Analytical and pharmacokinetic studies with 5-chloro-2′-deoxycytidine

Article abstract of DOI:10.1016/S0006-2952(02)01413-2

5-Chloro-2′-deoxycytidine (NSC 371331, CDC) is in development as a possible radiosensitizing agent for cancer treatment. Previous studies have been done to demonstrate the in vivo efficacy of CDC with various modulators of its metabolism. This paper descr

Full text of DOI:10.1016/S0006-2952(02)01413-2

Biochemical Pharmacology 64 (2002) 1493±1502
0
Analytical and pharmacokinetic studies with 5-chloro-2 -deoxycytidine
*
JodiAnne T. Hale, James C. Bigelow, Linda A. Mathews, John J. McCormack
Department of Pharmacology and Vermont Cancer Center, University of Vermont, B-322 Given Building, Burlington, VT 05405, USA
Received 16 July 2001; accepted 15 May 2002
Abstract
0
5
-Chloro-2 -deoxycytidine (NSC 371331, CDC) is in development as a possible radiosensitizing agent for cancer treatment. Previous
studies have been done to demonstrate the in vivo ef®cacy of CDC with various modulators of its metabolism. This paper describes our
preclinical studies to determine the pharmacokinetic properties of CDC and the disposition of the drug, both alone and in the presence of
the metabolic modulator tetrahydrouridine (THU), a cytidine deaminase inhibitor. Detection of the drug in biological ¯uids was performed
by HPLC analysis using a C-18 column, gradient elution with solvents composed of aqueous tri¯uoroacetic acid and acetonitrile, and
ultraviolet absorbance at 290 nm. Samples were processed by treatment with ammonium sulfate prior to injection into the HPLC system.
CDC was stable in aqueous solution and in mouse plasma. High doses of CDC (100 mg/kg) were given i.v. or i.p. to mice for the
determination of CDC plasma half-life (10 min). CDC was not detectable in plasma after oral administration. It was converted rapidly to
0
5
-chloro-2 -deoxyuridine (CDU) by cytidine deaminase, and CDU was readily discernable in plasma and urine samples collected after i.v.
and i.p. administration of CDC. When CDC in doses ranging from 5 to 100 mg/kg was given with 100 mg/kg of THU, increased plasma
levels of CDC were seen. CDC was eliminated through the kidneys, as well as by enzymatic deamination, and did not bind to plasma
proteins. The initial steps of the CDC metabolic pathway were determined in vitro with isolated enzymes. Cytidine deaminase from mouse
kidney converted CDC into CDU; thymidine phosphorylase converted CDU into 5-chlorouracil (5-CU). The conclusions of these studies
are: (a) CDC is a drug with a short half-life and (b) it is excreted through the kidney, mainly in metabolite form. Administration of THU
substantially increased the concentrations of CDC in mouse plasma, supporting proposals that the combination of THU with CDC should
be evaluated in clinical trials.
#
2002 Elsevier Science Inc. All rights reserved.
0
Keywords: 5-Chloro-2 -deoxycytidine; Tetrahydrouridine; Pharmacokinetics; Cytidine deaminase; Radiation sensitization; Pyrimidine nucleoside
metabolism
1
. Introduction
anabolized to the active form of CDU. Although CDU, like
IdU and BrdU, is incorporated into DNA, CDC itself is not
catabolized by uridine and thymidine phosphorylases nor
is it dehalogenated by thymidylate synthetase like IdU and
BrdU [2,3]. There is some evidence that CDC itself is
phosphorylated and incorporated into DNA despite being a
poor substrate for deoxycytidylate kinase [2], allowing for
the drug to possibly exert some biological actions inde-
pendent of those associated with the formation of CDU [4].
Studies done with human xenografts in mice have shown
that CDC is tumor selective because many tumor cells have
elevated amounts of deoxycytidylate deaminase, cytidine
deaminase, and deoxycytidine kinase compared to normal
cells [2]. Elevated enzyme levels allow more CDC to be
converted to its active form in the tumor cell than in normal
cells. Therefore, enzyme levels, rather than simple prolif-
erative status of the tissue, determine drug incorporation;
the CDU metabolite is formed by enzymatic deamination
CDC, like other 5-halogenated pyrimidine nucleosides,
is known to sensitize cells to ultraviolet and ionizing
radiation once it is incorporated into DNA. The related
compounds IdU and BrdU, to be effective for radiosensi-
tization, have to be infused continuously to overcome rapid
catabolism by uridine and thymidine phosphorylases, as
well as rapid dehalogenation. Another limitation of these
two drugs is their severe toxicity to normal tissues, which
includes bone marrow suppression and dermal phototoxi-
city [1,2]. CDC, on the other hand, is a pro-drug that is
*
Corresponding author. Tel.: ‡1-802-656-4494; fax: ‡1-802-656-4523.
E-mail address: john.mccormack@uvm.edu (J.J. McCormack).
0
Abbreviations: CDC, 5-chloro-2 -deoxycytidine, cytochlor; CDU, 5-
0
chloro-2 -deoxyuridine; THU, tetrahydrouridine; 5-CU, 5-chlorouracil;
BrdU, 5-bromo-deoxyuridine; IdU, 5-iodo-deoxyuridine; ara-C, cytosine
arabinoside; AUC, area under the curve.
