam, a second isomer of
hydroxymeloxicam and dihydroxylated meloxicam were detected for the first time in horse urine and the microbial incubations.
Phase II metabolites were not discovered in the C. elegans samples but hydroxymeloxicam glucuronide was detected intact
in horse urine for the first time in this study. Urine from six horses was further analyzed in a semi-quantitative sense and
5ꢀ-hydroxymethylmeloxicam gave peaks with much higher intensity compared to the parent drug and the other metabolites,
and was detected for at least 14 days after the last given dose in some of the horses. From the results presented in this article,
we suggest that analytical methods developed for the detection of meloxicam in horse urine after prohibited use should focus
on the 5ꢀ-hydroxymethyl metabolite and that C. elegans can be used to produce large amounts of this metabolite for potential
c
future use as a reference compound. Copyright ꢀ 2009 John Wiley & Sons, Ltd.
Keywords: meloxicam; metabolism; mass spectrometry; horse; Cunninghamella elegans
Introduction
groupmentioned(Fig. 1),resultingin5ꢁ-hydroxymethylmeloxicam
and 5ꢁ-carboxymeloxicam. These two metabolites have been
detected both in vitro and in vivo in different species such as
humans,[7,8] thoroughbred horses,[5,6] rats,[2,8] mice and mini-
pigs.[8] In earlier publications, the biotransformation of meloxicam
has been investigated by the use of radioactivity monitoring
after administration of 14C-labelled meloxicam[2,8] and high
performance liquid chromatography (HPLC) in combination with
ultraviolet diode array detection (UV-DAD)[2] or either triple
quadrupole mass spectrometry (MS)[8] or ion trap MS.[5,6]
Smith and Rosazza suggested that microorganisms could
be used as models of metabolism of xenobiotic substances
in mammals.[9] Since then, filamentous fungi such as Cun-
ninghamella elegans have been used both as a complemen-
tary in vitro model for drug metabolism[10,11] and to produce
Meloxicam [4-hydroxy-2-methyl-N-(5-methyl-2-thiazolyl)-2H-1,2-
benzothiazine-3-carboxamide-1,1-dioxide] is a non-steroidal anti-
inflammatory drug (NSAID) that is cyclooxygenase-2 (COX-2)
selective and belongs to the enolic acid group (Fig. 1). It is
closely structurally related to other oxicams such as sudoxicam,
piroxicam, isoxicam and tenoxicam[1]; although the methyl group
on the thiazolyl ring has facilitated its metabolism and hence
resulted in a shorter half-life compared to those of the others.[2]
NSAIDs are extensively used in veterinary medicine due to
their anti-inflammatory, analgesic and antipyretic properties.[3]
Meloxicam is one of the newest drugs in this group and it is
prohibited in the sports of horse racing.[4] For this reason it
is important to develop powerful analytical methods for the
detection of meloxicam in urine from race horses in the doping
control laboratories. The finding of a prohibited substance in a
sample could be either a finding of the substance itself or of
a metabolite or isomer of the substance or the metabolite.[4]
Some drugs are extensively metabolized in the body and it
might be more efficient to search for a metabolite in the sample
than for the parent compound. The metabolism of meloxicam
in horses is only briefly discussed in the literature by Dumasia
et al.[5] and de Kock et al.[6] The most commonly previously
reported metabolites are formed by oxidation of the methyl
∗
Correspondence to: Mikael Hedeland, Department of Chemistry, National
Veterinary Institute, SVA, SE-751 89 Uppsala, Sweden.
E-mail: Mikael.Hedeland@sva.se
†
a
Present address: Semcon AB, Uppsala, Sweden
Division of Analytical Pharmaceutical Chemistry, Uppsala University, Sweden
Department of Chemistry, National Veterinary Institute, Uppsala, Sweden
b
c
J. Mass. Spectrom. 2009, 44, 1026–1037
Copyright ꢀ 2009 John Wiley & Sons, Ltd.

A mass spectrometric study on meloxicam metabolism in horses
Drug administration to the horses and urine sampling
Metacam (meloxicam) was administered as an oral suspension
to six healthy horses at Bjerke Dyrehospital in Oslo, Norway. The
horses were standardbred trotters that weighed between 395 and
499 kg. The meloxicam dose given was 0.6 mg/kg bodyweight
and was administered once daily for 5 days. A urine sample was
collected 4 days prior to the first dose (the blank sample), on the
third day of administration and every day thereafter for 21 days.
The samples analyzed in this study were the ones taken day 5, 6,
8, 10, 13, 16 and 19. The urine samples were stored in −20 ◦C until
analysis. The company that was responsible for the clinical trial
also held the ethical permission for the study.
