Table IX-Extent of Complex Formation between 3,4-
`Dimethylphenol and Ethyl Myristate Determined by
`Partitioning Study
`
`Total Concentration
`of 3,4-Dimethylphenol,
`X 102 M
`
`Partition
`Coefficient
`
`[DMPCO"!Pl
`[DMPol
`
`0
`1.53
`3.07
`6.14
`9.21
`15.03
`30.00
`
`1.37
`1.73
`1.96
`2.61
`3.31
`4.57
`7.94
`
`0.27
`0.43
`0.91
`1.43
`2.35
`4.83
`
`for various cosolvent systems, and each showed a straight-line relation(cid:173)
`ship between the -two parameters (Fig. 4). The stability constants, K 1,1
`and K 1,2, for each ester were calculated from Fig. 4 and are shown in Table
`VII. It is evident from these results that 4-hexylresorcinol forms not only
`1:1 but also 1:2 complexes with esters in hexane. The stability constant
`values obtained for ethyl myristate were somewhat higher than those for
`the other esters. This result is probably due to the fact that ethyl my(cid:173)
`ristate has a larger hydrocarbon chain, which results in a better interac(cid:173)
`tion with the hydrophobic portion of phenols.
`To ascertain whether the formation of 1:2 complexes is due to the in(cid:173)
`volvement of the two hydroxy groups of 4-hexylresorcinol, the parti(cid:173)
`tioning study was repeated with 3,4-dimethylphenol. The data obtained
`from this study are shown in Tables VIII-X. A plot of [DMPcompJ/
`
`Table X-Extent of Complex Formation between 3,4-
`Dimethylphenol and Ethyl Pivalate Determined by Partitioning
`Study
`
`Total Concentration
`of 3,4-Dimethylphenol,
`X 102 M
`
`Partition
`Coefficient
`
`[DMPcomp]
`[DMPo}
`
`0
`3.82
`7.68
`15.36
`23.04
`38.40
`
`1.37
`1.78
`2.33
`3.33
`4.32
`6.63
`
`0.30
`0.70
`1.44
`2.17
`3.86
`
`•
`
`•
`.
`
`• e
`
`e
`
`•
`
`~16
`~
`0
`~ 12
`lU ::::
`Cl. 8
`E
`0 " ~ 4
`a
`
`60
`50
`40
`30
`20
`1 0
`0
`CONCENTRATION OF TOTAL ESTER,
`[Erl. IN HEXANE, X 102 M
`Figure 5-Plot of [DMPcompl I [DMPtreJ [ET] as a function of the total
`ester concentration, [ET], in hexane; [DMPcompJ and [DMPtreel rep(cid:173)
`resent the concentrations of complexed and free forms of 3,4-dimeth(cid:173)
`ylphenol, respectively. ([DMPcomp] = [DMPTJ - [DMPol). Key: ■ ,
`ethyl acetate; e, ethyl myristate; and .._, ethyl pivalate.
`
`[DMP01[Er] versus the total ester concentration is shown in Fig. 5;
`[DMPcomp] is the concentration of 3,4-dimethylphenol in the complex
`form, and [DMP0] is the concentration of the free form.
`As seen in Fig. 5, the monohydroxy compound forms only a 1:1 complex
`with the esters. The stability constants calculated from Fig. 5 are given
`in Table VII. The results of this study substantiate the conclusion that
`the diffusion of 4-hexylresorcinol through ethylene-vinyl acetate co(cid:173)
`polymers involved the formation of 1:1 and 1:2 complexes between the
`drug and the vinyl acetate portion of the copolymers.
`
`REFERENCES
`
`(1) E. Akaho, Ph.D. dissertation, University of Kentucky, Lexington,
`Ky. 1979.
`(2) H.B. Kostenbauder and T. Higuchi, J. Am. Pharm. Assoc. Sci. Ed.,
`45, 518 (1956).
`(3) T. Higuchi, J. H. Richards, S.S. Dacis, A. Kamada, J.P. How, M.
`Nakano, N. I. Nakano, and I. H. Pitman, J . Pharm. Sci., 58, 661
`(1969) .
