![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Vol. 26, Issue 5, 418-428, May 1998
Biomedical NMR Group, Interactions Moleculaires et Reactivite Chinique et Photochinique Laboratory, Université Paul Sabatier (C.J., R.M., V.G., M.M.-M.), Centre Claudius Regaud (P.C.), and ASTA Medica AG (U.N.).
 |
Abstract |
|---|
Top
Abstract
Introduction
Materials & Methods
Results
Discussion
References
|
|---|
Phosphorus-31 NMR spectroscopy was used to analyze urine samples from patients treated with cyclophosphamide (CP) on 2 consecutive days. CP and all of its known phosphorylated metabolites except the tautomeric pair 4-hydroxycyclophosphamide/aldophosphamide, i.e. carboxycyclophosphamide (CXCP), dechloroethylcyclophosphamide (DCCP), alcophosphamide, ketophosphamide, and phosphoramide mustard (PM), were determined. Several other signals corresponding to unknown CP-related compounds were observed. Seven of them were identified; all were hydrolysis products of CP or its metabolites (one from CP, two from CXCP, three from DCCP, and one from PM). Twenty-four-hour urinary excretion of unmetabolized CP was not significantly different on the first (17% of the daily administered dose) and second (16%) days of treatment. The amounts of phosphorylated metabolites excreted in 24-hr urine samples were much higher after the second CP dose (37%) than after the first (20%), suggesting autoinduction of CP metabolism. CXCP and its two degradation products (accounting for 7-10% of CXCP) were by far the major metabolites (11.5 and 23% after the first and second doses, respectively). DCCP plus its degradation products and alcophosphamide each represented 2-3% on the first day of treatment and 5% on the second day of treatment. Levels of PM and its degradation products were extremely low (0.3 and 0.6% after the first and second CP doses, respectively), as were those of ketophosphamide (0.4 and 0.6% on the first and second days of treatment, respectively). We noted only modest interpatient variation in excreted levels of CP and all of its metabolites.
| |
Introduction |
|---|
|
Top
Abstract
Introduction
Materials & Methods
Results
Discussion
References
|
|---|
CP1
was introduced in tumor therapy in 1958 and is
currently the most widely used alkylating agent in human drug therapy
(Brock, 1989
; Sladek, 1994). As shown in fig.
1, CP is a prodrug that requires
biotransformation to become cytotoxic (Moore, 1991; Sladek, 1988,
1994). Activation as well as detoxication pathways are mediated by
hepatic CYP enzymes. Multiple CYP forms, including CYP3A4, CYP2B6,
CYP2C8, and CYP2C9, are capable of activating CP in human hepatocytes
(Chang et al., 1997), whereas CYP3A enzymes catalyze >95%
of the CP detoxication reaction in rat liver microsomes (Yu and Waxman,
1996) and CYP3A4 is the major enzyme involved in CP detoxication in
human liver microsomes (Bohnenstengel et al., 1996). First,
hydroxylation of the oxazaphosphorine ring at the carbon-4 position
(activation pathway) leads to the formation of OHCP, which exists in
equilibrium with its ring-opened tautomer AldoCP. Spontaneous
-elimination of urotoxic acrolein (Brock et al., 1979)
from AldoCP yields PM, the active alkylating species. OHCP may be
partially deactivated to KetoCP by an alcohol dehydrogenase. AldoCP may
be either oxidized to inactive CXCP by an aldehyde dehydrogenase or
reduced to AlcoCP by an aldehyde reductase (Sladek, 1994) (fig. 1).
Second, N-dechloroethylation of CP (detoxification pathway)
produces DCCP and chloroacetaldehyde, a compound that may be
responsible for the oxazaphosphorine-induced neurotoxicity, urotoxicity, and cardiotoxicity (Goren et al., 1986; Pohl
et al., 1989; Joqueviel et al., 1997b).
|
Extensive pharmacokinetic data on CP itself have been published and
reviewed by Sladek (1988, 1994) and Moore (1991). However, much less is
known about the pharmacokinetics of CP metabolites. In many of the
early pharmacokinetic studies, total alkylating activity was measured
in the biofluid of interest by a colorimetric method based on the
ability of the alkylating species to react with NBP (Friedman and
Boger, 1961). This so-called NBP-alkylating metabolites assay (CP is
devoid of alkylating activity) (Bagley et al., 1973) does
not provide information about specific metabolites, because all of the
alkylating agents present in the sample are measured indiscriminately
(Bagley et al., 1973; Fuks et al., 1981; Egorin
et al., 1982; Wilkinson et al., 1983).
Until recently, there was no single method available for the specific
determination of CP and its main metabolites in body fluids. The
TLC-photographic densitometry technique described by Hadidi and Idle
(1988) and modified by Boddy et al. (1992) and Tasso
et al. (1992) was the first method that reliably detected, in a single assay, CP, CXCP, DCCP, and KetoCP, with concentration thresholds ranging from 2 to 8 µM, depending on the compound (Tasso et al., 1992). However, contrary to claims made in the
original report by Hadidi and Idle (1988), PM and its cleavage compound NNM cannot be assayed by this method (Yule et al., 1993),
which involves solid-phase extraction of the sample and NBP
derivatization. Almost all of the studies on CP metabolism have used
this method (Hadidi et al., 1988; Boddy et al.,
1992; Tasso et al., 1992; Yule et al., 1995).
Momerency et al. (1994) described a GC/MS method that
determines accurately and with very high sensitivity (
2 nM) CP,
CXCP, DCCP, KetoCP, AlcoCP, NNM, and its adduct with bicarbonate ion in
plasma, i.e. 3-(2-chloroethyl)-1,3-oxazolidin-2-one, but not OHCP or PM. This method requires two differential extraction treatments as well as one or two derivatization procedures, depending on the
compound. To our knowledge, no pharmacokinetic studies have been
carried out using this method.
