A comparative study of the kinetics of cytarabine hydrolytic deamination in aqueous solutions

Full text of DOI:10.1023/A:1005299914278

Pharmaceutical Chemistry Journal
Vol. 34, No. 8, 2000
A COMPARATIVE STUDY OF THE KINETICS OF CYTARABINE
HYDROLYTIC DEAMINATION IN AQUEOUS SOLUTIONS
I. V. Smirnov,1 G. G. Zav’yalov,2 and S. V. Eremin3
Translated from Khimiko-Farmatsevticheskii Zhurnal, Vol. 34, No. 8, pp. 53 – 56, August, 2000.
Original article submitted January 11, 2000.
In application to nucleosides, IPHPLC can be used in
several different modes [36], including chromatography un-
der socratic conditions (most frequently used for the analysis
of arabinosyl nucleosides) [20 – 22], gradient-elution chro-
matography, and ion-pair chromatography [29, 34, 35].
As is known [3], the maximum extinction coefficient of
cytosine nucleosides is observed at low pH. Therefore, using
these conditions we may reach a maximum sensitivity of
spectrophotometric detection. At the same time, all the mole-
cules of cytosine nucleosides studied under these conditions
are ionized and, hence, exhibit weak adsorption on the im-
mobile (nonpolar) phase that results in lower resolution.
Thus, optimum possibilities for controlling the interaction of
partitioned components with the immobile phase are pro-
vided by the method of ion-pair chromatography.
Cytarabine (1-b-D-arabinofuranosylcytosine, I) is a drug
belonging to the group of nucleoside antimetabolites [1]. At
present, the drug is produced abroad both in dry (lyophilized)
and liquid forms. The production of liquid forms is less la-
bor-consuming and more economically effective. At the
same time, it was reported [2 – 8] that cytarabine is subject to
hydrolytic deamination in aqueous media with the formation
of a therapeutically inactive 1-b-D-arabinosyluracil (II) and
ammonia (Fig. 1).
Our task was to study the stability of cytarabine in aque-
ous solutions during a preset storage time. For this purpose,
it was necessary to select a method that would be (i) capable
of studying the kinetics and rate of the cytarabine hydrolytic
deamination reaction and (ii) sufficiently simple and opti-
mum from the standpoint of sensitivity and reproducibility.
Various methods were developed for determination of
the content of I in medicinal preparations and for the study of
hydrolytic deamination of this compound. These include
TLC [9], RIA [10 – 13], mass spectrometry [14], gas chro-
matography in various modes, including those combined
with other techniques [15, 16], UV spectrophotometry [17],
extraction methods [18], and liquid chromatography at low
[19] and high pressures [20 – 35].
The method of direct UV spectrophotometry [17] is
based on the difference between the absorption spectra of I
and its decomposition product II (having absorption peaks at
280 and 260 nm, respectively, in 0.1 N HCl). This technique
is most simple and was widely used in the early investiga-
tions of the hydrolytic deamination reaction [2 – 4, 6 – 8].
In recent years, compounds I and II are more frequently
determined by HPLC techniques offering some advantages
over the other methods. Compounds belonging to arabinosyl
nucleosides were predominantly studied by inversed-phase
HPLC (IPHPLC) [20 – 22], although ion-exchange HPLC
[23] and combined techniques can be used as well [27].
MATERIALS AND METHODS
Cytarabine. 99.4% purity; Pharmacopoeial Clause FS
42-2982 – 93; produced by the experimental plant at the In-
stitute of Biological and Organic Chemistry, Academy of
Sciences of the Republic of Belarus, Belarus.
UV spectrophotometric determination of compounds
I and II [3, 4]. Using the optical density of a sample deter-
mined at 260 and 280 nm, the concentrations of compounds I
and II are calculated by the following formulas:
CI = (0.927D280 – 0.389D260) ´ 10 – 4
CII = (1.21D260 – 0.546D280) ´ 10 – 4
,
,
where CI and CII are the concentrations of I and II in solution
(mole/liter); D260 and D280 are the optical densities at 260
and 280 nm, respectively.