`(4) H. Fung and T. Higuchi, ibid., 60, 1782 (1971).
`
`High-Performance Liquid Chromatographic Analysis of
`Chemical Stability of 5-Aza-2' -deoxycytidine
`
`KUN-TSAN LIN*x, RICHARD L. MOMPARLER**, and GEORGES E. RIVARD*
`Received December 17, 1979, from the *Centre de Recherche Pediatrique, H6pital Sainte-Justine, and the I Departement de Pharmacologie,
`Universite de Montreal, Montreal, Quebec , Canada H3T !CS.
`Accepted for publication March 31, 1981.
`
`Abstract □ The chemical stability of 5-au-2' -deoxycytidine (I) in acidic,
`neutral, and alkaline solutions was analyzed by high-performance liquid
`chromatography. In alkaline solution, I underwent rapid reversible de(cid:173)
`composition to N-(formylamidino)-N'-P-D-2-deoxyribofuranosylurea
`(II), which decomposed irreversibly to form 1-P-n-2'-deoxyribofurano(cid:173)
`syl-3-guanylurea (III). The pseudo-first-order rate constants for this
`reaction were determined. The decomposition of I in alkaline solution
`was identical to that reported previously for the related analog, 5-aza(cid:173)
`cytidine. However, in neutral solution (or water), there was a marked
`difference in the decomposition of I and 5-azacytidine. The same de(cid:173)
`composition products were formed from 5-azacytidine in neutral solution
`
`as in alkaline solution. However, in neutral solution, I decomposed to II
`and three unknown compounds that were chromophoric at 254 nm.
`Compound I was most stable when stored in neutral solution at low
`temperature.
`
`Keyphrases Cl 5-Aza-2'-deoxycytidine-analysis of chemical stability
`using high-performance liquid chromatography □ Antileukemic
`agents-5-aza-2'-deoxycytidine, analysis of chemical stability using
`high-performance liquid chromatography □ High-performance liquid
`chromatography-analysis of chemical stability of 5-aza-2'-deoxycyti(cid:173)
`dine
`
`3). This antimetabolite is related to 5-azacytidine, an agent
`5-Aza-2'-deoxycytidine (I), a nucleoside antimetabolite,
`is a very active antileukemic agent in mice (1, 2) and a
`currently used in the clinical treatment of acute leukemia
`potent cytotoxic agent against neoplastic cells in vitro (2,
`(4).
`CELGENE 2021
`APOTEX v. CELGENE
`IPR2023-00512
`
`0022-354918111100-1228$01.0010
`© 1981, American Pharmaceutical Association
`
`1228 / Journal of Pharmaceutical Sciences
`Vol. 70, No. 11, November 1981
`
`

`

`NH2
`
`NJ_N
`
`NH~ 0
`J._ n
`N7 NHCH
`
`HO~
`
`HO~H
`
`OH H
`I
`
`OH H
`II
`
`NH2
`
`N~NH2
`
`HO~H
`
`OH H
`m
`
`Scheme I
`
`BACKGROUND
`
`One major problem encountered in the clinical formulation of 5-aza(cid:173)
`cytidine is its chemical instability, leading to solutions of decreasing
`potency on storage. The chemical stability of 5-azacytidine was first
`studied by Pithova et al. (5), who demonstrated that in alkaline solution
`the triazine ring of 5-azacytidine opens and loses a formyl group to form
`1-/1-D-ribofuranosyl-3-guanylurea. These workers proposed that the
`intermediate compound in this reaction was N-(formylamidino)-N'(cid:173)
`/1-D-ribofuranosylurea, but they were unable to isolate it using paper
`chromatography because of its chemical instability. Beisler (6), using
`high-performance liquid chromatography (HPLC), isolated and identi(cid:173)
`fied this intermediate compound and showed that it could be partially
`converted back to 5-azacytidine.
`The chemical decomposition of I at alkaline pH (Scheme I) presumably
`follows the same reaction steps as described previously for 5-azacytidine.
