Acrolein Mediated by Bifunctional Catalysts 4

[CANCER RESEARCH 42, 830-837, March 19821 0008-5472/82/0042-0000$02.00 Conversionof 4-Hydroperoxycyclophosphamideand4- HydroxycyclophosphamidetoPhosphoramideMustard and AcroleinMediated by BifunctionalCatalysts1

Joseph E. Low, Richard F. Borch, and N. E. Sladek2

Departments of Pharmacology (J. E. L., N. E. 5.1and Chemistry (R. F. B.], University of Minnesota, Minneapolis, Minnesota 55455

ABSTRACT in the urine as well as the pH of the urine may be important with regard to the potential of cyclophosphamide to induce, via The rates at which 4-hydroperoxycyclophosphamide and 4- acrolein, hemorrhagic cystitis. hydroxycyclophosphamide are converted to phosphoramide mustard and acrolein were determined as a function of buffer INTRODUCTION composition, buffer concentration, and pH. Conversion of 4- hydroperoxycyclophosphamide to 4-hydroxycyclophospham Cyclophosphamide is a prodrug widely used as an antitumor ide in 0.5 MTris buffer, pH 7.4, 37Â°,wasfirst-order (k = 0.016 and immunosuppressive agent; its chemistry, metabolism, and min@), but subsequent conversion of 4-hydroxycyclophos pharmacology have been reviewed (4, 21 , 22). A metabolic phamide to phosphoramide mustard and acrolein under these scheme summarizing the current understanding of its metabo conditions was negligible. Phosphoramide mustard and acro lism is presented in Chart 1. Cyclophosphamide is first hydrox lein were readily generated from 4-hydroperoxycyclophos ylated to 4-hydroxycyclophosphamide, also a prodrug, via a phamide or 4-hydroxycyclophosphamide when either of these reaction catalyzed by the microsomal mixed-function oxygen agents was placed in phosphate buffer. Conversion of 4-hy ase system of the liver and, to a lesser extent, lung. 4-Hydrox droxycyclophosphamide to phosphoramide mustard and ac ycyclophosphamide serves as a circulating (transport) form of rolein was first-order with respect to 4-hydroxycyclophospham cyclophosphamide and gives rise to aldophosphamide, be ide (k = 0.1 26 min' in 0.5 M phosphate buffer, pH 8, 37Â°)as lieved to be yet another prodrug (4, 16, 21, 39). Aldophos well as first-order with respect to phosphate serving as a phamide can, in turn, undergo /1elimination to generate phos catalyst. The rate-determining step in the reaction was pH phoramide mustard and acrolein. dependent only insofar as the hydrogen ion concentration Phosphoramide mustard released within cells from 4-hydrox governed the relative amounts of monobasic and dibasic phos ycyclophosphamide is believed to account for the bulk of the phate present. Pseudo-first-order rate constants were 0.045 cytotoxic action of cyclophosphamide to those cells (4, 21, M1min1 for monobasic phosphate and 0.256 M1min' for 49); unequivocalevidencechallengingthis view has yet to be dibasic phosphate. The role of phosphate in this reaction was presented. However, the basis for the relatively favorable ther as that of a bifunctional catalyst. The reaction was not subject apeutic index of cyclophosphamide remains obscure; we and to specific or general, acid or base, catalysis. Other bifunctional others believe that it resides with 4-hydroxycyclophosphamide catalysts such as glucose-6-phosphate and bicarbonate also and/or aldophosphamide (4, 21). It may depend, at least in catalyzed the reaction, albeit less efficiently. Aldophosphamide part, on the relative rates at which 4-hydroxycyclophospham apparently exists only transiently; its presence could not be ide is converted to phosphoramide mustard within sensitive established by 31Pnuclear magnetic resonance spectroscopy. and insensitive cells. Other than its apparent pH dependence We conclude that, in the reaction sequence 4-hydroxycyclo (4), little is known about the chemistryof this reaction or the phosphamide .-* aldophosphamide â€”@phosphoramidemustard factors that might affect the rate at which it proceeds. + acrolein, the conversion of 4-hydroxycyclophosphamide to The first reaction in the sequence of interest is the conversion aldophosphamide is rate limiting and is subject to bifunctional of 4-hydroxycyclophosphamide, a carbinolamide, to aldophos catalysis; this reaction can proceed efficiently only in the phamide, an amidoaldehyde. Carbinolamides and carbinolam presence of a bifunctional catalyst. Assuming that the onco ines are known to be subject to the action of polyfunctional toxic specificity of cyclophosphamide resides with 4-hydroxy catalysts, e.g., inorganic phosphate (6, 13, 14, 25, 33). Re cyclophosphamide and that its cytotoxic effect at therapeutic ports that the rate at which 4-hydroxycyclophosphamide was doses is largely mediated by phosphoramide mustard released converted to phosphoramide mustard and acrolein varied con within cells, these observations offer the possibility that the siderably with the nature of the buffer used (Ref. 43 quoted in intracellular concentration of bifunctional catalysts, whether in Ref. 4; 46), and our own preliminary observation that 4-hy the form of inorganic phosphates, organic phosphates, en droxycyclophosphamide was highly stable in Tris buffer, known zymes, or other species, serve as important determinants with to be incapable of polyfunctional catalysis, strongly suggested regard to the oncotoxic potential and specificity of cyclophos-â€¢that the first step in the reaction sequence was rate limiting phamide. Similarly, the concentration of bifunctional catalysts and that it was subject to polyfunctional, but not to specific or general, acid or base, catalysis. This hypothesis was pursued ,SupportedbyUSPHSGrantsCA26357andCA21737.Adescriptionof in the present investigation. parts of this investigation has appeared in abstract form (38). This is Paper 10 in @ a series on â€˜CyclophosphamideMetabolism.â€• 2 To whom requests for reprints should be addressed, at 3-260 Millard Hall, MATERIALS AND METHODS 435 Delaware Street SE., Minneapolis, Minn. 55455. Received August 13, 1981 ; accepted November 30, 1981. Materials. Triethyl phosphite, acrolein, m-aminophenol, and hydrox