0006-2952/02/$ ± see front matter # 2002 Elsevier Science Inc. All rights reserved.
PII: S 0 0 0 6 - 2 9 5 2 ( 0 2 ) 0 1 4 1 3 - 2

1494
J.T. Hale et al. / Biochemical Pharmacology 64 (2002) 1493±1502
and subsequently phosphorylated and incorporated into
DNA [5].
doses were injected into the lateral tail vein and i.p. doses
were injected laterally in the abdominal region.
Since many tumors have elevated cytidine deaminase
levels, THU, a cytidine deaminase inhibitor, can be used to
prevent systemic deamination of CDC in the serum and
liver with comparatively lesser inhibition of this pathway
in the tumor cells. THU does not add any toxicity or itself
enhance radiosensitization [5]. Even though THU limits
the cytidine deaminase pathway, the deoxycytidylate dea-
minase pathway could still activate CDC after initial
kinase-mediated phosphorylation. Because there are two
metabolic pathways with two distinct kinases leading to
active drug, it is unlikely for a cell to become resistant to
the drug. Although the mechanism of radiosensitization
produced by CDC treatment is not different from that
produced by IdU and BrdU, CDC does appear to have a
lower toxic potential in mice [2,6].
2.3. Collection of samples
Blood was obtained by retro-orbital puncture and col-
lected in heparinized capillary tubes. Samples were taken
at various points over a time course of 5 min to 24 hr.
Plasma was separated by centrifugation (6200 g for 5 min
at room temperature). Urine was collected by clean catch
during blood collection or through the use of a metabolism
cage for the 24-hr samples. Mice were killed by cervical
dislocation, organs were inspected for gross abnormalities,
and all samples were frozen at À208 until analysis.
2.4. HPLC conditions
Because CDC is a non-hypoxic radiosensitizing pro-
drug when it is given to mice with THU, developmental
research has been undertaken to assess its potential value as
part of cancer treatment regimens. In these preclinical
studies, we determined the pharmacokinetics of CDC alone
and with THU. Plasma binding and mode of excretion also
were determined. In vitro enzyme studies were conducted
to establish that the principal metabolite observed in the in
vivo studies was formed by the expected enzyme system.
Chromatographic separation was achieved using an
Econosphere C-18 reverse phase 250 mm  4:6 mm col-
umn from Alltech. Hardware consisted of a 20-mLinjec-
tion loop, a variable wavelength detector, gradient pump,
dynamic mixer, and integrator from Spectra-Physics. The
mobile phases consisted of 0.1% tri¯uoroacetic acid in
water (solvent A) and acetonitrile (solvent B). CDC and
metabolites were detected at 290 nm after gradient elution.
The gradient consisted of holding the aqueous phase at
100% for 5 min, then a linear decrease to 85% aqueous
over 5 min, and holding at 85% aqueous phase for 5 min
before returning to initial conditions. The total run time
was 25 min; the ¯ow rate was 1 mL/min. The identity of
CDC in the pertinent region of the chromatograms was
con®rmed by HPLC-MS, using a Finnigan MAT SSQ-
710C LC-MS system in the electrospray mode. Character-
istic peaks were noted at m/z 262.6 (protonated molecular
ion) and at m/z 284.6 and 300.6 (sodium and potassium
adducts, respectively). As expected, signals due to chlorine
2
. Materials and methods
2
.1. Chemicals
HPLC-grade tri¯uoroacetic acid, HPLC-grade water,
HPLC-grade acetonitrile, and ammonium sulfate were
purchased from Krackeler Scienti®c. CDU, 5-chloro-
uridine, cytidine, uridine, thymidine, thymidine phosphor-
ylase from Escherichia coli, and anhydrous dibasic sodium
phosphate were purchased from the Sigma Chemical Co.
Potassium phosphate was purchased from Mallinckrodt,
and dextrose from J.T. Baker. Mouse plasma used for
generating standard curves and for the measurement of
stability was purchased from Pel-freeze Biologicals. Cyti-
dine deaminase was isolated and partially puri®ed from
mouse kidney homogenate as previously described [7];
mouse kidney acetone powder was purchased from Sigma.
CDC and THU were provided by the National Cancer
Institute (NCI).
37
isotopes ( Cl) were also noted. The identity of the material
in the CDC peak was con®rmed by Photo Diode Array
spectroscopy using a Waters Millenium system.
2.5. Plasma assay
Plasma proteins were precipitated by mixing 50 mLof
plasma with 250 mLof 0.5 mg/mLammonium sulfate.
After vortexing, the samples were centrifuged at 8000 g
for 10 min at 48. The supernatant solutions were then ready
for HPLC analysis. Recovery of CDC from seeded samples
typically was in the 90% range. Concentration±peak area
relationships for CDC solutions in water and in mouse
plasma proved linear (correlation coef®cient 0.997) through
2
.2. Dosing
À6
À5
Normal male mice (CD F ; weighing approximately
2
the concentration range of 5 Â 10 to 5 Â 10 M. Same-
day accuracy and within- and between-day precision were
3, 4, and 8%, respectively. External standards of CDC and
CDU in water were used to determine the concentrations of
the drug in the various samples by comparison of the peak
areas. Estimates for derived pharmacokinetic parameters
1
2
5
0±25 g) were provided by the NCI. Mice were given
±100 mg/kg doses (bolus) of CDC in 0.2 mLvolumes.