The initial solutions were diluted 2000 times with 0.1 M
hydrochloric acid solution. The samples were measured to
determine the optical densities at 260 and 280 nm, with ref-
erence to pure 0.1 M HCl solution. The measurements were
performed in 10-mm-thick quartz cells on an SF-46
spectrophotometer (LOMO, Russia),
1
Hematological Research Center, Russian Academy of Medical Sciences,
Moscow, Russia.
2
Agat-Med, Joint-Stock Company, Moscow.
Lomonosov State Academy of Fine Chemical Technology, Moscow, Russia.
3
451
0091-150X/00/3408-0451$25.00 © 2001 Plenum Publishing Corporation

452
I. V. Smirnov et al.
O
N
NH2
HN
N
H2O
+ NH3
O
O
N
CH2OH
O
CH2OH
O
1
HO
HO
OH
OH
II
I
Fig. 1. Hydrolytic deamination of cytarabine (arabinosylcytosine).
2
IPHPLC determination of compounds I and II. The
HPLC measurements were performed in a system equipped
with a Model 64 high-pressure pump, Model 50 temperature
programmer, variable wavelength monitor, and D-1000 in-
jector with a 10 ml loop (all Knauer, Germany). The
chromatograms were recorded with an LKB TZ-4620 line re-
corder (Czech Republic) and processed using a Chromatopac
C-R3A integrator (Shimadzu, Japan). Sorbents and columns:
Sepharon SGX C18 (5 mm), 150 ´ 4 mm; Econosil C18 5U
(5 mm), 250 ´ 4.6 mm; Hypersil ODS (6 mm), 250 ´ 3 mm.
The HPLC analyses were performed at a flow rate of
1 ml/min. Mobile phase: trifluoroacetic acid (TFA) – water
(1 : 99), pH 1.31. Prior to analysis, the system was ultrasoni-
cally outgassed at low vacuum in a Bandelin Sonorex TK52
setup (Knauer, Germany). The useful signals were detected
at 260 nm.
3
Fig. 2. Chromatograms of a mixture of compounds I and II on
Sepharon SGX C18 (5 mm) sorbent eluted in a TFA – water system
with various component ratios: (1 ) 0.25 : 100; (2 ) 0.5 : 100; (3 )
1 : 100.
cessing, we obtained the following coefficients: a1 = 0.0248
(r2 = 0.998) for I and a1 = 0.0345 (r2 = 0.999) for II.
The sample concentrations were calculated by the for-
mula
I ´20
C =
,
a1 ´1000
Before sample application, the HPLC system was washed
with the mobile phase for 20 min. The sample volume was
10 ml. The sample solutions were prepared by 20-fold dilution
of the initial liquid preparation with distilled water.
where C is the concentration of cytarabine (arabinosyluracil)
in the sample (g/liter), I is the integral HPLC peak intensity
(mV ´ sec), a1 is the corresponding calibration coefficient,
and 20 is the sample dilution factor.
The concentrations of I and II were determined using a
column filled with Sepharon SGX C18. The corresponding
calibration plots were constructed [in peak area (mV ´ sec)
versus concentration coordinates] for aqueous solutions of
pure compounds I (concentration interval, 0 – 2 mg/ml) and
II (0 – 0.16 mg/ml). The results of the calibration measure-
ments were approximated using the least squares method by
linear equations of the type
Determination of ammonia. The content of ammonia
was determined by a colorimetric method based on a reaction
with the Nessler reagent. To 5 ml of a sample (diluted
50 – 200 times with distilled water) were added 0.25 ml of
the Nessler reagent. The mixture was thoroughly stirred, al-
lowed to stand for 10 min, and measured in the
spectrophotometer at 400 nm against the blank solution. The
corresponding calibration plot was constructed for ammo-
nium sulfate (reagent grade) solutions.