Chan et al. (1994) published a pharmacokinetic study of CP
and its metabolites using a GC/MS system and a stable isotope-labeled method, which adequately quantified CP, AlcoCP, PM, OHCP/AldoCP tautomers, and NNM, with detection limits ranging from 0.07 to 0.2 µM, depending on the compound. This analytical method has the
advantage of compensating for any procedural loss or decomposition during work-up, by addition of appropriate deuterium-labeled internal standards. The drawback is that the deuterium-labeled derivatives must
be synthesized first. Moreover, this technique requires two separate
clean-up procedures for the samples and a derivatization procedure.
Compared with these chromatographic methods, 31P
NMR presents significant advantages. It enables direct analysis of
crude urine samples, avoiding the problems encountered during clean-up
and derivatization procedures, as well as those stemming from the pH
sensitivity of many metabolites. It allows detection and
quantification, in a single assay, of all of the phosphorus-containing
compounds, because only the presence of a phosphorus atom is required
for detection. The only major difficulty is its relative insensitivity. Nevertheless, it has been used for quantitative study of the metabolism of IF (Martino et al., 1992; Gilard et al.,
1993a). We report here a direct qualitative and quantitative
31P NMR analysis of CP and its phosphorylated
metabolites in the urine of four patients treated with CP, with
elucidation of the structures of seven unknown compounds.
| |
Materials and Methods |
|---|
|
Top
Abstract
Introduction
Materials & Methods
Results
Discussion
References
|
|---|
Materials.
CXCP, DCCP, AlcoCP, KetoCP, PM, and DDCCP were generously supplied by
ASTA Medica AG (Frankfurt, Germany). CP and IF were obtained from ASTA
Medica (Bordeaux, France). Cr(acac)3 and MPA were
purchased from Spectrométrie Spin et Techniques (Paris, France)
and Aldrich (St. Quentin Fallavier, France), respectively. 9-OdAP,
PAE1, and PAE2 (fig. 2) were synthesized
as described by Gilard et al. (1994). PAmA, degradation
products of CXCP (PAEAc and PAEAm) (fig. 1), and DCPM
[P(O)(OH)(NH2)(NHCH2CH2Cl)]
were obtained using the methods described by Sheridan et al.
(1971), Joqueviel et al. (1997a), and Wang and Chan (1995),
respectively.
|
80°C.
Although 6-OAP (fig. 2) was not isolated, it represents 75% of the
reaction mixture (as estimated by 31P NMR) after
an aqueous solution of PAmAE1 is maintained at pH 8 for 3 days at room
temperature. PAmAE2 (fig. 2) was prepared by dissolving 12 mg of DDCCP
in 2 ml of water and adjusting the pH to 4.9. After 1.5 hr at room
temperature, the solution was neutralized with 0.1 M NaOH.
31P NMR showed that PAmAE2 represented 90% of
the reaction mixture. The structures of PAmAE1, 6-OAP, and PAmAE2 were
characterized by MS and 13C NMR.
Patients and Urine Sampling. Four women with breast cancer participated in the study. The median age was 48 years (range, 37-57 years). The patients received CP iv at a dose of 60 mg/kg/day (as a 3-hr infusion in 1 liter of isotonic glucose) for 2 consecutive days. The uroprotector mesna (2-mercaptoethane sulfonic acid) was administered at a dose of 60 mg/kg/day over 5 hr; its infusion was started 2 hr before the start of the CP infusion. Hydration of patients (3 liters of isotonic glucose/m2/day) was started 2 hr before the infusion of CP and was continued for the 2 days of CP treatment. The complete treatment required administration of mitoxantrone at a dose of 45 mg/m2 on day 1, CP on days 2 and 3, and melphalan (L-phenylalanine mustard) at a dose of 140 mg/m2 on day 4. No other phosphorylated drug was administered during the study.
Urine was collected, in 6-hr time periods, for 24 hr after the start of the two CP infusions. Urine volumes were recorded, and then the samples were immediately stored at20°C until the end of the 24-hr
collection period and, after that, at 80°C until NMR analysis,
which was performed within 1 month for samples at pH <6.0 and within
1.5 month for the other samples. Because the patients exhibited marked
diuresis (2.5-7.9 liters/day, primarily >5 liters/day), urine samples
were concentrated. Twenty milliliters of urine were freeze-dried and
resuspended in an accurately measured volume of
H2O (6 ml), with care being taken to wash any
residual pellet several times. The pH of the urine samples ranged from 5.1 to 7.9 and was little affected by the concentration procedure.
31P NMR Analysis.
1H-decoupled NMR spectra were recorded at 121.5 MHz with a Bruker WB-AM300 spectrometer, without nuclear Overhauser
effect. The magnetic field was shimmed on the free induction decay from H2O in the sample. 31P NMR
chemical shifts are reported in ppm with respect to 85% H3PO4, which was used as an
external reference. Spectra were acquired using 10-mm-diameter NMR
tubes under the following instrumental conditions: probe temperature,
4°C (25°C during hydrolysis of CXCP and PM); sweep width, 15,151 Hz; data points, 32,768 zero-filled to 65,536; pulse width, 5 µsec
(i.e. flip angle, 35°); RT, 3.08 sec or 6.08 sec for
qualitative or quantitative purposes, respectively; free induction
decay, processed by exponential multiplication with a line broadening
of 3 Hz; number of transients, 5300-8300. The urine samples were doped
at saturation (about 3 mM) with the paramagnetic agent
Cr(acac)3 to shorten the T1
relaxation times of the phosphorylated compounds. The concentrations of
all of the compounds detected were measured by comparing the areas of their 31P NMR signals with that of MPA, the
standard for quantification, which was placed in a sealed coaxial
insert; all signals were expanded on a 12-Hz/cm scale. The areas were
determined after the different signals had been cut out and weighed.