`In alkaline solutions, I undergoes a reversible hydrolytic reaction to form
`N-(~ormyla~idino)-N'-/1-D-2-deoxyribofuranosylurea (II), which, by
`the 1rrevers1ble loss of the formyl group, forms 1-/1-D-2' -deoxyribofura(cid:173)
`nosyl-3-guanylurea (Ill).
`The present study investigated the decomposition of I in acidic, neu(cid:173)
`tral, and alkaline solutions and at different temperatures to illustrate the
`chemical stability of this compound.
`
`EXPERIMENTAL
`
`5-Aza-2'-deoxycytidine (I) was synthesized1 by modification of an
`earlier method (7).
`Method of Analysis-HPLC2 was performed with a variable-wave(cid:173)
`le~gth detector by monitoring at 220 or 254 nm. Analytical and prepar(cid:173)
`ative work was accomplished with a 300 X 3.9-mm i.d. commercially
`packed octadecylsilane column3, which was eluted at a flow rate of 2
`ml/min with 10 mM potassium phosphate buffer (pH 6.8). For analytical
`and preparative work, 20- and 500-µl injector loops, respectively, were
`used.
`Stock solutions of 10 mM I in water were stored at -70°. Aliquots of
`this solution were thawed quickly and stored at 1-2°. The following
`buffers were used for the pH stability studies: phosphoric acid, pH 2.2;
`phosphoric acid- potassium phosphate, pH 3.2; sodium acetate buffer,
`pH 4.5 and 5.7; potassium phosphate buffer, pH 6.4 arid 7.0; and sodium
`borate, pH 8.5, 9.2, 9.8, and 10.4. Compound II was prepared by addition
`of 20 µl of 0.5 M sodium borate, pH 10.4, to a 1.0-ml solution of 10 mM
`I. The mixture was incubated for 2 min at 24° and then neutralized with
`
`i By Dr. A. Piskala,_Institute of Organic Chemistry and Biochemistry, Czecho(cid:173)
`slovak Academy of Science, 16610 Prague 6, Czechoslovakia.
`2 Model 110 pump and Hitachi variable-wavelength detector Altex Scientific
`Berkeley, Calif.
`'
`'
`3 µBondapak C1s, Waters Associates, Framingham, Mass.
`
`Omin
`
`I O.OlAU
`
`2min
`
`pH 10.4 24 °C '
`
`IOmin
`
`25min
`
`60min
`
`60min
`
`Omin
`
`I O.OIAU
`
`2min
`
`25min
`
`E
`C
`(h
`N
`w·
`CJ z
`Q:) er
`0
`CJ)
`CD
`<(
`
`<(
`
`E
`C
`
`~
`N
`w·
`CJ z
`CD er
`0
`CJ)
`CD
`<(
`
`<(
`
`Figure 1-Decomposition of I at pH 10.4 and 24°. Compound I (1 .0
`mM) was dissolved in 10 mM borate buffer (pH 10.4). At the indicated
`times, an aliquot of the solution was neutralized with phosphate buffer
`and analyzed by HPLC at 254 and 220 nm.
`
`20 µl of 1.0 M KH2P04, and a 500-µl sample was injected onto the column
`for isolation of TI.
`Kinetic Experiments-A series of tubes containing 10 µl of 0.5 M
`stock sodium borate buffer were incubated at 37° for 10 min. Then 100
`µl of isolated II (in 10 mM potassium phosphate, pH 6.8) was added to
`each tube to start the reaction. At timed intervals, an aliquot was with(cid:173)
`drawn, and 20 µI was injected onto the column. The absorbance at 220
`nm was recorded, and the relative concentration of the observed degra(cid:173)
`dation products was characterized by peak heights. Similar experimental
`procedures were followed for the kinetic study on 5-azacytidine4•
`
`RESULTS
`
`The decomposition of I at pH 10.4 and 24° as determined by HPLC
`is shown in Fig. 1. Measurement of the absorbance of the column eluate
`at 254 nm initially showed a single major peak (I) with a tR value of 5.5
`min. With time, a second peak (II) appeared with a tR value of 6.6 min.