830 CANCERRESEARCHVOL. 42

CICH2CH2 P*4.O.@2 \/ \ tation wavelength was 346 nm; the emission wavelength was 500 nm. N-P.O CH2 Authentic 4-hydroperoxycyclophosphamide was used as the standard.

CIC@42CPl2/\/ O@CH2 The lower limit of sensitivity was approximately 1 nmol of 4-hydrope cva..ot'[email protected]â‚¬ roxycyclophosphamide per ml of water. NADPH Acrolein Assay.The releaseof acroleinfrom 4-hydroperoxycyclo PARENT r@@[email protected]$e phosphamide or 4-hydroxycyclophosphamide was monitored in a Gil ford Model 2400 automatic recording spectrophotometer equipped -- --@@-@-----@ 0 with a Haake constant temperature circulator to maintain a reaction @ CICH2cM2 NH-CPI OcH2cH2 NH-C temperature of 37Â°(2).The reaction mixture routinely contained 70 to @ @@d-P'@ oldehyde @idose \/ \ @ /\/@ 100 nmol of 4-hydroperoxycyclophosphamide or 4-hydroxycyclophos @ OCH@O12 @0@CH2 C1CH2CM2 0-CH@ phamide per ml of buffer; NaCI was added to maintain a constant ionic @ 4-HVDRCX'tt'@ti@0PH0SPHAMI0( 4-KETocYCLOPHOSPHAMIDE strength of 1.6 (32). Buffers studied included phosphate, bicarbonate,

NAD borate, and glucose-6-phosphate; buffer concentrations and pH are TRANSPORT j N@@@?.@.,CICHarH21 NN@ indicated for each experiment. Reaction mixtures were sealed in 1-cm quartz cuvets and monitored at 210 nm. Authentic acrolein was used @ C;CH@CH@ NH@ [email protected]/massspectrometrywasusedto N-P-0 confirm the identity of the acrolein formed. The lower limit of UV CICH2CH2/\?I 0CH@CH@-C-H I sensitivity was approximately 5 nmol of acrolein per ml. Initial rates of 0*,_. ALDOPHOSPHAMIDE acrolein appearance were used to obtain the half-life of 4-hydroxycy clophosphamide; this was necessary because acrolein slowly disap @@iÃ§ peared (k = 0.0003 to 0.002 min â€˜depending on the pH and buffer composition of the reaction medium), presumably because it polymer