The external standard [MPA in deuterated water that had also been
doped at saturation with Cr(acac)3 to shorten its
T1 relaxation time, with the deuterated solvent
providing the field frequency lock for the spectrometer] was
calibrated against IF solutions of known concentrations, with recording
conditions [pulse width, 5 µsec (i.e. flip angle,
35°); RT, 6.08 sec] set to produce fully relaxed spectra (Martino
et al., 1992).
; Gilard et al.,
1993a). Because the immediate environment of the phosphorus atom in CP
and its phosphorylated metabolites is identical to that in IF and its
metabolites, the T1 values for CP and IF and
derivatives differed little. We therefore assumed that our previous NMR
recording conditions (flip angle, 35°; RT, 6.08 sec) would yield
fully relaxed spectra in which the peak areas were directly
proportional to concentration. This was verified for CP and CXCP,
representing cyclic and linear phosphorylated CP-related compounds,
respectively. Recording the spectra of solutions of these two compounds
under the conditions described and with a longer RT (10.08 sec), with
all other parameters left unchanged, did not alter the signal
intensities.
The accuracy and precision of the 31P NMR assay
were determined in several experiments. With the recording time used in
this study (9-14 hr), the accuracy and precision of seven assays of CP
at 10
3 M and of AlcoCP and DCCP at
104 M in human urine doped at saturation with
Cr(acac)3 and adjusted at pH 8 or 5 were less
than ±10%. Moreover, we prepared solutions of CP and CXCP at known
concentrations down to 105 M and of DCCP at
a concentration of 5 × 106 M in human
urine doped at saturation with Cr(acac)3, with
the pH adjusted to 7.0. Three to five assays per sample and two or three samples per concentration were analyzed. The accuracy and precision were less than ±10% for concentrations of 5 × 104, 104, and 5 × 105 M and approximately ±20% at
105 M. At 5 × 106
M, the signal-to-noise ratio ranged between 2:1 and 4:1, and this
concentration was the detection limit for our spectrometer. The
accuracy and precision were then approximately ±30-35%. Only the
signals for PAE2 (in one urine sample), PM or its degradation products
resonating at 13-16 ppm (in two urine samples), and unknown
compounds were detected at concentrations of
<105 M. We therefore think that the
quantification of all compounds with concentrations of 
5 × 105 M was accurate (less than ±10%) and that
for concentrations between 5 × 105 and
105 M was acceptable (approximately ± 20%). Only the quantitative data for the few compounds whose
concentrations were <105 M were estimates.
Because of the length of time required for quantitating CP and its
metabolites (9-14 hr), the NMR data were acquired in 2.5-hr blocks.
These blocks were then compared to check the stability of the
phosphorylated compounds detected during the period of NMR recording.
There were no significant differences with time even for the most
acidic urine samples, in which degradation was most rapid. Quantitation
was therefore carried out using spectra resulting from the sum of all
blocks.
The 31P 
values for the degradation products
of CXCP (PAEAc and PAEAm) are close to that of phosphate ion
(Pi), and their signals were thus observed as a
foot on the strong signal for Pi. We verified that standard PAEAc added (at a known concentration of 5 × 105 M) to human urine with the pH adjusted to 8 or 5 was quantified with accuracy and precision of less than ±10%
(mean of seven measurements at each pH).
We also verified that the 3-fold concentration of the urine samples did
not lead to any loss of CP and its metabolites. The amounts of CP,
AlcoCP+CXCP (whose signals were not separable in the crude urine sample
because of the high value for line broadening used to improve the
signal-to-noise ratio of the spectrum), DCCP, and 9-OdAP [the only
phosphorylated compounds, apart from Pi, detected
in a crude 0-6-hr urine fraction (pH 5.2)] were within ±10% of
those measured in the same sample after concentration. However, 11 other CP-related phosphorylated compounds were detected in the
concentrated sample, increasing by 41% the amount of CP metabolites
detected in the fraction. All analyses were therefore carried out with
the concentrated urine samples.
| |
Results |
|---|
|
Top
Abstract
Introduction
Materials & Methods
Results
Discussion
References
|
|---|
Identification of CP and Its Phosphorylated Metabolites in Urine
from Patients.
31P NMR spectra of 3-fold concentrated urine
samples from patients treated with CP at 60 mg/kg/day showed the
presence of up to 30 signals (fig. 3,
A and B). All values mentioned below were
measured at pH 5.8 and 4°C unless otherwise indicated.
|
3-fold concentrated) from
patients before CP infusion demonstrated the presence of the Pi signal and two faint signals in the
phosphomonoester region (2-6 ppm) (fig. 3C). The values
for these peaks were pH dependent, i.e. 1.29 (Pi), 2.45, and 2.68 ppm at pH 5.8 and 3.38 (Pi), 4.78, and 5.17 ppm at pH 7.8. The signal at
2.68 ppm (or 4.78 ppm) was attributed to phosphorylethanolamine and
that at 2.45 ppm (or 5.17 ppm) to glycerol-1-phosphate, as determined
by spiking urine samples, at both pH values, with the standards. The
possibility of interference of the NMR signals of CP-related
phosphorylated compounds with those of compounds from endogenous
metabolism arises only with Pi, by far the major
compound detected in urine. The two endogenous phosphomonoesters were
found only in low amounts and in only a few of the samples analyzed.
