91395y(l 974).
`(l544) L. W. Dittert, Drug Intel/. Clin. Pharm., 8, 222(1974).
`(1545) V. G. Hitzenberger, Wien . Med. Wochenschr., 123 (45),
`653(1973); through Chem. Abstr., 80, 33668d(1974).
`(1546) E. S. Vesell, J . Pharmacokinet. Biopharm., 1 (6),
`521(1973).
`(1547) M. Rowland and S. B, Matin, ibid., 1 (6), 553(1973).
`(1548) R. E. Notari, A. M. Burkman, and W. K. Van Tyle, J.
`Pharm . Pharmacol .. 26, 481(1974).
`(1549) H. G. Boxenbaum, S. Riegelman, and R. M. Elashoff, J.
`Pharmacokinet . Biopharm., 2 (2). 123(1974).
`(1550) A. J. Sedman and J, G. Wagner, ibid., 2 (2), 161(1974).
`(1551) D. Shen and M. Gibaldi, J . Pharm. Sci, 63, 1698(1974) .
`(1552) J. G. Wagner, C/in. Pharmacol. Ther., 16, 691(1974).
`(1553) C. D. Thron., Pharmacol. Reu., 26 (I), 3(1974); through
`Chem. Abstr. , 80, 127965a(l974).
`(1554) J. G. Wagner, P. D. Holmes, P. K. Wilkinson, D. C. Blair,
`and R. G. Stoll, Amer. Reu. Resp. Dis., 108 (3), 536(1973); through
`Chem. Abstr., 80, 63788k(l974).
`(1555) A. Ouchida, E. Mizuta, and T. Shima, Takeda Kenk(cid:173)
`(4) , 522(1973);
`through Chem. Abstr., 80,
`yusho Ho, 32
`ll 5946g(l 97 4).
`
`(1556) D. W. Vere, Sci. Basis Med., 1972, 363; through Chem.
`Abstr., 81, L30710q(1974J.
`(1557) W. J. Westlake, Curr. Cone. Pharm. Sci.: Dosage Form
`through Chem. Abstr., 81,
`Des. Bioauailability, 1973, 149;
`58063m(1974).
`(1558) P. T. Schoenemann, D. W. Yeeair, J. J. Coffey, and F. J.
`Bullock, Ann. N. Y . Acad. Sci., 226, 162(1973); through Chem.
`Abstr., 80, 78280n (1974).
`(1559) H. Braeunlich and F. K. Splinter, Acta Biol. Med. Ger.,
`31 (3), 435(1973); through Chem. Abstr., 80, 33884w(l 974).
`(1560) P , J. Niebergall, E. T. Sugita, and R. L. Schnaare, J.
`Pharm. Sci., 63, 100(1974).
`(1561) E. S. Vesell, C/in. Pharmacol. Ther, 16 (I, Pt. 2).
`135(1974); through Chem. Abstr., 81, 145558v(l974).
`(1562) M. Gibaldi, B. Grundhofer, and G. Levy, ibid., 16,
`520(1974).
`
`ACKNOWLEDGMENTS AND ADDRESSES
`Received from the • Institute of Pharmaceutical Sciences, Syn(cid:173)
`lex Corporation, Palo Alto, CA 94304, and 1Warner-Lambert Re(cid:173)
`search Institute, Morris Plains, NJ 07950
`• To whom inquiries should be directed.
`
`RESEARCH ARTICLES
`
`Kinetics and Mechanisms of Degradation of the
`Antileukemic Agent 5-Azacytidine in
`Aqueous Solutions
`
`ROBERT E. NOTARI x and JOYCE L. De YOUNG•
`
`Abstract C The hydrolytic degradation of 5-azacytidine was stud(cid:173)
`ied spectrophot.ometrically as a function of pH, temperature, and
`buffer concentration. Loss of drug followed apparent first-order
`kinetics in the pH region below 3. At pH <I, 5-azacytosine and 5-
`azauracil were detected; at higher pH values, drug was lost to
`products which were essentially nonchromophoric if examined in
`acidic solutions. The apparent first-order rate constants associated
`with formation of 5-azacytosine and 5-azauracil from 5-azacytidine
`are reported. Above pH 2.6, first -order plots for drug degradation
`are hiphasic. Apparent first-order rate constants and coefficients
`for the biexponential equation are given as a function of pH and
`buffer concentration. A reaction mechanism consistent with the
`data is discussed together with problems associated with defining
`
`the stability of the drug in aqueous solutions. At 50°, the drug ex (cid:173)
`hibited maximum stability at pH 6.5 in dilute phosphate buffer.
`Similar solutions were stored at 30° to estimate their useful
`shelflife. Within 80 min, 6 X 10-• M solutions of 5-azacytidine de (cid:173)
`creased to 90% of original potency based on assumptions related to
`the proposed mechanisms.
