TERATOLOGY 4591-103 (1992)

Effects of (R)-Deoxycoformycin (Pentostatin) on Intrauterine Nucleoside Catabolism and Embryo Viability in the Pregnant Mouse T.B. KNUDSEN, R.S. WINTERS, S.K. OTEY, M.R. BLACKBURN, M.J. AIRHART, J.K. CHURCH, AND R.G. SKALKO Department of Anatomy, Jefferson Medical College, 1020 Locust Street, Department of Philadelphia, Pennsylvania 19107 (T.B.K.,M.R.B.), Anatomy, James H . Quillen College of Medicine, East Tennessee State University, Johnson City, Tennessee 37614 (R.S.W.,S.K.O., M.J.A., J.K.C., R.G.S.)

ABSTRACT The viability of early mouse embryos is acutely sensitive to (R)-deoxycoformycin(pentostatin), a tight-binding inhibitor of adenosine deaminase (ADA). Previous studies have shown that a single 5-mg/kg dose on day 7 (plug = day 0) of gestation fully inhibits uteroplacental ADA activity within 0.5 h; causes massive cell death in the neural plate and primary mesenchyme by 6 h, major craniofacial anomalies by day 10, and resorption by day 12 [Knudsen et al., '89; Airhart et al., ,911. The present study has examined further the developmental toxicity and early effects of this inhibitor on ADA metabolism. (R)-Deoxycoformycinwas administered to pregnant CD-1 (ICR)mice as a single intraperitoneal dose of 0.5-10 mg/kg total body weight on days 6-11 of gestation. The major adverse effect, early resorption, was dose dependent and specific to day 7-8 exposure. Treatment with 5 mg/kg on day 7 resulted in 85% resorptions, 15% malformations, and a 24% reduction in mean fetal weight, whereas the same dose of (5')-deoxycoformycin had no effect. Levels of adenosine and 2'-deoxyadenosine, which are the endogenous substrates of ADA, were monitored in the embryo/decidual unit (E/D) by reversed-phase high-performance liquid chromatography (RP-HPLC). In response to the inhibitor, both nucleosides increased transiently in the antimesometrial compartment (antimesometrial decidua embryo). Peak levels (Cmax)of adenosine and 2'-deoxyadenosine were dose dependent over the range tested (0.05-10 mglkg). Exposure to 5 mgkg on day 7 raised adenosine levels within 0.5 h to 42-fold over the basal level of 0.06 nmol/mg protein. There was an even stronger effect on 2'-doexyadenosine levels, which were elevated 674-fold over the detection limit of 0.0005 nmol/mg protein. Direct exposure to the inhibitor in serum-free E/D culture produced similar results: 50 FM (R)-deoxycoformycinwithin 1h raised adenosine levels 26-fold and 2'-deoxyadenosine levels 410-fold. In vivo studies also showed a general correlation between embryolethality and the length of adenine nucleoside pool expansion, apparent for exposure on day 7,8, or 9 but not on day 6, suggesting that the embryo becomes sensitive to adenosine or 2'deoxyadenosine once the neural plate has formed.

+

Adenosine deaminase (ADA) (E.C.3.5.4.4) catalyzes the hydrolytic deamination of adenosine to inosine and 2'-deoxyadenosine to 2'-deoxyinosine. Much of the current interest in this enzyme has arisen from its to ment of the immune System (Kellems et al., 0 1992 WILEY-LISS. INC.

'85) and from prospects of somatic gene replacement therapy for ADA-SCID (severe combined immunodeficiency disease) (CulReceived June 24,1991; accepted August 27, 1991. Address reprint requests to Dr. Thomas B. Knudsen, Department of Anatomy,Jefferson Medical College, 1020 Locust Street, Philadelphia, PA 19107.

92

T.B. KNUDSEN ET AL.

