| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
University of Michigan Medical Center, Room 2G332, 1500 East Medical Center Dr., Ann Arbor, MI 48109-0054.
2
Pfizer Pharmaceutical, 2800 Plymouth Rd., Ann Arbor, MI
48105.
3
Department of Pathology, University of Utah, Salt Lake
City, UT 84132.
aAuthor for correspondence. Fax 734-763-4095; e-mail annesley{at}umich.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: The metabolite was isolated from uremic sera using solid-phase extraction and HPLC. Structural analysis was performed using HPLC–tandem mass spectrometry, nuclear magnetic resonance (NMR), deuterium exchange, and chemical derivatization. Immunoreactivity was evaluated using a fluorescence polarization immunoassay.
Results: The metabolite had a parent ion at m/z 457 in the negative-ion mode and fragmented to yield the m/z 251 of phenytoin, as well as other mass fragments of phenytoin. Mass fragments associated with glucuronic acid were also present. The chromatographic peak corresponding to this metabolite demonstrated immunoreactivity sufficient to lead to falsely increased reported values for phenytoin immunoassays. The observed immunoreactivity was also proportional to the relative concentration of the metabolite in collected fractions. Analysis by NMR indicated the presence of phenyl groups with chemical shifts identical to those of phenytoin, as well as the presence of a methylene bridge, which was consistent with the same methylene bridge present on the phosphate ester of fosphenytoin. Comparative analysis of serum samples from renally impaired patients receiving phenytoin vs fosphenytoin using multiple reaction monitoring quantification demonstrated that this metabolite was associated with fosphenytoin administration.
Conclusions: A unique immunoreactive oxymethylglucuronide metabolite derived from fosphenytoin has been isolated from sera from uremic patients receiving this prodrug.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The concept of "prodrug" development has been introduced as a solution to these types of problems. Prodrugs are structural analogs or derivatives of the desired active drug that endow the compound with properties that help mediate the delivery, pharmacokinetics, or action of the drug. Examples of recently introduced compounds include prodrug forms of platelet-activating factor antagonists (1), antiviral nucleoside analogs (2), cytoprotective agents (3), antineoplastic derivatives (4)(5)(6), neuroprotective drugs (7), and anti-influenza virus drugs (8).
One approach to increasing the solubility of drugs containing hydroxyl
groups, or drugs in which a hydroxyl group can be attached to a
chemical bridge, has been the addition of a phosphate group
(6)(9)(10). Fosphenytoin (Fig. 1
) is a phosphate ester prodrug of phenytoin that provides
improved efficacy and safety when given intravenously or
intramuscularly (11). After systemic administration, the
phosphate moiety of fosphenytoin is rapidly cleaved (12) to
form an unstable intermediate that breaks down to yield formaldehyde
and the active drug phenytoin. Subsequent metabolism and renal
clearance should follow the same hydroxylation and conjugation pathways
known to exist for phenytoin. A recent report (13) presented
data that supported the hypothesis of a unique fosphenytoin metabolite
found in renally impaired patients receiving fosphenytoin. This
purported metabolite demonstrated substantial cross-reactivity in
numerous commercial phenytoin immunoassays, leading to falsely
increased reported values for phenytoin.
|
In this study, we confirmed that a novel metabolite of fosphenytoin does indeed exist. We were able to initially identify, using gradient reversed-phase HPLC–tandem mass spectrometry (HPLC-MS-MS),1 a compound with a m/z 251 phenytoin ion and a parent m/z 457 ion. This metabolite was subsequently isolated and characterized by chemical derivatization, HPLC, MS-MS, exact mass analysis, isotope exchange, and nuclear magnetic resonance (NMR) spectroscopy. The metabolite is immunoreactive in the most widely used immunoassay for monitoring phenytoin and has been structurally identified as the N-3'-oxymethylglucuronide of phenytoin. To our knowledge, this also illustrates the first example of direct glucuronidation of the reactive site on a prodrug.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
drugs and reagents
Phenytoin, 5-(p-hydroxyphenyl)-5-phenylhydantoin
(HPPH), and N-3'-hydroxymethylphenytoin were obtained
from Pfizer Pharmaceutical. Phenytoin-N-glucuronide and
HPPH-glucuronide were synthesized as described previously
(17). TDx phenytoin and free phenytoin reagents were
purchased from Abbott Laboratories and were used according to the
manufacturer’s instructions. The methylating reagent
BF3–methanol was obtained from Pierce Chemical.