`Keyphrases a 5-Azacytidine-kinetics and mechanisms of degra(cid:173)
`dation in aqueous solutions, effect of pH, temperature, and buffer
`concentration a Hydrolysis, 5-azacytidine in aqueous solutions(cid:173)
`kinetics and mechanisms, effect of pH , temperature, and buffer
`concentration □ Antileukemic agents-5-azacytidine, kinetics and
`mechani&1!1S of degradation in aqueous solutions
`
`The synthesis of 5-azacytidine (I) was reported in
`1964 (1 ). The compound has antimicrobial and anti(cid:173)
`tumor activities (2-4) and is currently being evalu(cid:173)
`ated clinically in the treatment of human leukemias.
`The clinical use of this drug makes it desirable to
`determine kinetic data on its reactivity in aqueous
`
`1148 / Journal of Pharmaceutical Sciences
`
`solutions. Such information is especially important in
`consideration of its relative instability when com(cid:173)
`pared to cytidine itself. Especially notable are the
`ease with which the triazine ring hydrolyzes and the
`!ability of the sugar-triazine bond (5). Although most
`hydrolysis products of I have been identified (5), lit-
`CELGENE 2024
`APOTEX v. CELGENE
`IPR2023-00512
`
`

`

`Table I-Apparent First-Order Rate Constants (hour -• ) for
`Hydrolysis in Hydrochloric Acid at 70°
`
`Substrate
`
`5-Azacy tosineO
`
`[HCl],
`N
`
`0 .10
`0 .25
`[HCI],
`N
`
`hobs
`0.376
`0.238
`
`h,,
`
`0.121
`0.119
`
`k "
`0.255
`0.1 I 9
`
`h,.
`
`0.0495b
`0 .0636b
`
`hobs
`
`h"
`0.111
`0.136
`
`k .,
`
`h,.
`
`0.110
`0. 101
`
`1.77
`0.810
`
`tie has been done concerning the kinetics and mecha(cid:173)
`nism of its degradation.
`The purpose of this report is to describe the results
`of studies on the hydrolysis of I and the related com(cid:173)
`pounds 5-azacytosine (II) and 5-azauracil (III).
`
`EXPERIMENTAL
`
`Analytical Metbods--Beer's law plots were constructed for 5-
`azacytidine, 5-azacytosine, and 5-azauracil in 0.1 N HCI using a re(cid:173)
`cording spectrophotometer•. Molar absorptivities were 2.36 X 103
`for 5-azacytidine at 255 nm; 3.35 X 103, 4.33 X 10-\ and 5.98 X 103
`for 5-azacytosine at 230, 235, and 250 nm, respectively; and 3.67 X
`103, 3.80 X 103, and 1.50 X 103 for 5-azauracil at 230, 235, and 250
`nm, respectively.
`Concentrations of components in mixtures of 5-azacytosine and
`5-azauracil were calculated using:
`i03 [II] = 0.217 A ,:,o - 0.088A2,.
`
`(Eq. I)
`
`103 [III]= 0.353Az., - 0.198A,:;0
`
`(Eq. 2)
`
`which were derived from simultaneous equations for total absorb(cid:173)
`ance, A , at 230 and 250 nm for mixtures of 5-azacytosine and 5.
`azauracil in 0.1 N HCI. Good agreement was obtained between the
`results of the assay and the known concentrations in synthetic
`mixtures.
`Mixtures of 5-azacytidine and the two bases were assayed utiJiz.
`ing the differential stability of the components. The absorbance at
`255 nm of solutions containing all three components in 0.1 N HCl
`was determined. These solutions were then made 0.05 N in sodium
`hydroxide and allowed to stand at room temperature for 40- 50
`min. This period was sufficient to degrade 5-azacytidine complete(cid:173)
`ly to nonchromophoric products without loss of 5-azacytosine or
`5-azauracil. Samples were reacidified with 0.1 N HCI, and the ab(cid:173)
`sorbances at 235, 250, and 255 nm were determined. The 235- and
`250-nm values were used to calculate the concentrations of 5-aza(cid:173)
`cytosine and 5-azauracil:
`
`10" [II] - 0.234A250 - 0.092A235
`
`103 [III] - 0.~A,,, - 0.267 A 250
`
`(Eq.3)
`
`(Eq. 4)
`
`After correcting for dilution, the absorbances of the sample be(cid:173)
`fore and after treatment with alkali were used to calculate the 5-
`azacytidine concentration:
`
`(Eq. 5)
`
`This essay method was shown to be satisfactory using synthetic
`mixtures.
`Kinetics of 5-Azacytoslne and 5-Azauracil Hydrolyais in
`Hydrochloric Acid Solutions-Reaction solutions containing 2
`X 10- ◄ M 5-azacytosine or 3 X 10- ◄ M 5-azauracil in 0.1 and 0.25
`
`NH ,
`
`N~~
`0)-_NJ
`H
`II
`
`1 Cary 15.
`
`5-AzacytidineC
`
`0.10 1.98
`0.25
`1.05
`
`a See Sche me I. Concentration is 2 X 10-. ,\/ . b Determined from
`hydrolysis of 5-azauracil (3 X 10-< M). c See Sc hem,· II. Conccn tra(cid:173)
`tion is 2 X IO'"'M .
`
`N HCI were placed in constant-temperature baths. Samples of the
`5-azacytosine reactions were withdrawn as a function of time and
`assayed for substrate and 5-azauracil as described previously. The
`spectra of the 5-azauracil reactions decreased with time, with no
`evidence of formation of a new chromophore. Rate constants for its
`loss were calculated directly from absorbance data.
