382

Chem. Res. Toxicol. 1992,5, 382-385

inducer-receptor complex and evidence for ita nuclear translocation. J . Bid. Chem. 254, 11636-11648. (18) Gasiewicz, T. A., and Neal, R. A. (1982) The examination and quantitation of tissue cytosolic receptors for 2,3,7,8-tetrachlorodibenzo-p-dioxin using hydroxylapatite. Anal. Biochem. 124, 1-11.

(19) Zacharewski, T., Harris, M., and Safe, S. (1989) Induction of cytochrome P-450-dependent monooxygenase activities in rat hepatoma H-4-11 E cells in culture by 2,3,7,8-tetrachlorodibenzop-dioxin and related compounds: Mechanistic studies using radiolabeled congeners. Arch. Biochem. Biophys. 272, 344-355. (20) Scatchard, G. (1949) The attraction of proteins for small molecules and ions. Ann. N.Y. Acad. Sci. 51, 660-672. (21) Beck, J. S., and Goren, H. J. (1983) Simulation of association curves and Scatchard plots of binding where ligand and receptor are degraded or internalized. J . Recept. Res. 3, 561-577. (22) Bunce, N. J., Landers, J. P., and Safe, S. H. (1988) Kinetic with models for association of 2,3,7,8-tetrachlorodibenzo-p-dioxin the Ah receptor. Arch. Biochem. Biophys. 267, 384-397. (23) Kester, J. E., and Gasiewicz, T. A. (1987) Characterization of the in uitro stability of the rat hepatic receptor for 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). Arch. Biochem. Biophys. 252, 606-625. (24) Bradfield, C. A,, Kende, A. S., and Poland, A. (1988) Kinetic and equilibrium studies of Ah receptor-ligand binding: Use of [1251]-2-iodo-7,8-dibromodibenzo-p-dioxin. Mol. Pharmacol. 34, 229-237. (25) Farrell, K., Safe, L., and Safe, S. (1987) Synthesis and Ah receptor binding properties of radiolabeled polychlorinated di-

benzofuran congeners. Arch. Biochem. Biophys. 259, 185-195. (26) Harris, M., Zacharewski, T., Piskorska-Pliszczynska, J., Rosengren, R., and Safe, S. (1990) Structure-dependent induction of aryl hydrocarbon hydroxylase activity in C57BL/6 mice by 2,3,7,8-tetrachlorodibenzo-p-dioxinand related congeners: Mechanistic studies. Toxicol. Appl. Pharmacol. 105, 243-253. (27) Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) Protein measurements with the Folin phenol reagent. J . Biol. Chem. 193, 265-275. (28) Astroff, B., Zacharewski, T., Safe, S., Arlotto, M. P., Parkinson, A., Thomas, P., and Levin, W. (1988) 6-Methyl-l,3,8-trichlorodibenzofuran as a 2,3,7,8-tetrachlorodibenzo-p-dioxin antagonist: Inhibition of the induction of rat cytochrome P-450 isozymes and related monooxygenase activities. Mol. Pharmacol. 33, 231-236. (29) Bunce, N. J., Forber, C. L., Hutson, J. M., and McInnes, C. (1988) Single step methods for calculating activation parameters from raw kinetic data. J. Chem. Soc., Perkin Trans. 2, 363-368. (30) Mason, G., Sawyer, T., Keys, B., Bandiera, S., Romkes, M., Piskorska-Pliszczynska, J., Zmudzka, B., and Safe, S. (1985) Polychlorinated dibenzofurans (PCDFs): Correlation between in vivo and in vitro structure-activity relationships. Toxicology 37, 1-12.

(31) Mason, G., Farrell, K., Keys, B., Piskorska-Pliszczynska, J., Safe, L., and Safe, S. (1986) Polychlorinated dibenzo-p-dioxins: Quantitative in uitro and in uiuo structure-activity relationships. Toxicology 41, 21-31. (32) Bunce, N. J., Landers, J. P., Nakai, J. S., Winhall, M. J., and Safe, S. (1990) In uitro thermal inactivation of the Ah receptor from several species. Toxicol. in Vitro 4, 87-92.