I = a1 ´ C,
where I is the integral peak intensity (mV ´ sec) and C is the
component concentration in solution (mg/liter). Upon data pro-
The concentration of ammonia in the sample (mmole/li-
ter) was calculated by the formula
C = a ´ 0.298 ´ D,
where D is the optical density at 400 nm, a is the sample di-
lution factor, and 0.298 is the coefficient determined for the
calibration plot, which was linear in the 0.040 – 0.200
mmole/liter concentration interval.
Cytarabine decomposition kinetics. The kinetics of
cytarabine decomposition (deamination) was studied in drug
solution with a concentration of 20 g/liter and a pH adjusted
to 7.0 ± 0.2 with 0.01 M NaOH solution. The sample solu-
TABLE 1. The Parameters of Cytarabine Hydrolytic Deamination
Kinetics Determined by Various Methods
Method
k ´ 10 – 3, h – 1
CV (%)
r2
Loss of I (HPLC)
0.501
0.444
0.138
0.645
5.243
5.698
6.419
2.916
0.963
0.958
0.910
0.988
Loss of I (spectroscopy)
Formation of II (HPLC)
Loss of I (ammonia)

A Comparative Study of the Kinetics of Cytarabine Hydrolytic Deamination
453
Loss of I (by ammonia)
8.2
35
Loss of I (by HPLC)
30
25
Loss of I (by spectrophotometry)
Formation of II (by HPLC)
Formation of II (by spectrophotometry)
“
20
15
10
“
“
“
6.833
“
“
“
“
5
0
100
200
300 400
Incubation time, h
500
600
Fig. 3. A typical chromatogram of a cytarabine solution incubated
at 70°C.
Fig. 4. Variation of the composition of an aqueous solution of
cytarabine (I) with the time of incubation at 70°C by data of various
methods.
tion was poured by 3-ml portions into 5-ml ampules. The
ampules were sealed and incubated in a BT-120 thermostat
(Czech Republic) at 70 ± 0.3°C. After incubation for a cer-
tain time, four samples were taken and analyzed for ammo-
nia content.
lished data [2 – 4, 6, 8], we assumed that the hydrolytic
deamination of I is a first-order reaction. The rate constants
were calculated upon the least squares approximation of data
by an equation of the type
ln C = –k ´ t,
RESULTS AND DISCUSSION
For comparative analysis and verification of data ob-
tained by the direct spectrophotometric technique, we used
the results of ion-pair inversed-phase HPLC measurements.
where C is the current concentration of a given substance
(expressed as a percentage of the initial value), t is the sam-
ple incubation time (h), and k is the reaction rate constant.
The values of k calculated for all the methods employed are
listed in Table 1 together with data on the relative standard
errors CV (%) and correlation coefficients r2 for the
linearized experimental relationships.
The HPLC measurements were performed using
a
TFA – water system that is rather widely employed for
ion-pair IPHPLC chemical analysis [36]. The samples were
analyzed in a column filled with Sepharon SGX C18 sorbent.
In preliminary experiments on HPLC of a I + II mixture sep-
arated using three eluents with different compositions, we se-
lected a 1 : 99 TFA – water ratio (Figs. 2 and 3). Attempts at
increasing the resolution by using different columns (with
the same eluent) were unsuccessful.
As seen from the presented data, determination of the
loss of compound I by the spectrophotometric method and
HPLC leads to analogous results. The average variation coef-
ficients determined for a large degree of decomposition of
compound I amount to 20 and 12%, respectively. However,
the error of determination exhibits a manifold increase at
small decomposition levels (below 12 and 7% for the spec-
trophotometric method and HPLC, respectively) because the
small value (the yield of deaminated I) is determined as a dif-
ference of two close values (initial and final cytarabine con-
centrations).