THU was injected concomitantly at 100 mg/kg. Drug
solutions were prepared in water containing 5% dextrose.
Oral doses were administered with a feeding needle; i.v.

J.T. Hale et al. / Biochemical Pharmacology 64 (2002) 1493±1502
1495
were made using non-compartmental analysis where the
AUC is calculated by the linear trapezoidal rule and l , the
terminal phase rate constant, is determined from the linear
regression of time vs. log concentration (WinNonlin; Phar-
sight Corp.).
slightly higher than those reported in the literature
À5
ꢀꢀ5À7†  10 M† [8]. This discrepancy may re¯ect the
z
fact that the literature value is for mouse kidney acetone
powder, rather than for freshly isolated enzyme.
The second step of the metabolic process was observed
by spectroscopic analysis at 280 nm, using a Beckman DU-
70 system in the kinetics mode. In these experiments,
2
.6. Organ assay
À4
1
 10 M thymidine or CDU was incubated with
Selected organs (kidney and liver) were removed from
10 mLthymidine phosphorylase (1% dilution of the Sigma
stock in 50 mM phosphate buffer) and 0.5±50 mM phos-
mice post-mortem and frozen at À208 until processing.
Approximately 0.2 g of each thawed organ was homoge-
nized in 0.1 M potassium phosphate buffer (1 mL), and
centrifuged at 8000 g for 10 min at room temperature;
supernatant solutions were then analyzed as described
above for plasma.
phate buffer (pH 7). The Vmax and K of thymidine or CDU
m
with thymidine phosphorylase were calculated using Line-
weaver±Burk plots.
To study the deamination and phosphorolysis steps of
the cytidine metabolic pathway in sequence, 10 mLcyti-
dine deaminase and 10 mLthymidine phosphorylase
À4
2
.7. Stability
(diluted as above) were mixed with 1 Â 10 M CDC
and 0.5±50 mM phosphate buffer to a total volume of
1 mLand incubated in a water bath at 37 8. At various
time points, 25-mLsamples were removed and diluted 1:12
in 0.8 g/mLof ammonium sulfate. These fractions were
centrifuged at 8000 g for 10 min at room temperature prior
to HPLC analysis.
Solutions of CDC in water and in mouse plasma at a
concentration of 20 mM were incubated in a 378 water bath.
Samples (50 mL) were removed at various time points,
processed, and analyzed as described above for plasma.
2
.8. Plasma binding
Plasma binding was estimated by adding 200 mLplasma
3. Results
or water containing 20 mM CDC to Amicon centrifree
assemblies (MW cutoff of 30,000). The assemblies were
centrifuged in a Sorval centrifuge at 1000 g and 08 for
The chromatographic conditions described for analyzing
plasma resulted in the elution pattern seen in Fig. 1; in this
representative chromatogram the peak for CDC was at
9.7 min and that for the deamination product, CDU, at
10.2 min. There were no interfering peaks in processed
samples of control mouse plasma. THU does not absorb
UV light at the analytical wavelength used. Concomitant
administration of THU with CDC (Fig. 1C) resulted in an
increase in the characteristic peak for CDC and a decrease
in that for CDU relative to administration of CDC alone.
CDC was stable in plasma for at least 6 hr at 378, as judged
by recovery of 95% of the drug after incubation at this
temperature.
3
0 min. The ultra®ltrate volume was measured, and the
sample was diluted 1:5 with 0.1 M potassium phosphate
buffer (pH 7) prior to HPLC analysis.
2
.9. Urine assay
Urine was diluted 1:40 with water prior to HPLC
analysis. Solvents and the assay system were the same
as described above except that no tri¯uoroacetic acid was
used in solvent A, causing the CDC peak to elute after the
CDU peak.
Mice appeared to tolerate high doses of CDC alone and
of CDC with THU with no obvious adverse effects. The
mice were alert and exhibited normal activity following
drug administration. No gross abnormalities were observed
in organs such as liver and kidney. The 10-min plasma
concentration of CDC in mice administered either CDC
alone or CDC with 100 mg/kg of THU increased with
dose, as can be seen in Table 1. The pharmacokinetic
pro®les determined in these exploratory studies in mice
are summarized in Table 2. CDC was eliminated rapidly
from plasma after the i.v. administration of 100 mg/kg,
with an estimated elimination half-life of 10 min. Con-
comitant administration of 100 mg/kg of THU, a cytidine
deaminase inhibitor, increased the amount of CDC in the
plasma at any given time whether the drugs were given i.v.
or i.p. For both routes of administration, the averaged
2
.10. Isolated enzyme studies
All drug solutions (CDC, CDU, cytidine, uridine, and
thymidine) were made in HPLC-grade water. To determine
the effect of cytidine deaminase activity on CDC, 50 mLof
cytidine deaminase solution was added to 0.05 M phos-
phate buffer containing CDC (100 mM), and the mixture
(500 mLtotal volume) was incubated in a 37 8 water bath.