Figure 1. Chemical structure of meloxicam. The position of the three
deuterium atoms of CD3-meloxicam is on the N-methyl group in bold.
drug metabolites in amounts sufficient for complete structural
confirmation[12,13] or further toxicological testing.[14,15] Some
examples of drugs that have been metabolized by C. ele-
gans are the antihistamines clemastine,[16] brompheniramine,
chlorpheniramine and pheniramine[14]; the NSAIDs naproxen[17]
Microbial transformation of meloxicam
Cultures of C. elegans were grown on Sabouraud dextrose agar
plates (mycological peptone 10 g/l, dextrose 40 g/l, agar 15 g/l)
for 5 days in 27 ◦C and were then stored in 4 ◦C. The spores and
mycelia from one agar plate were blended with 75 ml of sterile
physiological saline solution in a sterile container. Five millilitres
of the mycelia suspension were transferred to 250 ml Erlenmeyer
flasks containing 30 ml of broth. Four different kinds of broth were
compared at pH 5.5 and 7.0, those were yeast extract broth (yeast
extract 8 g/l, sodium chloride 4.6 g/l, sucrose 20 g/l, monopotas-
sium phosphate 3.75 g/l, disodiumphosphate 7.09 g/l), peptone
broth (mycological peptone 8 g/l, sodium chloride 4.6 g/l, sucrose
20 g/l, monopotassium phosphate 3.75 g/l, disodiumphosphate
7.09 g/l), and two different Sabouraud dextrose broths (mycologi-
cal peptone 10 g/l and dextrose 20 g/l or 40 g/l, respectively). The
pH of the broth was adjusted from about 5.5 to 7.0 by the addition
of 1.0 M phosphate buffer pH 6.9 (10–1000 ml of broth) and, when
needed, 2.0 M sodium hydroxide. The different broths and pH were
chosen with guidance of the extensive comparison of different
cultivation media performed by Freitag et al.[22] The cultures were
incubated for 3 days in 27 ◦C before the broth was decanted, and
30 ml of fresh broth was added along with meloxicam or a mixture
of meloxicam and CD3-meloxicam (100 µl, 7.0 mM in dimethylfor-
mamide). After 7 days in 27 ◦C the incubation was terminated by
the addition of 25 ml of acetonitrile. The samples were stored at
−20 ◦C until solid phase extraction. The Sabouraud dextrose broth
with 20 g/l dextrose and pH 7.0 was used to prepare the cultures
for meloxicam metabolite isolation.
and indomethacin[18]
and mirtazapine[11]
;
antidepressants like amitriptyline[10]
and proton-pump inhibitors such as
;
pantoprazole[19] and omeprazole.[20] Both phase I and phase II
metabolism have been reported. Experiments with 18O2 labelling
and cytochrome P450 inhibitors have indicated that cytochrome
P450 monooxygenase is expressed in Cunninghamella.[10] Ac-
tivities of cytochrome P450 have been measured in micro-
somal preparations, activities of PAPS sulfotransferase, glu-
tathione S-transferase and UDP-glucosyltransferase have been
measured in the cytosol. Furthermore, UDP-glucuronyltransferase
has been shown to be active in both microsomal and cytosolic
preparations.[21] We wanted to examine if C. elegans is a useful
model of metabolism of meloxicam in the horse, and if so, use the
microorganism to produce larger amounts of the metabolites of
interest for future use as reference compounds in the laboratory
since drug metabolites often are not commercially available.
In summary, the aims of this study were: (1) to characterize
the meloxicam metabolites excreted in urine from standardbred
horses treated with meloxicam as well as the metabolites formed
in C. elegans incubations and to compare the findings, (2) to use
C. elegans for large-scale production of interesting metabolites
for future use as reference compounds in the doping control
laboratory, (3) to evaluate the relative concentrations of the drug
and metabolites in urine from the horses and estimate the time
after the last given dose of the parent drug and the main
metabolites could be detected. The information gained will be
useful in the development of new analytical methods for the
detection of meloxicam in doping samples from race horses.
Sample preparation
In vitro samples
The microbial samples were centrifuged (1000 g for 10 min) and
the supernatants were removed and diluted with 45 ml of 1.0%
formic acid in water. Varian abselut NEXUS solid phase extraction
cartridges(200 mg,6 ml,purchasedfromScantecLabABinPartille,
Sweden) were conditioned with 5 ml of methanol and 5 ml of Milli-
Qwaterandequilibratedwith5 mlof1.0%formicacid.Thesamples
were applied and the cartridges were washed with 5 ml of 1.0%
formic acid and left to dry under moderate vacuum for 5 min. The
analytes were eluted with 5 ml of ethyl acetate. The eluate was
evaporated to dryness under a stream of nitrogen gas at 60 ◦C.
Experimental
Chemicals
Meloxicam sodium hydrate was purchased from Sigma Aldrich
(Steinheim, Germany) and CD3-meloxicam was obtained from
Toronto Research Chemicals (ON, Canada). The water was purified
using a Milli-Q water purification system (Millipore, Bedford, MA,
USA) and all chemicals and solvents used were of analytical grade
or better and used without further purification. Cunninghamella
echinulatavarianselegans, American Type Culture Collection 9245,
lot no 3656357 was purchased from LGC Promochem (Borås,
Sweden). The Virkon-S, the saline solution, the Sabouraud agar
plates, and the different kinds of broth needed for the experiment
of C. elegans were obtained from the National Veterinary Institute
(SVA) in Uppsala, Sweden.
In vivo samples
The pH of the horse urine samples (10.0 ml) was adjusted to
2.5 0.3 with 1.0 M hydrochloric acid. Dichloromethane with 2%
c
J. Mass. Spectrom. 2009, 44, 1026–1037
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/jms

A. Tevell Åberg et al.
isopropanol (20 ml) was added and the samples were mixed on
a rotary mixer for 15 min followed by centrifugation at 1000 g for
15 min. The organic phase was transferred to a clean test tube
and evaporated to dryness under a gentle stream of nitrogen at
50 ◦C. In the semi-quantitative study, 100 µl of CD3-meloxicam
(10 µg/ml) was added to the urine before the pH adjustment.
the spray voltage was 5 kV and the nitrogen sheath and auxiliary
gas flow rates were 80 and 15 arbitrary units, respectively. The
capillary voltage was set to 21 V (positive ESI) or −47 V (negative
ESI). The ion trap was filled with helium gas and the MS1, MS2, MS3
and MS4 scan modes were used. The relative collision energy was
optimized for each mass transition and ranged from 22% for MS2
to 34% for MS4 The wide band activation parameter was not used
in these experiments.
