Chem. Res. Toxicol. 1994, 7,762-769
762
Metabolism and Disposition of 06-Benzyl-2’-deoxyguanosine in Sprague-Dawley Rats Demetrius M. Kokkinakis,**tRobert C. Moschel,$ Anthony E. Pegg,s M. Eileen Dolan,” and S. Clifford Schold, Jr.t Department of Neurology, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, Texas 75235-9036,National Cancer Institute-Frederick Cancer Research and Development Center, P.O. Box B, Building 538,Frederick, Maryland 21 702-1201,Departments of Cellular and Molecular Physiology and Pharmacology, Pennsylvania State University College of Medicine, The Milton S. Hershey Medical Center, P.O. Box 850,Hershey, Pennsylvania 17033,and Division of Hematology-Oncology,The University of Chicago Medical Center, 5841 South Maryland Avenue, Box MC2115, Chicago, Illinois 60637 Received May 19, 1994@
06-Benzyl-2’-deoxyguanosineis a potential antitumor drug modulator that is intended to reduce or eliminate 06-alkylguanine-DNA alkyltransferase activity in tumors prior to treatment with genotoxic chemotherapeutic alkylating agents. The rationale for using this compound instead of the more active 06-benzylguanine and its substituted benzyl derivatives a t the benzyl ring is its greater solubility in aqueous media and potential pharmacologic advantage. Metabolism and disposition of 06-benzyl-2’-deoxyguanosinewas determined in adult male Sprague-Dawley rats following an ip injection of 100 mg/kg. Under these conditions, the compound was partially metabolized to yield a glucuronic acid conjugate, which was secreted exclusively in the bile. Removal of the 2’-deoxyribose or the benzyl group to yield 06benzylguanine and 2’-deoxyguanosine, respectively, occurred to a lesser extent. Metabolism accounted for the clearance of at least 58% of the total dose and took place primarily in the liver. Direct excretion of unchanged drug, mainly in urine, accounted for the remainder of the dose. Analysis of venous blood showed the presence of 06-benzyl-2’-deoxyguanosineand 06-benzylguanine a t concentrations which are considered to be effective in depleting alkyltransferase activity. Levels of the nucleoside reached a maximum of 45 pM a t 2 h, while those of 06-benzylguanine peaked to 20 pM a t 4 h and remained a t that level for a t least 4 more in C6 glioma cells increased linearly with hours. Transport of 06-benzyl-2’-deoxyguanosine the extracellular concentration of the drug up to 600 pM. Intracellular levels of the drug reached 1.2 pmol per pM of extracellular compound per lo6 cells as soon as 30 s after exposure and remained as high for a t least 1h. Such levels indicate that entrapment of the nucleoside inside cells by either phosphorylation or other means is probably not a n important feature for this drug. The extensive glucuronidation of 06-benzyl-2’-deoxyguanosinemay result in the inactivation of the drug as a n alkyltransferase inhibitor, thus protecting the intestinal epithelium from being sensitized to alkylating agents. However, hydrolysis of the conjugate by bacterial /3-glucuronidases could restore the inhibitory effect of the drug in the colon, which could have pharmacologic implications in the treatment of colon cancers.
Introduction There is increasing evidence that alkylation of guanine residues in DNA at the 06-position is a critical cytotoxic event following treatment with alkylating agents (1-3) and that endogenous OWkylguanine-DNA alkyltransferase (AGTY gene expression may be a major factor in the resistance to such agents (4,5). The reaction between 06-alkylguanines and AGT is stoichiometric, during which one molecule of the protein is inactivated for each molecule of the adduct (6). Second-order chemical kinet-
* TOwhom correspondence should be addressed. Phone: (214)6486314;Fax (214)648-7992. + University of Texas Southwestern Medical Center. National Cancer Institute-Frederick Cancer Research and Development Center. Pennsylvania State University College of Medicine. The University of Chicago Medical Center. Abstract published in Advance ACS Abstracts, October 1, 1994. Abbreviations: 06-BzldGu0, 06-benzyl-2’-deoxyguanosine;0 6 BzlGua, 06-benzylguanine; Gua, guanine; dGuo, 2’-deoxyguanosine; AGT, OB-alkylguanine-DNA alkyltransferase; BCNU, carmustine; PEG 400,polyethylene glycol 400;AIDS, acquired immune deficiency syndrome; AZT, 3’-azido-3’-deoxythymidine.
* *
@
ics at rates near diffusion control ensures rapid and complete repair of these lesions (7). Mammalian cells show considerable variation in AGT activity. Many rodent lines such as V79 and CHO do not possess AGT activity and are defined as mer- (8,9).In animal tissues, AGT activity is dependent on the type of tissue, which varies with species and developmental stage (10,ll).The observations imply that the AGT gene is regulated at tissue, cellular, developmental, and possibly, the cell cycle stage levels. Modulation of AGT activity has been successful in hepatocytes of the rat liver following - hepatectomy or chronic administration of genotoxic agents; however, induction has not been observed in other tissues or in the liver of other species (12). AGT activity has also been induced in human hepatoma (LICH) and glioblastoma (T98G) cell lines following treatment with alkylating agents (131,which underlines the potentid of certain tumors to become more resistant as the result of repeated chemotherapy with such agents* The characterization of the promoter region of the human AGT
0893-228x/94/2707-0762$04.50/00 1994 American Chemical Society
Metabolism of 06-Benzyl-2'-deoxyguanosine in Rat indicates that modulation of AGT levels is probably subject to several regulatory elements (24). Thus, the probability of down-regulation of AGT specifically in tumors of the mer+ phenotype to allow sensitization of such tumors to alkylating agents is rather remote at present. Effective depletion of AGT activity has been achieved with compounds which react with the active site of the protein stoichiometrically, thus reducing its cellular concentration. 06-Methylguanine has been used to sensitize various cell lines to alkylating agents; however, due to this compound's slow reaction with AGT, sensitization is incomplete even in culture systems (25). Recently, a number of compounds, most of which are derivatives of 06-benzylguanine (06-BzlGua),have been synthesized that have the capacity to deplete AGT activity at micromolar concentrations (16, 2 7). Theoretically, the limiting factor in preventing 06-alkylguaninerelated cytotoxicity in human tissues during anticancer chemotherapies with DNA alkylating agents is the level of AGT in normal tissues related to that in the tumor. In normal human tissues, AGT activities vary widely among individuals but are usually much higher than those found in experimental animal models (18). Gerson (29) has reported AGT levels of 480 f 130 fmoYmg of protein in human liver and much lower activities in small intestine (130 f 801, colon, and brain (100 f 60). High levels of alkyltransferase activity have been found in certain cell lines derived from human tumors (20, 22); however, a direct comparison of such levels with those in normal human or rodent tissues cannot be made because of the wide diversity of protein levels among various cell types and the inaccuracy of available methodologies for assaying absolute AGT activity. It is encouraging, however, that a combination of 06-BzlGua and carmustine (BCNU) has been found to be effective against human tumor xenografts in nude mice even though AGT levels of such tumors apparently exceed those of normal tissues of the host (20). This paradox is probably the result of several factors such as differences in systemic availability, cell permeability, and metabolism of the AGT inhibitor in normal and malignant tissue. Understanding the pharmacokinetics and metabolism of 06-BzlGua and its analogs in experimental animals and humans will be important not only for elucidating the mechanism of the synergism of AGT inhibitors and alkylating drugs, but also for designing more effective drugs and regimens for the treatment of a wide variety of mer+ tumors. In this paper, we present data on the metabolism and disposition of 06-benzyl-2'-deoxyguanosine (06-BzldGuo). This compound, although less effective than 06-BzlGua in reacting with AGT, is more water soluble and more easily administered at selective sites.