`Both peaks showed a gradual decrease in peak size with time. At 220 nm,
`the column eluate initially showed a single major peak (I) with a tR = 5.5
`min. With time, peak II (tR = 6.6 min) and peak III (tR = 3.5 min) became
`evident. At 60 min, most of I decomposed to peak III.
`
`4 Drug SY!)thesis and Chemistry Branch, Division of Cancer Treatment, National
`Cancer Institute, Bethesda, Md.
`
`Journal of Pharmaceutical Sciences I 1229
`Vol. 70, No. 11, November 1981
`
`

`

`Omitt
`
`A
`
`~
`
`pH8 .J
`
`37°C
`
`E
`C:
`;1;
`N w
`() z
`
`<(
`a:i
`a:
`0
`Vl
`IX)
`<(
`
`Omin
`
`II
`
`pH 3.2 0°C
`
`0.5 min
`
`a
`
`e C:
`~
`w·
`u z
`a:i a: g
`
`<(
`
`Cll
`<(
`
`5 min
`
`30 mir.
`
`60min JlJ
`
`0
`Figure 2-Conversion of II to I at pH 8.3, 37° (A), and to unidentified compounds at pH 3.2, 0° (B). Peak II was isolated by HPLC and placed
`in 20 mM potassium phosphate buffer (at final pH 8.3 or 3.2). At the indicated times, aliquots of this solution were analyzed by HPLC.
`
`When II was isolated and placed at pH 8.3 and 37°, I appeared rapidly,
`indicating that II could be converted to I (Fig. 2A). However, when lI was
`placed in an acidic solution (pH 3.2), it did not produce I; three uniden (cid:173)
`tified peaks at 254 nm were observed. The latter reaction was very rapid,
`even at 0° (Fig. 2B).
`The decomposition of I and 5-azacytidine when stored in water and
`in pH 7.4 phosphate buffer at 24° for 24 hr is compared in Fig. 3. Ab(cid:173)
`sorbance measurements at 254 nm of the column eluate showed five major
`peaks for I and two major peaks for 5-azacytidine. At 220 nm, the column
`eluate showed seven major peaks for I and three major peaks for 5-aza(cid:173)
`cytidine.
`
`The decomposition rates of I in acidic, neutral, and basic solutions at
`24 and 37° are shown in Figs. 4A and 4B, respectively. Compound I de(cid:173)
`composed more rapidly at pH 2.2, 8.5, and 9.2 than at pH 5.7, 6.4, and 7.0.
`The decomposition rate of I was the slowest at pH 7.0. At pH 2.2 a peak
`with the same tR value (2.3 min) as 5-azacytosine appeared on the chro(cid:173)
`matogram. The decomposition rate of I was very temperature dependent.
`For example, at pH 7.0, the decomposition rate was about sevenfold
`greater at 37 than at 24°.
`A kinetic study at pH 8.1 and 9.5 was then carried out with II as the
`starting substance. According to Scheme I, the pseudo-first-order rate
`constants k1, k- 1, and k2 were determined using the following differential
`
`5-AZA-2 1-DEOXYCYTIDINE
`H20
`
`IOmin.
`
`Omin. :,
`"'
`N
`0.
`0
`
`"
`
`:,
`0
`o•
`
`pH7.4
`
`:,
`"'
`o.
`
`0
`
`)
`
`I
`
`/
`
`1
`I
`
`:,
`<(
`N
`0
`0
`
`:,
`"'
`6
`d
`
`:,
`"'
`6.
`0
`
`E
`C:
`fi3
`N
`w·
`() z
`ttJ a:
`0
`(I)
`IX)
`<(
`
`<(
`
`0
`
`E
`C:
`~
`N
`w·
`u z
`<(
`IX)
`a:
`0
`en
`a:,
`<(
`
`E
`C:
`o .
`N ,
`N
`w··
`u
`z
`<(
`ai
`a:
`0
`Vl
`tXl
`<(
`
`E
`C:
`;%
`N
`w·
`u
`z
`<(
`IX)
`a:
`0
`Vl
`IX)
`<(
`
`pH7.4
`
`5-AZACYTIDINE
`
`Omin
`
`IOmin.