High performance liquid chromatography
HPLC-MS/MS experiments
Triple quadrupole experiments
Solventsinthemobilephasewere0.2%formicacidinwater(A) and
acetonitrile (B) and two different gradients were used. Gradient
I: 20% B 0–5.0 min; 20–95% B 5.0–11.0 min; held at 95% B until
13.0 min; equilibration at 20% B 13.0–16 min. Gradient II: 20%
B 0–2.0 min; 20–35% B 2.0–25.0 min; 35–90% B 25.0–26.0 min;
held at 90% B until 40.0 min; equilibration at 20% B 40.0–50.0 min.
The mobile phase was delivered at a flow rate of 200 µl/min by
a quaternary Surveyor LC system (Thermo Electron Corp., San
Jose´, CA, USA) or a binary LC system of the 1100 series (Agilent
Technologies, Waldbronn, Germany). The analytical column was
a Phenomenex Luna 5 µ C18 (2), 150 × 2.00 mm, and the guard
column was an ODS-C18 (4.0 × 2.0 mm, length × inner diameter)
both from Scandinaviska Genetec AB (Va¨stra Fro¨lunda, Sweden).
A triple quadrupole TSQ Quantum Discovery instrument (Thermo
Electron Corp., San Jose´, CA, USA) with an electrospray ion source
was used in the full scan modes of MS, MS/MS and neutral
loss (NL) as well as selected reaction monitoring (SRM). ESI
in the positive and negative mode was used and the sheath
and auxiliary gas was nitrogen at 45–60 and 5–45 mTorr
pressure, respectively. In the different experiments the capillary
voltage was 3.5–4.8 kV and the temperature was 302–350 ◦C.
For collision induced dissociation (CID) the collision cell was
filled with argon at 1.5 mTorr and the collision energies used
will be presented in the figure legends. Instrument control,
data acquisition and data processing were carried out with
the Xcalibur software, version 2.0 SR2. The mass transitions
used in the SRM mode (positive ESI) were m/z 352 → 115 for
meloxicam, m/z 355 → 115 for CD3-meloxicam (the internal
standard), m/z 368 → 131 for 5ꢁ-hydroxymethylmeloxicam,
m/z 382 → 145 for 5ꢁ-carboxymeloxicam and m/z 384 → 131 for
dihydroxymeloxicam. The collision energy was set to 20 V in the
SRM mode.
Isolation of meloxicam metabolites
Three of the metabolites detected in this study were isolated
from the C. elegans cultures by the use of a Waters Fraction
CollectorIII(WatersCorp. Milford, MA, USA)inordertocharacterize
them further by the use of MSn experiments and nuclear
magnetic resonance spectroscopy (NMR, data intended for a
coming publication). The solvents in the mobile phase were the
same as above and delivered as gradient III: 20% B 0–2.0 min;
20–35% B 2.0–25.0 min; 35–90% B 25.0–26.0 min; held at 90% B
26.1–32.0 min;equilibrationat20%B32.1–40 min. Theseparation
was achieved on a Hypersil GOLD column, 3 µm particles,
150 × 4.6 mm (Thermo Electron Corp., San Jose´, CA, USA) and
the volume injected was 50 µl. The flow rate was 1.0 ml/min and
a split placed after the column directed at 15% of the flow to
the mass spectrometer (a Finnigan LCQ ion trap) and the rest to
the fraction collector. Three metabolites were collected in each
run (suggested to be dihydroxymeloxicam 11 min 50 s–13 min
10 s, 5ꢁ-hydroxymethylmeloxicam 14 min 0 s–17 min 30 s, and 5ꢁ-
carboxymeloxicam 20 min 0 s–23 min 0 s). The fractions collected
from a total of 14 C. elegans cultures were pooled and freeze dried.
For the MSn experiments a small portion of each metabolite was
dissolved in 0.2% formic acid in water and acetonitrile 50 : 50.
Results and Discussion
Fragmentation of meloxicam
To interpret the mass spectra of the meloxicam metabolites it
was necessary to first investigate the fragmentation of the parent
compound. The CID of meloxicam was studied on an ion trap
instrument during infusion of a meloxicam standard solution. In
the positive ion mode (Fig. 2), a loss of water from protonated
meloxicam ([M + H]+ m/z 352) resulted in m/z 334, but the three
major fragments in the MS2 spectrum contained the thiazolyl
functionandwerem/z115and141,bothpreviouslyreported,[5,6,23]
and m/z 184. MS3 of m/z 352 → 184 → scan showed that m/z
184 could be fragmented further into m/z 115, as well as the less
abundant ions m/z 127, 141 and 156. See Fig. 2(c) for proposed
structures of the fragments formed. The CID spectrum of CD3-
meloxicam, gave the same major fragments as meloxicam (m/z
115 and 141), as they did not contain the deuterated methyl group
(results not shown). However, the fragment m/z 184 of meloxicam
was shifted to m/z 187 for the isotopically labelled analog.