`TLC of Reaction Solutions of 5-Azacytosinc and 5-Azaura(cid:173)
`cil-Samples of the reactions of approximately 0.01 M 5-azacyto(cid:173)
`sine and 0.01 M 5-azauracil in 0. 10 N HCI at 70° were taken after
`about 50 and 100% of the substrate had reacted . These samples
`were neutralized and spotted along with authentic samples of 5-
`azacytosine and 5-azauracil on TLC plates coated with 0.25 mm
`silica gel GF2s4 • The plates were developed in 1-butanol-acetic
`acid-water (100:10:30) for a distance of about 10 cm, air dried at
`room temperature, and examined with a shortwave UV lamp.
`Kinetics of 5-Azacytidine Hydrolysis in Hydrochloric Acid
`Solutions- Reaction solutions containing 5-azacytidine in excess
`hydrochloric acid were prepared as shown in Tables I and II. The
`absorbance of reactions in solutions less concentrated than 0.1 N
`HCl was determined by diluting aliquots of the reaction with hy- ·
`drochloric acid to give a final concentration of 0.25 N HCI. In these
`solutions, the UV absorption spectrum of 5-azacytidine decreased
`with time without the appearance of any new absorbances. Ab(cid:173)
`sorbance data were used directly to calculate the apparent first(cid:173)
`order rate constants.
`In solutions containing 0.1 N or stronger hydrochloric acid,
`there was evidence that a fraction of the 5-azacytidine decomposed
`to form 5-azacytosine and 5-azauracil. In these cases, the reactions
`were assayed using the described differential stability method. At
`70°, Eqs. 3, 4, and 5 were used to ·assay for 5-azacytidine, 5-azacy(cid:173)
`tosine, and 5-azauracil, respectively. In the 50° reactions, the dif.
`ferential absorbance due to 5-azacytidine (Eq. 5) was used to cal(cid:173)
`culate the apparent first-order rate constants for loss of su bstrate.
`Kinetics of 5-Azacytidine Hydrolysis in Buffers-Reaction
`solutions containing 5-ezacytidine in excess buffer were prepared
`as shown in Table llI. Solutions were placed in constant-tempera·
`ture baths; aliquots were removed at appropriate times, chilled,
`and diluted with an equal volume of hydrochloric acid to give a
`final concentration of 0.25 N HCI. The absorbance of these solu(cid:173)
`tions was measured at 255, 260, and 265 nm . Since absorbance
`values decreased with time to become zero when the reaction was
`complete, absorbance data were used directly in the calculation of
`
`Table II-Apparent First-Order Rate Constants for
`Hydrolysis of 5-Azacytidinea at 50° in Hydrochloric Acid
`
`[HCI], N
`
`0.80
`0.50
`0. 25
`0.097
`0.0097
`0.005
`
`pHb
`
`0.22
`0.44
`0.73
`1.12
`2.06
`2.39
`
`kobs, hr - •
`
`0 . 0595
`0 .0764
`0 .1 18
`0 . 239
`1.12
`1.39
`
`a Reactions in 0.80, 0.50, 0. 25. an d 0.005 N HCI rnn tarn 4 X
`JO-< M substrate. Substrate conc en tration in 0.097 ,tnd 0.0097 ,'J
`HCI rea ctio ns is 6 X I O'"' M. b Th e pH of hydrochlnm acid so lu
`t ion s was c alcul ate d us ing activit y col'fficicnt va.hll·s from tht· lit era(cid:173)
`tur e (6).
`
`Vol. 64, No. 7, July 1975 / 1149
`
`

`

`Table III-Experimental Conditions and Apparent First-Order Rate Constants for Loss of 6 X 10--q M 5-Azacytidine in
`th1'! Presence of Buffers in the 7.6-2 .6 pH Range at 50°; µ = 0 .5
`
`pH
`
`[NaH,PO,)
`
`[Na,HPO,)
`
`Buffer
`
`7.59 ± 0.08
`
`6.52 ± 0.04
`
`5.4 2 ± 0.01
`
`5.60 ± 0.04
`
`4.62 ± 0.04
`
`3.58 ± 0.02
`
`2.58 ± 0. 06
`
`0.015
`0.010
`0.005
`0.003
`
`0.10
`0.08
`0 .05
`0.02
`
`0.30
`0.20
`0.10
`0.06
`
`[CH,COOH)
`0.040
`0.024
`0.016
`0 .008
`
`0.400
`0 .240
`0.144
`0.080
`
`0.40
`0.24
`0.16
`0.08
`
`[HCOOHJ
`0.40
`0.24
`0.16
`0.08
`
`0.15
`0.10
`0.05
`0.03
`Unbuffered
`0.10
`0.08
`0.05
`0 .02
`Unbuffered
`0.03
`0.02
`0.01
`0.006
`Unbuffered
`[CH0COONa)
`.40
`0.24
`0. 16
`0. 08
`Unbuffered
`0.400
`0.240
`0.144
`0.080
`Unbuffered
`
`0.040