According to the data reported in [2 – 8], the hydrolytic
deamination of cytarabine in aqueous solutions leads to the
formation of ammonia among other products (Fig. 1). We
have used this fact for the first time for monitoring the reac-
tion kinetics by measuring the ammonia concentration with
the aid of the Nessler reagent.
Figure 4 shows variation of the composition of an aque-
ous solution of cytarabine (I) with increasing incubation
time. Figure 5 presents data on the variation coefficients cal-
culated for cytarabine hydrolytic deamination kinetics stud-
ied by various methods and plotted versus the amounts of de-
termined substances.
In the initial stage of cytarabine hydrolytic deamination,
the reaction kinetics should be studied by determining the
content of the reaction products. The HPLC technique is ca-
pable of determining the increments of II at sufficiently high
precision (average variation coefficient is 7%) in the entire
concentration range studied, beginning with virtually zero
values. However, the rate of the formation of compound II is
almost four times smaller than the rate of the loss of com-
The experimental data were used to calculate the rate
constants for the reactions studied. According to the pub-

454
I. V. Smirnov et al.
350
2. R. E. Notari, J. Pharm. Sci., 56(7), 804 – 809 (1967).
3. R. E. Notari, L. C. Marylin, and A. Cardoni, Tetrahedron Lett,
40, 3499 – 3502 (1969).
4. R. E. Notari, L. C. Marylin, and A. Cardoni, J. Pharm. Sci.,
59(1), 28 – 32 (1970).
5. H. Lonnberg, P. Suokas, R. Kappi, et al., Acta Chem. Scand.,
B40, 798 – 805 (1986).
6. R. Shapiro and R. S. Klein, Biochemistry, 5, 2358 (1966).
7. E. R. Garret and J. Tsau, J. Pharm. Sci., 61, 1052 (1972).
8. R. E. Notari, M. L. Chin, and R. J. Wittebor, J. Pharm. Sci., 57,
1189 (1972).
Loss of I (by ammonia)
300
250
200
150
100
50
Loss of I (by HPLC)
Loss of I (by spectrophotometry)
Formation of II (by HPLC)
Formation of II (by spectrophotometry
“
9. B. Paw and G. Misztal, Acta Pol. Pharm., 52(6), 455 – 457
(1995).
10. T. Sato, M. Morozumi, K. Kodama, et al., Cancer Treat. Rep.,
68(11), 1357 – 1366 (1984).
“
“
“““
“
11. T. Okabayashi and J. G. Moffatt, Methods Enzymol., 84,
470 – 483 (1982).
0
5
10
15
20
25
30
35
12. T. Okabayashi, S. Mihara, D. B. Repke, and J. G. Moffatt, Can-
cer Res., 37(9), 3132 – 3135 (1977).
13. T. Okabayashi, S. Mihara, and J. G. Moffatt, Cancer Res.,
37(2), 625 – 628 (1977).
14. Y. Makino, Y. Matsubara, K. M. Watanabe, and M. Hirobe,
Yakugaku Zasshi, 102(1), 49 – 55 (1982).
% of initial I
Fig. 5. Plots of the variation coefficient (n = 4) for cytarabine
hydrolytic deamination kinetics versus the amounts of substances
determined by various methods.
15. J. Boutagy and D. J. Harvey, J. Chromatogr., 156(1), 153 – 166
(1978).
16. W. Sadee, C. Finn, and J. Staroscik, Adv. Mass Spectrom.
Biochem. Med., 1, 509 – 515 (1976).
17. R. L. Furner, R. W. Gaston, J. D. Strobel, et al., J. Nat. Cancer
Inst., 52(5), 1521 – 1528 (1974).
18. W. Kreis, C. Gordon, C. Gizoni, et al., Cancer Treat. Rep.,
61(4), 643 – 646 (1977).
19. B. C. Pal, J. Chromatogr., 148, 545 – 548 (1978).
20. A. A. Miller, J. A. Benvenuto, and T. L. Loo, J. Chromatogr.,
228, 165 (1982).