Acetonitrile was purchased from Mallinckrodt. Deuterated solvents were
from Cambridge Isotope Laboratories Inc. Oasis HLB 1-mL extraction
columns were obtained from Waters Corporation.
purification of metabolite from serum
Two 1.0-mL aliquots of pooled sera from uremic patients receiving
fosphenytoin (14) were placed into two separate
polypropylene tubes. Acetonitrile (4 mL) was added to each tube with
vigorous vortex-mixing for 60 s. The tubes were centrifuged, and
the acetonitrile supernatants from the two tubes were combined and
dried, in small additions, under nitrogen in a 1.5-mL polypropylene
microcentrifuge tube. The contents of the microcentrifuge tube were
reconstituted into 100 µL of methanol and vortex-mixed, followed by
the addition of 1 mL of water with the final vortex. The liquid was
transferred to a 30-mg Oasis extraction column that had been activated
with 100 µL of methanol and rinsed with water. The column was washed
with 300 µL of water–acetonitrile (95:5 by volume). Phenytoin
metabolites were eluted from the column with four serial additions of
water–acetonitrile (80:20 by volume). The individual eluates were
analyzed qualitatively for the presence of the m/z 457
metabolite using a Perkin-Elmer Series 200 autosampler and pump
connected to a Micromass Quattro Ultima tandem mass spectrometer.
Fractions containing the metabolite were pooled and dried under
nitrogen. The pooled solid-phase-extracted fractions containing the
metabolite were reconstituted in methanol–water (10:90 by volume), and
the m/z 457 compound was isolated chromatographically using
a 2.1 x 150 mm Zorbax RX-C18 column with a mobile phase of
acetonitrile–1 mL/L acetic acid in water (22:78 by volume) at a flow
rate of 0.3 mL/min. Fractions (1 min; 0.3 mL) were collected in 1.5-mL
polypropylene microcentrifuge tubes and analyzed by MS-MS for
phenytoin, HPPH, HPPH-glucuronide, phenytoin-N-glucuronide,
and the fosphenytoin metabolite. Phenytoin-N-glucuronide and
HPPH-glucuronide were monitored in the positive-ion mode because of
better response, and the other compounds were monitored in the
negative-ion mode for the same reason. Fractions containing the
fosphenytoin metabolite were pooled, evaporated, and reconstituted in a
small volume of methanol–water (10:90 by volume). Further purification
was performed using a 2.0 x 100 mm Betamax Acid column, using a
mobile-phase of acetonitrile–1 mL/L acetic acid plus 10 mmol/L
ammonium acetate in water (85:15 by volume) at a flow rate of 0.3
mL/min. As above, 1-min fractions were collected and analyzed by MS-MS.
Fractions containing the metabolite were pooled and evaporated. To
remove any residual ammonium acetate, a final solid-phase extraction,
similar to that described above, was performed.
immunoassay reactivity
During chromatographic separation and purification of the
m/z 457 fosphenytoin metabolite, the 1-min (0.3 mL)
fractions were analyzed using the TDx free phenytoin fluorescence
polarization immunoassay. A 150-µL aliquot of each eluate fraction
was placed in a 0.5-mL polypropylene microcentrifuge tube and
evaporated under nitrogen to remove mobile phase containing
acetonitrile. The dried tube was reconstituted with 150 µL of 9 g/L
NaCl, vortex-mixed, and analyzed on the TDx instrument.
exact mass analysis
Exact mass analysis was performed on a Mariner (PE Biosystems)
time-of-flight mass spectrometer. The instrument was tuned and
calibrated on HPPH-glucuronide and phenytoin-N-glucuronide.