`0.024
`0.016
`0 .008
`Unbuffered
`[HCOONa]
`0.040
`0.024
`0.016
`0.008
`Unbuffered
`
`[NaCl]
`
`0 .035
`0.19
`0.345
`0 .407
`
`0.10
`0.18
`0.30
`0.42
`
`0.11
`0.24
`0.37
`0.42
`
`[NaCl]
`0.10
`0.26
`0.34
`0.42
`
`0.100
`0.260
`0.356
`0.4 20
`
`0.46 0
`0.476
`0.484
`0.492
`
`[NaCl]
`0.460
`0.476
`0.484
`0.492
`
`ka", hr -•
`
`hba, hr-•
`
`0.74
`0.62
`0 .50
`0.44
`0 .37b
`0.29
`0 .24
`0.18
`0.12
`0 .076b
`0.41
`0 .29'
`0 .18
`0.13
`0.065b
`
`0 .15
`0.12
`0.099
`0.080
`0.064b
`0.48
`0 .41
`0.37
`0.27
`0 .25b
`
`15.0
`13.2
`11.2
`10.6
`9.42b
`5.49
`5.05
`4.38
`3.34
`2.91b
`5.51
`4.87
`5.19
`5.03
`5.}5C
`
`4.39
`4.83
`4.72
`4.38
`4.58C
`2.15
`3.04
`3.70
`4.11
`4.57b
`hobs
`:.82d
`1.79d
`1.80d
`1. 80d
`1.80C
`
`1.57e
`1.58'
`l.6oe
`1.69'
`l.6F
`
`A"
`
`0.17
`0.17
`0.20
`0.18
`0 .18C
`0.19
`0.21
`0.20
`0.18
`0. 2()C
`0.28
`0.25
`0.21
`0.22
`0.24C
`
`0.26
`0.23
`0.24
`0.25
`0.24C
`0.60
`0.47
`0.32
`0.32
`0.23b
`
`B"
`
`0.83
`0.83
`0.80
`0.82
`0.82C
`0.81
`0. 79
`0. 79
`0.82
`0 BOC
`0. 72
`0. 75
`0.79
`0. 78
`0.76"
`
`0. 74
`0.77
`0.76
`0.75
`0.76C
`0.40
`0.53
`0.68
`0.68
`0.77b
`
`a See text for explana l ion of these constants and the method used to calculate them. Average of results for two o r three wavckngths. b Ob(cid:173)
`tained from. intercept of plot of k ve rsus buffer concentrati on. c Average of values in prcsc11cc of buffer . d It was not feasibl e to sc parat1..' the
`two phases for analysis. T he rate constants given wen: obtained from the fir.st l 7(¾1 of the reaction. Set: text for furthc: r deta ils. e Rt.'anion
`appears fi rst or der over at least two half-lives.
`
`rate constants. In phosphate buffers of pH 7.6 and 6.5, spectra of
`nonacidified samples were also obtained. At wavelengths greater
`than 240 nm, there was eventual loss of the UV absorption spec(cid:173)
`trum. Absorbance data in this region were used directly for the cal(cid:173)
`culation of rate constants.
`In two additional experiments, samples of 5-azacytidine (l0- 2
`M) were reacted in buffer for a sufficient time to ensure that the
`initial rapid phase of the reaction was completed. Both reactions
`were conducted at 70°, one in 0.33 M phosphate buffer at pH 7.59
`for 10 min and the other in 0.04 M phosphate buffer at pH 6.52 for
`1 hr. Aliquots from each reaction were added to 10 times their vol(cid:173)
`ume of hydrochloric acid at 70° to yield a final reaction condition
`of 0.1 or 0.25 N HCI (Table IV). These reactions were then fol(cid:173)
`lowed at 70° using the differential assay for 5-azacytidine de(cid:173)
`scribed earlier (Eq. 5).
`Evidence for 5,6-Addition to 5-A:zacytidine- The UV spectra
`
`Table IV-Apparent First-Order Rate Constants•
`(hour-') in Hydrochloric Acid at 70° for 5-Azacytidine
`Compared to Those Obtained when 5-Azacytidine was
`Previously Reacted in Phosphate Bufferb
`
`[HCI], N
`
`0 .10
`0.25
`
`Unreacted
`Sample
`
`1.98
`1.05
`
`Phosphate Bu fferb
`
`pH 6.5
`
`1.86
`1.07
`
`pH 7 .6
`
`1.95
`
`of 5-azacytidine and 5-azacytosine were recorded at each of three
`conditions: 0.1 N HCl , 0.001 N HCI , and pH 6.5 phosphate buffer.
`Each solution was then altered to match one of the others so that
`the sequence represents all possihle combi nations of the three con(cid:173)
`ditions. Precautions were taken to eliminate any changes in ab(cid:173)
`sorbance due to passage of time or dilution of sample. In this way,
`UV spectra for solutions of authentic samples were compared to
`those that had been previously prepared and recorded at one or
`more of the other conditions as described under Results.
`Solutions of 5-azacytidine (6 X 10-5 M) and 5-azacytosine ( 1.5 X
`10-6 M) were prepared in 0.025. 0.00/i, and O.!lOI M aqueous bisul (cid:173)
`fite previously adjusted to pH 3.67 ± 0.06. A fourth solution of 5.