ver et al., '91; Ferrari et al., '91). ADA is expressed at low levels in most mammalian tissues and, in addition, is highly expressed in some differentiating tissues (Aronow et al., '89; Chinsky et al., '89). One potentially important site of elevated expression is the uteroplacental unit (Brady and O'Donovan, '65; Sim and Maguire, '70; Knudsen et al., '88; Chinsky et al., '90). The cellular localization of ADA in the pregnant uterus of the mouse (Knudsen et al., '91) and in the pseudopregnant uterus of the rat (Hong et al., '91) has revealed an intricate pattern of overlapping decidual and placental phases of expression. In the mouse, ADA levels first increase on days 6-9 of gestation, mainly as a result of synthesis and accumulation in the antimesometrial decidua, but the center shifts to the basal zone of the placenta by day 13. The role of ADA a t the maternalembryonal interface has not been determined. A logical approach to aid the study of the intrauterine function of ADA is to inhibit the enzyme pharmacologically at specific stages of gestation. The most potent ADA inhibitor, (R)-deoxycoformycin (pentostatin), is a transition-state nucleoside analogue with an equilibrium dissociation constant of 2.5 x lop1' M (Agarwal et al., '77). Intrauterine viability of early mouse embryos is acutely sensitive to (R)-deoxycoformycin as treatment of pregnant dams with a single 5-mg/kg dose results in a high level (61-78%) of early resorptions specific to exposure on gestational day 7-8 (Knudsen et al., '89). Early features of day 7 treatment include a rapid inhibition of decidual ADA activity (0.5 h postinjection) and massive cell death in the neural plate and primary mesenchyme (3-6 h postinjection) (Knudsen et al., '89). Recent studies have suggested that early resorption stems from an all-or-none effect on the embryo. Embryos exposed to (R)-deoxycoformycin on day 7 are alive on day 10 but show severe malformations consistent with precursory damage to the anterior neural plate. Resorption and a corresponding decrease in malformation rate occurs by day 12 (Airhart et al., '91). Understanding why (R)-deoxycoformycin is toxic to the trilaminar embryo may provide new insights into the metabolic basis of defects involving the anterior neural axis. ADA is not essential for the growth of most cells in culture. However, genetic de-

ficiency or pharmacologic inhibition of the enzyme enhances the cytotoxicity of exogenous adenine nucleosides (Henderson and Smith, '81; Kantoff et al., '86; Kellems et al., '89; Kohn et al., '89). This may be relevant to the mechanism by which (R)-deoxycoformycin is lethal to the embryo because (1) exogenous adenosine is toxic to early postimplantation mouse embryos a t relatively high concentrations in vitro (Spindle and Pedersen, '77); (2) analogues of adenosine resistant to deamination by ADA are lethal to midgestational mouse embryos at low dosages in vivo (Clark et al., '87); (3) 2'deoxyadenosine, the toxic nucleoside that accumulates in the plasma of ADA-SCID infants (Carson et al., '87), precipitates growth retardation and developmental anomalies in the embryonic chick (Karnofsky and Lacon, '61); and (4) 2'-deoxyadenosine becomes markedly elevated in mouse plasma within minutes after exposure to 5 mg/kg @)-deoxycoformycin (Kuttesch and Nelson, '82). For these reasons, the present study was undertaken to characterize further the developmental toxicity of (R)-deoxycoformycin in relation t o early alterations in adenosine and 2'-deoxyadenosine homeostasis in the mouse. A dose-related accumulation of both endogenous ADA substrates was found in the antimesometrial compartment of the embryo/decidual unit (embryo + antimesometrial decidua) and was directly proportional to embryolethal potential on day 7. These results suggest a mechanistic involvement of endogenous adenosine and/or 2'-deoxyadenosine in early resorption. MATERIALS AND METHODS

Deoxycoformycin administration CD-1 (ICR) mice were purchased from Charles River Breeding Laboratories (Wilmington, MA). Nulliparous females 3-5 months of age were mated to males overnight and the finding of a vaginal plug in the morning signified day 0 of gestation. (R)-Deoxycoformycin [co-vidarabine; (8R)3-(2-deoxy-~-D-erythro-pentofuranosyl)-3, 6, 7,8-tetrahydroimidazo [4,5-d1[1,31diazepin8-01; pentostatin] was a gift from Dr. David Herzig (Parke-Davis Pharmaceutical Research Division, Warner-Lambert Company, Ann Arbor, MI). (S)-Deoxycoformycin was kindly provided by Dr. Fred Rudolph (Department of Biochemistry, Rice Univer-

ADENOSINE METABOLISM IN MICE

sity, Houston, TX). Stock solutions were prepared gravimetrically in water; the precise concentration was determined usin a molar extinction coefficient of 8.2 x 10Bat 282 nm (Schramm and Baker, ’85). (S)Deoxycoformycin was pure as judged by high-performance liquid chromatography (HPLC), and we are aware of the potential problem of spontaneous epimerization t o trace levels of the 8R-isomer under storage and assay conditions (Schramm and Baker, ’85). Working dilutions were prepared in water to deliver the appropriate dose a t an injection volume of 0.2 m1/30 g total body weight. Pregnant dams were treated by a single intraperitoneal injection at 0900 h on the gestational day indicated. Untreated or vehicle-treated dams provided controls.