Identification of Cyclophosphamide-DNA Adducts in Rat Embryos Exposed in Vitro to 4-Hydroperoxycyclophosphamide Philip E. Mirkes,*pt Nigel A. Brown,$ Mahmud Kajbaf,§ John H. Lamb,§ Peter B. Farmer,§ and Stephen Naylor**§J Division of Embryology, Teratology and Congenital Defects, Department of Pediatrics, University of Washington School of Medicine, Seattle, Washington 98195, MRC Experimental Embryology and Teratology Unit, St. George's Hospital Medical School, University of London, London, S W l 70RE, U.K., and MRC Toxicology Unit, Carshalton, Surrey, SM5 4EF, U.K. Received November 12, 1991 Cyclophosphamide and other bifunctional alkylating agents are potent animal teratogens inducing a variety of malformations. Although cyclophosphamide-induced DNA damage is implicated as a primary mechanism underlying the teratogenesis initiated by cyclophosphamide, additional insights into the complex nature of the teratogenic process have been hampered by the inability to analyze the primary teratogenic lesions, Le., cyclophosphamide-DNA adducts. Using tandem mass spectrometry, we show that the monofunctional adduct N-(2-chloroethyl)-N-[ 2- (7-guaninyl)ethyl]amine (NOR-G) and bifunctional adduct N,N-bis[ 2 4 7guaniny1)ethyllamine (G-NOR-G) can be detected in the DNA of organogenesis-stage rat embryos after an in vitro exposure to an embryotoxic concentration of activated cyclophosphamide, i.e., 4- hydroperoxycyclophosphamide.

Introduction Cyclophosphamide (Figure 1)is a bifunctional alkylating agent that is widely used as an anticancer drug. In addition, this drug is a teratogen causing birth defects in animals and humans (1). In order to achieve its antineoplastic and teratogenic effects, cyclophosphamide must be metabolically activated (2). This activation involves the formation of 4-hydroxycyclophosphamideand its equili-

* Authors to whom correspondence should be addressed. 'University of Washington School of Medicine. * University of London. 8 MRC Toxicology Unit. 11 Present address: Department of Biochemistry and Molecular Biology, Mayo Foundation, Rochester, MN 55905.

bration with the acyclic tautomer, aldophosphamide. The latter spontaneously breaks down to form phosphoramide mustard (Figure l),which is believed to be the principal metabolite responsible for the antineoplastic and teratogenic properties of activated cyclophosphamide (2). Although the mechanisms underlying the cytotoxic and teratogenic properties of cyclophosphamide are not completely understood, it is generally believed that cyclophosphamide-induced toxicity is related to the induction bf DNA-DNA interstrand cross-links by phosphoramide mustard (3). The initial reaction leading to the formation of DNA-DNA cross-links is alkylation at the N-7 position of guanine. This initial phosphoramide-DNA adduct is relatively unstable and converted to the CorresPonding nornitrogen mustard adducts, one of which is the G-

0893-228x/92/2705-0382$03.00/00 1992 American Chemical Society

Chem. Res. Toxicol., Vol. 5, No. 3, 1992 383

Identification of Cyclophosphamide-DNA Adducts H

H

cyclophosphamide

phosphoramide mustard

NOR-G-OH

NOR-G H CH,CH,NCH&H,

H N A N I H2N

~

N/>

A

0 A

4~ I

~

A NHZ

H

I

30

I

I

130

230

Time (minutes)

G-NOR-G

Figure 2. Chromatogram of 4hydroperoxycyclophosphamide/rat

Figure 1. Structures of cyclophosphamide, phosphoramide

embryo DNA reaction products after acid hydrolysis. Products were run on a Sephadex G-10 column, eluted at a flow rate of 2.5 mL/min, and monitored at 260 nm. A = adenine; G = guanine. Fractions designated 1-3 were subsequently analyzed by tandem mass spectroscopy.

mustard, and the guanine adducts NOR-G, NOR-G-OH, and G-NOR-G.

NOR-G1 cross-link (4). Tentative evidence, based upon alkaline elution studies, suggests that DNA cross-links are responsible for the embryotoxic effects of cyclophosphamide (5). Although alkaline elution studies have implicated DNA cross-links in cyclophosphamide-induced embryotoxicity, no direct evidence is available to confirm the presence of DNA cross-links in embryos exposed to embryotoxic levels of cyclophosphamide. Recently we reported that not only the G-NOR-G adduct but also the monoadducts NOR-G and NOR-G-OH could be detected in calf thymus DNA alkylated in vitro with phosphoramide mustard using tandem mass spectrometric parent ion scanning (6). The purpose of the present studies was to determine the types of adducts formed in rat embryos exposed to embryotoxic levels of 4-hydroperoxycyclophosphamide,a preactivated analogue of cyclophosphamide, using this same methodology. In tandem mass spectrometry (7) two mass spectrometers are employed in sequence. The first mass spectrometer is used to ionize and separate compounds on a mass to charge basis. It is then possible to select an ion of interest (usually the molecular ion) and introduce that into a collision cell containing an inert gas such as argon, xenon, or helium. The subsequent collision(s) that the molecular ion undergoes (undergo)with the inert gas induces (induce) collisionally activated dissociation (CAD)l resulting in fragmentation. These fragments, or daughter ions, are mass analyzed in the second spectrometer. In the present study we demonstrate the utility of tandem mass spectrometry to identify the initial teratogen-DNA adducts formed when early postimplantation rat embryos are exposed to teratogenic concentrations of cyclophosphamide.