`
`H20
`
`:,
`
`"'
`"' q
`
`0
`
`:,
`
`"' "'
`
`0.
`0
`
`0
`0
`Figure 3-Decomposition of I and 5-azacytidine in pH 7.4 phosphate buffer and water. 5-Azacytidine (0.5 mM) and I (0.5 mM) were dissolved
`in 10 mM potassium phosphate (pH 7.4) and water and incubated at 24° . At the indicated times, aliquots of these solutions were analyzed by HPLC
`at 2.54 and 220 nm.
`
`'
`
`1230 / Journal of Pharmaceutical Sciences
`Vol. 70, No. 11, November 1981
`
`

`

`Table I-Rate Constants for the Alkaline Decomposition of I and
`5-Azacytidine at 37°
`
`Compound and pH 0
`
`5-Azacytidine, 9.5
`5-Azacytidine, 8.1
`I, 9.5
`
`Pseudo-First-Order Rate Constant, min- 1
`k1
`k-1
`k2
`
`2.0 X 10-1
`1.7 X 10-2
`2.2 X 10-1
`
`2.2 X 10-1
`2.0 X 10-2
`2.8 X 10-1
`
`6.0 X 10- 1
`5.6 X 10-2
`7.5 X 10- 1
`
`0 The pH value was determined at 37° in a separate tube by mixing one volume
`of0.5 M stock sodium borate buffer with 10 volumes of 10 mM potassium phosphate
`(pH 6.8).
`
`9.2
`
`8.5
`
`80
`
`z
`0
`j'.:
`u
`<{
`a:
`u..
`w 40
`...I
`0
`:::E
`
`75
`
`* 6
`z
`z
`~ 50
`~
`w
`a:
`
`25
`
`10L
`
`75
`
`'#.
`(.!) z
`~ 50
`<{
`~
`w
`a:
`
`25
`
`25
`
`30
`
`35
`
`0
`
`5
`
`10
`
`20
`15
`HOURS
`B
`Figure 4-Ef f ect of pH and temperature on the stability of I. Com(cid:173)
`pound I (1.0 mM) was dissolved in 10 mM of the appropriate buffer with
`the indicated pH and incubated at either 24° (top) or37° (bottom). At
`the indicated times, an aliquot of the solution was neutralized with
`phosphate buffer and analyzed by HPLC.
`
`0
`
`20
`
`60
`40
`MINUTES
`Figure 5-Time-concentration profile of decomposition for I at pH 9.5
`and 37°. Key: .&, normalized data for I; ■ , normalized data for II; and
`•• normalized data for III. Solid lines were generated by computer for
`.&, ■, and • using rate constants determined independently.
`
`80
`
`100
`
`the kinetic study of 5-azacytidine, and the results are shown in Table I
`and Figs. 6 and 7.
`
`DISCUSSION
`
`The hydrolysis of 5-azacytidine in alkaline solution and water results
`in the opening of the triazine ring between C-6 and N-1 to form N(cid:173)
`(formylamidino)-N'-/3-D-ribofuranosylurea in a reversible reaction; when
`this latter compound loses the formyl group, it irreversibly forms 1-
`/3-D-ribofuranosyl-3-guanylurea (5, 6). By using quantum chemical cal-
`
`100
`
`80
`
`z
`0
`~ 60
`<{
`a:
`u..
`w
`...I 40
`0
`:::E
`
`20
`
`0
`
`20
`
`0 ~-~-=-::::i;=:::::I:=== ........ .._
`40
`80
`60
`MINUTES
`Figure 6-Time-concentration profile of decomposition for 5-azacy(cid:173)
`tidine at pH 9.5 and 37°. Key: •• normalized data for 5-azacy tidine; ■,
`normalized data for N-(formylamidino)-N'-/3-D-ribofuranosy lurea; and
`• normalized data for 1-/3-D-ribofuranosyl-3-guanylurea. Solid lines
`were generated by computer for •• ■, and • using rate constants de(cid:173)
`termined independently.