The reaction mixture was analyzed by HPLC as described
above for plasma. Inhibition of cytidine deaminase was
evaluated by adding THU (100 mM) to the enzymatic
reaction mixture. The K value for cytidine in the cytidine
m
deaminase preparations from freshly isolated mouse kid-
ney, determined by UV spectrometry and calculated using
Lineweaver±Burk plots, was 9:2 Â 10À5 M. This value is

Fig. 1. HPLC analysis of mouse plasma samples. (A) Chromatogram of a blank plasma sample; (B) chromatogram 10 min after 20 mg/kg of CDC was given i.v. to normal mice; and (C) chromatogram 10 min after
0 mg/kg of CDC and 100 mg/kg of THU were given i.v. to normal mice.
2

J.T. Hale et al. / Biochemical Pharmacology 64 (2002) 1493±1502
1497
Table 1
Plasma concentrations in mice of CDC administered with and without
THU
studies). With 100 mg/kg of THU, the average half-life
of these three doses was 28.6 min. Bioavailability after
i.p. administration was approximately 80% when CDC was
given alone. The one pilot study of CDC given with THU
i.p. showed a signi®cant increase in AUC. A pilot study in
which 100 mg/kg of CDC was given i.v. 15 min after
Dose of CDC
CDC alone
CDC ‡ l00 mg/kg
THU (mM)
(mg/kg)
(mM)
5
4.8
20.5
1
2
0
0
14.09
19.7
N.A.
58.3
1
00 mg/kg of THU given i.p. showed an even greater
increase in the AUC (64,479 mMÁmin), compared to the
i.v. studies. The half-life value found in the preincubation
study (23.2 min) was approximately the same as when
CDC and THU were given in the same dose. CDC was not
detected in mouse plasma after oral administration.
THU inhibited CDC metabolism in vivo, as seen in
Fig. 2. When a dose of 5 mg/kg of CDC was given alone,
the plasma AUC was 113.4 mMÁmin; with 5 or 100 mg/kg
of THU, the AUC values were 426.4 and 1179.7 mMÁmin,
respectively. The half-life, volume of distribution, and
clearance also increased with the increased dose of THU.
A pilot study was done in which 100 mg/kg of THU was
given alone to determine if endogenous levels of cytidine
were increased by treatment with this inhibitor. No sub-
stantial accumulations of cytidine were observed. Aliquots
of plasma samples obtained in this study were incubated
100
119.1 Æ 17
214.3 Æ 105
Plasma concentrations in mice 10 min after i.v. administration of CDC
or CDC with 100 mg/kg of THU. Data for the 5, 10, and 20 mg/kg doses
represent data from pilot studies, while the data for 100 mg/kg of CDC
alone is the average Æ SD of four studies and 100 mg/kg CDC with
1
00 mg/kg THU is the average Æ SD of three studies. The study of 10 mg/
kg of CDC with 100 mg/kg of THU is not available (N.A.). Five to nine
mice were used in each study.
pharmacokinetic data show that, with THU, the area under
the curve increased substantially, the half-life approxi-
mately doubled, volume of distribution decreased, and
the rate of plasma clearance decreased. CDC appeared
to be eliminated from plasma by ®rst order kinetics as the
half-lives of the drug given alone in doses of 5, 20, and
1
00 mg/kg were comparable (average 9.7 min; pilot
Table 2
CDC pharmacokinetic data in mice
AUC (mMÁmin)
T1/2 (min)
Volume of distribution (mL)
Plasma clearance (mL/min)
CDC
Intravenous
Intraperitoneal
a
3050 Æ 265
2450 Æ 365
10.3 Æ 1.3
10.9 Æ 1.6
38.9 Æ 7.2
44.1 Æ l.0
2.4 Æ 0.3
2.9 Æ 0.3
b
d
CDC and THU
c
Intravenous
7880 Æ 1920
21.6 Æ 1.1
29.8 Æ 6.4
0.9 Æ 0.2
Intraperitoneal
27,314
26.3
10.5
0.3
Pharmacokinetic data were obtained after the i.v. or i.p. administration of 100 mg/kg of CDC with and without 100 mg/kg of THU. Calculations were
done using non-compartmental analysis (WinNonlin). There were 5±9 mice in each study.
a
Average Æ SEM of four studies.
b
Average Æ range of two studies.
c
Average Æ SEM of three studies.
d
Pilot study.
Fig. 2. Concentration±time relationship for 5 mg/kg of CDC following i.v. administration with and without THU. Representative data from one study; each
point is indicative of one mouse.

1498
J.T. Hale et al. / Biochemical Pharmacology 64 (2002) 1493±1502
Fig. 3. Excretion of CDC and its primary metabolite, CDU, after 100 mg/kg i.v. administration in normal male mice. Representative data from one study;
each point is indicative of one mouse.
with CDC and cytidine deaminase to ascertain (via HPLC
analysis) the effective duration of the THU exposure.
There was no CDU formation until the 1-hr plasma sample
when the THU concentration was low enough to permit
Fig. 5, was mediated by the expected enzymes. The con-
version of CDC to CDU mediated by cytidine deaminase
from fresh mouse kidney progressed at a rate of approxi-
mately 8 mM/min. This reaction was inhibited by THU,
resulting in a deamination rate of approximately 3.5 mM/
min. Preincubation with THU was not tested, as it is not
reported to increase the inhibition of mouse kidney cyti-
dine deaminase [9]. These studies were also performed
with the enzyme preparation from mouse kidney acetone
0
6
.6 mM CDU to be formed per minute compared with
.8 mM/min in control plasma, as analyzed by HPLC.