Materials and Methods Drugs and Chemicals. 2'-Deoxyguanosine and calf thymus DNA were obtained from Sigma Chemical Co., Ltd. (Poole, U.K.). Bis(2-chloroethy1)amine hydrochloride, glycerol, p-toluenesulfonic acid, 2,2,2-trifluoroethanol, and triethylamine (Gold label) were purchased from Aldrich Chemical Co., Ltd. (Gillingham, U.K.). 4-Hydroperoxycyclophosphamidewas a gift from Dr. Michael Colvin, Johns Hopkins University, Baltimore, MD. Abbreviations: G-NOR-G, N,N-bis[2-(7-guaninyl)ethyl]amine; NOR-G, N-(2-chloroethyl)-N-[2-(7-guaninyl)ethyl]amine; NOR-G-OH,

N-(2-hydroxyethyl)-N-[2-(7-guaninyl)ethyl]amine; CAD, collisionally activated dissociation; FAB, fast atom bombardmenc SDS, sodium dodecyl sulfate; MCA, multichannel analysis.

Synthesis of Adducts. NOR-G, G-NOR-G, and NOR-G-OH were prepared by a modification of the procedure of Hemminki (4).

In Vitro Embryo Culture. Day 11rat conceptuses (embryo

+ yolk sac and amnion) were removed from the uterus, explanted,

and cultured in vitro as described previously (8). Conceptuses were cultured for 1h in medium consisting of 50% immediately centrifuged rat serum and 50% Waymouth's and then exposed to 20 pM 4-hydroperoxycyclophosphamide for 5 h. At the end of this exposure, conceptuses were removed from culture and washed through three changes of Hanks balanced salt solution. The yolk sac and amnion were then removed, and embryos from approximately 80 day 11 conceptuses were frozen a t -70 "C. Isolation of DNA-DNA Adducts. Embryos were homogenized in digestion buffer containing 100 mM NaC1, 25 mM EDTA, 0.5% SDS, 0.1 mg/mL proteinase K, and 10 mM %is-HC1, pH 7.5. This homogenate was then incubated at 50 "C for 18 h and extracted with an equal volume of phenol/chloroform/iyl alcohol (25241), and the DNA was precipitated from the aqueous phase with ethanol. Purified DNA was then hydrolyzed by the addition of 1 M HCl and incubation at 70 "C for 60 min. The solution containing hydrolyzed DNA was then centrifuged at 4000 rpm for 10 min and the supernatant decanted. This supernatant was then neutralized with NaOH, applied to a 30- x 400-mm Sephadex G-10 column, and eluted with 100 mM ammonium formate in double-distilled water at a flow rate of approximately 2.5 mL/min. Three different fractions (see Figure 2) were collected and lyophilized prior to analysis by mass spectrometry as described previously (6). Control Samples. Calf thymus DNA (40 mg) was vortexmixed with double-distilledwater and left at 4 "C overnight. The DNA was then subjected to acid hydrolysis and applied to a Sephadex G-10 column prior to being analyzed by mass spectrometry. Mass Spectrometry. All mass spectra were obtained on a VG-70-SEQ (VG Analytical Ltd., Manchester, U.K.) of EBQIQ, configuration, where E is an electrostatic analyzer, B is the magnet, Q1 is an rf-only quadrupole collision cell, and Qz is a mass filter quadrupole. EB and Qz correspond to mass spectrometer one (MSJ and two (MS,), respectively. The samples were ionized by positive ion fast atom bombardment mass spectrometry (FAB-MS) (9, 10). Xenon atoms, accelerated to 8.5 keV, from a saddle-field fast atom gum (B11N from Ion Tech, Teddington, U.K.) were used to ionize the sample. The secondary ions produced by the impact of the fast xenon atoms were accelerated out of the source region to an energy of 8 keV. (a) Daughter Ion Spectra. Molecular ions (also known as parent ions) were selected with a resolution of -1000 using EB (MS,) and subjected to CAD in Qz, using argon as the collision gas at a pressure of 1.5 X lo-' mbar with -20-eV collision energy