93

sayed for ADA enzymatic activity as described by Aganval et al. (’77). Inhibition constants (Ki) were estimated by use of 150 analysis (Aganval et al., ’77). Enzyme-specific activity is given as nmol adenosine converted to inosine per minute per microgram protein (nmollminfpg protein).

Nucleoside extraction E/D were freeze-fixed on a stainless steel plate that had been precooled with Dry Ice. They were cut, while freezing, into mesometrial (E/D-m) and antimesometrial (E/D-a) compartments. The latter consisted of antimesometrial decidua + embryo. Frozen tissues (10 per sample) were pooled in microcentrifuge tubes and stored on Dry Ice for up to 1 h. Sample collection required approximately 6 min per litter. Ice-cold 0.4 N Evaluation of developmental toxicity perchloric acid (PCA, 850 p1) was added to Two rounds of studies were performed. In the frozen tissue, followed by sonication (30 the first round, pregnant dams were treated sec at 30 W). Tubes were placed on ice for 15 with a single injection of 0.5,5, or 10 mg/kg min, 100 pl was removed for protein deter(R)-deoxycoformycinon gestational days 6 - mination, and the homogenate was centri11. In the second round, dams were treated fuged a t 12,OOOg (5 min at 4°C). The superwith 1, 3, or 5 mg/kg on days 7 or 8. Ten natant (710 pl) was transferred to a clean litters from each dose group and a parallel tube and neutralized with 0.6 M KHCO,/ untreated control group were autopsied on 0.72 M KOH (356 11.1).Phenol red indicator gestational day 17. The parameters re- (40 pl of a 0.1-mg/ml solution) was added, corded were resorption incidence, fetal and samples were acidified with 0.18 M amweight, external malformations, and mater- monium dihydrogen phosphate (pH 5.1, 111 nal weight gain during pregnancy. The p1) and 1 drop of dilute phosphoric acid. “litter” represented the sampling unit for They were clarified by centrifugation and both resorptions (percentage of total im- stored at 4°C for HPLC analysis within 24 plantation sites resorbed in the litter) and hr or a t -20°C for longer periods. fetal weight (mean gram weight of fetuses HPLC conditions in the litter). Data were evaluated using a The system consisted of a model U6K one-way (dose) or two-way (dose, stage) analysis of variance (ANOVA). If dose-de- sample injector, two model 501 HPLC pendent data exhibited a significant trend pumps, model 680 Automated Gradient or a significant stage by dose interaction, Controller, and model 990 + photodiode arspecific pairwise comparisons were carried ray detector (Waters Associates, Inc., Millipore Corp., Milford, MA). Separation was out by Student’s t-test. performed on a prepacked Whatman PartiADA assay Sphere C,, reversed-phase cartridge column Embryo/decidual units (EID) were re- (4.6 ID x 12.5 cm) protected with a 1.7-cm moved from the uterus and rinsed in ice-cold reversed-phase guard cartridge (Whatman Hank’s balanced saline solution (HBSS). International Ltd., Maidstone, England). The antimesometrial decidua (with adher- The mobile phase was 0.02 M NH,H2P0,, ent trophoblast) was isolated and homoge- pH 5.1 and a superimposed methanol gradinized in 1 ml ice-cold 10 mM Tris-HC1, pH ent with the following shape: 0% (isocratic) 7.5, containing 75 mM KC1, 10 mM MgCl,, for 0-4 min, 0-8% (convex curve no. 4) for and 1mM dithiothreitol. Samples were cen- 4-6 min, 8-20% (concave curve no. 9) for trifuged for 1 min a t 12,OOOg; the superna- 6-8 min, and 20% (isocratic) for 8-18 min. tant was further purified by 55-90% ammo- The column was equilibrated to starting nium sulfate fractionation (Knudsen et al. conditions for 12 min before injecting the ’88). Aliquots of 0.1-50 pg protein were as- next sample. A constant flow rate of 1.5 ml/