To assign some of these signals, urine samples were spiked with CP and
its metabolites and degradation products of CP and its metabolites. The
data are listed in table 1, and fig. 1
illustrates our results. CP (16.24 ppm), CXCP (20.78 ppm), DCCP (15.73 ppm), and AlcoCP (20.92 ppm) were detected in all of the samples.
KetoCP (8.44 ppm) was observed in most of the samples (24 of 27 with pH
values of 
7.5). It was not detected in the five samples at pH
7.8-7.9, probably because it was not sufficiently stable at that pH.
The cytotoxic alkylating agent PM (12.94 ppm) was detected in urine
from the four patients but only in a few samples (7 of 32).
|
3.5-8.5), hydrolysis of CP leads to
9-OdAP, the phosphorus-nitrogen bond breakdown of which is acid
catalyzed, giving rise to PAE1 (fig. 2) (Gilard et al., 1994). Addition of PAE1 to several urine samples produced a new signal
resonating at 2.95 ppm at pH 5.4, demonstrating its absence in the
urine samples.
In acidic medium, DCCP is hydrolyzed in a three-step pathway (fig. 2)
(Gilard et al., 1993b). The first step is the breakdown of
the phosphorus-nitrogen bond of the oxazaphosphorine ring. In the
second step, the linear phosphoramidic acid ester (PAmAE1) is cyclized
with concomitant breakdown of the phosphorus-nitrogen chloroethyl bond,
leading to the six-membered ring compound 6-OAP. The third step is the
breakdown of the phosphorus-nitrogen bond of 6-OAP, yielding PAE2.
Spiking urine samples with PAmAE1, 6-OAP, and PAE2 demonstrated that
the signal resonating at 9.49 ppm, which was observed in most of the
samples (24 of 32), corresponds to PAmAE1. The signals at 6.43 ppm (pH
7.8) and 2.77 ppm correspond to 6-OAP and PAE2, respectively; one or
both signals were observed in all samples except four, in which the
presence of PAmAE1 was noted.
6-OAP and PAE2 can also be derived from acid hydrolysis of DDCCP
(formed from N-dealkylation of DCCP), according to the same degradation pathway as for DCCP (fig. 2) (Gilard et al.,
1993b). The presence of DDCCP and its first degradation product,
PAmAE2, was examined in urine samples. Addition of these two compounds led to the appearance in the 31P NMR spectrum of
two new signals, at 17.62 ppm (DDCCP) and 10.72 ppm (PAmAE2),
demonstrating their absence in the samples. Therefore, the degradation
products 6-OAP and PAE2 were thought to be derived from hydrolysis of
PAmAE1, the first degradation product of DCCP.
The two signals at 1.64 ppm and 1.77 ppm at pH 5.5 and 25°C were
attributed to degradation products of CXCP because their intensities
increased with time, while that of the CXCP signal decreased. These
signals, which were observed in most of the samples (25 of 32), were
identified by spiking urine samples with the two isolated and
characterized products derived from degradation of CXCP in acidic
medium. At 4°C and pH 5.8, the upfield signal (1.90 ppm) corresponds
to PAEAc and the signal at 2.07 ppm corresponds to PAEAm (fig. 1).
To characterize the 31P NMR signals produced by
hydrolysis of PM, patient urine samples at acidic and neutral pH were
spiked with authentic compound. After 13 hr at 25°C in the NMR probe, the signal for PM was markedly reduced. Six signals at pH 5.5 and nine
at pH 6.8, resulting from PM degradation, were observed, of which three
were present in the urine samples before spiking, i.e.
13.11, 3.11, and 1.87 ppm at pH 5.5 and 15.82, 14.85, and 1.31 ppm
at pH 6.8. Only the signal upfield from the
H3PO4 external reference
was identified by spiking with authentic material. It corresponds to
PAmA (fig. 1), giving a signal at 2.63 ppm at pH 5.8 and 4°C.
PM and its degradation products were observed in 25 of 32 samples. PAmA
was observed in 16 (of 18) of the samples with pH values of 6.9. It
was not detected in samples with pH values of 7.0. PM degradation
products resonating (at 4°C) at 13.2 ppm (and also, in a few
samples, at 3.4 ppm) for pH 6.3 or at 16.0 ppm and 15.1 ppm
for pH 6.6 were observed in 18 samples, together with the signals
from PAmA in nine of the samples. These signals consistently
accompanied those of PM. Although these compounds were not identified,
those producing signals between 13 and 16 ppm, i.e. close to
the PM signal, were assumed to contain structures whose environment
around the phosphorus atom was not significantly different from that of
PM. They were thought to be the compounds formed by successive
hydrolysis of chloroethyl groups that had been identified by Watson
et al. (1985):
|
5 × 106 M), despite the 3-fold concentration.
31P NMR Chemical Shifts of Phosphorylated
Compounds as a Function of Urine pH.
It can be seen from the results presented in table
2 that the values for CP and most of
its metabolites (CXCP, DCCP, AlcoCP, 9-OdAP, PAmAE1, 6-OAP, and KetoCP)
were altered little in the range of pH values found in the urine
samples (pH 5.1-7.9) or at which the compounds were detected. The
signals for these compounds could thus be unambiguously assigned in all
spectra. On the other hand, the signals for Pi
and all of the compounds of the same type, i.e. the
phosphoric acid esters PAE2, PAEAc, and PAEAm, exhibited marked
variability in values in this pH range (2.5-3 ppm), because of
protonation of the phosphate group. It should be remembered that the
second pKa of phosphoric acid is 6.7. Although the strong signal from Pi was readily
identified, the signals of the three other compounds were assigned only
after addition of standards to the urine samples at the different pH
values.
|
values between 25°C and 4°C (
0.7 to 0.9 ppm) in
this pH range. The signal was thus assigned only after addition of the
standard to several samples at 4°C. However, compared with phosphoric
acid ester compounds, the signal for PAmA was more readily assigned
because it appears in a region containing few other signals.