`azocytidine in 0.001 M bisulfitc containing 0.024 M NaCl was also
`prepared. Immediately ofter dissolution of the substrate, UV spec·
`tra were recorded as a function of time .
`Spectrophotometric Determination of pKa Values for 5-
`Azacytosine- The pH of o solution containing 10-• M 5-azacyto(cid:173)
`sine in buffer (0.08 M acetic ocid- 0.008 M sodium acetat.e- 0.492 M
`sodium chloride) was measured , and the UV spectrum of the solu (cid:173)
`tion was obtained. The pH was then adjusted using hydrochloric
`acid or sodium hydroxide to obtain a series of spectra ranging from
`the cationic form of 5-azacytosine to the anionic form. To check
`for degradation, a sample at high pH was adjusted to a previously
`obtained lower pH value and the two spectra were compared to
`show that no irreversible changes had occurred.
`
`RESULTS
`
`a Average of val u es o btained ar two or three wavclengchs. b St:~
`Experimental for reaction co ndit ions.
`
`Apparent First-Order Rate Constants in Hydrochloric
`Acid Solutions-Good first-order plots were ohtained when ex-
`
`1150 / Journal of Pharmaceutical Sciences
`
`

`

`!II~ NCP
`
`k::,
`II--+-
`
`i k,,
`NCP
`
`Scheme I
`
`perimental data were graphed according to:
`In X = In X., -
`
`k, ,.t
`
`(Eq.6>
`
`where t is time, X is the concentration of unreacted 5-azacytidine
`or 5-azacytosine, and X o is the initial concentration. When appro(cid:173)
`priate, data were plotted according to:
`
`NCP
`
`NCP
`
`·~ 1
`k., 1
`I~ II ~ I I I~
`NCP
`~rv-~f
`
`Scheme 11
`
`identical conditions. The sum of the two areas provides the total
`area, (area III),. The area under the concentration versus time
`curve for hydrolysis of an equimolar solution of 5-azauracil, (area
`Illl2, was also calculated from Eq. 14. Since [IJJo = [III[o, Eqs. 11
`and 13 may be combined to yield:
`
`In A= In A., -
`
`k .. ,, .. t
`
`(Eq. 7)
`
`(area IlI),/(aree ID)., -
`
`/
`
`(Eq.15)
`
`where A is the absorbance due to unreacted substrate at 255, 260,
`or 265 nm for 5-azacytidine or at 230 or 250 nm for 5-azauracil.
`The apparent first-order rate constants for loss of 5-azacytidine
`and 5-azacytosine (k 0 b,l and 5-azauracil (k35) are listed in Table I.
`TLC of Reaction Solutions of 5-Azacytosine and 5-A:r.aura(cid:173)
`cil in 0.10 N Hydrochloric Acid - The 50 and 100%-reacted sam(cid:173)
`ples of 5-azacytosine showed faint spots at the same R1 value as an
`authentic sample of 5-azauracil. The 100%-reacted sample showed
`no trace of starting material. The 5-azauracil reaction showed a
`disappearance of the spot due to the uracil without development of
`any other compounds absorbing UV light. The R1 values obtained
`were 0.38 for 5-azacytosine and 0.58 for 5-azauraci I.
`Kinetics of 5-Azacylosine and 5-Azauracil Hydrolysis in
`Hydrochloric Acid Solutions- The observed spectral changes
`during the reactions and the results of the TLC experiments point(cid:173)
`ed to the applicability of Scheme I to 5-azacytosine hydrolysis.
`Nonchromophoric products are designated as NCP. A value for ka6
`may he obtained directly from fi-azauracil hydrolysis, while k2:1
`and k 25 were obtained from k.,,,. by:
`
`(Eq_ 8)
`
`and:
`
`r.,, = f k,,,,,
`
`where f is the fraction of 5-azacytosine that deaminates to 5-aza(cid:173)
`uracil. The method used to calculate / is described below. In the
`case of a sequential reaction with parallel first-order loss, as in
`Scheme I, the concentration of Ill as a function of time is given by:
`
`The rate constants obtained are listed in Table I. These were used
`in an analog computer2 program to simulate the time course of the
`various reaction components. A typical plot is shown in Fig. 1,
`which shows good agreement between observed and computer -gen(cid:173)
`erated values.
`Kinetics of 5-Azacytidine Hydrolysis in Hydrochloric Acid
`Solutions-The proposed path for 5-azacytidine hydrolysis in
`0.1-0.8 N HCI is shown in Scheme II. The apparent first-order rate
`constants k 12, k 14 , and k 13 were determined in a manner similar to
`that used for k 23 and k2s. For example:
`k,, = f(k""")
`
`(Eq. 16)
`
`where/ is the fraction of 5-azacytidine that hydrolyzes to 5-azacy(cid:173)
`tosine . This fraction is obtained using Eq. 15, replacing (area III),
`with the area under the concentration curve for 5-azacytosine pro(cid:173)
`duced from 5-azacytidine and replacing (area Jill, with the area
`starting with 5-azacytosine itself.