CDC was detectable in urine samples collected after
administration of CDC to mice; the most prominent species
detected in urine samples was the deaminated metabolite
CDU (Fig. 3). Peak urinary concentrations of parent drug
and metabolite were observed around 20 min after admin-
istration, and both could be detected in urine for at least
4
hr in three studies in which 100 mg/kg of CDC was
administered. When 5 mg/kg of THU was administered
concurrently with 5 mg/kg of CDC, the expected preserva-
tion of CDC was noted, as exempli®ed by observed con-
centrations of 870 mM CDC in urine at 120 min as
compared with no detectable drug at 90 min after injection
of 5 mg/kg of CDC alone (Fig. 4).
CDU was found in the kidneys at a concentration of
56 mg/g 5 min after 100 mg/kg of CDC was given i.v., and
1
the peak due to CDU gradually decreased until it was
indistinguishable from small peaks due to endogenous
substances. CDC could not be detected in processed kidney
homogenates because of interference from endogenous
substances. We were unable to detect CDC or CDU in
liver samples, because of the presence of endogenous
peaks in processed liver in the chromatographic areas of
interest.
Plasma protein binding was determined by comparing
the concentration of free drug in plasma after ultra®ltration
to an aqueous standard containing the drug at the same
concentration as the plasma. CDC binding to plasma
proteins was very low in triplicate determinations, with
virtually all the drug (>99%) in the free form.
Fig. 4. Comparison of the preservation of CDC, administered with and
without THU, in the urine. (A) Urine data after the i.v. administration of
5
mg/kg of CDC. (B) Urine data after the i.v. administration of 5 mg/kg of
In vitro studies were done for comparison with the in
vivo studies to con®rm that the metabolic pathway, seen in
CDC and 5 mg/kg of THU. Urine samples were ``clean-catch.'' Each bar
represents an individual mouse.

J.T. Hale et al. / Biochemical Pharmacology 64 (2002) 1493±1502
1499
Fig. 5. First two steps of the CDC metabolism pathway.
powder. Similar results were seen with THU inhibiting the
production of CDU, but the enzyme functioned at a slower
rate than it did in the fresh preparation. Thymidine phos-
phorylase converts CDU to 5-CU by phosphorolytic
removal of the sugar substituent. The Km and Vmax for
thymidine and CDU are reported in Table 3 at three
different concentrations of phosphate buffer, pH 7. The
complete progression of the conversion of CDC to 5-CU
over time can be seen in Fig. 6, while the rates for CDC
disappearance and CDU or 5-CU appearances in each
concentration of phosphate buffer are summarized in
Table 4.
Table 3
Table 4
Rates for thymidine phosphorylase
Rate of CDC disappearance
Sample
K
m
(mM)
Vmax (mM/min)
PO
4
(mM)
CDC (mM/mm)
CDU (mM/min)
5-CU (mM/min)
Thymidine in 50 mM PO
4
0.317 Æ 0.01
0.255 Æ 0.01
0.728 Æ 0.4
0.110 Æ 0.01
0.140 Æ 0.03
0.079 Æ 0.01
166 Æ 15
127 Æ 4
50
5
À3.90
À3.95
À5.16
0.0016
0.37
0.61
3.17
2.07
1.44
Thymidine in 5 mM PO
Thymidine in 0.5 mM PO
CDU in 50 mM PO
CDU in 5 mM PO
CDU in 0.5 mM PO
4
4
131 Æ 56
143 Æ 19
225 Æ 79
68.5 Æ 0.09
0.5
4
À4
CDC (1 Â 10 M) was incubated with cytidine deaminase and
thymidine phosphorylase in 0.5±50 mM phosphate buffer, and the rates
of drug disappearance and metabolite formation were measured. Data are
from one experiment at each concentration.
4
4
K
m
and Vmax values and for thymidine and CDU with thymidine
phosphorylase when incubated at 378 5±50 mM phosphate buffer. Data are
average Æ SEM of three experiments.
Fig. 6. Kinetics of CDC degradation and metabolite formation after 1 Â 10À4 M CDC was incubated at 378 with cytidine deaminase and thymidine
phosphorylase in 50 mM phosphate buffer. Representative data from one study.

1500
J.T. Hale et al. / Biochemical Pharmacology 64 (2002) 1493±1502
4
. Discussion
related to the onset of inhibition of CDC metabolism and/
or altered distribution. It is interesting to note, in this
regard, that Kreis and his colleagues [16] have reported
that THU affects not only the deamination of ara-C but also
its distribution. The half-life does not change in this case
due to higher initial plasma concentrations of CDC related
to the altered distribution of the drug, while the elimination
process remains the same. The net effect is that THU
signi®cantly increases the effective plasma concentration
of CDC. THU has a comparable effect in humans after i.v.
administration to patients receiving ara-C, re¯ected in
substantial increases in the plasma concentration of ara-
C compared to those observed with ara-C alone [16].
CDC was not detected in plasma after oral administration.