Furthermore, there is a relatively small variation in values
(0.8 ppm) in the pH range where it is detected (5.1-6.9), because
of the fact that its two pKa values (3.0 and 8.3 at 4°C) lie outside this range (Gamcsik et al.,
1993).
Influence of Urine pH on Levels of Hitherto Unidentified
Degradation Products.
Because all of the new compounds identified were derived essentially
from hydrolysis (in acidic medium) of CP and some of its metabolites,
we examined whether there was a relationship between the levels of
these compounds and the pH of the urine samples. The percentage of
9-OdAP, with respect to CP, did not depend on the pH of the urine. The
marked differences (table 3) were the
result of the fact that, in the 18-24-hr samples obtained on the
second day of treatment, the amounts of CP excreted were low and highly
variable, leading to marked variations in the levels of 9-OdAP,
i.e. 9.2% (pH 6.9) to 90.7% (pH 6.3). Apart from these four samples, the levels of 9-OdAP did not depend significantly on the
pH of the samples (8%) (table 3). The urinary excretion of 9-OdAP
by the patients represents 6.1% of the CP excreted on the first day of
treatment (5.2-7.6%, depending on the patient) and 7.1% excreted on
the second day (5.5-10.7%).
|
; Joqueviel et al., 1997a).
Despite the precautions taken during collection and storage of the
samples, the degradation products of CXCP represented 40% of CXCP
at pH 5 and 20% at pH 6. At pH values greater than pH 6,
there was relatively little degradation of CXCP in urine (table 3).
Urinary excretion of the degradation products of CXCP by the patients represented 10.4% (range, 1.6-35.5%) of the CXCP excreted on the first day of treatment and 6.6% (range, 3.2-13.6%) excreted on the
second day. The higher percentage observed on the first day was the
result of the fact that the samples collected in the first 12 hr were
the most acidic.
The percentages of the degradation products of DCCP, with respect to
DCCP, did not depend on the pH in the range of 5.8-7.9. On the other
hand, the percentages were markedly higher at pH 5.1-5.3 (57% of DCCP
vs. 22% at pH 5.8) (table 3). The first intermediate
in the acidic degradation of DCCP, PAmAE1 (fig. 2), is hydrolyzed
directly to PAE2 at the most acidic pH values (5.1-5.8). In contrast,
at pH 7.5, PAmAE1 is hydrolyzed only to 6-OAP. At intermediate pH
values (5.9-6.9), we observed 6-OAP with PAE2 (in six of nine samples)
or PAE2 alone. Approximately 40% of PAmAE1 was hydrolyzed in the urine
samples, i.e. 43.7 ± 13.0% into PAE2 at pH 5.8
(N = 6), 37.5 ± 5.6% into 6-OAP at pH 7.5
(N = 5), and 17.8 ± 12.8% into 6-OAP
(N = 6) and 18.5 ± 9.2% into PAE2 (N = 9) at pH 5.9-6.9. Inclusion of these DCCP
degradation products led to an 18.6% increase (range, 12.1-25.9%) in
the amount of DCCP excreted on the first day of treatment and only an
11.0% increase (range, 7.5-22.1%) in that for the second day. This
discrepancy stems essentially from the fact that six of the eight urine
samples in which no degradation products of DCCP were detected were
collected on the second day of CP treatment.
Because PM was always observed with its degradation products in the
13-16 ppm range, it was always quantified with them. PAmA is the major
hydrolysis product of PM at acidic pH. It was the only hydrolysis
product of PM detected in seven samples, whereas it represented 57%
of PM and its degradation products in nine other samples. At pH 7.0,
only PM and its degradation products with signals at 15-16 ppm were
detected. The ratio of PAmA to PM plus PM degradation products at
13-16 ppm was thus a function of the sample pH. For example, the ratio
was 0.1 for one patient (the pH values for most of the urine samples
from this patient were 7.5) and 5.0 for another patient (the pH
values for the urine samples obtained on the first day of treatment
ranged from 5.1 to 6.1).
Urinary Excretion of CP and Its Phosphorylated Metabolites. Table 4 summarizes the data on the 24-hr urinary excretion of CP and its phosphorylated metabolites after the start of each 3-hr CP infusion for the four patients treated on 2 consecutive days. The measured values for four successive 6-hr aliquots after the beginning of each CP infusion were summed to give the total amounts of CP and metabolites excreted in each 24-hr period. Fig. 4 is a stacked bar graph showing the urinary recovery of CP and its phosphorylated metabolites in each 24-hr period; it includes individual data and mean data.
|
|
16% unchanged CP was excreted on day 1 and 15% on day
2. The hydrolysis product of CP, 9-OdAP, represented 1% of the
delivered dose on both day 1 and day 2. This increased the percentage
of unmetabolized CP to approximately 17% of injected CP on day 1 and
approximately 16% on day 2. It can be seen, therefore, that the amount
of unmetabolized CP (either alone or together with its degradation
product 9-OdAP) did not differ significantly (p > 0.1) between the first and second days of treatment.