94

T.B. KNUDSEN ET AL

min and injection volume of 200 p1 was used. Peaks were tentatively identified on the basis of co-retention and spectral congruency with reference standards purchased from Sigma (St. Louis, MO). Those relevant to this study, inosine (HR), 2'-deoxyinosine (HdR), adenosine (AR), and 2'deoxyadenosine (AdR), were verified by enzymatic peak shift assays using ADA (12 U/ml) or purine nucleoside phosphorylase (PNP, 0.5 U/ml) incubation for 30 min at 37°C prior to HPLC analysis. Sample concentration was calculated from peak area at the appropriate absorbance maximum with reference standards calibrated by molar absorbancy coefficients: HR and HdR = 12.2 x lo3 a t 249 nm; AR and AdR = 15.4 x lo3 at 260 nm. Short-term E/D culture Day 7 explants were grown for up to 1.5 h in rotation culture. Five E/D were placed in a 70-ml bottle containing 10 ml of medium (serum-free DME/F12 mixture, Sigma) equilibrated a t 37°C. The medium contained various concentrations of (R)deoxycoformycin up to 50 pM. Bottles were sealed, gassed for 45 sec with a N,-0,-CO, mixture (74%-21%-5%), and rotated on a miniroller apparatus a t 37"C, 10 rpm. Cultures were gassed each 30 min of incubation. At the end of the culture period, E/D were rinsed in HBSS and quickly processed for nucleoside analysis (5 per sample). Extracts were further purified by passage over a Prep-Sep C,, minicolumn (Fisher Scientific, Fair Lawn, NJ), and adenine nucleosides were eluted from the minicolumn with 4 mM ammonium phosphate in 50% methanol before HPLC analysis under isocratic conditions (9% methanol in 0.02 M NH,H,P04, pH 5.1). RESULTS

(R) -Deoxycoformycin induces early resorptions: stage specificity a nd dose dependency Results from the previous study (Knudsen et al., '89) demonstrated sensitivity to (R)deoxycoformycin with regard to embryolethality on gestational days 7 and 8; however, only one dose level (5 mg/kg) was tested during the day 6-11 period. In the present study, data on embryo viability, malformation rate, mean fetal weight, and maternal weight gain were obtained following a single injection of 0.5, 5, or 10 mgikg

during the gestational day 6-11 period. Significant effects were obtained with treatments on days 7 and 8 only, confirming the narrow sensitive period (Fig. 1). A second round of treatments was carried out in which dams received a single injection of 1, 3, or 5 mg/kg (R)-deoxycoformycin during days 7 and 8 in order to complete the doseresponse relationships shown in Table 1. Statistical evaluation of the data revealed significant dose-dependent effects on embryo viability and maternal weight gain following day 7 and day 8 treatments and on mean fetal weight following day 7 treatment only. The major adverse effect induced at either stage was early resorption. Resorptions became significantly increased over the background incidence at a dose level of 3 mg/kg, and the effect was dose dependent between 0.5 mg/kg (an insignificant trend) and 10 mg/kg (nearly complete embryolethality). It did not differ significantly between treatment days (Table 1). Fetal weight was reduced by up to 24% following day 7 treatment only, and the effect became significant a t 1mg/kg. The stage-specific effect on maternal weight gain correlated inversely with resorption incidence and thus appeared to be the consequence of pregnancy loss. Severe craniofacial malformations were observed in a low percentage of fetuses surviving treatment; however, malformed fetuses comprised 16% of the survivors at most, and the dose effect was not shown to be statistically significant (Table 1). Developmental toxicity is specific for the 8R-stereoisomer of 2' -deoxycoformycin Pregnant mice were exposed to 5 mg/kg (R)-deoxycoformycin or (5')-deoxycoformycin, or were left untreated, on day 7 of gestation and autopsies were performed on day 17. The 8R-isomer (pentostatin) precipitated a high resorption incidence (92% resorptions), a low rate of external malformations (14% of survivors), and a statistically significant reduction in mean fetal weight (24% vs. controls), as expected. No significant adverse effects were observed among litters treated with the same dose of (5')deoxycoformycin (Table 2). Inhibition of decidual ADA activity was investigated as a further measure of enantiomer selectivity. Inhibition constants estimated by use of I,, analysis revealed the expected result for (R)-deoxycoformycin (K, = 1.3 x lo-'' M);

95

ADENOSINE METABOLISM IN MICE

control

z 0 I-

0.5 rngikg

< I-

5.0 rng/kg

4 a

10 rngikg

z

malformed

5

se

s

cont.

6

7

8

9

10

11

GESTATIONAL DAY OF TREATMENT

Fig. 1. Stage response to (R)-deoxycoformycin in pregnant CD-1 (ICR) mice. Dams were treated with a single injection of (R)-deoxycoformycin a t 0900 h on the gestational day indicated and were autopsied on day 17. The percentage of total implantation sites resorbed (hatched) and malformed (solid) is shown for three dose levels (mean per litter fSE for n = 10). Statistical eval-

uation of the data by two-way analysis of variance revealed significant dose effect (P sO.OOl), stage effect (P 50.01), and dose by stage interaction (P ~ 0 . 0 0 1 with ) respect to resorptions. An asterisk (*) indicates a significant difference in resorption incidence from untreated controls by the paired t-test, P ~ 0 . 0 0 1 .