Experimental Procedures Materials. 06-BzldGuo and 06-BzlGua were synthesized as previously described (22, 23). Their W spectra in 10 mM phosphate buffer (pH 7.0) recorded on a Perkin-Elmer Lambda 4 spectrophotometer (06-BzldGuo: €243 = 9860 M-l, €282 = 10 500 M-l; 06-BzlGua: €240 = 7300 M-I, €281 = 8300 M-l) were similar to previously published spectra (16, 22). [3Hl-06-BzldGuowas prepared by Amersham (Arlington Heights, IL) by tritium exchange. The tritiated compound was purified by highperformance liquid chromatography (HPLC), and its specific activity was determined as 1.17 CVmmol. Guanine (Gua) and 2'-deoxyguanosine (dGuo) were purchased from Sigma (St. Louis, MO).
Chem. Res. Toxicol., Vol. 7, No. 6, 1994 763 Animals. Male Sprague-Dawley rats weighing 200 g were purchased from Charles River (Wilmington, DE) and fed a 7001 Teklad rodent diet (Teklad, Madison, WI). All animals used in this work were 300 f 5 g. Urine Collection. Three animals were injected ip with 100 mg/kg [3H]-O6-BzldGuo (sp ac. 1.96 CVmol) dissolved at room temperature in 30% polyethylene glycol 400 (PEG 400) in buffered saline (pH 7.4). The animals were placed in Nalgene metabolic cages (Fisher Scientific), and urine was collected for 24 h in plastic bottles containing 20 mg of ascorbic acid and 20 mg of sodium azide. Urines were filtered immediately after collection and frozen to -70 "C until used. Bile Collection. Five animals were anesthetized with methoxyflurane inhalation. A midline incision was made, and the common bile duct was ligated near the pancreas. A second ligature was partially tied around the duct near the hilum of the liver. A 0.28 mm i.d. and 0.61 mm 0.d. polyethylene tube was inserted in the bile duct above the lower ligature and pushed toward the liver past the loose second ligature. The catheter was secured by tightening the ligature. The end of the catheter was moved through the abdominal wall and the abdominal incision was closed. Animals were placed in Centrap restraining cages (Fisher Scientific) and were given free access to water and food. The restraining device allowed the collection of urine. All animals were allowed to recover completely from anesthesia (4 h). Three of the animals, showing complete recovery from anesthesia and excreting similar levels of bile (0.6-0.8 m m ) , were injected with 100 mgkg [3Hl-06-BzldGuo dissolved in 30% PEG 400 in buffered saline. The rest were injected with the carrier alone. Bile was collected in tubes containing 3 mL of 0.05 M phosphate buffer (pH 7.0) at 0 "C at hourly intervals for 24 h. Bile samples were frozen at -70 "C until used for analysis. Urine was also collected and stored at -70 "C. Blood Collection. Fifteen animals were injected ip with 100 mgkg [3H]-06-BzldGuo(sp act. 0.19 CUmol) in 30% PEG 400. Five additional animals were injected with the carrier alone. Three treated animals and one control were anesthetized with methoxyflurane at 1, 2, 4, 6, or 8 h after the injection. While under anesthesia, 3 mL of blood was removed from the vena cava and placed in heparinized tubes. One hundred microliters of blood were immediately placed in scintillation vials, bleached with 0.5 mL of 30% H202, acidified with 0.1 mL of glacial acetic acid, mixed with 15 mL of 3a70B scintillation fluid (r.p.i., Mount Prospect, IL), and counted for tritium. For HPLC analysis, 1 mL of the blood was mixed with 3 volumes of absolute ethanol. Precipitated material was extracted three additional times with 70% ethanol, the ethanolic fractions were combined, and their volume was reduced to 0.3 mL under vacuum in a rotatory evaporator and was increased to 1 mL with 30% acetonitrile. The solution was passed through a 3 mL Supelclean LC-18 cartridge (Supelco Inc., Bellefonte, PA) which was then washed with 3 mL of 30% acetonitrile in water. All the radioactivity in the sample was recovered, but lipids were retained by the cartridge. Washes were combined, their volume was reduced to 0.5 mL under vacuum, and they were stored at -70 "C until used. Urine, bile, and blood samples were analyzed by HPLC directly without additional treatment. High-Performance Liquid Chromatography. 06-BzldGuo and its metabolites were chomatographed using a PerkinElmer Series 2 system. Ultraviolet absorbance, fluorescence, and radioactivity were determined in eluates simultaneously using a Perkin-Elmer LC-85 W detector set at 254 nm, a Spedrovision FD-300 dual monochromator fluorescence detector set at 305 (excitation) and 370 (emission) nm, and a Radiomatic Flow-One radioactivity detector set on the tritium channel, serially connected to a Supelcosil-LC-18 DB analytical 5-pm column (Supelco Inc., Bellefonte, PA). The column was eluted at a rate of 1.5mumin with 0.05 M aqueous sodium phosphate (pH 3.8) (0-10 min) followed by a linear gradient of 2.2%/min acetonitrile in the aqueous sodium phosphate (10-20 min) and 22% acetonitrile in the same buffer (20-35 m i d . Under the above conditions, Gua eluted at 5.1 min, dGuo at 21.5 min, 06BzlGua at 31.2 min, and 06-BzldGuo at 34.4 min. Ratios of