Direct syringe injection of samples was performed. The elemental
composition software used was Masslynx, Ver. 3.3.
nmr analysis
NMR data were acquired on a Varian Inova 600 HPLC-NMR, using the
VNMR 6.1B program, equipped with a
1H-{15N,
13C} flow cell with a 60-µL active volume, in
series with a Waters ZMD single quadrupole mass spectrometer. To
alleviate the effects of exogenous impurities, HPLC isolation of the
m/z 457 metabolite for online NMR was performed on a Zorbax
XDB-C8 column (2.1 mm x 15 cm) in series with an Upchurch Uptight
guard column (Upchurch Scientific) packed with
C18 pellicular (40 µm particle size).
Separation was achieved using a mobile phase of deuteroacetic acid (1
mL/L) in deuterium oxide–deuteroacetonitrile (80:20 by volume; flow
rate, 0.3 mL/min). The peak corresponding to the metabolite was
monitored using m/z 461 (negative-ion mode; four
exchangeable protons by 2H). NMR data were
acquired at the apex of the peak. The 1D NMR spectra were generated
using fast Fourier transformation of the free induction decay, which
were the sums of 18 352 transients. Each transient was induced using a
nonselective 90° 1H pulse with a selective
presaturation pulse applied during the relaxation decay to the HDO
resonance. The resulting time-averaged free induction decay was
multiplied by an exponential decay function (line broadening, 3 Hz) to
enhance the signal-to-noise ratio.
measurement of metabolite in serum
Relative amounts of the metabolite were assayed in sera from 12
uremic patients receiving fosphenytoin and 12 uremic patients receiving
phenytoin. Uremic patient serum (3 µL) plus blank plasma (10 µL),
to increase the mixing volume, were placed into a 0.5-mL polypropylene
microcentrifuge tube. This was followed by the addition of 300 µL of
acetonitrile with vigorous vortex-mixing for 30 s. After
centrifugation for 5 min, the acetonitrile supernatant was transferred
to a second 0.5-mL tube and evaporated under nitrogen. The residue was
reconstituted with 5 µL of methanol, vortex-mixed, and 45 µL of
water was added for a total volume of 50 µL. After vortex-mixing, the
contents were transferred to a glass injection vial, and 10 µL was
analyzed by HPLC-MS-MS. Chromatography was performed on a 2.1 x
150 mm Zorbax RX-C18 column, with a mobile phase of acetonitrile–1
mL/L acetic acid in water (15:85 by volume) at a flow rate of 0.3
mL/min. The integrated area response of the metabolite was monitored in
the multiple reaction monitoring mode with the transition of
m/z 457 to 193 using the Micromass Quattro Ultima tandem
mass spectrometer.
deuterium exchange experiments
Individual solutions (50-µL aliquots) of phenytoin, HPPH,
HPPH-glucuronide, phenytoin-N-glucuronide, and the purified
fosphenytoin metabolite were incubated with 200 µL of deuterium oxide
for 1 h at ambient temperature. After this incubation, 10-µL
volumes were injected directly into the Micromass Quattro Ultima mass
spectrometer, using the Perkin-Elmer autosampler, in a mobile phase of
acetonitrile–deuterium oxide (50:50 by volume). No analytical column
was used because no separations were needed.
carboxyl group methylation
A solution (10 µL) containing purified fosphenytoin metabolite
was incubated overnight with an excess of BF3 in
methanol. The mixture was then evaporated, reconstituted in 100 mL/L
methanol, and injected onto a Betamax Acid column using a mobile phase
of acetonitrile–1 mL/L acetic acid plus 10 mmol/L ammonium acetate in
water (85:15 by volume) at a flow rate of 0.3 mL/min.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A complete mass spectrum of the isolated fosphenytoin metabolite,
obtained with the Micromass Ultima MS-MS system, is shown in Fig. 2
. As illustrated in Fig. 2
, there are several special features
to this mass spectrum: The parent m/z 457 ion matches the
accurate mass observed with the Mariner time-of-flight instrument. In
addition, several of the daughter masses can be attributed to phenytoin
via comparison with reference spectra generated for phenytoin using the
same tandem mass spectrometer. These include m/z 251, 131,
103, and 77. The m/z 193, 175, and 113 fragments cannot be
demonstrated to be products of phenytoin, but they can be attributed to
parent and daughter fragments known to exist for glucuronic acid.