TABLE I . Embrvotoxicitv of (R)-deoxvcoformvcinduring the early iJostimplantation period'

2'

Treatment Dose Implantation sites (stage) (mg/kg) litters Total Res. Malf. Untreated Day 7

Day 8

0 0.5 1.0 3.0 5.0 10.0 0.5 1.0 3.0 5.0 10.0

20 10 10 10 20 10 10 10 10 20 10

239 135 130 130 267 117 133 135 118 270 135

12 12 17 35 222 105 5 8 34 194 133

0 0 4 7 7 0 0 2 1 2

0

% resorbed (mean tSD)'

4.8 t- 7.1% 9.2 f 10.1% 12.7 +- 16.5% 26.5 2 81.3 +- 24.7%5 91.1 t 20.2%5 3.9 t 5.5% 6.5 t 5.4% 27.8 f 22.7%5 71.6 3.3%5 98.2 f 3.9%5

*

Maternal Fetal weight weight gain Survivors in g in g malformed (mean tSD)2-4 (mean fSD)2

0% 0% 3.5% 7.4% 15.6% 0% 0% 1.6% 1.2% 2.6% 0%

1.13 t 0.15 1.09 t 0.08 0.98 2 0.055,6* 0.93 t 0.0S5,'* 0.86 t 0.105,'** 0.99 t- 0.20 1.05 t- 0.05 1.04 t 0.06 1.07 f 0.12 1.05 f 0.14 0.95 f 0.25

'Pregnant CD-1 (ICR) mice were treated at 0900 h on day 7 or 8 of gestation and autopsied on day 17. 'Analysis of variance (dose),P 50.001. 3Analysis of variance (stage),P 50.05. 4Analysis of variance (dose, stage), P ~0.001. 5Significantly different from controls by the unpaired t-test; P ~0.001. 'Significantly different between stages by the unpaired t-test, *P 50.05; **P ~0.005.

22.9 t 6.4 24.0 t 3.8 21.9 -t 5.2 19.0 t_ 5.8 9.8 f 5.85 5.4 t 5.75 24.5 t 3.8 24.7 t 4.4 18.9 f 3.8 12.6 t 6.75 7.7 ? 3.25

96

T.B.KNUDSEN ET AL. TABLE 2. Earlv DostimDlantation embrvolethalitv stereoselective for the 8R-isomer fmntostatin)'

%

Fetal weight implantation sites % resorbed in g Treatment litters Total Res. Malf. (mean tSD)' (mean ?SD)' Untreated 6 8 0 0 0 0% 1.01 ? 0.06 (R)-Deoxycoformycin,5 mg/kg 6 80 73 l4 91.6? 17.1%5**,60.77 2 0.015'.6 (S)-Deoxvcoformycin,5 mgikg 6 78 3 0 4.6 2 8.1% 1.13 ? 0.07