384 Chem. Res. Toxicol., Vol. 5, No. 3, 1992 (i.e., at the first transmission node of the parent ion). The resulting fragment (daughter) ions were mass analyzed in Qz, and the daughter ion spectrum was acquired by scanning Qz over the mass range 40-450 daltons, with 15 scans being collected in the MCA mode. (b) Parent Ion Spectra. In this scan node, MSI is scanned over an appropriate mass range, allowing each set of ions of a specific mass into Q1. However, MS2 is adjusted to only transmit selected ion(s) derived from the parent ions originally subjected to the CAD process in the collision cell Q1. Detection of the specific ion (in this case, either m / z 152 or 221) results in determination by computer control of which ion entered into Q1 that ultimately gave rise to the daughter ion. Parent ion spectra were acquired by scanning the magnet over the mass range 450-100 daltons a t 20 s/decade and subjecting ions to CAD a t 5 X mbar argon with a collision energy of -20 eV. Qzwas selected to alternatively transmit only ions at m / z 152 and 221. Thirty scans were obtained in the MCA mode. (c) Sample Preparation. The lyophilized fractions from the Sephadex G-10 column were ultimately dissolved in 5 rL of 1 M HC1, and 2 pL of this solution was added to 1.5 pL of 0.2 M p-toluenesulfonic acid in glycerol (used as the FAB matrix) and thoroughly mixed on the stainless steel probe tip. The sample was subsequently inserted into the mass spectrometry source and analyzed.

Results and Discussion In previous work we described conditions that allowed us to detect both monofunctional (NOR-G and NOR-GOH) and cross-linked adducts (G-NOR-G) of phosphoramide derivatives with calf thymus DNA that had been alkylated in vitro (6). It is not possible to detect directly phosphoramide mustard adducts using this method because of the instability of such adducts (+ = 2-3 h) (11). By using a suitable FAB matrix (0.2 M p-toluenesulfonic acid in glycerol) in conjunction with parent ion scanning mass spectrometry of ions at m / z 152 and 221, it was possible to detect down to approximately 40 pmol of GNOR-G on the FAB probe tip. In the present study, DNA was isolated from day 11rat embryos that had previously been exposed in vitro to 4hydroperoxycyclophosphamide, a preactivated form of cyclophosphamide. Hydrolyzed DNA was fractionated on a Sephadex G-10 column, and fractions 1-3 (Figure 2) were then subjected to concomitant parent ion scanning of ions at m/z 152 and 221. Nothing of interest was detected in the first two fractions; however, previous results with calf thymus DNA indicated that cyclophosphamide-DNA adducts eluted in fraction 3 (6). Samples from fraction 3 exhibited parent ions at m / z 2571259 (corresponding to NOR-G) and m/z 372 (corresponding to G-NOR-G) as shown in Figure 3. These ions were not detected in control spectra or in fractions 1 and 2. Numerous other ions, marked by 0 in Figure 3, were detected in the control spectra, including an ion at m/z 239 which could also denote the presence of NOR-G-OH. Thus, parent ion scanning is a useful technique for rapidly screening complex biological mixtures that have only undergone minor purification procedures (12). In addition, on the basis of the parent ion scans, the presence of NOR-G-OH (mlz 239), NOR-G (mlz 257/259), and G-NOR-G (mlz 372) can be tentatively identified. In particular, the ion at m/z 259 (approximately 30% of mlz 257) strongly suggests the presence of chlorine in the structure of the ion at mlz 257 (13). To c o n f i i that such ions correspond to guanine adducts of cyclophosphamide, daughter ion spectra were acquired on the ions detected at m / z 239, 257, and 372 (from the parent scan). These spectra were compared with the daughter ion spectra of synthetic standards of NOR-G-OH, NOR-G, and G-NOR-G reported previously (6). These

Mirkes et al.

loo

1

324

50-

369

Table I. Daughter Ion Spectra of Synthetic Standards of G-NOR-G,NOR-G, and NOR-G-OH,as Well as Daughter Ions Observed for Ions at m / z 372,257, and 239 from Rat Embryo DNA Samples ions derived from synthetic standards" rat embryo DNA daughter daughter compound MH+ ions MH+ ions' G-NOR-G 372 221 (20) 372 (100) 221 (15) NOR-G 257 (100) 178 (15). 257 (100) 178 (15). 152 (40) 152 (35) NOR-G-OH 239 (100) 152 (20), 239 (100) d 87 (20) a Results obtained from ref 6. Values in parentheses are ion abundance5 relative to the parent ion at 100% abundance. Ions