The major metabolite (12% on day 1 and 23% on day 2) was CXCP,
taking into account its degradation products (PAEAc and PAEAm). DCCP is
the major metabolite in the deactivation pathway of CP, accounting for
2.9% (3.4%, including its degradation products) on the first day of
treatment and 4.5% (5.0%, including its degradation products) on the
second day. AlcoCP was excreted at similar levels, i.e.
2.3% on day 1 and 4.6% on day 2. These percentages are significantly higher than those for KetoCP and PM, which accounted for <0.5% on day
1 and 0.6% on day 2. Excretion of the other unknown phosphorylated compounds derived from CP amounted to 1.6% (day 1) and 2.3% (day 2)
of the injected dose. The urinary excretion of the metabolites of the
activation pathway (CXCP, AlcoCP, PM, KetoCP, and their degradation
products) was significantly higher (3.5-fold on day 1 and 5-fold
on day 2) than that of the metabolites of the deactivation pathway
(DCCP and its degradation products as well as 9-OdAP, the degradation
compound of CP).
Time Profiles for Excretion. The time profiles for excretion of CP and its identified metabolites, including their degradation products, and the sum of unidentified phosphorylated compounds derived from CP are shown in fig. 5 as percentages of the daily administered dose in each urine fraction. The same metabolites were found in each 6-hr fraction. On the first day of treatment, the excretion of CP reached a maximum (in three of four patients) in the 6-12-hr fractions, whereas CP metabolites (except KetoCP, which was at maximal levels in the 6-12-hr fraction) peaked in the 12-18-hr fraction. Excretion subsequently decreased (more for CP than for its metabolites) in the 18-24-hr fraction. The amount of excreted CP fell below that of the sum of its metabolites in the 12-18-hr fraction. On the second day of treatment, we found a marked increase in the urinary excretion of CP, as well as its metabolites, in the 0-6-hr fraction. Excretion of CP and its metabolites peaked in the first 6 hr (except for one patient, whose excretion peaked in the 12-18-hr fraction) and then declined gradually (except for KetoCP, the excretion of which was constant in the first three fractions). Very small amounts of CP (0.4%) and small amounts of CP metabolites (2.8%) were measured in the last collection period (18-24 hr). We thus concluded that excretion of intact CP is negligible 24 hr after the start of the 3-hr infusion and that the metabolites of CP are also nearly all excreted in this period.
|
| |
Discussion |
|---|
|
Top
Abstract
Introduction
Materials & Methods
Results
Discussion
References
|
|---|
All of the phosphorylated CP metabolites thus far described,
except for the tautomer pair OHCP/AldoCP, which we did not attempt to
determine, could be assayed by 31P NMR. For
accurate quantification, OHCP and AldoCP, which are highly unstable,
must be trapped by the addition of appropriate stabilizing reagents,
preferably at the bedside (Chan et al., 1994; Ludeman
et al., 1995).
Seven new phosphorylated compounds were identified in the urine samples
(fig. 1). 9-OdAP is derived from the hydrolysis of CP, and its levels
in urine were found not to depend on pH (table 3). Because CP is little
hydrolyzed at room temperature between pH 3.4 and pH 8.6 (3% after 4 days at 20°C) (Gilard et al., 1994), 9-OdAP may be derived
from CP degradation in vivo rather than degradation during
collection and storage of urine samples.
PAEAc and PAEAm are derived from the hydrolysis of CXCP (Joqueviel
et al., 1997a), in amounts inversely proportional to the pH
of the urine samples (mean, 40% of CXCP at pH 5, <10% at pH >6.0) (table 3). This degradation may occur in vivo and
during collection of urine samples but not during storage at 80°C
for <6 weeks, in which time CXCP has been shown to be little degraded (Joqueviel et al., 1997a). Therefore, urinary measurement of
CXCP does not reflect systemic production if the urine pH is acidic (6).
PAmAE1, 6-OAP, and PAE2 are derived from the acid hydrolysis of DCCP.
Overall, these products made up approximately 20% of the DCCP
detected, and this level was found not to depend on urine pH in the
range of 6-8 (table 3). Because DCCP is stable for at least 3 days
at room temperature at pH 6.5 (Martino et al., 1992),
degradation was thought to have occurred in vivo. On the other hand, the percentage of DCCP degradation was markedly increased at pH 5, which may have occurred during collection and possibly also
during storage of the urine samples.
The absence of DDCCP and its primary hydrolysis product (PAmAE2)
indicated the absence of a sequential N-dechloroethylation route for CP, in contrast to that observed for IF in both humans (Martino et al., 1992; Gilard et al., 1993a) and
rats (Wang and Chan, 1995). We failed to detect DCPM, a compound
produced from DCCP by the same process of metabolic activation as that
transforming CP into PM. This process has been demonstrated by Wang and
Chan (1995), who detected DCPM in the urine of rats treated with IF.
PM is extensively degraded in urine, and its degradation products were
detected with very little of the parent compound. Only the compound
derived from cleavage of the phosphorus-nitrogen bond leading to
liberation of NNM, PAmA, was formally identified. Overall, the
degradation products of CP or its metabolites that were newly
identified in the present study accounted for 3% of administered CP
and made up approximately 15% of the excreted metabolites of CP on day
1 and 10% on day 2.