OCH@CH2 N2 I CARSOXYPHOSPHAMIDE ized under the reaction conditions. 31pNMR@'Spectrometry.3'P NMR spectra were obtainedwith a CIDCH;)@@Z@oH 7 Varian XL-100 muitinuclear spectrometer operating in the Fourier PHOSPHORAMIDEMUSTARD transform mode with external â€˜9Flockand proton decoupling; pulse acquisition time was 400 msec, and 500 transients were collected @ TOX@ t INACTIVE giving an elapsed acquisition time of 3.3 mm. Hexamethylphosphor CICH2CH2 I amide (1 % in water) was used as a coaxial standard, and both chemical \ shifts and peak intensities were measured from this standard. The RF frequency was 40.5066 MHz with a spectral width of 5000 Hz; the norHN2 reference peak appeared at 361 1 Hz, and the compounds of interest Chart 1. Proposed route of cyclophosphamide metabolism. appeared in the region 2300 to 3200 Hz. Identification was based on signals obtained with authentic compounds. The chemical shifts of ylamine hydrochloridewere purchasedfrom Aldrich ChemicalCo., phosphate and phosphoramide mustard were highly dependent upon Milwaukee, Wis. Acrolein was distilled before use. Glucose-6-phos PH;largervalueswereobtainedasthe pHwasdecreased. phatedisodiumwas purchasedfrom SigmaChemicalCo., St. Louis, Data Analysis. Values were submitted to linear regression analysis Mo.AmericanChemicalSocietyreagent-gradeacetonewaspurchased to obtainall linearfunctions. fromSpectrumChemicalManufacturingCorp.,RedondoBeach,Calif., and was dried and distilled immediatelybeforeuse. Linearhigh-per RESULTS formancethin-layerchromatography(LHP-K)plateswere purchased from Whatman,Inc., Clifton, N. J. 4-Hydroperoxycyclophosphamide Initial experiments established the stability of 4-hydroxycy wassupplied by Dr. A. Takamizawa, Shionogi and Co., Fukushima-ku, clophosphamide in Tris buffer. In these experiments, the de Osaka,Japan. composition of 4-hydroperoxycyclophosphamide at 0Â°in 0.5 4-Hydroxycyclophosphamidewaspreparedin acetonefrom 4-hy M Tris buffer was first examined via 31P NMR spectroscopy. No droperoxycyclophosphamide and an equivalent amount of triethyl change in the 31Pspectrum was observed at pH 6 or 7. At pH phosphite at â€”20Â°as described by Struck (40); the reaction was allowedto proceedfor 5 mmafterwhichthe solutionwasevaporated 7.6, very slow conversion (t112> 8 hr) to 4-hydroxycyclophos to an oilyconsistency.Theresiduewasdissolvedinchloroform,and phamide was observed; increasing the pH to 8 resulted in a separationfrom tnethyl phosphate,excess triethyl phosphite,and small increase in the conversion rate. When the temperature unreacted4-hydroperoxycyclophosphamidewasachievedby placing was raised to 20Â°(pH 8), apparent first-order conversion to 4- the solution onto 2 Sep-Pak silica cartridges (Waters Associates, Inc., hydroxycyclophosphamide was observed with a rate constant Milford, Mass.) arranged in tandem and eluting successively with of 0.0135 min' and a t,,2 of 51 mm (r > 0.99). A similar chloroform,acetone:chloroform(3:1, v/v), and methanol.Fractions experiment at pH 7.4, 37Â°,gavea rate constant of 0.01 6 min@ containing4-hydroxycyclophosphamidewerepooledand takento an and a t,,2 of 43 mm (r > 0.99) (Chart 2). At temperatures of oily conststency. 4-Hydroxycyclophosphamide was dissolved in water 20Â°or below, 4-hydroxycyclophosphamide was the only prod before use; quantificationwas via the fluorescentassay described uct observed. At pH 7.4 and 37Â°,a small peak exhibiting a below.Yieldsweretypicallygreaterthan85%. 4-HydroxycyclophosphamideAssay.A fluorometricassay(1, 44) chemical shift characteristic of phosphoramide mustard ap wasusedto quantifythe amountof 4-hydroxycyclophosphamidegen peared, but even after 200 mm, this peak accounted for less erated. Fifty-1LIsamples of the solution containing 4-hydroxycyclophos than 10% of the 4-hydroxycyclophosphamide present (Chart phamide were placed into tubes containing 200 @lofm-aminophenol 2). (5 mg/mI) andhydroxylaminehydrochloride(6mg/mI) in 3 NHCIand Having established the stability of 4-hydroxycyclophospham 300 @zIofwater. The mixture was then placed into a boiling water bath ide in Tris buffer, we next investigated the stability of 4-hydro for 15 mm, after which it was cooled to room temperature in the dark. peroxycyclophosphamide and 4-hydroxycyclophosphamide in Theconcentrationoftheproduct,7-hydroxyquinoline,wasdetermined

in an Aminco-Bowman Model J4-8950 spectrofluorometer. The exci 3 The abbreviation used is: NMR, nuclear magnetic resonance.

MARCH 1982 831

phosphate buffer (Charts 3 to 6). Appearance of acrolein was monitored by UV spectroscopy for this purpose. The rate at which 4-hydroperoxycyclophosphamide was converted to ac rolein increased with either increasing pH or phosphate con centration (Charts 3 and 4, respectively); mixed kinetics were observed.The rate at which 4-hydroxycyclophosphamide was converted to acrolein also increased as either the pH or the phosphate concentration of the incubation medium increased (Charts 5 and 6, respectively). However, conversion was ap parently first-order with respect to 4-hydroxycyclophospham

5mm MINUTES

Chart 4. Releaseof acrolein from 4-hydroperoxycyclophosphamide at 37Â°as a function of time and phosphate concentration. Reactions were monitored at 210 nm by UV spectroscopy as described in â€˜Materialsand Methods.â€œThe number identifying each curve refers to the phosphate concentration of the medium.

A Chart 2. 3P NMR spectra showing the stability of 4-hydroxycyclophospham ide in Tris buffer at 37Â°.4-Hydroperoxycyclophosphamide (20 mM)was placed into 0.5 M Tris-HCI in water solution, pH 7.4. Spectra were obtained at various times thereafter as described in â€˜Materialsand Methods.â€˜â€˜A,1% hexamethyl phosphoramide in water serving as the coaxial standard: B, 4-hydroxycyclo phosphamide: C. the aziridinium zwitterion of phosphoramide mustard; 0, phos phoramide mustard: E, 4-hydroperoxycyclophosphamide. Selected time points are presented. Spectra are slightly offset for clarity of presentation. I .9 MINUTES Chart 5. Release of acrolein from 4-hydroxycyclophosphamide at 37Â°as a function of time and pH. Reactions were monitored at 210 nm by UVspectroscopy as described in â€œMaterialsandMethods.â€˜â€˜Thenumber identifying each curve refers to the pH of the medium.