ADA activity (mean ?SD)'z3 0.43 ? 0.16 0.01 ? 0.005**26 0.34 ? 0.11

~~~

'Pregnant dams were treated at 0900 h on day 7 and autopsied on day 17. 'Analysis of variance (treatment), P SO.001. 3nmol/mini~gprotein in centrifuged homogenate of the antimesometrial decidua (n = 4) 0.5 h postinjection. 4Exencephaly. 'Significantly different from the control group by the unpaired t-test, *P ~ 0 . 0 0 5**P ; ~0.001. 'Significantly different from S-isomer group by the unpaired t-test, P 50.001.

however, the value measured for (5')-deoxycoformycin (Ki = 3.5 x lop9 M) was about 4,500-fold lower than expected. Spontaneous epimerization of inactive to active forms of the isomer (Schramm and Baker, '85) probably accounted for the discrepancy. With this caveat, extracts of the antimesometrial region were assayed for ADA enzymatic activity in untreated dams and 0.5 h after treatment with 5 mg/kg active or inactive isomers (Table 2). Decidual ADA activity was almost totally inhibited by (R)deoxycoformycin. The same dose of (S)-deoxycoformycin yielded an insignificant residual effect that could be explained by contamination with the active isomer. These results demonstrate stereoselectivity at the level of the putative biochemical target for pentostatin. Alterations in purine nucleoside pool sizes reflect the inhibition of A D A activity Extracts of the antimesometrial compartment (E/D-a) were analyzed by HPLC following maternal administration of (R)deoxycoformycin on day 7. This approach made it possible to monitor nucleosides in the environment of the embryo without the limitation of sample size. Steady-state pool sizes reflected a normal balance toward the products of ADA catalysis that was abruptly reversed upon exposure to a nearly complete embryolethal dosage of (R)-deoxycoformycin (Figs. 2, 3). The effect on adenosine and 2'-deoxyadenosine can be compared to other components in the cell extracts in Fig. 2. Adenosine, normally a minor component, was a major component in the antimesometrial compartment in dams treated with a nearly complete embryolethal dose of inhibitor. Little or no effect was obtained on components other than the four direct metabolites of ADA (Figs. 2, 3). The following

nucleoside profile (nmol/mg protein *SE) was observed in control samples: inosine (1.43 ? 0.07); adenosine (0.059 ? 0.039); and 2'-deoxyinosine (0.027 2 0.007). Basal 2'-deoxyadenosine was not detected at a sensitivity of 0.0005 nmol/mg protein. Exposure to 10 mg/kg (R)-deoxycoformycin raised adenosine and 2'-deoxyadenosine levels within 0.5 h to 3.71 nmol/mg protein and 0.419 nmol/mg protein, respectively, and almost completely depleted the region of inosine and 2'-deoxyinosine. These alterations were directly proportional to inhibitor dose (0.05-10 mg/kg) on a semilog scale (Fig. 4). To determine the duration of imbalance, nucleoside profiles were monitored 3 and 6 h after exposure to 0.05,0.5, and 5 mg/kg (R)deoxycoformycin on day 7. These points were selected to reveal differences preceding the earliest histological abnormalities in the embryo (Knudsen et al., '89). Adenine nucleoside concentrations were maximal (Cmax)0.5 h postinjection and subsequently declined during a 3- to 6-h recovery phase. Data for the antimesometrial compartment (EID-a) were compared with the mesometrial compartment (E/D-m) and maternal spleen (Table 3). In general, the rank order of recovery was tissue dependent (E/D-a > spleen > EID-m) and dose related (0.05 mg/kg > 0.5 mg/kg > 5 mg/kg) and differed according to the particular nucleoside (2'-deoxyadenosine > adenosine). Consequently, the effect of 0.5 mgikg was not apparent in the antimesometrial compartment at 3 h postinjection, whereas that of 5 mgkg lasted through 3 h. The correlation between embryolethal potential and 3-h pool sizes on day 7 raised the question of whether the same might exist at other stages of gestation. To address this question, dams were given 5 mg/kg (R)-

97

ADENOSINE METABOLISM IN MICE

Fig. 2. HPLC analysis of the antimesometrial compartment on day 7 of gestation. Reversed-phase HPLC analysis of EID-a harvested from (A) an untreated litter, and (B) 0.5 h following maternal administration of 10 mg/kg (R)-deoxycoformycin.Peaks were identified as inosine (l), 2'-deoxyinosine (2), adenosine (3), and 2'-deoxyadenosine(4).