Table 5 summarizes the data obtained for
the urinary excretion of CP and its metabolites using various
analytical techniques, including NMR. In our study, unmetabolized CP
was the major compound recovered in urine on day 1 of the treatment
(16% of administered dose) and the second major compound on day 2 (14.7%). These amounts are in accord with the values reported by other
authors (11-20%) (table 5; see also Sladek et al., 1980;
Fasola et al., 1991; Chen et al., 1995) but are
much higher than those reported by Jarman et al. (1979) and
Boddy et al. (1992), who administered much lower doses of
CP. The amount of CP excreted on the second day of treatment was
slightly less than that excreted on day 1, in agreement with the
findings of Fasola et al. (1991). For 19 patients treated
with a dose of 60 mg/kg/day for 2 days, those authors observed urinary
excretion of CP of 15.6 ± 6.3 and 12.0 ± 8.4% on days 1 and 2, respectively. However, the difference between the 2 days was not
statistically significant in our study (p > 0.1) and was at the limit of significance in that of Fasola et
al. (1991) (p = 0.03). In both studies,
some patients were found to excrete more CP on the first day of
treatment and others on the second day.
|
The major metabolite (and on day 2 the major compound) was CXCP. It can
be seen from table 5 that the levels we observed were the highest of
those reported for adults and close to the values observed for children
(Tasso et al., 1992; Yule et al., 1995), in whom
CP is metabolized more rapidly (Sladek, 1988, 1994; Moore, 1991).
However, it should be noted that our results included the degradation
products of CXCP, which were not included in other reports. The
rigorous analytical procedure (storage of samples at 80°C, storage
for no more than 6 weeks, and 31P NMR carried out
at 4°C) limited degradation to 1.1% of the injected dose on day 1 and 1.4% on day 2 (table 4). Under less stringent conditions, there
might have been more degradation, with a consequently lower percentage
of excreted CXCP. We observed less interpatient variation in the
amounts of CXCP excreted than in the studies listed in table 5, except
for that of Jarman et al. (1979). Those authors studied
catheterized patients, whose urine samples were left for a maximum of 2 hr at room temperature before being cooled to 4°C to the end of the
collection period (i.e. 24 hr) and then stored at 30°C.
Such collection conditions, which minimize the degradation of CXCP,
could account for the relatively low interpatient variability. The
small amounts of CXCP (5.5%) detected by those authors most likely
resulted from the significant degradation of the compound upon
extraction with ethyl acetate at pH 2, a pH at which it is highly
unstable. The marked variability among the other studies might be
derived, at least in part, from variable degradation resulting from the
pH of the samples, the conditions of urine collection and storage, and
the duration of urine storage before sample analysis. The fact that our
four patients were being treated for the same type of cancer with the
same therapeutic regimen could account for the low interpatient
variability we noted in the urinary excretion of CXCP and the other
metabolites (table 4). CP, which is known to be relatively stable in
urine, showed much lower interpatient variation than that observed by Chen et al. (1995) or that in most of the studies listed in
table 5 or reviewed by Sladek (1994).
The amounts of DCCP excreted were comparable to those found in children
by Yule et al. (1995) and much higher than those detected by
Boddy et al. (1992) in adults and by Tasso et al.
(1992) in children. The percentage of excretion of AlcoCP was
considerably higher (2.3 ± 0.7%) than that reported by Chan
et al. (1994) (0.4 ± 0.4%), with less interpatient
variation. The amounts of KetoCP were markedly lower than those
reported by most other authors (0.4% vs. 1%) (Jarman
et al., 1979; Hadidi et al., 1988; Yule et
al., 1995) but were close to those obtained by Boddy et
al. (1992) and Tasso et al. (1992) (0.4-0.5%).
We found very little PM (essentially in the form of degradation
products) in the urine samples, and levels were much lower than those
reported by other authors. For example, Chan et al. (1994)
reported PM as a major excretion product in urine (39% of the
delivered dose of CP). In fact, quantitation of PM is not straightforward (Jardine et al., 1978; Phillipou et
al., 1993; Momerency et al., 1994; Yule et
al., 1993, 1995), and the results reported by Jardine et
al. (1978), as well as those obtained using the TLC-photographic
densitometry method are considered not to be accurate (Jardine et
al., 1978; Yule et al., 1995). PM is no longer
quantified by this method (Yule et al., 1995). Only the results reported by Chan et al. (1994) are discussed here,
because they were obtained using a deuterium-labeled PM internal
standard, which should circumvent the problem of decomposition of PM
during the assay procedure. Because of the length of time required for the quantitation of PM in urine using 31P NMR
(9-14 hr), we thought that PM would be degraded during the recording
even at 4°C. The half-life of PM at pH 7.4 and ambient temperature is
approximately 2 hr (Boal et al., 1989; Dirven et al., 1994). The ultimate hydrolysis product of PM is
Pi (Engle et al., 1982). We found in
preliminary experiments that approximately one third of the PM was
hydrolyzed within 15 hr at 4°C in cacodylate buffer at pH values
between 5 and 7. Pi, the only hydrolysis product that cannot be detected in urine by using 31P
NMR, accounted for only approximately 5% of the initial concentration of PM under our conditions. It may be that PM is degraded to
Pi during the freeze-drying stage, although
neither PM nor its hydrolysis products were detected in the
nonconcentrated urine samples. If PM were the major compound in urine,
as claimed by Chan et al. (1994), it should have been
detected by 31P NMR at 4°C with its
phosphorylated degradation products. The very small amounts of PM and
its degradation products in urine are therefore not the result of
extensive degradation of PM during NMR recording or freeze-drying.
Nevertheless, degradation of PM during collection and storage of urine,
as well as in vivo, could not be ruled out.
The total amount of excreted metabolites (20% on day 1) was of the
same order of magnitude as values reported by other authors (Bagley
et al., 1973; Jardine et al., 1978; Hadidi
et al., 1988; Tasso et al., 1992; Yule et
al., 1995) but higher than the values reported by Boddy et
al. (1992) and Jarman et al. (1979) and markedly lower
than the values obtained by Chan et al. (1994),
i.e. 42.9%, with 39.0% for PM alone (table 5). However,
the results are not readily comparable because not all of the same
metabolites were measured in the different studies.