764 Chem. Res. Toxicol., Vol. 7, No. 6, 1994 fluorescence versus U V absorbance were characteristic for each of the four compounds. Such ratios were 3.2, 0.7, 98, and 18 for peaks of Gua, dGuo, 06-BzlGua, and 06-BzldGuo, respectively. Mass Spectra. Metabolites were purified from bile using a semipreparative Supercosil LC-l&DB reverse phase column (25 cm x 10 mm) (Supelco, Inc., Bellefonte, PA). One milliliter of diluted bile ( 1 : l O in 0.05 M phosphate buffer, pH 4.0) was injected onto the column and eluted with the same sequence of phosphate buffer-acetonitrile gradient used in analytical applications at a flow rate of 4 mumin. Peaks not present in control bile were collected, neutralized with sodium bicarbonate, and concentrated to about 1mL volume by rotatory evaporation at 30 "C. Major radiolabeled metabolites were purified to homogeneity by reinjecting onto the column and eluting with 0.05 M ammonium bicarbonate (pH 7.8) (0-10 min), followed by a linear gradient of 2.2%/min acetonitrile (10-20 min) and 22% acetonitrile in the same buffer (20-40 min). Metabolite peaks were collected, quantitated using the specific activity of the parent compound, and evaporated to near dryness at 30 "C using a rotatory evaporator. The residue was redissolved in absolute ethanol, and it was then re-evaporated to dryness. The compounds were finally dissolved in HPLC-grade acetonitrile at estimated concentrations of 100 nmol/mL. Electrospray mass spectra were acquired with a VG 30-250 quadropole mass spectrometer (VG-Biotech, Manchester, GB). Samples (5 pL) were injected into direct flow of 50% methanol in water containing 1%acetic acid set at 6 pumin. Samples were ionized by positive ion electrospray (VG-Biotech, Manchester, GB). In Vitro Metabolism. Livers from three male SpragueDawley rats were used to prepare microsomes according to previously published methods (24).Glucuronidation of 06BzldGuo was assayed by incubating microsomes in 0.05 M TrisHCl (pH 7.4) containing 0.05%Triton X-100,lO mM MgC12, and 1mM tritiated substrate for 15 min at 32 "C. The reaction was started by adding uridine diphosphate glucuronic acid at a final concentration of 3 mM. Reactions were allowed to continue for an additional 20 min and were terminated by adding 2 volumes of cold ethanol. Precipitated proteins were removed by centrifugation at 12000g, while supernatants were removed and concentrated t o about 100 pL under vacuum. Samples were analyzed by analytical HPLC as described above. In order to test for the presence of glucoronic acid conjugates, 0.1 mL of urine or bile from [3H]-06-BzldGuotreated animals was incubated with 1000 units of /3-glucuronidase (Sigma) in 0.9 mL of 50 mM phosphate buffer, pH 6.8, for 2 h at 35 "C. Radiochromatograms of urines and biles before and aRer /3-glucuronidase treatment were compared to determine the presence of glucuronic acid conjugates. Transport. The method of Fariss et al. (25) was used for cell transport measurements. C6 rat glioma cells (ATCC, Rockville, MD) were trypsinized from cultures and suspended in buffered saline (pH 7.4) at a density of about 20 x lo6 cells/ mL. Cell numbers were determined using a Coulter counter (Coulter Electronics, Inc., Hialeah, FL). Cell suspensions and [3Hl-06-BzldGuosolutions in buffered saline were incubated at 37 "C separately for at least 5 min and then rapidly mixed and layered onto 0.7 mL of dibutyl phthalate which had been previously layered over 0.4 mL of 10% perchloric acid in microfuge tubes. The tubes were rapidly centrifuged a t 15000g, which forced the cells through the dibutyl phthalate cushion and into the acid where they were lysed, and macromolecules and cell debris were packed in the bottom of the tube. An aliquot of 250 pL was removed from the perchloric acid layer with a Hamilton microsyringe and counted for radioactivity.
Results Aqueous solutions of labeled 06-BzldGuo were stable at p H 7.0 a n d room temperature for at least 96 h. The compound, however, rapidly decomposed at acid pH. The half-life of t h e 06-BzldGuo was 7 a n d 60 min at 65 "C at
Kokkinakis et al.
P
0
4
12
8
16
20 2 4
TIME FROM INJECTION (HI Figure 1. Cumulative excretion of L3H1label in urine of three animals with (A, V, +) or without (A, V, 0 ) bile cannulas. Cumulative secretion of label in the bile of the first group of animals is also shown (0)as mean SD from the 3 animals of the same group.
*
50,
!
1
BILE
I
50, URINE
20
I ELUTION TIME (MIN)
ELUTION TIME (MINI
Figure 2. HPLC radiochromatography of 06-BzldGuometabolites in the bile and urine of cannulated rats. Separation of metabolites was accomplished on a Supelcosil LC-18-DB semipreparative column (25 cm x 10 mm) at pH 4.0 according t o the methods listed in Experimental Procedures. p H values of 3 a n d 4, respectively. At these p H values, dGuo a n d 06-BzlGua were formed at similar rates. Further decomposition of t h e above products yielded radiolabeled Gua. Retention of all t h e tritium originally present in 06-BzldGuo by Gua during acid hydrolysis indicated that 06-BzldGuo was labeled exclusively at t h e 8-position of Gua. 06-BzldGuo yielded a simple mass spectrum composed of t h e protonated parent ion (M 1 = 358.391, the protonated ion of 06-BzlGua (M 1= 242.321, a n d the C7H7 fragment (M = 91.13). The amounts of combined radioactivity in the urine and bile of animals with cannulated bile duct collected within 24 h after treatment with [3Hl-06-BzldGuowere 43 f 4% a n d 24 f 3%, respectively. In noncannulated animals, t h e amount of label excreted in urine within t h e same time interval was 49 f 3% of t h e total. The rates of excretion a n d secretion of radiolabel in urine a n d bile were nearly constant for t h e first 8 h. After that time, they sharply declined a n d reached a plateau at 24 h (Figure 1). A small b u t consistent difference in the amount of label excreted in the urine of the two groups (with a n d without bile cannulas) may be due to limited reabsorption of the parent compound or its metabolites from t h e gut. HPLC analysis of metabolites in urine a n d bile showed dramatic differences in composition in these two fluids (Figure 2). There were at least 11radiolabeled compounds in each of t h e fluids, but t h e relative concen-
+
+
Metabolism of O~-Benzyl-2'-deoxyguanosine in Rat
Chem. Res. Toxicol.,Vol. 7, No. 6, 1994 765 "1
3El 1. 0 El Bl
0
1. 0. 0. 0.