ide concentration. The initial â€˜â€˜lagperiod'â€˜observed when 4- hydroperoxycyclophosphamide was used probably reflects the rate at which 4-hydroperoxycyclophosphamide is converted to 4-hydroxycyclophosphamide (46). The half-lives of 4-hydroxycyclophosphamide were calcu lated from the data obtained in these and additional experi ments and are presented in Charts 7 and 8 as a function of the pH and of the phosphate concentration of the incubation me dium. Examination of these charts reveals that the values expected when the phosphate concentration is 70 mt@iandthe pH of the medium is 5, 6, or 7 are similar to those obtained MINUTES under these conditions by Voelcker (Ref. 43 quoted in Ref. 4). Chart 3. Release of acrolein from 4-hydroperoxycyclophosphamide at 37Â°as Also apparent is that there is a log linear relationship between a function of time and pH. Reactions were monitored at 210 nm by UV spectros copy as described in â€˜Materialsand Methods.â€œThenumber identifying each the rate of the reaction and the phosphate concentration (Chart curve refers to the pH of the medium. 8). Such a relationship was not observed between the rate of

832 CANCERRESEARCHVOL. 42

l000

.@ioo I io I

MINUTES Chart6. Releaseofacroleinfrom4-hydroxycyclophosphamideat37Â°asa function of time and phosphate concentration. Reactions were monitored at 210 0@0I 0.1 nm by uv spectroscopy as described in â€œMaterialsandMethods.â€˜â€˜Thenumber [PO4@,M identifying each curve refers to the phosphate concentration of the medium. Chart 8. t /2 of 4-hydroxycyclophosphamide at 37Â°as a function of pH and phosphate concentration. Reactions were monitored at 210 nm by UV spectros copy, and t1/25 of 4-hydroxycyclophosphamide were calculated as described in , â€˜Materials and Methods. â€˜â€˜The number identifying each curve refers to the pH of the medium.

1000

@l00

C E U) 0

I I I I 5 6 7 8 pH Chart 7. t112of 4-hydroxycyclophosphamide at 37Â°as a function of pH and phosphate concentration. Reactions were monitored at 210 nm by UV spectros [@o4],M copy, and t112'sof 4-hydroxycyclophosphamide were calculated as described in Chart 9. Dependence of the rate constant of acrolein release from 4-hydrox â€œMaterialsandMethods.â€•The number identifying each curve refers to the ycyclophosphamide on buffer concentration at constant pH. Values were calcu phosphate concentration of the medium. Iated from the data presented in Chart 8. The number identifying each curve refers to the pH of the medium. the reaction and pH. The change in half-life was relatively small when the hydrogen ion concentration was changed in the pH to phosphate, and supporting the conclusion made previously range 7 to 8; however, it was more substantial in the pH range that the reaction is not subject to specific base catalysis in this 5 to 6 (Chart 7). pH range. In theory, the participation of the phosphate buffer The present observations indicate that the phosphate buffer in the reaction could only be as that of a general base or participates directly in the rate-determining step ofthe reaction. bifunctional catalyst. However, the reaction cannot be subject At a given pH, the phosphate buffer cannot be acting to to base catalysis because Tris buffer does not catalyze it. facilitate any specific base catalysis that may be effected by Therefore, phosphate must be acting as a bifunctional catalyst. an increased hydroxide ion concentration because the latter The slopes of the plots presented in Chart 9 give the pseudo remains constant with a change in the buffer concentration. An first-order rate constants of the buffer-catalyzed reaction at arithmetic plot of the reacton rates as a function of phosphate each pH. The way in which these slopes change as the com - concentration is presented in Chart 9. The relationship between position of the buffer changes reveals whether the monobasic the 2 parameters was linear and passed through the origin for and/or dibasic components of the buffer are the active catalytic each pH, indicating that the reaction is first-order with respect species. A plot of the pseudo-first-order rate constants as a

MARCH 1982 833

function of the fraction of the phosphate buffer that is present as dibasic phosphate (Chart 10) shows that the dibasic com ponent is a better catalyst than is the monobasic component by a factor of approximately 5.7. The intercept on the right ordinate, 0.256 M@min', gives the catalytic constant for dibasic phosphate catalysis. The fact that the catalytic rate does not approach zero as the fraction of dibasic phosphate in the buffer is decreased means that monobasic phosphate is also an active catalyst; the intercept on the left ordinate, 0.045 M1min1, gives the catalytic constant for monobasic phos phate catalysis. The rate law for the reaction under investigation is then

kOb8 kHpo42-[HP042] + kH2p@4-[H2P04]

In the absence of specific base catalysis, it can be concluded that the rate-determining step in the overall reaction is pH MINUTES dependent only insofar as the hydrogen ion concentration Chart 11. Release of acrolein from 4-hydroxycyclophosphamide at 37Â°as a governs the relative amounts of mono- and dibasic components function of time and buffer composition. Reactions were monitored at 210 nm by @ that are present. Uv spectroscopyasdescribedin â€˜MaterialsandMethods.â€˜Bufferconcentra tions were 0. 1 M.The identity of each of the buffers is given with the appropriate Several other agents have been found to serve as bifunc curve. tional catalysts in the breakdown of carbinolamines and carbi nolamides, namely, arsenate, cacodylate, bicarbonate, and carboxylic acids (6, 13, 14, 25). None was as active at pH 7 as was phosphate. This may be related to their relatively unfavor able intrinsic activity and/or pK8's. For example, carboxylic acids possessed the highest intrinsic activity but were pre sumed to be much less effective than phosphate at pH >7 because of the low concentration of the protonated form pres ent at these pH values (13, 14). In our system, glucose-6- phosphate, bicarbonate, and borate were less effective than phosphate in catalyzing the conversion of 4-hydroxycyclo i8O@@ phosphamide to acrolein at pH 7 (Chart 11). In the experiments described thus far, the reactions were followed by monitoring acrolein as the end product. Presum ably, identical results would have been obtained if phosphor amide mustard formation had been measured. 31PNMR exper iments indicated that this was indeed the case (Chart 12). 22J Addition of adequate amounts of phosphate resulted in the disappearance of 4-hydroxycyclophosphamide and the ap pearance of phosphoramide mustard and the corresponding aziridinium zwitterion. These observations are in agreement