|
chromatographic characteristics
The retention properties under several different chromatographic
conditions were also consistent with the proposed structure of the
fosphenytoin metabolite. Using conventional reversed-phase
chromatography on a C18 column, the metabolite
demonstrated a shorter retention time similar to the polar phenytoin
glucuronide metabolites (all times <7 min). In contrast, phenytoin was
highly retained on the column with a retention time of 14 min.
Retention properties were also checked using a Betamax Acid column. This column uses a special embedded polar group in a long alkyl chain and is designed to have a special affinity for carboxylic acid groups. As expected, phenytoin and HPPH did not interact strongly with the column. In contrast, the fosphenytoin metabolite, HPPH-glucuronide, and phenytoin-N-glucuronide were observed to have stronger interactions and longer retention times. Retention of the fosphenytoin metabolite and phenytoin-N-glucuronide, which have a free carboxylic acid (18)(19), were very close; these compounds were also the last eluting peaks.
immunoreactivity of fosphenytoin metabolite
The immunoassay response plotted against chromatographic time for
a serum specimen from one of the uremic patients receiving fosphenytoin
is shown in Fig. 3
. The chromatography time was expanded to maximize separation.
The retention times of the various hydantoin compounds are also
indicated in Fig. 3
, illustrating that two predominant peaks with
measurable immunoreactivity were observed. The first peak (10–14 min)
was identified as phenytoin, and the second peak (26–32 min) was
identified as the m/z 457 fosphenytoin metabolite. This
demonstrates that the fosphenytoin metabolite has a readily observed
cross-reactivity in this particular immunoassay.
|
During various stages of purification of the metabolite from uremic
sera, as well as during analyses of acetonitrile precipitates of serum
samples, chromatographic eluate fractions containing the metabolite
were assayed to demonstrate a correlation between the relative
concentration of the metabolite and the phenytoin concentration as
measured by immunoassay. A direct correlation exists and confirms that
the metabolite does indeed cross-react with the phenytoin immunoassay,
as illustrated by Fig. 4
.
|
nmr structural analysis
The NMR data obtained for the metabolite and two reference
compounds are illustrated in Fig. 5
. Phenytoin-N-glucuronide (top spectrum) demonstrated
two salient features by NMR. The two phenyl rings showed two distinct
resonances at 7.3 and 7.4 ppm, whereas the glucuronide moiety produced
the five resonances (1'–5') consistent with the presence of such a
carbohydrate structure. For hydroxymethylphenytoin (Fig. 5
, middle
spectrum), the phenyl rings also showed the same resonance pattern
because they were not modified or conjugated in any way. Because no
glucuronide is present on hydroxymethylphenytoin, none of the
resonances associated with the carbohydrate was observed. However, the
added CH2 group (peak A) produced a predicted
single resonance at 5.0 ppm. The NMR pattern for the fosphenytoin
metabolite (Fig. 5
, bottom spectrum) showed resonances consistent with
both of the other reference compounds. The resonance signals for the
two phenyl rings are identical at 7.3 and 7.4 ppm, showing that the
point of conjugation is not on either of the phenyl rings. The
resonances 1'–5' between 3.0 and 4.5 ppm were consistent with what
would be expected for a glucuronide. The CH2
bridge (peak A) was present, but was split because the two hydrogens
were under the influence of two different elements: the hydantoin ring
and the glucuronide ring.