The drug is stable in dilute hydrochloric acid, pH 1, at 378
for 5 hr, suggesting that failure to detect the drug in plasma
after oral administration is due to ®rst pass metabolism and
not to degradation in the stomach. Analysis of urine 60 min
after oral administration of CDC showed the presence of
CDU, in the 40 mM range, but no detectable CDC, support-
ing the idea that the drug is absorbed from the gastrointest-
inal tract, but is metabolized extensively in the ®rst pass
through the liver. This conclusion is consistent with obser-
vations that ara-C also has poor oral bioavailability due to
rapid hepatic deamination [17]. Another study involving
ara-C and ara-U looked at the activity of cytidine deaminase
in liver and kidney. This study found that, although cytidine
deaminase activity is higher in mouse kidney (639.5 IU/g of
wet tissue), it is also signi®cant in the liver (35.2 IU/g of wet
tissue) [18]. Because THU is reported to have only 10% oral
bioavailability, we did not give CDC orally with THU,
although it may have been helpful, as there are bacterial
gut ¯ora having signi®cant cytidine deaminase activity that
may also decrease absorption [16]. It is possible that if the
drug is given with zebularine, an orally active cytidine
deaminase inhibitor [19], CDC might be detectable in
plasma after oral administration.
Comparatively low doses of THU have substantial
effects on the metabolism of CDC as evidenced by the
similar effects observed after administration of THU at
doses of 5 and 100 mg/kg. Administration of THU alone
did not show a detectable increase in endogenous cytidine
levels in plasma up to 1 hr after administration. Plasma
from mice treated with THU inhibited CDC deamination
by cytidine deaminase in vitro, and THU remained in the
mouse plasma at concentrations suf®cient to completely
inhibit production of CDU for at least 1 hr. Even in the 1-hr
plasma sample, there was still enough THU present to only
allow a small amount of CDC to be metabolized. The
elimination half-life of THU was reported to be 40 min in
mice after 50 mg/kg was injected i.p. [20]. A previous
study showed that 12.5±100 mg/kg of THU could com-
pletely inhibit cytidine deaminase in rhesus monkeys for
up to 2 hr [16].
CDC and its primary metabolite, CDU, have similar UV
spectroscopic properties, which is advantageous in that
each is readily detected at the analytical wavelength of
2
90 nm. CDC and CDU are similar chemically, but resolu-
tion of HPLC peaks was generally good even though
retention times were quite close. Although we evaluated
a large number of compounds, pyrimidine and purine
nucleosides as well as other compounds, we were unable
to identify an appropriate internal standard and, therefore,
used external standards of CDC and CDU for estimation of
drug concentrations in the processed samples. CDC did not
bind to a signi®cant extent to plasma proteins. The lack of
signi®cant binding of CDC to plasma proteins parallels
0
0
observations with the nucleoside gemcitabine (2 ,2 -
di¯uorodeoxycytidine), which also has negligible plasma
binding to mouse, rat, and dog proteins [10].
We chose to use 100 mg/kg of CDC and THU for the
majority of our in vivo studies. Pharmacodynamic studies
in the literature have reported using 25 mg/kg of THU
to inhibit CDC catabolism, but these studies also used
higher doses of CDC (167±444 mg/kg) along with other
metabolic modulators [2,6,11]. One study used an infusion
of CDC at 628.08 mg/kg with and without THU at
1
256.16 mg/kg over the course of 3 days [5]. It is hypothe-
sized that low doses of THU may be more effective in
enhancing CDC ef®cacy since higher doses may also
inhibit the tumor enzymes [12]. CDC has good bioavail-
ability when given to mice i.p., as seen by the similar AUC
and half-life values for i.v. and i.p. administration of the
drug given alone. The estimated elimination half-life of
9
1
.7 min is comparable to the alpha phase half-life of
2 min reported for the analogous compound, ara-C, in
humans [13]. Gemcitabine also has a similar plasma half-
life of 17 min in humans and mice [10,14]. THU appears
to have a greater effect on inhibiting metabolic processes
with i.p. administration than with i.v. administration. The
half-life of the combination was 30% longer and clearance
was reduced in the i.p. dosing when compared with the i.v.
administration. The volume of distribution remained
approximately the same between the two modes of com-
bined drug administration. Although preincubation of
THU with cytidine deaminase does not enhance its inhi-
bitory effects in vitro [9], our 15-min ``preincubation'' in
vivo, with THU administered i.p. prior to i.v. administra-
tion of CDC, showed signi®cant effects on the AUC, while
the half-life remained the same compared to concomitant
administration. A study of the interaction of ara-C and
THU in humans involved administration of a portion of the
THU dose prior to administration of ara-C, because the
onset of inhibition is slow, reversible, and time-dependent
even though the compound binds tightly to hepatic cytidine
deaminase [15]. It is possible that the signi®cant differ-
ences seen in the AUC when THU was given i.p. as a
preincubation step, relative to the i.v. combination, are
It is not surprising that the majority of the drug excreted
through the kidneys was the deaminated metabolite. The

J.T. Hale et al. / Biochemical Pharmacology 64 (2002) 1493±1502
1501
kidneys contain large amounts of cytidine deaminase,
which is why kidneys were used to isolate the enzyme
for in vitro work. As THU is also excreted by the kidneys
phosphate buffer, our data appear to be comparable to
published results. CDU is reported to have a Km of
0.186 mM when thymidine phosphorylase is in 0.1 M
potassium phosphate at pH 6.0 [24], which is similar to
our results. Three different concentrations of phosphate
buffer were used because the related enzyme, uridine
phosphorylase, shows complex product inhibition in the
presence of phosphate buffer at concentrations greater than
3 mM [25]. However, product inhibition was not observed
with thymidine phosphorylase using either thymidine or
CDU in any of the three concentrations of phosphate
buffer.