1
I
I
534.45
20
0 100
-0-
8 4.01 I
4
i e'
8
12
16
20
24
i
I
TIME OF SECRETION IN BILE (H) Figure 3. Cumulative secretion of 06-BzldGuoand its metabolites in bile during a 24 h period, following ip injection of 100 mgkg of the drug. (A) Hydrophilic metabolites 1 ( 0 ) , 2 (O),
and 3 (A) exchange their label during isolation. (B)More hydrophobic metabolites 4 (A), 7 (+), 9 (01,and the parent compound 10 (0) retain their label at the 8-position of the guanine. Points: mean from 3 animals.
I
H O . o GJy : ' ' ~ ' : : ~ ; : : : a
242.75
0
4
8
12
16
20
24
TIME OF EXCRETION IN URINE (H) Figure 4. Cumulative excretion of 06-BzldGuo(inset)and 06BzldGuo metabolites 1 (0))4 (A),5 (HI, 8 (VI,and.9 !O! in urine of noncannulated rats during a 24 h period after ip injection of 100 mgkg of the drug. Data from two animals was used to compose this graph. Urines were collected at 2,6,9,13,18,and 22 h from the first animal and at 4, 8,11, 16,20, and 24 h from the second animal.
trations of these metabolites varied widely, as seen in Figures 3 and 4. The most striking difference between urine and bile was the presence of metabolite 7, which was almost exclusively found in the latter fluid. In bile, most metabolites plateaued at about the same time (8 h), except for metabolites 1-3 which continued to be secreted at significant levels, apparently even beyond a 24 h period. Metabolite 1 was the major component in rat urine between 20 and 24 h. Mass spectroscopy of component 10 in either bile or urine yielded a spectrum identical to that of 06-BzldGuo. Bile metabolites 4 and 9 had mass spectra identical to those of authentic dGuo and 06-BzlGua, respectively.
200
3w
mlz
400
500
600
Figure 5. Mass spectrum of metabolite 7.
Levels of 06-BzlGua were similar in bile and urine (0.4 f 0.1% of the dose of 06-BzldGuo). However, such levels were significantly lower than those of the parent compound, especially in urine. During isolation of metabolites 1 and 3, the tritium was removed by evaporation, indicating a complete exchange of the label with solvent. The mass spectra of the isolated unlabeled metabolites 1 and 3 could not be interpreted as retaining an intact 06-BzlGua moiety, and therefore, identification of these metabolites was not pursued further. Metabolite 7, which was the major radiolabeled component in bile, showed different mobility during HPLC runs at pH 7 and 4, which indicated the presence of an ionic group. Its mass spectrum (Figure 5) indicated a parent molecular ion of 534.45 and a major fragment at 242.75. These values correspond to the glucuronic acid conjugate of 06BzldGuo (MW 533.51) plus a proton, and the 06-BzlGua (MW 241.26) plus a proton. The identity of metabolite 7 as the glucuronic acid conjugate of 06-BzldGuo was verified by digestion with @-glucuronidaseand also by enzymatic synthesis. Treatment of the bile with p-glucuronidase resulted in nearly complete suppression of the W, fluorescence, and radioactivity associated with peak 7. At the same time, an increase of the relative concentration of the parent compound eluting at 36.6 min (peak 10) was observed. Treatment of the urine with p-glucuronidase did not have any significant effect on the relative concentrations of any of the metabolites in this fluid. Incubations of [3H]-06-BzldGuo (99.9% pure) with the microsomal fraction of rat liver and UDP-glucuronic acid yielded only one radiolabeled product coeluting with 'metabolite 7 under two different HPLC systems using either sodium phosphate (pH 4) or ammonium bicarbonate (pH 7) buffers to elute the metabolite. This enzymatic synthesis yielded only 1.8% of the conjugate, despite the high concentrations of UDP-glucuronic acid (3 mM) and 06-BzldGuo (1mM) used. The glucuronic acid conjugate of 06-BzldGuo had a half-life of 60 min at pH 4 and 65 "C. Under these conditions, one major product of hydrolysis containing 42% of total label was identified as the glucuronic acid conjugate of dGuo by electrospray mass spectroscopy (M 1 = 444.40). A fragment ion a t M 152 which corresponds to guanine confirmed this assignment. An additional hydrolysis product containing 15%of the label was identified as 06-BzldGuo. More prolonged hydrolysis did not change the relative amounts of these products, but resulted in the accumulation of Gua and concomitant decline of the radioactivity associated with metabolite 7. The above indicated a strong chemical bond between the glucuronic acid and dGuo which was more resistant to hydrolysis as compared to either the 06-benzyl bond or the 2'-deoxyribose-guanine glycosidic bond. The absence of a radiolabeled peak other than those of 06-BzldGuo,
+
766 Chem. Res. Toxicol.,Vol. 7,No.6, 1994
Kokkinakis et al.