Mass spectrometry
Ion trap experiments
Figure 3 illustrates the MS2, MS3 and MS4 spectra of meloxicam
in the negative ion mode. The m/z for deprotonated meloxicam
[M−H]− was 350 and the most intense peaks in the MS2 spectrum
were m/z 286 (loss of SO2), m/z 210 (deprotonated benzothiazine)
and m/z 146. This fragmentation has been proposed earlier by
de Kock et al.[6] Further, MS3 of 350 → 286 → scan (Fig. 3(b))
as well as 350 → 210 → scan (spectrum not shown) proved
that the m/z 146 fragment could be formed from either m/z
286 or 210. The m/z 252 fragment was formed from m/z 286
most probably by the loss of H2S (34 Da). In the MS4 spectrum
m/z 350 → 286 → 146 → scan, m/z 128 (loss of water) and m/z
A Finnigan LCQ ion trap mass spectrometer (Thermo Electron
Corp., San Jose´, CA, USA) equipped with an electrospray interface
was used in this study. Instrument control, data acquisition and
data processing were carried out with the Xcalibur software,
version 1.3. The electrospray ionization (ESI) and MS parameters
were manually optimized for sensitivity during a 3 µl/min infusion
of a meloxicam solution in methanol/water. The solution was
mixed with the LC flow (0.2% formic acid in water/acetonitrile,
50 : 50, 200 µl/min) through a connecting T. In both positive and
negative ion modes the capillary temperature was set to 225 ◦C,
c
www.interscience.wiley.com/journal/jms
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2009, 44, 1026–1037

A mass spectrometric study on meloxicam metabolism in horses
Figure 2. Positive ESI MSn analysis of meloxicam standard. Data was collected for 60 s during a constant infusion of the solution into the ion trap
instrument. (a) MS2 spectrum of m/z 352, the collision energy was 22%. (b) MS3 spectrum of m/z 352 → 184 → scan, the collision energies were 22 and
25%, respectively. For details, see section on Experimental. (c) Proposed structures of the fragments.
c
J. Mass. Spectrom. 2009, 44, 1026–1037
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/jms

A. Tevell Åberg et al.
c
www.interscience.wiley.com/journal/jms
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2009, 44, 1026–1037

A mass spectrometric study on meloxicam metabolism in horses
c
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2009, 44, 1026–1037
www.interscience.wiley.com/journal/jms

A. Tevell Åberg et al.
119 were the most abundant ions. Structural suggestions of the
majorfragmentsarepresentedinFig. 3(d).ForCD3-meloxicam,the
major fragments in the CID spectrum contained the deuterated
methyl group, resulting in a shift of 3 m/z units compared to
meloxicam, giving the peaks m/z 149, 213 and 289 (results not
shown).
The negative ion mode MS/MS spectrum of the major peak is
shown in Fig. 4(e) and (f), fungus and horse urine respectively.
The spectra had peaks at the same m/z values as the previously
published ones for 5ꢁ-hydroxymethylmeloxicam.[6]
The MS/MS spectrum for the minor peak in the negative ion
mode is shown in Fig. 4(g) (fungus) and (h) (horse urine). The m/z
302 fragment probably corresponded to meloxicam +16 –SO2,
but the most abundant fragment was m/z 252 (m/z 255 for the
deuterium labelled metabolite) which may be derived from the
cleavage of the amide bond. The hydroxylation has most likely
occurred on the benzothiazine moiety, but the exact position
could not be concluded. Similar metabolites of piroxicam have
been discovered in rat urine but the sites of the hydroxylation for
the two isomers were not determined closer than in position 5, 6 or
7 on the benzothiazine ring.[25] An m/z 252 fragment was found in
meloxicam standard as well (Fig. 3, suggested to be derived from
meloxicam –SO2 –H2S) and most likely these are two completely
different fragments although they have the same nominal m/z
value. However, a fragment corresponding to the cleavage of
the amide bond (m/z 236) could not be found in the meloxicam
standard. The reason for this might be that the aromatic oxidation
in the metabolite shifted the localization of the deprotonation and
hence affected the fragmentation. This is one example of a case
where it was useful to be able to compare the spectra with those
of the deuterium labelled meloxicam to make sure that the m/z
252 fragment was really derived from the meloxicam metabolite
although the two m/z 252 fragments could not be identical.
Meloxicam metabolites in C. elegans incubations and horse
urine
Awaytofacilitatetheidentificationofdrugmetabolitesincomplex
matrices is to administer a mixture of the drug and its deuterium
labelled analog and then look for the resulting doublet peaks
in the full scan mass spectra.[24] The presence of the mass shift
between the fragmentions of the drug and the deuterium labelled
drug can also make the interpretation of noisy mass spectra easier.
We applied this approach in our search for phase I metabolites
produced by C. elegans and added a one-to-one mixture of
meloxicam and CD3-meloxicam to the microbial incubations. The
result was a shift of three in the m/z of the molecular ions and
their fragment ions for the biodegradation products formed from
meloxicam and deuterium labelled meloxicam. The result was a
faster sifting of ions of interest from the many spectra produced.
Four kinds of broth were compared at pH 5.5 and 7.0 in
order to find the optimum cultivation medium for meloxicam
metabolite production. Approximately 75–90% of the meloxicam
added to the C. elegans cultures was biotransformed into several
metabolites. The different kinds of broths did not result in any
major differences in meloxicam metabolite formation except for
the yeast broth at pH 7.0 that resulted in significantly lower yield of
metabolites. The Sabouraud dextrose broth (mycological peptone
10 g/l and dextrose 20 g/l, pH 7.0) was chosen for the metabolite
isolation cultures.