`Since the values for k25, k23, k3s, and the time course for II are
`all known (Scheme II), they can be employed to define the time
`course for [Ill]n (the 5-azauracil arising from 5-azacytosine). The
`difference between the observed values for III as a function of
`time, [IIIJ, and the expected values, [lll]n, defines the time course
`for III arising from IV uia k 13:
`
`[IllJ,v = [ID) -
`
`[I11) 11
`
`(Eq. 17)
`
`The fraction of I converted to Ill by this route was used to calcu(cid:173)
`late the value for k 1a; this value, in turn, was employed to calculate
`k" from:
`
`k,. = k,.. - k,, - k,.,
`
`(Eq. 18)
`
`where [11)0 is the initial concentration of 5-azacytosine. The area
`under a plot of [IIIJ versu.s time is found by integrating Eq. 10
`from zero to infinity:
`
`(are.a Ill), = /(ll 1Jk, ..
`where f is the same as in Eq. 9. Similarly, for the simple first-order
`loss:
`
`(Eq. 11)
`
`Ill ~ NCP
`Integrating tho equation for III yields:
`
`(Eq. 12)
`
`z
`0
`t(cid:173)
`<(
`a:
`t(cid:173)z
`w u
`z
`0 u
`
`•
`
`(area □ I), =(ni l,/k"
`
`(Eq. 13)
`
`Application of Eqs. I I and I:l to experimental data allows the
`calculation of f. The concentration of 5-azauracil produced from
`5-azacytosine was plotted as a function of time. During formation
`frnm 5-azacytosine, the area under this curve was estimated by the
`method of trapezoids. After the 5-azacytosine was exhausted, the
`remaining area was calculated from :
`
`(area),- oo = [Ill ]tfk ..
`
`(F.,q.14)
`
`5
`
`10
`
`25
`
`30
`
`35
`
`20
`15
`HOURS
`Figure I-Time course for 5-azacytosine (•), 5-azauracil (■),
`and nonchromophoric products ( It.) during the hydrolysis of 5-
`azacytosine in 0.1 N HCI at 70°. The lines were {ienerate_d using
`an analog computer programmed with rate constants list ed in
`Table I.
`
`where Jill], is the concentration of 5-azauracil at time t, and kar. is
`the apparent first -order rate constant for 5-azauracil loss under
`
`2 F.AI, model TR 20.
`
`Vol. 64, No. 7, July 1975 I 1151
`
`

`

`2
`0
`r(cid:173)
`<l'.
`
`a: r(cid:173)z
`UJ u
`2
`0 u
`
`X
`0
`
`4 ,0
`
`3 .0
`
`2 .0
`
`1.0
`
`4
`•
`o!Li.____L_--L:!:::::r::~=~==~==:,
`3
`2
`4
`5
`6
`7
`8
`9
`10 11
`0
`HOURS
`Figure 2-Time course for 5-azacytidine (e), 5-azacytosine (■) ,
`and 5-azauracil ( •J during the hydrolysis of 5-azacy tidine in 0.25
`N HCI at 70°, showing the agreement between the experimental
`points and the computer-generated lines. The scale for 5-azacy(cid:173)
`tosine and 5-azauracil is given on the right ordinate.
`
`These analyses were carried out for the reaction conditions list(cid:173)
`ed in Table I. Typical plots are shown in Fig. 2. The k 0 ti. values ob(cid:173)
`tained at 50° are listed in Table II.
`Kinetics of 5-Azacytidine Hydrolysis in Buffer Solutions(cid:173)
`In buffers at pH 4.6 or higher, the plots of absorbance versus time
`were distinctly biphasic. Acidified samples gave plots similar to
`Fig. 3, showing a rapid initial decrease in absorbance, followed by a
`much slower second phase. When the spectra of buffered samples
`were determined , there was a rapid initial increase in absorbance
`followed by a slower decrease (Fig. 4).
`The biphasic curves were analyzed by feathering the data using
`commonly employed methods3• Treated in this way, each reaction
`produced two linear first -order plots (Figs. 3 and 4). If the data are
`normalized relative to the initial absorbance values, a single equa(cid:173)
`tion, wherein the algebraic sign for the coefficient A is negative in
`buffer and positive in acid, may be used to describe the biphasic
`
`0 . 8
`
`0 .6
`
`0.4
`
`~ 0.2
`U'l
`U'l
`N
`u.i
`U 0.10
`2
`<l'. 0 .08
`OJ
`gj 0 .06
`
`(/)
`OJ
`<l'. 0. 04
`
`0.02
`
`0 .01 ..,_ _ __ _ ,._ _ _ _ _._ _ _ __
`2
`0
`HOURS
`Figure 3- Absorbance of acidified samples at 255 nm as a func (cid:173)
`tion of time ('Y) for 5-azacytidine hydrolysis al 50° , pH 6.52,
`showing lines af slope k" ( e) and kb (..,) obta ined by feathering.
`
`. J __
`
`3
`
`3 ~•or a typical example of this method of data analysis, see Ref. 7.
`
`1152 / Journal of Pharmaceutical Scien ces
`
`0.8
`
`0.6
`
`0.4
`
`0 .2
`
`E
`C:
`0
`<t
`N
`w'
`u
`2
`<l'.