Table 1. Levels of 06-BzldGuoand of 06-BzlGuain Blood of Os-BzldGuoTreated Sprague-DawleyRats and Relative Elimination of These Two Compounds by the Kidney
1 2
4 6
8 10 12 14
40 f 5a 45 4 36 f 3 13 f 2 1114 NDb ND ND
*
10 f 3
15 f 2 20 f 2 19 1 3 18 f 4 ND ND ND
4.0 3.0 1.8
0.7 0.6 ND ND ND
Scheme 1. Pathways in the Metabolism of 06-BzldGuo
ND 234 f 84c 75 f 17 38 f 11 29 f 9 23 f 6 10 i 5 4 f 3
OH
" T ? ?--
+
Mean value from 3 animals f SD. ND: not determined. Determined from data shown in Figure 4. Mean from 3 animals fSD.
Discussion A number of 06-BzlGua derivatives are currently being tested to determine their efficacy in sensitizing tumors to the toxicity of alkylating agents (16, 17). The 06BzlGua and some para-substituted derivatives are the most promising because of their efficient reaction with
I
ofDebenzylation and/or AGT
a
Gua, and the glucuronic acid conjugate of dGuo indicated that glucuronic acid is conjugated to the sugar rather than the Gua. HPLC analysis of whole blood at 1, 2, 4,6, and 8 h after the injection of 06-BzldGuo resulted in the separation of six major radioactive peaks. Two of the peaks were eluted near the solvent front ( t R = 2.5 and 3.2 min). Two additional metabolites were tentatively identified as 5.3 and Gua and dGuo from retention times alone ( t ~ = 21.5 min). The remaining two were positively identified 31.2) and 06-BzldGuo( t =~34.4) from as 06-BzlGua ( t ~ = their U V spectra and relative fluorescence. Blood concentrations of the above metabolites were determined by HPLC analysis (Table 1). Levels of 06-BzldGuoreached a maximum of 45 pM in venous blood a t 2 h and gradually declined to 11pM at 8 h after treatment. High levels of 06-BzldGuo during the first 4 h possibly indicated continuous slow absorption of the compound from the site of injection. On the other hand, levels of 06-BzlGua increased, reaching a maximum around 4 h, and they did not decline for at least 8 h after injection. Since ratios of 06-BzldGuo to 06-BzlGua in blood are much lower than those found in urine (Table 11, it is evident that 06-BzldGuois excreted more effectively than 06-BzlGua. The observed low clearance of 06-BzlGuavia the urinary tract may be due to the binding of the compound with red blood cells (RBCs) and/or plasma, in which case the transport of this metabolite from blood to tissues may also be impaired. Transport of 06-BzldGuo into cells was studied using rat C6 glioma cells. These cells did not metabolize the compound except for the slow formation of 06-BzlGua with a V, of 7 pmol/(h.106 cells). The amount of 06BzldGuo transported into cells reached a maximum within 30 s from incubation and remained a t that level for at least 1 h. Transported 06-BzldGuo increased linearly with the cell number (up to 20 x lo6 cells) and substrate concentration (up to 600 pM). The amount transported was estimated as 1.2 pmoY(uM substrate106 cells). The above suggests a rapid simple diffusion process for the transport of 06-BzldGuo a t least in this tumor cell line.
I I 7
H6
Or
*
label OH
t OH
+
Uric Acid (loss of tritium)
AGT in vitro or in cell cultures. However, the poor solubility of these compounds in aqueous media makes systemic or local administration of these chemicals impractical without concomitant use of high concentrations of an organic carrier. This problem could possibly be minimized by using a more soluble 06-BzlGua derivative, e.g., a 2'-deoxyribonucleoside, that can still deplete AGT levels at concentrations which are easily attainable in the circulation of experimental animals and presumably in man. We have, therefore, examined the biological fate of 06-BzldGuo. The major products of 06-BzldGuo metabolism found in bile, urine, and blood are the glucuronic acid conjugate of the parent compound, 06-BzlGua, dGuo, and a number of metabolites that exchange their label during isolation. The latter do not retain an intact Gua ring structure. The glucuronidation of 06-BzldGuo was confirmed by mass spectroscopy, enzymatic synthesis, and degradation by ,&glucuronidase, The glucuronic acid was conjugated at the sugar moiety of 06-BzldGuo as shown by kinetic analysis of acid hydrolysis of the [3Hl-06-BzldGuoglucuronide, labeled a t the Gua moiety. Such hydrolysis yielded first 06-BzldGuo and the glucuronic acid conjugate of dGuo, and subsequently 06-BzlGua, dGuo, and Gua. During hydrolysis, there was no evidence for the intermediate formation of any additional labeled product, and thus conjugation of the glucuronic acid at the amino group of Gua must be excluded. Involvement of the sugar hydroxyl groups of dGua rather than the amino group of Gua appears to be favored in conjugating derivatives of dGua with glucuronic acid since this is also the case with carbovir, a carbocyclic derivative of Gua (26). The formation of 06-BzldGuo glucuronide is shown in Scheme 1 as involving the 5' hydroxyl group of the parent drug. This assignment is tentative, and the possibility of the involvement of the 3' hydroxyl is not excluded at the present time. Glucuronidation occurs primarily in hepatic tissue and is catalyzed by a microsomal glucuronyltransferase. Other tissues such as the kidney, lung, and brain may contribute to the glucuronidation of the compound; however, the extent of extrahepatic glucuronidation is expected to be low (27,28) and probably is of no biological significance. The isozyme responsible for the glucuronidation of 06-BzldGuo is probably the hy-
Metabolism of 06-Benzyl-2'-deoxyguanosinein Rat droxysteroid glucuronyltransferase which also has been implicated in detoxication of xenobiotics with free aliphatic hydroxyls (29,30). The importance of this major product of metabolism in depleting the AGT activity at the site of its formation, the liver, or in other tissues following its transport, has not been elucidated. The 06BzldGuo glucuronide is not secreted into the circulation and is almost exclusively eliminated by the biliary route. High concentrations of the glucuronide in the intestinal tract may result in complete depletion of the alkyltransferase in the brush epithelium and possibly sensitization of that epithelium to alkylating agents. However, the ionic character of this compound argues against an effective transport into cells in the small intestine. Incomplete depletion of the AGT activity in rat and mouse live9 shortly after the administration of 50 mg/ kg 06-BzldGuo indicates that this compound is rapidly inactivated in this tissue in terms of its ability to deplete AGT. The same conclusion can be drawn from structureactivity relations, demonstrating that modification of 06BzlGua by introducing ionic groups results in dramatic reduction of the effectiveness of the product to deplete AGT activity (16, 17). A possible role of 06-BzldGuoglucuronide, other than as an inactivation product, could be to deliver the conjugated drug to the colon, where the parent compound could be released a t high concentrations as the result of bacterial j3-glucuronidase (31, 32). In this regard, the glucuronidation of 06-BzldGuoor other related drugs in humans may be a useful pathway for transporting high concentrations of the drug into colon tumors, thus sensitizing