`
`100
`
`Journal of Pharmaceutical Sciences I 1231
`Vol. 70, No. 11, November 1981
`
`equations (8):
`Ii= L1II0(-l-e-At +-1-e-Bi)
`B-A
`A-B
`IIi = Ilo (hi -A e-At + k1 - Be-Rt)
`B-A
`A-B
`Ill = II (l - k2(k1 - A) -At_ k2(k1 - B) -Bi)
`o
`A(B-A) e
`B(A-B) e
`Here A and B are roots of the following quadratic equation taken with
`the reverse signs:
`
`(Eq. l)
`
`(Eq. 2)
`
`(Eq. 3)
`
`1
`
`The rate constants thus obtained are presented in Table I and were used
`to generate time-concentration profiles for I-III. Normalized experi(cid:173)
`mental data obtained by HPLC were superimposed on these generated
`time-concentration profiles (Fig. 5). Similar procedures were used for
`
`(Eq. 4)
`
`

`

`100
`
`80
`
`z
`0
`~ 60
`u
`<{
`ex:
`u.
`w
`...J 40
`0
`~
`
`20
`
`z
`0
`~ u
`<{
`ex:
`u.
`w
`...J
`0
`~
`
`80
`
`100
`
`120
`
`0
`
`20
`
`40
`
`60
`I1/IINUTES
`Figure 7-Time- concentration profile of decomposition for 5-azacy(cid:173)
`tidine at pH 8.1 and 37°. Key:• • normalized data for 5-azacytidine; ■,
`normalized data for N-(formylamidino)-N'-{3-o-ribofuranosy lurea; and
`• • normalized data for J-{3-D-ribofuranosyl-3-guanylurea. Solid lines
`were generated by computer for • • ■, and • using rate constants de(cid:173)
`termined independently.
`
`culations, it was shown (5) that the electron density at C-6 of 5-azacy(cid:173)
`tosine is much lower than cytosine, making this position more susceptible
`to nucleophilic attack by a hydroxyl ion.
`From the structural similarity of 5-azacytidine and I, it would seem
`that I would decompose in alkaline solution according to Scheme I. This
`study supports this reaction scheme in alkaline solution. As shown in Fig.
`1, the decomposition of peak I (tR = 5.5 min) at pH 10.4 produced peak
`II (tR = 6.6 min) when the column eluate was monitored at 254 nm. Peak
`III was not observed on this chromatogram because guanylurea deriva(cid:173)
`tives are nonchromophoric at 254 nm (5, 6).
`However, at 220 nm, peak III (tR = 3.5 min) became apparent following
`the decomposition of 11 to III. When peak II was isolated, it had the same
`UV max of 238 nm as reported previously (6) for N-(formylamidino)-N'(cid:173)
`{3-ribofuranosylurea. The hydrolysis of I in alkaline solution produced
`an initial increase in absorbance5 at 238 nm, supporting the formation
`of II, since the formylguanylurea derivative has a much higher extinction
`coefficient than 5-azacytosine derivatives at this wavelength (6). The
`conversion of II to I (Fig. 2A) indicates that this reaction is reversible,
`which is in agreement with other observations (5, 6) for 5-azacytidine.
`There were marked differences in the decomposition of I and 5-aza(cid:173)
`cytidine in phosphate buffer (pH 7.4) and water (Fig. 3). In these solvents,
`5-azacytidine decomposed to form formylguanylurea and guanylurea
`derivatives, as reported previously (6). However, the decomposition of
`I in phosphate buffer or water was very different from that of 5-azacyti(cid:173)
`dine as shown by the presence of five (254 nm) and seven (220 nm) peaks
`on the chromatograms of I. Two of these peaks represent I and II, whereas
`the other peaks were not identified. These unidentified peaks could also
`be produced rapidly by placing II at pH 3.2 (Fig. 2B), even at 0° . These
`unidentified compounds were not the intermediates that eventually
`formed III, since they were not converted to III when placed in alkaline
`buffer, as would happen to I and II under the same conditions.