The C. elegans samples were cleaned up and concentrated by
solid phase extraction, and the urine samples by liquid–liquid
extraction, before analysis with the triple quadruple mass
spectrometer (for details see section on Experimental). No
significant differences were observed between the MS/MS spectra
of meloxicam produced with the triple quadrupole compared to
the ion trap instrument, in neither the positive nor the negative
ion mode and thus, the identification of the metabolites was
achieved by the interpretation of the mass spectral data from the
triple quadrupole mass spectrometer and comparison with the
fragmentation of the meloxicam standard solution from the ion
trap instrument. The results are summarized below.
5ꢁ-Carboxymeloxicam
Another main metabolite of meloxicam that has been previously
reported in humans, rats, mice, mini-pigs[8] and horses[5,6]
is 5ꢁ-carboxymeloxicam, formed by further oxidation of 5ꢁ-
hydroxymethylmeloxicam. In Fig. 5 the positive and negative
electrospray MS/MS spectra of the peaks corresponding to
[M + H]+ m/z 382 and [M − H]− m/z 380, respectively, from a C.
elegans sample are presented. The positive electrospray spectra
from C. elegans (Fig. 5(a)) and the horse urine samples (results
not shown) were well in agreement with the one published
by Dumasia et al., demonstrating that C. elegans can metabolize
meloxicam into 5ꢁ-carboxymeloxicam.[5] Also in the negative ion
mode, the spectra were in agreement with that found in the
literature (Fig. 5(b) cf. ref. [6]).
Hence, meloxicam could be metabolized to 5ꢁ-carboxymelo-
xicam in both horses and C. elegans. Even though we identified
both 5ꢁ-hydroxymethylmeloxicam and 5ꢁ-carboxymeloxicam in
both species we did not detect any intermediate aldehyde
metabolite.
Two isomers of hydroxymeloxicam
One of the main metabolites reported in both humans[7] and
horses[5,6] is 5ꢁ-hydroxymethylmeloxicam ([M + H]+ m/z 368 and
[M − H]− m/z 366). HPLC-MS/MS analysis in the positive mode
of m/z 368 of the C. elegans and horse urine samples, gave two
peaks in the total ion chromatogram (TIC) that were suspected
to represent hydroxylated metabolites (not shown). The spectra
from the major peak (C.elegans Fig. 4(a) and horse urine 4(b)) were
well in agreement with those of the 5ꢁ-hydroxymethylmeloxicam
metabolite, previously described in the literature.[5,6] The most
abundant fragments in the spectrum of the minor peak were m/z
115 and m/z 141 (C. elegans Fig. 4(c) and horse urine 4(d) cf. 2(a))
were the same as for meloxicam, indicating a second isomer of
oxidized meloxicam where the oxidation was not on the thiazolyl
moiety.
Dihydroxymeloxicam
C. elegans is known to oxidize xenobiotic substances
extensively[10,11] and therefore we were encouraged to look for
other oxidized metabolites as well. Interestingly, MS/MS of m/z
384 ([M + H]+ for meloxicam + 32) resulted in a peak in the TIC
for both the C. elegans sample and the horse urine sample. This
metabolite has never been reported before and we used a fraction
collector to isolate it from the C. elegans cultures, together with
the two main metabolites, in order to confirm the structures by
the use of NMR (a future study) and maybe be able to use the
metabolites as reference material in the doping control laboratory.
The isolated metabolites were also analyzed in the MS2 and MS3
c
www.interscience.wiley.com/journal/jms
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2009, 44, 1026–1037

A mass spectrometric study on meloxicam metabolism in horses
Figure 5. Triple quadrupol MS/MS spectra and fragmentation suggestions of 5ꢁ-carboxymeloxicam from a C. elegans sample in positive (a) and negative
(b) ESI. The precursor ions were m/z 382 and 380, respectively, the collision energy was 15 V, and HPLC gradient I was used. For further information, see
section on Experimental.
modes of the ion trap mass spectrometer, and the MS2 spectra
of the suspected dihydroxylated metabolite in positive and neg-
ative ESI are illustrated in Fig. 6(a) and (b), respectively. The ions
m/z 131, 157 and 200 (Fig. 6(a)) were the same as those found
from 5ꢁ-hydroxymethylmeloxicam (cf. Figure 4(a)) and the other
oxidation must consequently be on the benzothiazine moiety. In
Fig. 6(b) the main fragments were m/z 162, m/z 226, m/z 252,
and m/z 318 (302 + 16 m/z units, cf. Figure 4(g)). When the horse
urine samples were analyzed on the triple quadrupole instrument
the most abundant ions in the spectra were m/z 131 and 157 in
positive ESI (Fig. 6(c)) and m/z 252 and 318 in negative ionization
(Fig. 6(d)). Thus, itappearstobepossiblefortheoxidativeenzymes
in C. elegans and in the horse to either perform oxidation in one of
the two positions in meloxicam or in both. We find it interesting
that the fungus C. elegans appeared to metabolize this drug in the
same way as horses, a fact that can be useful in future studies on
drug metabolism in horses.
scans of m/z for possible conjugates were applied and in Fig. 7 the
TIC, extracted ion chromatograms for m/z 544 and 368, and the
spectrumfromthepeakat28.3 min(gradientII)afterMS/MSof m/z
544 in positive ESI are presented. Hydroxymeloxicam glucuronide
would result in [M + H]+ m/z 544, and as can be seen in the
spectrum the m/z 544 ion disintegrated into m/z 368 (which
corresponds to hydroxymeloxicam) upon CID. Thus, these data
clearly indicates that a glucuronic acid conjugate of meloxicam
can be formed in the horse. Neutral loss scans of 80 and 176 Da
(for sulfate and glucuronic acid metabolites, respectively) did not
result in any additional findings in the horse urine samples, and no
conjugates could be detected in the C. elegans samples that had
been extracted with SPE.