Measurement of the radioactivity excreted in urine after treatment with
14C-labeled CP enables determination of the total
excretion of CP. Although this method avoids the problems of
instability and detection of metabolites, it cannot discriminate
between intact CP and its metabolites. The value (62%) obtained by
Bagley et al. (1973) is much higher than values reported by
other authors (37%), apart from Chan et al. (1994) (54%)
(table 5). The difference between the urinary excretion of CP
determined by 31P NMR (36%) and that determined
by measurement of radioactivity (62%) was assumed to result from
degradation of CP or one of its metabolites to
Pi. The Pi was thought not
to be derived from CP, DCCP, or CXCP, the degradation of which does not
lead to formation of Pi at the pH found in urine,
or OHCP/AldoCP or KetoCP, which are present at only low levels in urine
(Sladek, 1988) and the degradation of which could therefore not account
for the magnitude of the difference. A more likely possibility is PM,
the final hydrolysis product of which is Pi.
Several studies have shown that the urinary elimination of CP and its
metabolites is almost complete 24 hr after the start of treatment
(Bagley et al., 1973; Jardine et al., 1978;
Sladek et al., 1980). This was supported by the present
findings. Only 6% of the injected dose on day 1 and 3% on day 2 were recovered in the urine samples collected 18-24 hr after the
beginning of the CP infusion. The marked increase in urinary excretion
of metabolites on the second day of treatment (37.3% vs.
20.4% on day 1) (table 4) can be accounted for by induction of the
metabolism of CP. Autoinduction of the metabolism of CP was first
described by Bagley et al. (1973), who noted more rapid
metabolism of CP from the second day to the fifth day of treatment for
patients receiving daily doses of CP (ranging from 6 to 80 mg/kg/day)
for 5 consecutive days. The plasma half-life of CP was found to be
shorter and the levels of alkylating metabolites in plasma higher on
the fifth day than on the first day of treatment. These findings were
subsequently confirmed by numerous authors (Sladek et al.,
1980; Graham et al., 1983; Fasola et al., 1991)
from the decrease in the plasma half-life of CP. Fasola et
al. (1991) showed that the decrease in half-life (from 8.7 hr on
day 1 to 3.6 hr on day 2) was not accompanied by a significant drop in
the urinary excretion of CP. This indicated that the increase in the
urinary excretion of phosphorylated metabolites of CP is accompanied by
induction of the metabolism of CP, even in the absence of alterations
in the excretion of CP. Very recently, studies by Waxman and co-workers showed that CP induced one of the CYP enzymes (CYP3A4) involved in CP
4-hydroxylation, thus demonstrating an underlying metabolic basis for
this autoinduction phenomenon (Chang et al., 1997).
| |
Footnotes |
|---|
Received September 30, 1997; accepted January 28, 1998.
This study was supported by grants from the Association pour la Recherche sur le Cancer (Grant 6635) and Ligue Nationale Française contre le Cancer (Comité des Hautes-Pyrénées). This study was presented in part at the 88th Annual Meeting of the American Association for Cancer Research (San Diego, CA, April 12-16, 1997).
Send reprint requests to: Dr. R. Martino, Groupe de RMN Biomédicale, Laboratoire des IMRCP, Université Paul Sabatier, 118, route de Narbonne, 31062 Toulouse, France.
| |
Abbreviations |
|---|
Abbreviations used are: CP, cyclophosphamide; CYP, cytochrome P450; OHCP, 4-hydroxycyclophosphamide; AldoCP, aldophosphamide; PM, phosphoramide mustard; KetoCP, ketophosphamide; CXCP, carboxycyclophosphamide; AlcoCP, alcophosphamide; DCCP, dechloroethylcyclophosphamide; NBP, 4-(p-nitrobenzyl)pyridine; NNM, nor-nitrogen mustard; DDCCP, didechloroethylcyclophosphamide; IF, ifosfamide; Cr(acac)3, chromium(III) acetylacetonate; MPA, methylphosphonic acid; 9-OdAP, oxadiazaphosphacyclononane; PAE1, phosphoric acid ester 1; PAE2, phosphoric acid ester 2; PAmA, phosphoramidic acid; DCPM, dechloroethylphosphoramide mustard; PAEAc, phosphoric acid ester of 3-hydroxypropanoic acid; PAEAm, phosphoric acid ester of 3-hydroxypropanamide; PAmAE1, phosphoramidic acid ester 1; 6-OAP, oxazaphosphacyclohexane; PAmAE2, phosphoramidic acid ester 2; RT, repetition time.
| |
References |
|---|
|
Top
Abstract
Introduction
Materials & Methods
Results
Discussion
References
|
|---|
This article has been cited by other articles:
|
|
 |
|
  C. M. Szymanski, F. St. Michael, H. C. Jarrell, J. Li, M. Gilbert, S. Larocque, E. Vinogradov, and J.-R. Brisson Detection of Conserved N-Linked Glycans and Phase-variable Lipooligosaccharides and Capsules from Campylobacter Cells by Mass Spectrometry and High Resolution Magic Angle Spinning NMR Spectroscopy J. Biol. Chem., June 27, 2003; 278(27): 24509 - 24520. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Wei, D. Fabris, and C. Fenselau Covalent Sequestration of Phosphoramide Mustard by Metallothionein---An In Vitro Study Drug Metab. Dispos., July 1, 1999; 27(7): 786 - 791. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
|
 |
 |
 |