`Determination of the pseudo-first-order rate constants for I and 5-
`azacytidine at pH 9.5 and 37° demonstrated that both compounds de(cid:173)
`composed at comparable rates, although the formylguanylurea derivative
`for ribose appeared to be slightly more stable than that of deoxyribose
`(Table I). A 22-fold decrease in the hydroxyl-ion concentration (pH
`9.5- 8.1) reduced all three rate constants for 5-azacytidine equally by
`about 11-fold. Lowering the pH from 9.5 to 8.1 also greatly reduced the
`conversion rate of lI to I and to III (Figs. 5 and 8). The kinetic data ob(cid:173)
`t.ained for I at pH 8.1 did not quite fit in Eqs. 1-3 because apparently some
`of the decomposition did not follow Scheme I precisely due to the for(cid:173)
`mation of minor unidentified peaks. Thus the rate constants were not
`determined. These unidentified peaks became increasingly prominent
`as the pH was lowered, and there was virtually no conversion of II to I at
`pH <4.6. (Fig. 2B) .
`
`0
`
`20
`
`40
`
`60
`MINUTES
`Figure 8-Time-concentration profile of decomposition for I at pH 8.1
`and 37° . Key: •• normalized data for I; ■, normalized data for II; and
`•• normalized data for III.
`
`80
`
`Comparison of the overall stability profile of I in aqueous buffer so(cid:173)
`lution at various pH levels showed that I was most stable at neutral pH
`and at low temperature (Fig. 4). The instability of I at high pH could be
`explained on the basis that the opening of the triazine ring (k 1) and the
`breakage of an amide bond (k 2) is facilitated by a hydroxyl ion. Reduction
`in pH would slow down both these events and thus stabilize I. However,
`I became increasing unstable as the pH was gradually reduced below 7 .0.
`This result might be due to the increase in the rate of ring opening as well
`as to the instability of II in acidic solutions; thus, once ring opening took
`place and produced II, the latter broke down to the unidentified peaks
`more rapidly and minimized the reversal back to I (Fig. 2B) . In strong
`acidic solution (with pH <2.2), breakage of the glycolytic bond of I oc(cid:173)
`curred, producing 5-azacytosine6 as reported previously for I (9) and for
`5-azacytidine (5, 10).
`The effect of pH and temperature on the chemical stability of 5-aza(cid:173)
`cytidine was studied by Chan et al. (11), who observed that this com(cid:173)
`pound is most stable at pH 7 .0 and that its rate of decomposition increases
`in solutions of high or low pH and with temperature.
`
`REFERENCES
`
`(1) F. Sorm and J. Vesely, Neoplasma, 15,339 (1968).
`(2) R. L. Momparler and F. A. Gonzales, Cancer Res., 38, 2673
`(1978).
`(3) R. L. Momparler and J. Goodman, ibid., 37, 1636 (1977).
`(4) W. R. Volger, D. S. Miller, and J . W. Keller, Blood, 48, 331
`(1976).
`(5) P. Pithova, A. Piskala, J. Pitha, and F. Sorm, Collect . Czech.
`Chem. Commun., 30, 2801 (1965).
`(6) J . A. Heisler, J . Med. Chem., 21, 204 (1978).
`(7) J. Pliml and F. Sorm, Collect. Czech. Chem. Commun., 29, 2576
`(1964).
`(8) N. M. Rodiguin and E. N. Rodiguina, "Consecutive Chemical
`Reactions," Van Nostrand, Princeton, N.J., 1979.
`(9) A. Piskala, M. Synackova, H. Tomankova, P. Fiedler, and V.
`Zizkovsky, Nucleic Acid Res., 4, s109 (1978).
`(10) R. E. Notari and J. L. DeYoung, J. Pharm. Sci., 64, 1148
`(1975).
`(11) K. K. Chan, D. D. Giannini, J. A. Staroscik, and W. Sadee, ibid.,
`68,807 (1979).
`
`ACKNOWLEDGMENTS
`
`Supported by U.S. Public Health Service Grant CA23340 from the
`National Cancer Institute.
`The authors thank Suzanne Beaudet for assistance in preparation of
`this manuscript.
`
`5 R. L. Momparler, unpublished data.
`
`6 Unpublished observations.
`
`1232 / Journal of Pharmaceutical Sciences
`Vol. 70, No. 11, November 1981
`
`