them to a subsequent treatment with alkylating chemotherapeutic agents. In general, conjugation with glucuronic acid is a major pathway for detoxication of xenobiotics; however, with the exception of 3'-azido-3'-deoxythyidine (AZT)and of carbovir, a carbocyclic guanosine derivative (26), glucoronidation does not play a significant role in the metabolism of nucleosides used as antitumor or antiviral drugs (33-36). Glucuronic acid conjugates of 6-methoxypurine arabinoside have been tentatively identified in the urine of monkeys (Mucacafascicularis) treated with high doses of the drug a t amounts totaling 0.7% of the dose (33). Although metabolites of 6-methoxypurine have not been measured in the bile, the almost quantitative excretion of the parent compound and other metabolites in urine argues against extensive formation and subsequent secretion of a glucuronide of 6-methoxypurine in the biliary tract. Similarly, glucuronidation of pyrimidinebased drugs, such as fluorouracil, contributes minimally in their metabolism, possibly because of the rather hydrophilic nature of these drugs (34). On the other hand, the relatively hydrophobic AZT is glucuronidated up to 85% of the dose in acquired immunodeficiency syndrome (AIDS) patients (35) and somewhat less extensively in experimental animals, including primates (36). Apparently, the extent of glucuronidation of nucleosides is determined by many factors, including the species (26) and the hydrophobicity of the parent compound. It is, however, uncertain what determines disposition of the glucuronidated compound by urinary excretion or secretion in the bile. Extensive secretion of the 06-BzldGuoglucuronide in bile, but not in urine, is probably not a unique property of this compound. Other glucuronidated nucleosides disposed via the biliary route possibly have 2
D. M. Kokkinakis and S. C. Schold, unpublished observations.
Chem. Res. Toxicol., Vol. 7, No. 6, 1994 767 been missed since bile is usually not examined during pharmacokinetics or metabolic studies. Removal of the sugar moiety to yield 06-BzlGua, as shown in Scheme 1, contributes significantly to the metabolism of 06-BzldGuo and possibly to its biological effect, as it is determined by the presence of this metabolite in blood at concentrations comparable to those of its substrate for prolonged time periods. The ability of C6 glioma cells to catalyze formation of 06-BzlGua suggests that the removal of the sugar is not confined to liver, but it also occurs in other tissues. 06-BzldGuoand 06-BzlGua reach maximum levels in circulation at distinctly different time points following ip administration of the former drug, suggesting that clearance of 06BzldGuo differs from that of its product. Debenzylation of 06-BzldGuoto yield dGuo, as shown in Scheme 1,occurs readily, and the product is found in all three fluids examined. Removal of the benzyl group cannot result only from the reaction of the parent compound with AGT, and it probably involves additional pathways. The presence of dGuo in the bile suggests that debenzylation of 06-BzldGuoprobably occurs in the liver, but the product is also secreted in the circulation because of its relative hydrophilic nature. Reactions which result in degradation of the guanine structure and the consequent leakage of tritium were functional during the entire observation interval (0-24 h). The excretion of volatile tritium compounds and other metabolites which exchange their tritium with solvent requires previous degradation of the guanine ring. Catabolic reactions occur early following injection of the 06-BzldGuo, but become progressively more prominent compared to reactions yielding metabolites which retain the label. For example, the rate of excretion of metabolite 1trebles after 12 h compared to the rate observed between 4 h and 8 h. A similar increase is observed in bile. Metabolite 1 is the major or only radioactive component of urine 20 h after injection of the drug. The above suggests that reactions leading to the decomposition of the guanine structure are not likely to occur with 06-BzldGuo as substrate. Catabolism of guanosine in mammals involves the removal of the sugar by purine nucleoside phosphorylase, followed by oxidation reactions to yield uric acid (33). A similar pathway for the loss of tritium is tentatively proposed here. Under the ip route of administration of 06-BzldGuo, blood concentrations of the parent compound were maintained high for at least 3 h and then rapidly declined in a fashion paralleling the decline of the excretion and secretion of its metabolites in urine and bile, respectively. The high levels of 06-BzldGuo excreted in the urine as compared to those in plasma indicate that the association of the drug with plasma or RBCs is weak. This implies that drug transport into cells from the circulation is efficient. On the other hand, 06-BzlGua may be bound by plasma proteins which restrict transport of the compound into tissues or its elimination by the kidney. Transport of 06-BzldGuois at least 5 times greater than the transport of 5-fluoro-2'-deoxyuridine and 6-methoxypurine arabinoside in RBCs (37,38). However, in murine and human tumor cells, the transport of the above two drugs is at least an order of magnitude greater than that of OB-BzldGuo(37, 39). The above differences between 5-fluoro-2'-deoxyuridineand 6-methoxypurine arabinoside, on one hand, and 06-BzldGuo,on the other, are to be expected if one takes into account the complexity of nucleoside transport (40). Equilibrative and concentra-
768 Chem. Res. Toxicol., Vol. 7,No.6,1994 tive systems are both involved in the transport of nucleosides in mammalian cells, but the type of system and its sensitivity to inhibitors depends greatly on the type of cell and the structure of the nucleoside. In addition, the degree of nucleoside to nucleotide conversion exhibits a strong effect on the transport process and the levels of intracellular concentrations. Although the transport of 06-BzldGuo is not saturable up to 600 pM, the process still may be facilitated by low afXnity carriers. Such carriers with K, of 4.5 mM have been described for 5-fluorouracil (41).However, the greater hydrophobicity of 06-BzldGuo as compared to other deoxynucleosides may eliminate the need of a carrier system for its transport. The studies presented here support further consideration of 06-BzldGuoas a candidate for sensitizing tumors prior to treatment with chemotherapeutic alkylating agents. In spite of its ip administration, the compound and its active metabolite 06-BzlGua reached appreciable concentrations in blood and appear to be easily transported in tissues. Lack of significant secretion of the parent compound in bile curtails sensitization of the small intestine to subsequent treatment with alkylating agents. On the other hand, extensive glucuronidation of the compound and subsequent secretion of the glucuronide in the bile may have applications in targeting specific segments in the intestinal tract. Finally, the diversity of pathways involved in metabolism of 06BzldGuo is encouraging since bioavailability of this compound and its metabolites may be effectively modulated in vivo.