deoxycoformycin as a single dose during days 6-9 of gestation, and nucleoside profiles were monitored in the antimesometrial compartment at 3 h postinjection (Table 4). Basal levels of adenosine were highest on day 6 (0.15 nmol/mg protein) and lowest on day 9 (0.04 nmol/mg protein), and 2'-deoxyadenosine was not normally detected. After treatment, adenosine was elevated more than 30-fold over normal on days 6 and 7, and the response weakened to less than 7fold on days 8 and 9.2'-Deoxyadenosine was elevated more than 200-fold over the detection limit and the response weakened to 78fold on day 9. Therefore, embryolethal potential was not simply related to a 3-h pool expansion. Finally, E/D were explanted on day 7 and incubated in serum-free culture exposing them t o various concentrations of (R)-deoxycoformycin. In the absence of inhibitor, adenosine levels remained stable in culture (0.15 nmol/mg protein in the whole E/D) through 1.5 h and 2'-deoxyadenosine was not detected. An accumulation of both adenine nucleosides occurred in the presence of (R)-deoxycoformycin(Fig. 5). The effect was

concentration dependent (1-50 pM), linear through 1.5 h, and comparable to what was observed in vivo. For example, adenosine levels were raised to 3.92 nmol/mg protein and 2'-deoxyadenosine to 0.205 nmolimg protein when E/D were cultured for 1 h in the presence of 50 p M (R)-deoxycoformycin. These results suggest an intrinsic capacity to generate adenosine and 2'-deoxyadenosine. DISCUSSION

Results from the present study confirm that (R)-deoxycoformycinis embryolethal in mice during the early postimplantation period. In addition, the present study demonstrates (1) an apparent dose-response relationship, with no significant embryolethality a t 0.5 mg/kg and nearly complete litter resorptions at 10 mg/kg; (2) a sensitive period of days 7-8 of gestation; (3) embryotoxicity stereospecific for the 8R-isomer; and (4)a dose-related increase of adenosine and 2'-deoxyadenosine pool sizes in the antimesometrial compartment of the embryo/decidual unit (E/D). These results are

86

99

ADENOSINE METABOLISM IN MICE

AR

AdR HR

70

900 0.4

-

HdR 1.o

\ \ \

h

.-0

60-

3 -

0.8

P c

0.3

E =-

B$

- 600 0.6

40-

v c

0.2

I?

z

300.4

4

- 300

w

200.1

!5!

%z

0.2 10-

0

0.0 .01

.1

1

10

100

DOSE ( m g k g ) Fig. 4. Effects of (R)-deoxycoformycin on ADA metabolism in the antimesometrial compartment. Results from HPLC analysis 0.5 h after maternal administration of (R)-deoxycoformycin on day 7. Nucleoside levels were averaged from two separate experiments and expressed as a ratio of treatedkontrol. Control values (nmolimg protein) were: adenosine (AR, 0.06); 2’-deoxyadenosine (AdR, 0; not detected a t a sensitivity of 0.0005 nmolimg protein); inosine (HR, A; 1.43); and 2’-deoxyinosine (HdR, A; 0.027).

.;

2

Dose (mgikg) Control 0.05 mgkg 0.5 mgikg

5.0 mg/kg

Interval (h)

0.5 3 6 0.5 3 6 0.5 3 6

TABLE 3. Duration of adenine nucleoside pool expansion on day 7l Adenosine (nmolimg protein) 2’-Deoxyadenosine (nmolimg protein) EID-a 0.06 i 0.04 0.32 f 0.02 0.13 ? 0.03 0.23 f 0.03 2.17 f 0.98 0.45 ? 0.21 0.36 2 0.01 2.50 f 0.79 2.06 i 0.41 0.69 i 0.18

EID-m 0.28 rt 0.08 1.91 rt 0.17 1.28 2 0.41 1.62 2 0.03 3.60 2 0.43 2.29 i 0.06 2.75 2 0.89 4.05 rt 0.99 3.12 i 0.10 4.72 i 1.36

Spleen 0.08 ? 0.01 1.71 2 0.05 1.10 2 0.51 0.50 i 0.15 3.56 ? 0.01 2.39 2 0.35 1.99 2 0.45

EID-a ND3 ND ND ND 0.107 5 0.01 ND ND 0.337 5 0.04 0.114 2 0.02 0.006 2 0.01

EID-m ND ND ND ND 0.123 2 0.08 0.023 ? 0.02 0.003 ? 0.00 0.296 ? 0.05 0.155 2 0.06 0.053 2 0.01

Spleen 0.013 t 0.00 -

0.030 ? 0.01 0.026 ? 0.04 ND 0.236 i 0.05 0.150 t 0.07 0.128 t 0.07

‘Analysis of antimesometrial (E/D-a) and mesometrial (E/D-m)compartments,and of maternal spleen. ’Hours elapsed since treatment at 0900 h; each data point is from two independent litters (mean ?SE). 3ND, not detected (detection limit 0.0005 nmol/mg protein).

apparent at the level of decidual ADA. The high structural specificity for R-stereochemistry at the C-8 position of the diazepine ring points to, but does not prove, a primary mechanistic involvement of ADA

inhibition in the events underlying early resorption induced by pentostatin. The effect of (R)-deoxycoformycinon embryo viability became significant at 3 mgl kg, although the trend of early resorptions

100

Stage of treatment' Day 6 Day 7 Day 8 Day 9

T.B. KNUDSEN ET AL.