(roughly 90±100% within 24 hr) [16,20], it can inhibit
enzyme found in the kidney as well as cytidine deaminase
in other tissues, giving rise to the increased concentration
of parent drug in the urine observed when CDC and THU
are given together. Gemcitabine is excreted via the kidney
mainly in the uracil metabolite form in the mouse and dog,
but the parent drug is the main compound found in rat urine
[10]. Similarly, ara-C is excreted mainly in the ara-U form
in mice, monkeys, and humans [20].
Tissue distribution studies of ara-C and its water-soluble
0
Both cytidine deaminase and thymidine phosphorylase
were incubated with CDC so that both steps of the meta-
bolic progression could be seen at once. Although some
variation was observed in the rates of enzymatic activity
associated with the different concentrations of phosphate,
the metabolic process seen in the mouse studies was
effectively modeled in vitro.
In conclusion, CDC has potential as a radiosensitizing
drug. It has a short plasma elimination half-life and is
excreted by the kidney, mainly in metabolite form. THU
substantially increased plasma levels of CDC in the mice,
supporting proposals that the combination should be eval-
uated in clinical trials of CDC as a radiosensitizing agent.
relative, cyclocytidine (2,2 -O-cyclocytidine), gave results
comparable to observations in our pilot tissue study. After a
dose of 34 mg/kg, cyclocytidine was found at the highest
concentrations in the mouse kidney 10 min after injection,
with a concentration of 150 mg/g [21]. The concentration of
ara-C found in mouse kidney depends upon the breed.
Swiss mice have high kidney pyrimidine nucleoside dea-
minase activity, and the parent compound is not detected.
BDF1 and AKD2F mice are presumed to have lower
amounts of deaminase, as ara-C could be detected at
low levels. All three types of mice have higher levels of
the metabolite ara-U (compared with ara-C) in the kidneys
than in other tissues; these levels were lower when ara-C
was given with THU [20].
It is possible that some of the parent and metabolite
compounds in the tissues converted to their nucleotides. A
study of the pharmacokinetics of ara-C demonstrated that
small amounts of nucleotides were located in AKD2F
mouse livers and kidneys after 2 hr, and there were appre-
ciable concentrations of nucleotides in the livers of Swiss
mice after 30 min [20]. In each case, however, the con-
centration of ara-C or ara-U, depending on the presence of
THU, was much greater than the concentration of nucleo-
tides [20].
Acknowledgments
This work was supported by NCI Contract N01-CM-
27713 and by CA-22435.
References
[1] Mitchell JB, Russo A, Cook JA, Straus KL, Glatstein E. Radiobiology
and clinical application of halogenated pyrimidine radiosensitizers. Int
J Radiat Biol 1989;56:827±36.
HPLC methodology was required for the in vitro enzyme
work with CDC and its metabolite, CDU, because both
compounds have similar ultraviolet absorption properties,
thus obviating the use of conventional spectroscopic ana-
lysis. In the presence of THU, the rate of enzymatic
deamination of CDC was decreased by nearly 50% for
both the fresh kidney and the acetone powder enzyme
preparations. These observations parallel those made in
vivo where equal amounts of CDC and THU nearly
doubled the half-life of CDC.
Thymidine phosphorylase is reported to have its max-
imal activity at a pH of 6.0 in vitro. Thymidine phosphor-
ylase from Lactobacillus casei in 300 mM sodium
phosphate buffer, pH 6.0, gave a Km value of 1.32 mM
for thymidine [22]. Thymidine phosphorylase from horse
liver with thymidine is reported to have a Km value of
[
2] Santos O, Perez LM, Briggle TV, Boothman DA, Greer SB. Radiation,
0
pool size and incorporation studies in mice with 5-chloro-2 -deox-
ycytidine. Int J Radiat Oncol Biol Phys 1990;19:357±65.
[3] Lawrence TS, Davis MA. Selective radiosensitization and cytotoxicity
of human melanoma cells using halogenated deoxycytidines and
tetrahydrouridine. Int J Radiat Oncol Biol Phys 1989;16:1243±6.
0
[
4] Perez LM, Greer S. Sensitization to X-ray by 5-chloro-2 -deoxycy-
tidine co-administered with tetrahydrouridine in several mammalian
0
cell lines and studies of 2 -chloro derivatives. Int J Radiat Oncol Biol
Phys 1986;12:1523±7.
[
[
5] Russell KJ, Rice GC, Brown JM. In vitro and in vivo radiation
0
sensitization by the halogenated pyrimidine 5-chloro-2 -deoxycyti-
dine. Cancer Res 1986;46:2883±7.
6] Greer S, Schwade J, Marion HS. Five-chlorodeoxycytidine and bio-
modulators of its metabolism result in fifty to eighty percent cures of
advanced EMT-6 tumors when used with fractionated radiation. Int J
Radiat Oncol Biol Phys 1995;32:1059±69.