Meloxicam and its metabolites in urine from healthy horses
Nine healthy horses were a part of a pharmacokinetic study of
orally administered meloxicam at the Bjerke Veterinary Clinic
in Norway. The horses were divided in groups where one
served as control (three horses) and the remaining six were
treated with meloxicam. In this study we were interested in
comparing the levels of meloxicam (M) and its main metabolites
5ꢁ-hydroxymethylmeloxicam (M1), 5ꢁ-carboxymeloxicam (M2) and
dihydroxymeloxicam (M3) in urine from the six horses, as well as
investigating for how long time after the treatment was ended
the substances could be traceable in the urine samples. The
only substances available as synthetic standards were meloxicam
and CD3-meloxicam why a proper quantitative analysis of the
metabolites was not possible. Instead we made a relative estimate
based on the ratios between the area of the SRM-peak of the
substance M, M1, M2 or M3 and the SRM-peak of the internal
standard (CD3-meloxicam). This approach worked well for the
purpose of this study.
The urine samples were cleaned up and concentrated by
liquid–liquid extraction before they were injected into the LC-
MS/MS system. The retention times for the substances with
gradient II were 29.7 min (M and CD3-meloxicam), 28.4 min (M1),
30.8 min (M2) and 25.5 min (M3), see Fig. 8.
The six horses in this study showed great similarities in their
meloxicam metabolism and excretion patterns. In Fig. 9 the peak
area ratio was plotted against time for a representative horse
in the study. The maximum concentration of the parent drug
and the two main metabolites 5ꢁ-hydroxymethylmeloxicam and
5ꢁ-carboxymeloxicam in horse urine seemed to be reached on
Other phase I metabolites
According to the literature, NSAIDs of the oxicam family can
be biotransformed to metabolites where the amide bond has
been broken, e.g. a carboxybenzothiazine and a benzothiazine
metabolite of piroxicam have been detected in rat, dog, monkey
and man.[1] Another metabolite that is considered typical for
oxicams is the oxoacetic acid metabolite.[1,7] However, in this
study none of these metabolites was detected.
Hydroxymeloxicam glucuronide
In 2002, Dumasia et al. concluded that some increase in the
concentration of the two main metabolites of meloxicam (i.e. 5ꢁ-
hydroxymethylmeloxicam and 5ꢁ-carboxymeloxicam) could be
observed after enzymatic and base hydrolysis.[5] In another
publication, de Kock et al. could not observe any difference
in the yield of meloxicam or its metabolites after treatment
of the samples with arylsulfatase and/or β-glucuronidase.[6] O-
glucuronides have been detected in bile from humans and
rats after oxidation of the thienothiazine moity of tenoxicam
resulting in 7- or 8-O-glucuronyltenoxicam and in human urine
after oxidation of the benzothiazine group in piroxicam yielding
5ꢁ-O-glucuronylpiroxicam.[1]
In our study, the horse urine was diluted and injected without
further sample preparation into the triple quadrupole mass
spectrometer in order to search for phase II metabolites. MS/MS
c
J. Mass. Spectrom. 2009, 44, 1026–1037
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/jms

A. Tevell Åberg et al.
Figure 6. MS2 spectra and fragmentation suggestions of dihydroxymeloxicam isolated from C. elegans cultures in positive (a) and negative (b) ESI. The
precursor ions were m/z 384 and 382, respectively. The isolated metabolite was injected into the ion trap instrument as a constant infusion (8 µl/min).
Spectra from MS/MS of [M + H]+ m/z 384 (c) and [M − H]− m/z 382 (d) of a horse urine sample in the triple quadrupole instrument (HPLC gradient II).
For more information, see section on Experimental.
day 5 or 6, i.e. the last day of treatment or the day thereafter.
Interestingly, the intensity of the 5ꢁ-hydroxymethylmeloxicam
peak is about three times higher than that of the parent drug
meloxicam, a fact that could be useful in the development of
new analytical methods for the detection of this drug, e.g. for
doping control purposes. Toutain et al. report that the meloxicam
concentration in horse urine was below the limit of quantification
for the analytical method within 3 days after the final dose.[26] In
our study the main part of both the metabolites and meloxicam
seem to be excreted 3 days after the treatment is ended, however
the 5ꢁ-hydroxymethyl metabolite was detectable for at least
14 days after the last administration of the drug in some of
the horses.