Acknowledgment. We appreciate the expert advice of Dr. Clive Slaughter and technical assistance of Mr. Bikash C. Pramanik, Howard Hughes Medical Institute, The University of Texas Southwestern Medical Center at Dallas, Dallas, TX, in obtaining mass spectra of metabolites. This work was supported by the National Cancer Institute [CA57725 (A.E.P., M.E.D.), CA47228 (M.E.D.), and DHHS, under Contract N01-C074101 with ABL (R.C.M.)] and the National Institutes of Health [NS20581 (S.C.S.11. References (1) Kalamegham, R., Warmels-Rodenhiser, S., MacDonald, H., and Ebisuzaki, K. (1988) 06-methylguanine-DNA methyltransferasedefective human cell mutant: 06-methylguanine, DNA strand breaks and cytotoxicity. Carcinogenesis 9, 1749- 1753. (2) Loechler, E. L., Green, C. L., and Essigmann, J. M. (1984)I n vivo mutagenesis by 06-methylguanine built into a unique site in a viral genome. Proc. Natl. Acad. Sci. U S A . 81, 6271-6275. (3) Dunn, W. C., Tano, K., Horesovsky, J.,Preston, R. J., and Mitra, S. (1991) The role of 06-alkylguaninein cell killing and mutagenesis in Chinese hamster ovary cells, Carcinogenesis 12, 83-89. (4) Brent, T. P., Houghton, P. J., and Houghton, J . A. (1985) 06-Alkylguanine-DNA alkyltransferase activity correlates with the therapeutic response of human rhabdomyosarcoma xenograh to l-(2-chloroethyl)-3-(trans-4-methylcyclohexyl)-l-nitrosourea. Proc. Natl. Acad. Sci. U S A . 82, 2985-2989. ( 5 ) Schold, S. C., Jr., Brent, T. P., von Hofe, E., Friedman, H. S., Mitra, S., Bigner, D. D., Swenberg, J. A., and Kleihues, P. (1989) 06-Alkylguanine-DNA alkyltransferase and sensitivity to procarbazine in human brain-tumor xenografts. J. Neurosurg. 70, 573-577. (6) Pegg, A. E. (1983)Alkylation and subsequent repair of DNA after exposure to dimethylnitrosamine and related carcinogens. In Reviews in Toxicology (Hodgson, E., Bend, R., and Philpot, R. M., Eds.) Vol. 5 , pp 83-133, Elsevier-North Holland Biomedical Press, New York. (7) Graves, R. J., Li, B. F., and Swann, P. F. (1989) Repair of OBmethylguanine, 06-ethylguanine, 06-isopropylguanine and 04methylthymine in synthetic oligodeoxynucleotides by Escherichia
Kokkinakis et al. coli ada gene 06-alkylguanine-DNA-alkyltransferase. Carcinogenesis 10, 661-666. (8) Sklar, R. M., and Strauss, B. S. (1981) Removal of O‘j-methylguanine from DNA of normal and Xeroderma pigmentosum-derived lymphoblastoid lines. Nature 289, 417-420. (9) Yarosh, D. B., Foote, R. S., Mitra, S., and Day, R. S., I11 (1983) Repair of 06-methylguanine in DNA by demethylation is lacking in Mer- human tumor cell strains. Carcinogenesis 4, 199-205. (10)Pegg, A. E. (1990) Mammalian 06-alkylguanine-DNA alkyltransferase: regulation and importance in response to alkylating carcinogenic and therapeutic agents. Cancer Res. 50,6119-6129. (11) Washington, W. J., Foote, R. S., Dunn, W. C., Generson, W. M., and Mitra, S. (1989)Age-dependent modulation of tissue-specific repair activity for 3-methyladenine and 06-methylguanine in DNA in inbred mice. Mech. Ageing Dev. 48, 43-52. (12) Montesano, R., Bresil, H., Planche-Martel, G., Margison, G. P., and Pegg, A. E. (1980) Effect of chronic treatment of rats with dimethylnitrosamine on the removal of 06-methylguanine from DNA. Cancer Res. 40, 452-458. (13) Lefebvre, P, Zak, P., and Laval, F. (1993) Induction of 06methylguanine-DNA-methyltransferaseand N3-methyladenineDNA-glycosylase in human cells exposed to DNA-damaging agents. DNA Cell Biol. 12, 233-241. (14) Harris, L. C., Potter, P. M., Tano, K., Shiota, S., Mitra, S., and Brent, T. P. (1991) Characterization of the promoter region of the human 06-methylguanine-DNA methyltransferase gene. Nucleic Acids Res. 19, 6163-6167. (15) Gerson, S. L., Trey, J. E., and Miller, K. (1988) Potentiation of nitrosourea cytotoxicity in human leukemic cells by inactivation of 06-alkylguanine-DNA alkyltransferase. Cancer Res. 48,15211527. (16) Moschel, R. C., McDougall, M. G., Dolan, M. E., Stine, L., and Pegg, A. E. (1992) Structural features of substituted purine derivatives compatible with depletion of human 06-alkylguanineDNA alkyltransferase. J . Med. Chem. 35, 4486-4491. (17) Chae, M. Y., McDougall, M. G., Dolan, M. E., Swenn, K., Pegg, A. E., and Moschel, R. C. (1994) Substituted benzylguanine derivatives and their inactivation of human 06-alkylguanineDNA alkyltransferase. J . Med. Chem. 37,342-347. (18) Gerson, S. L., Trey, J. E., Miller, K., and Berger, N. A. (1986) Comparison of 06-alkylguanine-DNA alkyltransferase activity based on cellular DNA content in human, rat and mouse tissues. Carcinogenesis 7, 745-749. (19) Gerson, S. L., Miller, K., and Berger, N. A. (1985) 06-Alkylguanine-DNA alkyltransferase activity in human myeloid cells. J. Clin. Invest. 