21. L. D. Kisinger and N. L. Stemm, J. Chromatogr., 353,
309 – 318 (1986).
22. E. B. Rochard, D. M. C. Barthes, and P. Y. Courtois,
Am. J. Hosp. Pharm., 49, 619 – 623 (1992).
23. M. G. Pallavicini and J. A. Mazrimas, J. Chromatogr., 183(4),
449 – 458 (1980).
24. H. Breithaupt and J. Schick, J. Chromatogr., 225(1), 99 – 106
(1981).
25. P. Lenssen, A. Drenthe-Schonk, H. Wessels, et al., J.
Chromatogr., 223(2), 371 – 378 (1981).
26. M. Tuncel, R. E. Notari, and L. Malspeis, J. Liq. Chromatogr.,
4(5), 887 – 896 (1981).
27. J. A. Sinkule and W. E. Evans, J. Chromatogr., 274, 87 – 93
(1983).
28. G. G. Liversidge, T. Nishihata, T. Higuchi, et al., J.
Chromatogr., 276(2), 375 – 383 (1983).
29. R. L. Schilsky and F. S. Ordway, J. Chromatogr., 337(1),
63 – 71 (1985).
30. A. Riccardi, T. Servidei, A. Lasorella, et al., J. Chromatogr.,
497, 302 – 307 (1989).
31. J. R. Wermeling, J. M. Pruemer, F. M. Hassan, et al., Clin.
Chem., 35(6), 1011 – 1015 (1989).
pound I. According to the published data [2, 4, 5], we may
suggest that compound II exhibits subsequent conversion
with the formation of optically inactive products. This is con-
firmed by the fact that the increment of II exhibits a tendency
to slow down, whereas the cytarabine deamination proceeds
linearly (or even slightly increases). For this reason, we may
conclude that variation of the content of compound II does
not adequately reflect the kinetics of the hydrolytic
deamination of cytarabine.
The spectrophotometric determination of II in a mixture
with I by the method described above also gives results that
are rather inconsistent with the HPLC data.
However, the increment of ammonia – another product
of hydrolytic deamination – is well correlated with the de-
gree of cytarabine deamination. Determination of the
hydrolytic deamination kinetics by measuring the ammonia
increment is sufficiently precise (the variation coefficient not
exceeding 7% in the entire range of concentrations), while
being faster and simpler as compared to the HPLC proce-
dure. However, it must be noted that study of cytarabine
hydrolytic deamination kinetics by the “ammonia” technique
leads to a somewhat (up to 30%) overstated estimate of the
reaction rate. Additional ammonia production is probably
also related to the conversion of compound II.
Thus, the proposed methods of evaluation of cytarabine
stability in aqueous solutions, based on HPLC determination
of compounds I and II and the colorimetric analysis of am-
monia, can be used for drug stability monitoring in develop-
ment of liquid medicinal forms of cytarabine.
32. S. V. Galushko and I. P. Shishkina, Khim.-Farm. Zh., 24(1),
85 – 86 (1990).
33. S. V. Galushko and I. P. Shishkina, J. Pharm. Biomed. Anal.,
10(10 – 12), 1093 – 1095 (1992).
34. M. Burk, M. Volmer, K. Fartash, et al., Arzneim.-Forsch., 45(5),
616 – 619 (1995).
35. J. Braess, J. Pfoertner, C. C. Kaufmann, et al., J. Chromatogr.,
B: Biomed. Appl., 676(1), 131 – 140 (1996).
36. L. A. Osterman, The Chromatography of Proteins and Nucleic
Acids [in Russian], Nauka, Moscow (1985), p. 203.
37. O. V. Travina, A Guide for Biochemical Investigations [in Rus-
sian], Medgiz, Moscow (1955), pp. 126 – 129.
REFERENCES
1. Medicinal Preparations of the Upjohn Company, A Handbook,
Upjohn Company Moscow Office, Moscow (1993).