TABLE 4 . Stage-related expansion of antimesometrial nucleoside pools 3 h postinjection' Adenosine (nmol/mg protein) 2'-Deoxyadenosine (nmol/mg protein) Control (R)-Deoxycofonnycin Control (R)-Deoxycoformycin % resorptions4 0.15 & 0.03 8.43 ? 2.32 ( 5 6 . 2 ~ ) ND3 0.134 2 0.134 (268 x ) 4.4% 0.06 4 0.04 2.06 ? 0.41 (34.3 x ) ND 0.114 +- 0.015 (228 x ) 85.7% 0.11 2 0.03 0.48 ? 0.18 ( 4 . 4 ~ ) ND 0.102 t 0.079 (204 x ) 70.1% 0.04 +- 0.01 0.27 +- 0.07 (6.8x ) ND 0.039 t 0.010 (78 x ) 5.4%

'HPLC analysis of nucleoside extracts for two litters per data point (mean 5SEL 'Values for day 7 from Table 3. 3ND, not detected ( ~ 0 . 0 0 0 5nmolimg protein). 4Data from Fig. 1;background incidence, 5%.

AdR

AR

AdR

-

- 600

25-

- 500

- 500

; 20-

- 400

- 400

15-

- 300

- 300

10-

- 200

- 200

5-

- 100

- 100

30

.0 c ,m 9 c

r 600

3iri

-9? c

Y w

>

!kl u

0

8 3

0 3

Z

0

1

'

0

and fetal weight reduction was apparent following 0.5- to 1-mgkg treatment on day 7. It is therefore reasonable to consider the lower dosages as bordering a "threshold" response with respect to embryotoxicity. Absence of significant embryotoxicity below 3 mgtkg is in agreement with other studies using pentostatin in the pregnant mouse (Leubke et al., '87). The effects became strong at 5 mglkg, and nearly complete embryolethality was obtained a t 10 mglkg.

-

0

Lymphocytopenia and reduced hematocrit have been observed 1 or 2 days after administration of 10 mglkg (R)-deoxycoformycin to nonpregnant adult female mice (Smith et al., '80; Ratech et al, '85). Superimposition of embryolethality and potential maternal toxicity renders unlikely the possibility that the inhibitor interfered with processes specific to the embryo. Rather, the metabolic consequences of ADA inhibition may be the same for the embryo as for lymphocytes and

ADENOSINE METABOLISM IN MICE

erythrocytes (Carson et al., '78; Siaw et al., '80; Smith et al., '80; Ratech et al., '81; Hershfield et al., '82). It will be of interest to determine why the embryo is sensitive to (R)-deoxycoformycin during a relatively narrow window of development. Embryolethality as a consequence of the inhibition of ADA metabolism could arise from hypoxanthine starvation-the likely outcome of inosine deprivation, or from the accumulation of endogenous adenosine and 2'-deoxyadenosine. Studies with rat embryos in culture have shown that relatively normal organogenesis can occur in the absence of measurable quantities of salvageable purine bases or nucleosides (Rowe and McEwen, '83). Therefore, hypoxanthine starvation of the embryo is not a compelling model for the embryolethal effect of pentostatin. A more tenable hypothesis is the build-up of endogenous substrate nucleosides in the antimesometrial compartment. Adenosine (Spindle and Pedersen, '771, 2'deoxyadenosine (Karnofsky and Lacon, '61), and adenosine agonists (Clark et al., '87) have been clearly shown to effect embryonic development adversely by mechanisms that are not currently known. The present study has demonstrated that physiologic adenosine concentration is low in the antimesometrial compartment (0.06 nmol/ mg protein), whereas 2 '-deoxyadenosine was not detected at a sensitivity of 0.0005 nmol/mg protein. Exposure to an embryolethal regimen of (R)-deoxycoformycin resulted in substantial and rapid accumulation of both nucleosides. Two important correlates of embryolethal potential were the peak tissue concentration (Cmax)of the nucleosides reached at 0.5 h postinjection and the duration of the subsequent recovery phase. Following 5 mg/kg, for example, C,, was 2.50 nmol/mg protein for adenosine (42-fold elevation) and 0.337 nmol/mg protein for 2'-deoxyadenosine (674-fold elevation). The rank order of recovery directly reflected the tissue levels of normal ADA expression (E/D-a > spleen > E/D-m) (Knudsen et al., '88; Chinsky et al., '90; Knudsen et al., '91); the relative substrate affinity for ADA (2'-deoxyadenosine > adenosine) (Agarwal et al., '77); and the pharmacokinetics of the inhibitor (urinary excretion accounts for over 90%of the dose administered within 2 h) (McConnell et al., '78). These results are consistent with normally rapid turnover of potentially em-

101

bryotoxic quantities of adenine nucleosides in the early gestation site. Are adenosine and 2'-deoxyadenosine generated locally in the E/D, or does their accumulation reflect a systemic increase in the maternal plasma? Kuttesch and Nelson ('82) examined plasma nucleoside levels in the adult mouse and found that basal concentrations of adenosine (13 p M ) and 2'deoxyadenosine (