`OJ 0.1
`a:
`0 0 .08
`en
`OJ
`<t 0 .06
`
`0 .04
`
`0 .02
`
`0
`
`2
`HOURS
`Figure 4-Absorbance at 240 nm as a functi on of time ( Y ) for
`nonacidified samples of 5-azacytidine at 50°, pH 6.52, showing
`lines of slope k8 ( e) and kb(..,) obtained by feathering.
`
`3
`
`curves:
`
`(Eq . 191
`
`where F(A) is the fraction of absorbance remaining; A and -k0 are
`the intercept and slope of the rapid phase, respectively; and B and
`- kb are the intercept and slope of the slow phase, respectively.
`Rate constants are independent of the pH of analysis . Their
`values, together with the coefficients in acid, are given in Table III.
`Tobie IV compares first-order ra te constants for hydrolysis of
`&-azacytidine in hydrochloric acid at 70° to those obtained for &(cid:173)
`azacytidine previously reacted in phosphate buffer. Although the
`buffer reactions had proceeded into the second phase, the ob (cid:173)
`served rate constants after transfer to hydrochloric acid were the
`same as those for authentic samples.
`At pH 3.58, the plots of In (A) versus time showed definite cur(cid:173)
`vature but were not distinctly biphasic. In fact, the plots were ade(cid:173)
`quately described by a first-order equation over approximately the
`first half-life. Rate constants, estimated using data representing
`the first 17% of the reaction , are reported in Table III.
`Hydrolysis was apparent first order at pH 2. 58, and the rate con (cid:173)
`stants were obtained using Eq. 7 (Table Ill).
`Effect of Buffer on Hydrolysis Rate- Plots of k0 , k 6 , or k 0 .,.
`
`T 5/
`
`.l: 4
`"' ""3
`
`•
`
`• e -··
`•
`
`•
`
`5
`
`T
`~4
`.s:::
`"' ""3
`
`0
`
`0. 1
`0. 2
`TOTAL SUFFER
`la)
`
`0 0 .1 0. 2 0 .3 0.4 0.5
`TOTAL BUFFER
`(b)
`
`T
`.l: 3
`4~
`"' "" 2
`
`0
`
`0 .6 0 .8
`0 .2 0 .4
`TOTAL BUFFER
`(c )
`
`Figure 5- Plots of k. ver(cid:173)
`sus total buffer concentra(cid:173)
`tion for azacytidine hydrol(cid:173)
`ysis at .50° in ( a) pH 6.52
`phosphate buffer, (b) pH
`5.42 phosphate buffer ( e)
`and pH 5.60 acetat e buffer
`(■) , and ( c) pH 4.62 ace(cid:173)
`tate buffer.
`
`

`

`9
`
`la
`
`lb
`
`le
`
`3
`
`9
`
`2a
`
`2b
`
`2c
`
`3
`
`210
`
`230
`
`250
`
`270
`
`290
`
`210
`
`270
`:250
`230
`WAVELENGTH, nm
`Figure 6-- Pluts of apparent molar absorptiuity versus wauelength for 5-azacytidine in .solutions that were originally acid (Ia), then ad(cid:173)
`justed to neutrality (lb), and reacidified (le), or that were prepared at neutral pH (2a), were acidified (2b), and were ren eutra/ized (2c)
`
`290
`
`210
`
`230
`
`250
`
`270
`
`290
`
`uersus buffer concentration were extrapolated to zero buffer using
`least-squares linear regrnssion analysis to obtain the values for
`these constants in the absence of buffer (Table III). In all cases,
`buffer plots of kb were linear and showed a positive slope. But
`plots for k 0 had positive slopes at pH 6.52 and 7.59, showed no
`buffer catalysis at pH 5.60 and 5.42, and indicated inhibition by
`the buffer at pH 4.62 (Fig. 5). No buffer sensitivity was seen for
`kot,, at pH 3.58 or 2.58.
`Evidence for Addition across 5,6-nouble Rond of 5-Azacyti(cid:173)
`dine-If the UV spectrum of a 5-azacytidine solution is recorded
`as a function of pH , that spectrum is dependent on the order of the
`pH adjustment as well as on the pH itself (Fig. 6). For example, a
`neutral solution (2a in Fig. 6) has a spectrum that is less intense
`than that of a solution of equal concentration and pH that has
`been prepared from a previously acidic solution (1 b in Fig. 6).
`Clearly, acid at least partially converts the 5-azacytidine to a
`form that has a greater absorptivity at neutral pH than the 5-aza(cid:173)
`cytidine itself. However, this rapid conversion does not occur at
`neutral pH since acidification of an originally neutral sample re(cid:173)
`sults in a solution whose spectrum is identical to that of an authen (cid:173)
`tic sample in acid. Furthermore, if the neutral solution that was
`made acidic is readjusted to neutrality, the spectrum resembles
`one from an authentic sample in acid adjusted to neutrality (lb
`and 2c in Fig. 6).
`The degree of difference between a neutralized solution (such as
`lb in Fig. 6) and a neutral solution (2a in Fig. 6) depends upon the
`acidity of the initial solution . In Fig. 6 the initial acid concentra (cid:173)
`tion was 0. I N HCI, and the apparent absorptivity at the absorp(cid:173)
`tion maximum for the solution adjusted to pH 6,5 was 9.03 X J0-1.