|
deuterium exchange
The results of hydrogen-deuterium (H-D) exchange helped confirm
the identity of the fosphenytoin metabolite. Because a hydrogen (in
this case deuterium after exchange) molecule usually is lost in the
negative-ion mode, one of the exchanged deuteriums will be lost;
therefore, the observed m/z for the compound will be 1 mass
unit less than the actual m/z value. It should be noted that
C-H bonds are not readily exchanged, and the major H-D exchanges will
be observed for N-H and O-H bonds. The data listed in Table 1
show that the predicted and observed changes in mass
composition because of deuterium exchange are all consistent with the
structures of the hydantoin compounds, including the proposed structure
of the fosphenytoin metabolite.
|
carboxyl group derivatization
Our chromatographic studies demonstrated that the fosphenytoin
metabolite was retained on the Betamax Acid HPLC column, which was
presumed to result from the presence of a carboxyl group on the
molecule. Derivatization with BF3–methanol,
which selectively methylates any carboxyl group present, led to an
observed loss of the m/z 457 metabolite at the expected
chromatographic retention time and the appearance of a new compound
with a m/z of 471 that eluted near the void volume. This
change of 14 mass units was consistent with conversion of a single COOH
group to COOCH3.
metabolite concentrations in uremic patient sera
The comparison of sera from uremic patients receiving phenytoin vs
fosphenytoin is summarized in Table 2
. Sera samples (n = 12) in each group were all from
different patients. The metabolite was found in the sera from the
patients who had received fosphenytoin, but was not readily
identifiable above the background response in the sera from
phenytoin-treated patients.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Although HPLC-MS-MS is a valuable tool for obtaining structural information, H-D exchange methods are useful for the determination of exchangeable functional groups on parent and metabolite structures and therefore as an aid in structure validation (20). In our experiments, deuterated solvents were used in the HPLC mobile phase to prevent back-exchange. Phenytoin has two labile hydrogen atoms at the secondary amino groups and will exchange these for two deuterium atoms. With the loss of one of these during negative-ion-mode analysis, the m/z value for phenytoin would be predicted to change from 251 to 252 after H-D exchange, which is what we observed. If the fosphenytoin metabolite were a hydroxymethylglucuronide, we would predict a change of 5 mass units from exchangeable hydrogen atoms (4 on the glucuronide, 1 on the hydantoin ring). The loss of one deuterium atom in the negative-ion mode would yield a mass of 461, which was exactly what we observed.
The H-D exchange experiments were consistent with our hypothesis that the fosphenytoin metabolite was a unique glucuronide conjugate. These experiments did not, however, demonstrate whether the conjugation was through the hydroxyl or the carboxyl group of glucuronic acid. This question was solved through our chromatography data, as well as through the BF3–methanol derivatization data. If the carboxyl group were intact, then it would be methylated with this carboxyl-derivatizing reagent, yielding a compound with a parent mass 14 units higher. Additionally, if the carboxyl group were derivatized, then the glucuronide would be poorly retained on the Betamax Acid HPLC column. The results of our experiment showed that (a) before BF3 derivatization, the m/z 457 metabolite was retained on the Betamax Acid column with a retention time similar to that of phenytoin-N-glucuronide; (b) after derivatization, the m/z 457 peak disappeared, and a compound of m/z 471 was observed; and (c) the m/z 471 compound was poorly retained on the column. These observations were consistent with the conjugation involving one of the hydroxyl groups on glucuronic acid.
With the HPLC-MS-MS, H-D exchange, exact mass, elemental composition, and derivatization data, we were able to account for the presence of both phenytoin and glucuronic acid as part of the structure of the fosphenytoin metabolite. These two entities would account for 444 mass units from the overall 458 mass for the complete metabolite, leaving 14 mass units to be explained. The possibilities allowed by the elemental composition software calculations showed that only a CH2, and not an N, could account for this mass. The NMR data were invaluable in confirming that the two phenyl groups were not conjugated and that the CH2 was present on a bridge between the phenytoin and glucuronide.
Because this is a new metabolite and insufficient purified reference material was available for true quantification, we were only able to measure the relative amount of the oxymethylglucuronide metabolite in sera from uremic patients receiving fosphenytoin vs phenytoin. Our data demonstrate that the metabolite is present in the sera from patients receiving fosphenytoin. These results, in conjunction with our NMR data showing that the methylene bridge from fosphenytoin is still present on the metabolite, make us confident that fosphenytoin is the source of the new metabolite.