There were two differences in our findings compared to
the Dumasia and de Kock publications. Firstly, the main
metabolite in urine from all six horses in this study was 5ꢁ-
hydroxymethylmeloxicam, not 5ꢁ-carboxymeloxicam as in the
refs. [5,6]. Secondly, unchanged meloxicam was not the ma-
jor component found in urine from the horses in our study,
5ꢁ-hydroxymethylmeloxicam was. These differences might be
c
www.interscience.wiley.com/journal/jms
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2009, 44, 1026–1037

A mass spectrometric study on meloxicam metabolism in horses
Figure 7. Chromatograms of MS/MS of m/z 544 of a horse urine sample (a–c) and mass spectrum of the peak at 28.3 min (d). Gradient II was used, the
instrument was the triple quadrupole, the ionization was positive ESI, and the collision energy was 20 V. For other details, see section on Experimental.
explained by the use of different races since the horses in our
study were standardbred trotters while Dumasia and de Kock used
thoroughbred race horses.
Additionally, a semi-quantitative analysis of meloxicam, 5ꢁ-
hydroxymethylmeloxicam, 5ꢁ-carboxymeloxicam and dihydrox-
ymeloxicam was conducted with urine from six standardbred
trotters that had received meloxicam orally for 5 days. Interest-
ingly, 5ꢁ-hydroxymethylmeloxicam resulted in SRM peaks with
three times larger area than that of the parent drug and also
proved to be detectable in urine from some of the horses for at
least 14 days after the treatment ended with the analytical method
used.
Adding up the results presented in this article, it seems
advantageous to choose the 5ꢁ-hydroxymethyl metabolite for
detection instead of meloxicam itself in the development of
analyticalmethodsintheanti-dopingprogrammesinhorseracing.
Furthermore, the fungus C. elegans can be used to produce drug
metabolites for reference compound usage and appears to be a
promising model for phase I drug metabolism in horses.
Conclusions
Four phase I metabolites of meloxicam were detected in horse
urine and C. elegans cultures in this study. A second isomer of
hydroxylated meloxicam and the dihydroxylated metabolite are
described for the first time in this publication together with
the two main metabolites, 5ꢁ-hydroxymethylmeloxicam and 5ꢁ-
carboxymeloxicam, that have been previously reported in horse
urine. To our knowledge, the metabolism of meloxicam has not
been studied in the fungus C. elegans before and we found
it interesting that the metabolites detected were the same in
both species. Thus, the fungus proved to be a good model of
the phase I metabolism of meloxicam in horses and it was also
illustrated that C. elegans can be used to produce larger quantities
of the metabolites in a cheap and simple manner. Concerning
phase II metabolism, an intact glucuronic acid conjugate of
hydroxymeloxicam was detected in horse urine for the first time
in this study, but conjugates were not detected in the microbial
samples.
Acknowledgements
The authors would like to thank the Swedish Foundation of Equine
Research for the financial support to this project (grant number
H0547146), Boehringer Ingelheim Vetmedica for supporting the
clinical trial, and Dr Arne Holm and the staff at the Bjerke
Dyrehospital in Oslo, Norway, for conducting it.
c
J. Mass. Spectrom. 2009, 44, 1026–1037
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/jms

A. Tevell Åberg et al.
Figure 8. SRM chromatograms for (from top to bottom) meloxicam, CD3-meloxicam (the internal standard), 5ꢁ-hydroxymethylmeloxicam, 5ꢁ-
carboxymeloxicam and dihydroxymeloxicam from a urine sample from one of the horses in the study. The HPLC gradient II was used, for details,
see section on Experimental.
Figure 9. A plot with SRM-peak area ratio versus time for meloxicam, 5ꢁ-hydroxymethylmeloxicam, 5ꢁ-carboxymeloxicam and dihydroxymeolxicam for
one of the horses in the study. The blank sample was collected 4 days prior to the first dose, the first dose is administered at day 1 in the graph and the
first sample analyzed is the one taken on the fifth day of administration.
References
14th, International Conference of Racing Analysts and Veterinarians,
D. W. Hill, W. T. Hill (eds). R&W Publications: Newmarket, 2003,
343–348.
[1] T. F. Woolf, L. L. Radulovic. Oxicams: metabolic disposition in man
[6] S. S. de Kock, J. P. Rodgers, L. Jonker. LC/MS/MS study of the
metabolism and excretion of meloxicam in the horse. In
Proceedingsfromthe14th,InternationalConferenceofRacingAnalysts
and Veterinarians, D. W. Hill, W. T. Hill (eds). R&W Publications:
Newmarket, 2003, 135.
[7] J. Schmid, U. Busch, G. Heinzel, G. Bozler, S. Kaschke, M. Kummer.
Pharmacokinetics and metabolic pattern after intravenous infusion
and oral administration to healthy subjects. Drug Metabolism and
Disposition 1995, 23, 1206.
[8] U. Busch, J. Schmid, G. Heinzel, H. Schmaus, J. Baierl, C. Huber,
W. Roth. Pharmacokinetics of meloxicam in animals and the
relevance to humans. Drug Metabolism and Disposition 1998, 26,
576.
and animals. Drug Metabolism Reviews 1989, 21, 255.
[2] J. Schmid,
U. Busch,
G. Trummlitz,
A. Prox,
S. Kaschke,
H. Wachsmuth. Meloxicam: metabolic profile and biotrans-
formation products in the rat. Xenobiotica 1995, 25, 1219.
[3] C. Igualada, F. Moragues, J. Pitarch. Rapid method for the
determination of non-steroidal anti-inflammatory drugs in animal
tissue by liquid chromatography-mass spectrometry with ion-trap
detector. Analytica Chimica Acta 2007, 586, 432.