`When the initial acidity was 0.001 N HCI, the maximum absorp(cid:173)
`tivity at pH 6.5 was 7.'.l7 X 103. It was consistently ohserved that.
`solutions made neutral and then reacidified showed decreased
`maximum absorptivities compared to the original acidic spectra as
`illustrated for the example in Fig. 6. These changes cannot be at(cid:173)
`trihuted to time or dilution factors . This decrease was not oh(cid:173)
`served when solutions were adjusted from 0. I to 0.001 N HCI and
`back to 0.1 N.
`
`When experiments identical to those illustrated in Fig. 6 were
`carried out using 5-azacytosine, all spectra in 0.1 N HCI were iden(cid:173)
`tical, as were all spectra at pH 6.5.
`Increasing concentrations of bisulfite brought about a progres(cid:173)
`sive loss in intensity of spectra of hoth f\-azacytidine and !i-azacy(cid:173)
`tosine. When spectra changed with time, initial absorbances were
`estimated by extrapolating plots of absorbance uer.su.s lime to zero
`lime. When these estimates were used, 5-azacytidine showed ap(cid:173)
`parent molar absorptivities at 240 nm of 5.78 X JO". '.l.60 X 103,
`and 1.38 X 103 in 0.(JOI, 0.005, and 0.025 M hisulfite, respectively.
`Increasing the ionic strength by addition of sodium chloride did
`not alter the spectrum of the 0.001 M solution. In 0.001, 0.00,5, and
`0.025 M bisulfite, 5-azacytosine had apparent molar ahsorptivities
`of 3.65 X 103, 1.55 X 103, and 0.37 X 103, respectively.
`Spectrophotometric Determination of pKa Values for 5-
`Azacytosine- Absorbances at Arnn for the anionic form were plot(cid:173)
`ted uersus pH. Absorbances were corrected for dilution caused by
`the pH adjustment using:
`
`(Eq. 20)
`
`where Vis the volume after addition of acid or base, and Vo is the
`original volume. Figure 7 shows the plot obtained. The two pKa
`values were estimated from the midpoints or the curves and by use
`of Eqs. 21 and 22 for the pKa's of the conjugate acid and neutral
`molecule , respectively (8):
`pKa- pH + log [(A 1 - Al/(A - A_,.1]
`
`<Eq. 21)
`
`pKa=pH + log[(A -A,l/(A 1 -A>)
`
`(Eq. 22)
`
`The absorbance of the solution when all the s ubstrate is ionized is
`represented by A1; AN represents a solution of the neutral mole (cid:173)
`cule; and A is the observed absorbance at the given pH . Two or
`three pH values within 0.5 pH unit of the pKa's were used. The av(cid:173)
`erage pKa values were 2.64 for the protonated molecule and 8. 10
`for the neutral species.
`
`Vol. 64, No . 7. July 1975 / I 153
`
`

`

`0 .8
`
`0.7
`
`E
`C:
`0
`ln
`N
`w·
`~ 0 .6
`<!
`CD
`~ 0.5
`(/)
`CD
`<!
`
`0 .4
`
`0
`
`2
`
`3
`
`4
`
`5
`pH
`Figure 1-Absorbance at 250 nm versus pH for I X 10-• M 5-
`azacytosine.
`
`6
`
`7
`
`8
`
`9 10
`
`DISCUSSION
`
`Kinetics or 5-Azacytidine Hydrolysis in Hydrochloric Acid
`Solutions- Scheme 11 outlines the proposed path for 5-azacyti (cid:173)
`dine loss in acid. The formation of -~-ozacytosine and 5-azauracil
`was expected from previously reported results (5) and was con(cid:173)
`firmed by UV assay. TLC and spectral changes demonstrated that
`5-azauracil was produced from the hydrolysis of 5-azacytosine.
`This was the expected source for the 5-azauracil in the nucleoside
`hydrolysis.
`However, as may be seen in Fig. 2, the data indicate the immedi (cid:173)
`ate and rapid production of 5-azauracil. If it were produced solely
`from 5-azacytosine, a lag phase would be evident. Therefore, it was
`necessary to postulate the production of 5-ozauracil by another,
`more rapid route. The most logical path involves formation of 5-
`azauridine as a reactive intermediate. It is proposed to have only a
`transitory existence and a low concentration and may be said to be
`in the steady state. This is supported by the absence of a lag time
`for the formation of 5-azauracil from 5-azacytidine. A lag phase
`would be expected if 5-azauridine had an appreciable lifetime.
`Moreover, there was no spectral evidence for the existence of ii(cid:173)
`azauridine in the reaction samples.
`The steady-state assumption requires that 5-azauridine be very
`unstable in the reaction system. This assumption is reasonable in
`view of what is known about the reactivity of the molecule. It is re(cid:173)
`ported to cleave easily in hydrochloric acid at room temperature to
`form 5-azauracil and ribose. In dilute aqueous ammonia, 1/l-D-ri(cid:173)
`bofuranosylbi