A B CD Chart 12. 31PNMR spectra showing the conversion of 4-hydroxycyclophos phamide to phosphoramide mustard at 37Â°.4-Hydroperoxycyclophosphamide (20 mM) was placed into 0.5 M Tris-HCI in water solution, pH 7.4, at t â€”0. Inorganic phosphate (sodium salt) was added at t 106 mm(final concentration, 2 mM), at t = 181 mm (final concentration, 10 mM), and at t â€”226 mm (final concentration, 30 mM): the pH was maintained at approximately 7.4 throughout the experiment. Spectra were obtained as described in â€˜Materialsand Methods. @â€˜ A, 1% hexamethylphosphoramide in water serving as the coaxial standard; B, 4- hydroxycyclophosphamide; C, the aziridinium zwitterion of phosphoramide mus tard; 0, phosphoramide mustard; E, 4-hydroperoxycyclophosphamide. Selected time points are presented.

with the conclusion reached previously that a bifunctional catalyst is required for efficient conversion of 4-hydroxycyclo phosphamide to phosphoramide mustard and acrolein. In all Chart 10. Dependence of the apparent rate constant of buffer-catalyzed 31 p NMR experiments, a peak corresponding to aldophospham acrolein release from 4-hydroxycyclophosphamide on buffer composition. Values ide could not be detected; thus, it can be presumed that were calculated from the data presented in Charts 8 and 9. The Henderson aldophosphamide undergoes elimination to phosphamide mus Hasselbaich relationship was used to calculate the fraction of HP042 ; pK, = 7.21 . â€¢,pH8: A. pH 7; U, pH 6; V. pH 5. tard at a rate faster than the aldehyde is generated.