[4] European Trotting Union. Chapter IV: Regulations regarding anti-
doping rules. International Agreement on Trotting Races, http://
uet-trot.net/UET/ (2008-12-31), 2009. (in press).
[5] M. C. Dumasia, S. Bruce, E. Schofield. Detection of the COX-2
selective NSAIDS: Biotransformation and identification of in vivo
metabolites of meloxicam in equine urine. In Proceedings from the
c
www.interscience.wiley.com/journal/jms
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2009, 44, 1026–1037

A mass spectrometric study on meloxicam metabolism in horses
[9] R. V. Smith, J. P. Rosazza. Microbial models of mammalian
metabolism. Aromatic hydroxylation. Archives of Biochemistry and
Biophysics 1974, 161, 551.
[18] P. Zhang,L. H. Lin,H. H. Huang,H. Y. Xu,D. F. Zhong.Biotransforma-
tion of indomethacin by the fungus Cunninghamella blakesleeana.
Acta Pharmacologica Sinica 2006, 27, 1097.
[10] D. Zhang, F. E. Evans, J. P. Freeman, B. Duhart Jr, C. E. Cerniglia.
Biotransformation of amitriptyline by Cunninghamella elegans.
Drug Metabolism and Disposition 1995, 23, 1417.
[19] Z. Y. Xie, H. H. Huang, D. F. Zhong. Biotransformation of
pantoprazole by the fungus Cunninghamella blakesleeana.
Xenobiotica 2005, 35, 467.
[11] J. D. Moody, J. P. Freeman, P. P. Fu, C. E. Cerniglia. Biotransforma-
tion of mirtazapine by Cunninghamella elegans. Drug Metabolism
and Disposition 2002, 30, 1274.
[20] C. M. Pearce, M. V. Lushnikova. Microbiological production of
omeprazole metabolites by Cunninghamella elegans. Journal of
Molecular Catalysis B: Enzymatic 2006, 41, 87.
[12] S. Pilard, F. Caradec, P. Jackson, W. Luijten. Identification of an
N-(hydroxysulfonyl)oxy metabolite using in vitro microorganism
screening, high-resolution and tandem electrospray ionization
mass spectrometry. Rapid Communications in Mass Spectrometry
2000, 14, 2362.
[13] D. Zhang, H. Zhang, N. Aranibar, R. Hanson, Y. Huang, P. T. Cheng,
S. Wu, S. Bonacorsi, M. Zhu, A. Swaminathan, W. G. Humphreys.
Structural elucidation of human oxidative metabolites of
muraglitazar: use of microbial bioreactors in the biosynthesis of
metabolite standards. Drug Metabolism and Disposition 2006, 34,
267.
[21] D. Zhang, Y. Yang, J. E. Leakey, C. E. Cerniglia. Phase I and phase II
enzymes produced by Cunninghamella elegans for the metabolism
of xenobiotics. FEMS Microbiology Letters 1996, 138, 221.
[22] D. G. Freitag, R. T. Foster, R. T. Coutts, M. A. Pickard, F. M. Pasutto.
Stereoselective metabolism of rac-mexiletine by the
fungus Cunninghamella echinulata yields the major human
metabolites hydroxymethylmexiletine and p-hydroxymexiletine.
Drug Metabolism and Disposition 1997, 25, 685.
[23] Y. Yuan, X. Chen, D. Zhong. Determination of meloxicam in human
plasma by liquid chromatography-tandem mass spectrometry
following transdermal administration. Journal of Chromatography
B, Analytical Technologies in the Biomedical and Life Sciences 2007,
852, 650.
[24] K. Iyer. Mass spectrometry – assisted applications of stable-labeled
compounds in pharmacokinetic and drug metabolism studies. In
Mass Spectrometry in Drug Discovery, D. Rossi, M. Sinz (eds). CRC
Press: New York, 2002, 337.
[25] D. C. Hobbs, T. M. Twomey. Metabolism of piroxicam by laboratory
animals. Drug Metabolism and Disposition 1981, 9, 114.
[26] P. L. Toutain, N. Reymond, V. Laroute, P. Garcia, M. A. Popot,
Y. Bonnaire, A. Hirsch, R. Narbe. Pharmacokinetics of meloxicam in
plasma and urine of horses. American Journal of Veterinary Research
2004, 65, 1542.
[14] E. B. Hansen Jr,
B. P. Cho, W. A. Korfmacher, C. E. Cerniglia.
Fungal transformations of antihistamines: metabolism of
brompheniramine, chlorpheniramine, and pheniramine to N-oxide
and N-demethylated metabolites by the fungus Cunninghamella
elegans. Xenobiotica 1995, 25, 1081.
[15] W. Herath, D. Ferreira, S. I. Khan, I. A. Khan. Identification and
biological activity of microbial metabolites of xanthohumol.
Chemical and Pharmaceutical Bulletin (Tokyo) 2003, 51, 1237.
[16] A. Tevell Åberg, H. Lo¨fgren, U. Bondesson, M. Hedeland. Structure
elucidation of N-oxidized clemastine metabolites by HPLC-MS/MS
andtheuseofCunninghamellaeleganstofacilitatedrugmetabolite
identification. In manuscript 2009.
[17] D. F. Zhong, L. Sun, L. Liu, H. H. Huang. Microbial transformation of
naproxen by Cunninghamella species. Acta Pharmacologica Sinica
2003, 24, 442.
c
J. Mass. Spectrom. 2009, 44, 1026–1037
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/jms