76, 2106-2114. (20) Schold, S. C., Jr., Brent, T. P., von Hofe, E., Friedman, H. S., Mitra, S., Bigner, D. D., Swenberg, J. A,, and Kleihues, P. (1989) 06-Alkylguanine-DNA alkyltransferase and sensitivity to procarbazine in human brain-tumor xenografts. J. Neurosurg. 70, 573-577. (21) Ostrowski, L. E., von Wronski, M. A., Bigner, S. H., Rasheed, A., Schold, S. C., Jr., Brent, T. P., Mitra, S., and Bigner, D. D. (1991) Expression of 06-methylguanine-DNA methyltransferase in malignant human glioma cell lines. Carcinogenesis 12, 17391744. (22) Pauly, G. T., Powers, M., Pei, G. K., and Moschel, R. C. (1988) Synthesis and properties of H-ras DNA sequences containing 06substituted 2’-deoxyguanosineresidues at the first,second, or both positions of codon 12. Chem. Res. Toxicol. 1, 391-397. (23) Robins, M. J., Khwaja, T. A., and Robins, R. K. (1970) Purine nucleosides. ZWX. The synthesis of 2’-deoxy-~-adenosineand 2’deoxy-L-guanosine and their a-anomers. J . Org. Chem. 35, 636639. (24) Kokkinakis, D. M., Scarpelli, D. G., Subbarao, V., and Hollenberg, P. F. (1987) Species differences in the metabolism of N-nitroso(2-hydroxypropyl)(2-oxopropyl)amine. Carcinogenesis 8,295-303. (25) Fariss, M. W., Brown, M. K., Schmitz, J. A,, and Reed, D. J. (1985) Mechanism of chemical-induced toxicity. I. Use of a rapid centrifugation technique for the separation of viable and nonviable hepatocytes. Toxicol. Appl. Pharmacol. 79, 283-295. (26) Patanella, J. E., Walsh, J. S., Unger, S. E., Miwa, G. T., Parry, P. S., Daniel, M. J., and Evans, G. L. (1990) Identification of a glucuronide conjugate of the carbocyclic nucleoside, carbovir, isolated from Marmoset urine. Drug Metab. Dispos. 18, 10921095. (27) Minn, A., Ghersi-Egea, J., Perrin, R., Leininger, B., and Siest, G. (1991) Drug metabolizing enzymes in the brain and cerebral microvessels. Brain Res. Rev. 16,65-82. (28) Hook, G. E. R., Haseman, J. K., and Lucier, G. W. (1975) Induction and suppression of hepatic and extra-hepatic microsomal foreigncompound-metabolizing enzyme system by 2,3,7,8-tetrachlorodibenzo-p-dioxin. Chem.-Biol. Interact. 10, 199-214.
Metabolism of 06-Benzyl-2’-deoxyguanosine in Rat (29) Owens, I. S., and Ritter, J. K. (1992) The novel bilirubidphenol UDP glucuronosyltransferases, In Frontiers in Bwtransformution: Principles, Mechanisms and Biological Consequences of Induction (Ruckpaul, K, and Rein, H., Eds.) pp 93-108, VCH, New York. (30) Mackenzie, P. I. (1990) The cDNA sequence and expression of a variant 17heta-hydroxysteroidUDP-glucuronosyltransferase.J. Biol. Chem. 265, 8699-8703. (31) Bock, K. W. (1991) Roles of UDP-glucuronosyltransferases in chemical carcinogenesis. Crit. Rev. Biochem. Mol. Biol. 26, 129150. (32) Weisburger, J. H. (1971) Colon carcinogens: their metabolism and mode of action. Cancer 28, 60-70. (33) Burnette, T. C., Koszalka, G. W., Krenitsky, T. A., and de Miranda, P. (1991) Metabolic disposition and pharmacokinetics of the antiviral agent 6-methoxypurine arabinoside in rats and monkeys. Antimicrob. Agents Chemother. 35, 1165-1173. (34) Weckbecker, G. (1991) Biochemical pharmacology and analysis of fluoropyrimidines alone and in combination with modulators. Pharmacol. Ther. 50, 367-424. (35) Good, S. S., Koble, C. S., Crouch, R., Johnson, R. L., Rideout, J. L., and de Miranda, P. (1990) Isolation and characterization of an ether glucuronide of zidovudine, a major metabolite in monkeys and humans. Drug Metab. Dispos. 18,321-326.
Chem. Res. Toxicol., Vol. 7, No. 6,1994 769 (36) de Miranda, P., Burnette, T. C., and Good, S. S. (1987) Disposition and pharmacokinetics of the antiviral drug 3’-azido-3’-deoxythymidine (RetroviB) in monkeys and rats. In Program and Abstracts of the 27th Interscience Conference on Antimicrobial Agents and Chemotherapy, p 162, American Society for Microbiology, Washington, DC. (37) Uchida, M., Ho, D. H. W., Kamiya, K., Yoshimura, T., Sasaki, K., Tsutani, H., and Nakamura, T. (1989) Transport and intracellular metabolism of fluorinated pyrimidines in cultured cell lines. Adv. Exp. Med. Biol. 253B, 321-326. (38) Prus, K. L., Heidenreich, A. C., and Zimmerman, T. P. (1992) Transport of the anti-varicella-zoster virus agent 6-methoxypurine arabinoside and its 2’-0-valerate prodrug into human erythrocytes. Antimicrob. Agents Chemother. 36, 283-286. (39) Biron, K. K., de Miranda, P., Burnette, T. C . , and Krenitsky, T. A. (1991) Selective anabolism of 6-methoxypurine arabinoside in varicella-zoster virus-infected cells. Antimicrob. Agents Chemother. 35, 2116-2120. (40) Crawford, C. R., Ng, C. Y. C., Noel, L. D., and Belt, J. A. (1990) Nucleoside transport in L1210 murine leukemia cells. Evidence for three transporters. J. Biol. Chem. 265, 9732-9736. (41) Domin, B. A., Mahony, W. B., and Zimmerman, T. P. (1993) Transport of 5-fluorouracil and uracid into human erythrocytes. Biochem. Pharmacol. 46, 503-510.