834 CANCER RESEARCH VOL. 42

DISCUSSION concluded that the rate-limiting step in the overall reaction sequence was the conversion of aldophosphoramide to phos Conversion of 4-hydroxycyclophosphamide to acrolein and phoramide mustard. However, the identity of the material that phosphoramide mustard is believed to proceed via the inter they designated as aldophosphoramide has been challenged mediate, aldophosphamide (Chart 1). The present experiments (41). Moreover, we were unable to detect by 31p NMR spec demonstrate conclusively that the rate-limiting step in the over troscopy the formation of aldophosphoramide. In any case, all reaction is subjectto bifunctional catalysis but not to specific their conclusion, relative to the rate-limiting step, differs from or general, acid or base, catalysis in the pH range studied. ours. Most importantly, the reaction is not â€˜â€˜spontaneous'â€˜;it will not Any molecule that is capable of acting on another molecule proceed in the absence of a suitable bifunctional catalyst. The to essentially synchronously abstract a proton from one site rate-limiting step of the overall reaction sequence is pH de and protonate it at another site serves, by definition, as a pendent only insofar as pH affects the ionization state of the bifunctional catalyst (3, 6, 13, 14, 24, 25, 33, 42). Such an bifunctional catalyst. The reaction is first-order with respect to agent could be as complex as an enzyme or as simple as a 4-hydroxycyclophosphamide and with respect to the catalyst; polybasic inorganic acid. In addition to inorganic phosphate, a thus, it is second-order overall. Aldophosphamide apparently number of molecules, both inorganic and organic, could poten exists only transiently; we were unable to detect its presence tially function in this capacity with regard to the conversion of by 31PNMR spectroscopy. 4-hydroxycyclophosphamide to phosphorarnide mustard and On the basis of these observations, we postulate the scheme acrolein in vivo; some may even be more active in this regard presented in Chart 13 as one that describes the overall reaction than is inorganic phosphate. The implications of these consid sequence. According to this proposal, 4-hydroxycyclophos erations for enzymatic catalysis have not escaped our attention. phamide gives rise to aldophosphamide in the first of 2 steps Some cells are differentially sensitive to the cytotoxic action in the sequence; this step is subject to bifunctional catalysis of cyclophosphoramide even though they are not differentially and is rate limiting. Aldophosphamide then undergoes rapid sensitive to the cytotoxic action of other nitrogen mustards. base-catalyzed /.@elimination to give rise to phosphoramide The basis for this phenomenon has yet to be clearly estab mustard (and acrolein); this reaction is irreversible. Subse lished. However, the biological fate of 4-hydroxycyclophos quently, phosphoramide mustard gives rise to the correspond phamide would seem to be important in this regard. Thus, it ing aziridinium zwitterion, the ultimate alkylating species. may be that sensitive cells more readily convert 4-hydroxycy Voelcker and coworkers (Ref. 43 quoted in Ref. 4; 45, 46) clophosphamide to phosphoramide mustard because of a rel reported that conversion of 4-hydroxycyclophosphamide to atively greater bifunctional catalytic (enzymatic or chemical) acrolein in 70 mM phosphate buffer exhibited first-order reac activity. Alternatively, insensitive cells, due to a higher content tion kinetics and that the rate increased with increasing pH. of NAD-linked aldehyde dehydrogenase, aldehyde oxidase, or They observed that the reaction proceeded at a slower rate in alcohol dehydrogenase activities, may be more capable of buffered Krebs-Ringer solution. The slower rate was attributed inactivating the drug by converting it to carboxyphosphamide, to the presence of magnesium and calcium ions in buffered 4-ketocyclophosphamide, or alcophosphamide (8, 10â€”12,15, Krebs-Ringer solution which they suggested stabilized 4-hy 17, 18, 37). Selective uptake and sulfhydryl content may also droxycyclophosphamide. An alternative explanation may be be important determinants (4, 19, 20, 23). that the phosphate concentrations differed; unfortunately, the Intracellular phosphate levels apparently vary according to phosphate concentration of the Krebs-Ringer solution was not cell type. Values of about 10 and 20 mr@ihavebeen reported specified, but typically, it is less than 16 m@. These investiga for hepatocytes and Ehrlich ascites cells, respectively (7); tors also reported half-lives of 20 and 80 mm in 70 mp,i presumably, most of the phosphate is present in the organic phosphate buffer, pH 7.4, (37Â°)for materials designated as 4- state (36). Comparative studies relative to intracellular phos hydroxycyclophosphamide and aldophosphoramide, respec phate concentrations in cells sensitive and insensitive to cyclo tively (46). On the basis of these values, they quite logically phosphamide, but less differentially sensitive to other nitrogen mustards, have not been made. Many tumor cells exhibit greater glycolysis relative to that observed in normal cells (31, 48); it is conceivable that intracellular levels of inorganic and organic phosphates are elevated in these cells (27). Relative to the ionization state of phosphate, pH would be an important @@@ __ H2N0 Ii â€”N(CH2CH2cl)@ @. _ _â€˜-Pâ€”N(CH2CH2CI)2 determinant. Differences in intracellular pH between normal I ? (6 and tumor cells might also be expected if the metabolic activity H@[email protected]@ of normal and tumor cells differs; differences in pH have been @gH reported (28â€”31, 34, 47). However, our findings indicate that changes in intracellular pH compatible with survival are not H@CH= CH2 likely to be of sufficient magnitude to substantially affect the rate at which 4-hydroxycyclophosphamide is converted to @ci@cii2o phosphoramide mustard. Plasma pH (7.39 to 7.46) and plasma inorganic (0.8 to 1.3 @ ______4 HaN-'Pâ€”N:'@-.@CH24I,,.CH2cH2CI 00 6e @â€œciitâ€• mM) and organic (2.0 to 3.5 mM) phosphate concentrations apparently remain relatively constant in the healthy human Chart 13. Proposed mechanism by which bifunctional catalysts such as in organic phosphate facilitate the conversion of 4-hydroxycyclophosphamide to male (35). Conversion of 4-hydroxycyclophosphamide to ac phosphoramide mustard and acrolein. rolein and phosphoramide mustard at these levels proceeded