[
[
7] Tomchick R, Saslaw LD, Waravdekar VS. Mouse kidney cytidine
deaminase. J Biol Chem 1968;243:2534±7.
0
[
.11 mM in 0.1 M potassium phosphate buffer at pH 7.0
23]. Despite the fact that we used a pH of 7.0, which is
closer to the body's pH, and lower concentrations of
8] McCormack JJ, Marquez VE, Liu PS, Vistica DT, Driscoll JS.
Inhibition of cytidine deaminase by 2-oxopyrimidine riboside and
related compounds. Biochem Pharmacol 1980;29:830±2.

1502
J.T. Hale et al. / Biochemical Pharmacology 64 (2002) 1493±1502
[
9] Voytek P, Beisler JA, Abbasi MM, Wolpert-DeFillippes MK. Com-
[17] Rentsch KM, Schwendener RA, Schott H, H aÈ nseler E. Pharmacoki-
4
parative studies of the cytostatic action and metabolism of 5-azacy-
tidine and 5,6-dihydro-5-azacytidine. Cancer Res 1977;37:1956±61.
netics of N -octadecyl-1-b-D-arabinofuranosylcytosine in plasma and
whole blood after intravenous and oral administration to mice. J Pharm
[
[
10] Shipley LA, Brown TJ, Cornpropst JD, Hamilton M, Daniels WD,
Culp HW. Metabolism and disposition of gemcitabine, an oncolytic
deoxycytidine analog, in mice, rats, and dogs. Drug Metab Dispos
Pharmacol 1997;49:1076±81.
[18] Chandrasekaran B, Capizzi RL, Kute T, Morgan T, Dimling J.
Modulation of the metabolism and pharmacokinetics of 1-b-D-arabi-
nofuranosylcytosine by 1-b-D-arabinofuranosyluracil in leukemic
mice. Cancer Res 1989;49:3259±66.
1992;20:849±55.
11] Greer S, Santos O, Gottlieb C, Schwade J, Marion HS. 5-Chlorodeox-
ycytidine a radiosensitizer effective against RIF-1 and Lewis lung
carcinoma, is also effective against a DMBA-induced mammary
adenocarcinoma and the EMT-6 tumor in BALB/c mice. Int J Radiat
Oncol Biol Phys 1992;22:505±10.
[19] Driscoll JS, Marquez VE, Plowman J, Liu PS, Kelley JA, Barchi Jr JJ.
Antitumor properties of 2(1H)-pyrimidinone riboside (zebularine) and
its fluorinated analogues. J Med Chem 1991;34:3280±4.
[20] El Dareer SM, Mulligan Jr LT, White V, Tillery K, Mellett LB, Hill
3
[
[
12] Mekras JA, Boothman DA, Greer SB. Use of 5-trifluoromethyldeox-
ycytidine and tetrahydrouridine to circumvent catabolism and exploit
high levels of cytidine deaminase in tumors to achieve DNA- and
target-directed therapies. Cancer Res 1985;45:5270±80.
DL. Distribution of [ H]cytosine arabinoside and its products in mice,
dogs, and monkeys and effect of tetrahydrouridine. Cancer Treat Rep
1977;61:395±407.
[21] Hoshi A, Iigo M, Kuretani K, Kanai T, Ichino M. Distribution of
cyclocytidine in tissues. Chem Pharm Bull Tokyo 1975;23:725±8.
[22] Avraham Y, Grossowicz N, Yashphe J. Purification and characteriza-
tion of uridine and thymidine phosphorylase from Lactobacillus casei.
Biochim Biophys Acta 1990;1040:287±93.
13] Breithaupt H, Pralle H, Eckhardt T, von Hattingberg M, Schick J,
L oÈ ffler H. Clinical results and pharmacokinetics of high-dose cytosine
arabinoside (HD ARA-C). Cancer 1982;50:1248±57.
[
14] Storniolo AM, Allerheiligen SRV, Pearce HL. Preclinical pharmaco-
logic, and phase I studies of gemcitabine. Semin Oncol 1997;24:S72±7.
15] Kreis W, Chan K, Budman DR, Schulman P, Allen S, Weiselberg L,
Lichtman S, Henderson V, Freeman J, Deere M, Andreeff M, Vinci-
guerra V. Effect of tetrahydrouridine on the clinical pharmacology of
[23] Wataya Y, Santi DW. Continuous spectrophotometric assay of thy-
0
[
midine phosphorylase using 5-nitro-2 -deoxyuridine as substrate. Anal
Biochem 1981;112:96±8.
[24] Nakayama C, Wataya Y, Meyer Jr RB, Santi DV, Saneyoshi M, Ueda
0
1-b-D-arabinofuranosylcytosine when both drugs are coinfused over
T. Thymidine phosphorylase. Substrate specificity for 5-substituted 2 -
three hours. Cancer Res 1988;48:1337±42.
deoxyuridines. J Med Chem 1980;23:962±4.
[25] Krenitsky TA. Uridine phosphorylase from Escherichia coli:
kinetic properties and mechanism. Biochim Biophys Acta 1976;
429:352±8.
[
16] Kreis W, Woodcock TM, Gordon CS, Krakoff IH. Tetrahydrouridine:
physiologic disposition and effect upon deamination of cytosine
arabinoside in man. Cancer Treat Rep 1977;61:1347±53.