Considering the lower cross-reactivity of HPPH-glucuronide in commercial immunoassays, why would this metabolite seem to be sufficiently immunoreactive to cause the large false increases reported for phenytoin (13) in some patients receiving fosphenytoin? One contributing factor may be the fact that, for this novel metabolite, the glucuronide is conjugated onto the hydantoin ring of phenytoin rather than the phenyl ring. Another possibility is that the ester bridge may allow a better structural separation of the glucuronide and phenytoin portions of the compound, thus allowing a better conformational rotation and better reactivity with the antibody.
The implications of our identification of this novel metabolite cross
several areas, including analytical effects (13), drug
metabolism, and pharmacodynamics. This study demonstrates the direct
glucuronidation of a prodrug reactive site. This same sort of
reactive-site glucuronidation (or other type of conjugation) could
theoretically occur with other prodrugs, yielding a route of metabolism
that differs from conventional predictions. There currently are several
types of chemical derivatives (e.g., prodrugs) of existing drugs that
could be candidates for the same type of glucuronidation. In the area
of anticancer therapy, benzyl phosphate derivatives of
-N-heterocyclic carboxyaldehyde thiosemicarbazone
ribonucleotide reductase inhibitors have been synthesized and used in
clinical investigations (21). In a manner similar to that
for fosphenytoin, there is a proposed bioactivation pathway for these
anticancer drugs where cleavage of the phosphorus–oxygen bond leads to
an unstable intermediate, followed by a cascade fragmentation effect
where quinone methide and CO2 are lost to yield
the desired active drug species.
Phosphate monoesters, such as fosphenytoin, are not the only group of prodrugs that contain a reactive site that could be chemically modified. For drugs that have a primary or secondary amine, aromatic amine, phenol, or sterically hindered secondary alcohol, phosphoryloxymethyl carbonate and carbamate prodrugs have been synthesized (22). These compounds actually yield two potentially reactive intermediates after hydrolysis by phosphatase enzymes. The first intermediate, a reactive hydroxymethyloxycarbonyl group that has a terminal hydroxyl group, decomposes to yield formaldehyde and a second unstable intermediate containing a terminal oxycarbonyl group. This secondary spacer group is believed to ultimately release CO2 and the parent drug. Yet another approach to developing a hydrolyzable promoiety has been the development of acyloxymethyl and carboxyalkyl substituents (23)(24). Reactive intermediates can arise from prodrugs that contain ionizable promoieties other than phosphate (24)(25). Sulfonamidomethyl ester prodrugs of benzylpenicilloates have been synthesized and characterized for reactivity and chemical hydrolysis, and a dissociative mechanism for sulfonamide ester hydrolysis has been proposed (24).
From a pharmacodynamic perspective, an important question relates to whether this novel glucuronide metabolite possesses any pharmacologic activity. Generally, glucuronidation is believed to yield inactive metabolites that should be cleared via renal excretion. However, this is not always the case. An acylglucuronide conjugate of mycophenolic acid has been identified (26) that possesses pharmacologic potency comparable to that of the parent drug (27). Morphine-6-glucuronide is considered to be even more active than the parent drug (28). Thus, a definitive answer awaits synthesis of sufficient compound to perform these experiments for our described metabolite.
In summary, we have identified a unique metabolite isolated from sera from uremic patients receiving fosphenytoin. Using multiple analytical techniques, we have verified the structure as a glucuronide conjugate of the hydroxymethyl prodrug intermediate. We believe that the demonstration of direct conjugation of a prodrug reactive site should be considered a possible metabolic route for numerous structural classes of prodrugs.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  M. Oellerich and V. W. Armstrong Prodrug Metabolites: Implications for Therapeutic Drug Monitoring Clin. Chem., May 1, 2001; 47(5): 805 - 806. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, MS-MS integration.