MARCH 1982 835

very slowly in our experiments. Urine pH (4.8 to 7.5) and urine 15. Domeyer, B. E., and Sladek, N. E. Potential for in vivo inactivation of a cyclophosphamide metabolite (aldophosphamide) by tumor enzymes. Fed. inorganic (15.5 to 35.2 mM) and organic (2.0 to 35 mM) Proc.,33:581, 1974. phosphate concentrations in the healthy human male are more 16. Domeyer, B. E., and Sladek, N. E. Kinetics of cyclophosphamide biotrans variable (35). The urotoxic effects of cyclophosphamide are formation in vivo. Cancer Res., 40: 174â€”180,1980. 17. Domeyer, B. E.. and Sladek, N. E. Metabolism of 4-hydroxycyclophospham largely effected by the acrolein generated in the urine from 4- ide/aldophosphamide in vitro. Biochem. Pharmacol., 29: 2903â€”2912, hydroxycyclophosphamide (5, 9). It follows that these effects 1980. are more likely to be expressed in patients who put out an 18. Domeyer, B. E., and Sladek, N. E. Inhibition by cyanamide of 4-hydroxycy clophosphamide/aldophosphamide oxidation to carboxyphosphamide. Bio alkaline urine containing relatively high concentrations of phos chem. Pharmacol., 30: 2065-2073, 1981. phate. In support of this contention, we have noted that coad 19. Draeger, U., and Hohorst, H.-J. Permeation of cyciophosphamide (NSC 26271 ) metabolites into tumor cells. Cancer Treat. Rep., 60: 423â€”427, ministration of diuretics, such as furosemide, which have little 1976. effect on phosphate and bicarbonate excretion, prevented 20. Draeger, U., Peter, G., and Hohorst, H.-J. Deactivation of cyclophosphamide cyclophosphamide-induced bladder edema, whereas coadmin (NSC-26271) metabolites by sulfhydryl compounds. Cancer Treat. Rep., 60: 355â€”359,1976. istration of acetazolamide, a diuretic that is phosphaturic, mark 21. Friedman, 0. M., Myles, A., and Colvin, M. Cyclophosphamide and related edly increases bicarbonate excretion, and alkalinizes the urine, phosphoramide mustards. Current status and future prospects. In: A. Ro did not (26). Moreover, toxicity effected by 4-hydroperoxycy sowsky (ed), Advances in Cancer Chemotherapy, Vol. 1, pp. 143â€”204. New York: Marcel Dekker, Inc., 1979. clophosphamide placed into rat bladders increased as the pH 22. Hill, D. L. A Review of Cyclophosphamide. Springfield, III.: Charles C and phosphate concentration of the vehicle used to introduce Thomas,Publisher,1975. the drug increased.4 23. Hohorst, H.-J.. Draeger, U., Peter, G., and Voelcker, G. The problem of oncostatic specificity of cyclophosphamide (NSC-26271 ): studies on reac Whether intracellularly or in the urine, the environmental tions that control the alkylating and cytotoxic activity. Cancer Treat. Rep., composition will govern the rate at which 4-hydroxycyclophos 60: 309-315, 1976. 24. Jencks, W. P. Mechanism and catalysis of simple carbonyl group reactions. phamide is converted to phosphoramide mustard and acrolein Prog. Phys. Org. Chem., 2: 63-128, 1964. and would seem to be an important determinant with respect 25. Lee, Y.-N., and Schmir, G. L. Concurrent general acid and general base to the ultimate action of cyclophosphamide. Should differences catalysis in the hydrolysis of an imidate ester. 2. Blfunctional catalysis. J. Am. Chem. Soc.,101: 3026-3035, 1979. in intracellular bifunctional catalytic activity prove to be signifi 26. Low, J. E.. Coveney, J. R., Grage, G. A., and Sladek, N. E. Influence of cant with respect to the relatively favorable therapeutic index urine pH on cyclophosphamide-induced cystitis. Pharmacologist, 22: 177, of cyclophosphamide, a novel basis for selective toxicity would 1980. 27. Marx, J. L. Tumor viruses and the kinase connection. Science (Wash. D. be newly recognized. Such recognition could, in turn, serve as C.),211: 1336â€”1338,1981. the basis for the synthesis of new classes of antineoplastic 28. Poole. D. T. Intracellular pH of the Ehrlich ascites tumor cell as it is affected agents that fully exploit the hypothesized biological differences. by sugar and sugar derivatives. J. Biol. Chem., 242: 3731 â€”3736,1967. 29. Poole, D. T., and Butler, T. C. Effects of inhibitors of glycolysis on intracel lular pH and on accumulation of glycolytic intermediates in the Ehrlich REFERENCES ascites tumor cell. J. NatI. Cancer Inst., 42: 1027â€”1033, 1969. 30. Poole, D. T., Butler, T. C., and Waddell, W. J. Intracellular pH of the Ehrlich 1. Alarcon, R. A. Fluorometric determination of acrolein and related compounds ascites tumor cell. J. NatI. Cancer Inst.. 32: 939â€”946,1964. with m-aminophenol. Anal. Chem., 40: 1704-1708, 1968. 31 . Racker, E. Bloenergetics and the problem of tumor growth. Am. Sci., 60: 2. Alarcon, R. A. Acrolein. IV. Evidence for the formation of the cytotoxic 56â€”63,1972. aldehyde acrolein from enzymatically oxidized spermine or spermidine. 32. Remington's Pharmaceutical Science, Ed. 15, p. 253. Easton, Pa.: Mack Arch. Biochem. Biophys., 137: 365-372, 1970. Publishing Co., 1975. 3. Blackburn, G. M., and Jencks, W. P. The mechanism of the aminolysis of 33. Sayer, J. M., and Conlon, P. The timing of the proton-transfer process in methyl formate. J. Am. Chem. Soc., 90: 2638â€”2645,1968. carbonyl additions and related reactions. General-acid-catalyzed hydrolysis 4. Brock, N., and Hohorst, H.-J. The problem of specificity and selectivity of of imines and N-acylimines of benzophenone. J. Am. Chem. Soc., 102: alkylating cytostatics: studies on N-2-chloroethylamido-oxazaphosphorines, 3592-3600, 1980. z. Krebsforsch.,88:185-215,1977. 34. Schloerb, P. R., Blackburn, G. L., Grantham, J. J., Mallard, D. S., and Cage, 5. Brock, N., Stekar, J., PohI, J., Niemeyer, U., and Scheffler, G. Acrolein, the G. K. Intracellular pH and buffering capacity of the Walker 256 carcinoma. causative factor of urotoxic side-effects of cyclophosphamide, ifosfamide, Surgery (St. Louis), 58: 5-1 1. 1965. trofosfamide, and sufosfamide. Arzneim. Forsch., 29: 659â€”661,1979. 35, Scientific Tables, Ed. 7. Ardsley, N. V.: Geigy Pharmaceuticals, 1975. 6. Chaturvedi, R. K., and Schmir, G. L. The hydrolysis of N-substituted acetim 36. Simonsen, L. 0., and Cornelius, F. Inorganic phosphate in Ehrlich ascites idate esters. J. Am. Chem. Soc., 90: 441 3-4420, 1968. tumor cells and its distribution across the cell membrane. Biochim. Biophys. 7. Coe, E. L., and Lee, I.-Y. Phosphorylation of 2-D-deoxyglucose and associ Acta,511: 213â€”223,1978. ated inorganic phosphate uptake in ascites tumor cells. Biochemistry, 8: 37, Sladek, N. E. Bioassay and relative cytotoxic potency of cyclophosphamide 685-693, 1969. metabolites generated in vitro and in viva. Cancer Res., 33: 1150â€”1156. 8. Connors, T. A., Cox, P. J., Farmer, P. B., Foster, A. B., and Jarman, M. 1973. Some studies of the active intermediates formed in the microsomal metab 38. Sladek, N. E., Borch, R. F. , and Low, J. E. Phosphate catalyzed conversion olism of cyclophosphamide and isophosphamide. Biochem. Pharmacol., 23: of 4-hydroperoxycyclophosphamide and 4-hydroxycyclophosphamide to 115â€”129,1974. phosphoramide mustard and acrolein. Proc. Am. Assoc. Cancer Res., 22: 9. Cox, P. J. Cyclophosphamide cystitisâ€”identification of acrolein as the 214, 1981. causative agent. Biochem. Pharmacol., 28: 2045â€”2049,1979. 39. Sladek, N. E., and Powers, J. F. Cytotoxic activity and 4-hydroxycyclophos 10. Cox, P. J., Farmer, P. B., and Jarrnan. M. The microsomal metabolism of phamide and phosphoramide mustard concentrations in the plasma of some analogues of cyclophosphamide: 4-methylcyclophosphamide and 6- cyclophosphamide-treated rats. Pharmacologist, 22: 175, 1980. methylcyclophosphamide. Bichem. Pharmacol., 24: 599â€”606,1975. 40. StÃ±ick.R. F. Isolation and identification of a stabilized derivative of aldo 11. Cox, P. J., Phillips, B. J., and Thomas, P. The enzymatic basis of the phosphamide, a major metabolite of cyclophosphamide. Cancer Res., 34: selective action of cyclophosphamide. Cancer Res., 35: 3755â€”3761.1975. 2933-2935. 1974. 12. Cox, P. J., Phillips, B. J., and Thomas, P. Studies on the selective action of 41 . Struck, R. F. Aldophosphamide: synthesis, characterization, and comparison cyclophosphamide (NSC-2627 1): inactivation of the hydroxylated metabolite with â€œHohorst'saldophosphamide.'â€˜CancerTreat. Rep., 60: 317â€”319, by tissue-soluble enzymes. Cancer Treat. Rep., 60: 321 â€”326,1976. 1976. 13. Cunningham, B. A., and Schmir, G. L. Iminolactones. II. Catalytic effects on 42. Swain, C. G., and Brown, J. F., Jr. Concerted displacement reactions. VIII. the nature of the products of hydrolysis. J. Am. Chem. Soc., 88: 551â€”558. Polyfunctional catalysis. J. Am. Chem. Soc.. 74: 2538-2543. 1952. 1966. 43. Voelcker, G. Untersuchungen Uber Reaktionen von 4-Hydroxycyclophos 14. Cunningham, B. A., and Schmir, G. L. Hydroxyl group participation in amide phamid, einem Metaboliten des Cyclophosphamid und die Identifizierung hydrolysis. The influence of catalysts on the partitioning of a tetrahedral von Cyclophosphamid-Metaboliten in vivo. Inaugural-Dies., Fachbereich intermediate. J. Am. Chem. Soc., 89: 917-922, 1967. Biochemie und Pharmazie, Frankfurt: 1976. 44. Voelcker, G., Haelsperger, R., and Hohorst, H.-J. Fluorometrische Beitimung 4 P. C. Smith and N. E. Sladek, unpublished observations. von â€˜â€˜Activiertem'â€˜Cyclophosphamidund Ifosfamid rn Blut. J. Cancer Res.

836 CANCERRESEARCHVOL. 42

clln. Oncol., 93: 233-240, 1979. 1172â€”1176,1974. 45. Voelcker,G.,Wagner,T.,andHohorst,H.-J.Identificationandpharmaco 47. Waddell, W. J., and Bates, R. G. Intracellular pH. Physiol. Rev., 49: 285â€” kinetics of cyclophosphamide (NSC-26271 ) metabolites in vivo. Cancer 329, 1969. Treat.Rep.,60: 415â€”422,1976. 48. Warburg, 0. The Metabolism of Tumors. London: Constable, 1930. 46. VÃ¶lker,G., DrÃ¤ger,U., Peter, G. , and Hohorst, H.-J. Studien zum Spontan 49. Wrabetz, E., Peter, G., and Hohorst, H.-J. Does acrolein contribute to the zerfall von 4-Hydroxycyclophosphamid und 4-Hydroperoxycyclophos cytotoxicity of cyclophosphamide? J. Cancer Res. Clin. Oncol., 98: 1 19â€” phamid mit Hilfe der DUnnschichtchromatographie. Arzneim. Forsch., 24: 126, 1980.

MARCH1982 837

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1982 American Association for Cancer Research. Conversion of 4-Hydroperoxycyclophosphamide and 4-Hydroxycyclophosphamide to Phosphoramide Mustard and Acrolein Mediated by Bifunctional Catalysts

Joseph E. Low, Richard F. Borch and N. E. Sladek

Cancer Res 1982;42:830-837.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/42/3/830

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/42/3/830. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.