Studies on aromatase inhibition with 4-androstene-3,6,17-trione: Its 3β-reduction and time-dependent irreversible binding to aromatase with human placental microsomes

J. steroid Eiochem. Vol. 28, No. 3, pp. 337-344,

1987

0022-4731/87

Printed in Great Britain. All rights reserved

$3.00 + 0.00

Copyright 0 1987Pergamon Journals Ltd

STUDIES ON AROMATASE INHIBITION WITH 4-ANDROSTENE-3,6,17-TRIONE: ITS 3/bREDUCTION AND TIME-DEPENDENT IRREVERSIBLE BINDING TO AROMATASE WITH HUMAN PLACENTAL MICROSOMES MITSUTERU NUMAZAWA*, MASACHIKA TSUII and AYAKO MUTSUMI Tohoku College of Pharmacy, 4-1 Komatsushima-4-chome, Sendai 983, Japan (Received 20 January 1987) Summary-The metabolism of 4-androstene-3,6,17-trione (AT), previously described as a suicide substrate for aromatase, and its irreversible binding to aromatase were studied by using human placental microsomes. AT was rapidly converted into 3/?-reduced metabolite (3-OHAT) with an enzyme other than aromatase in the microsomes in the presence of NADPH under either aerobic or anaerobic conditions. The conversion was efficiently prevented by a steroid Sa-reductase inhibitor. 3-OHAT was characterized as a competitive (K, = 6.5 PM) and irreversible inhibitor of aromatase. Both W-labeled AT and 3-OHAT were demonstrated to be irreversibly bound to aromatase probably through a sulfur atom of the enzyme in time-dependent manners in the presence of NADPH, being accompanied with time-dependent losses of the enzyme activity. It was shown that the process of an apparent time-dependent loss of aromatase activity caused by AT even under conditions allowing its 38-reduction should principally depend on the action of the parent inhibitor AT itself and not on that of the metabolite 3-OHAT.

been definitely clarified so far because the radioactive AT was not available. Our interest is focussed on further study on the aromatase inhibition using 14Clabeled AT and human placental microsomes. We report here results which demonstrate that AT is readily converted by the action of a non-aromatase enzyme into 3/I-hydroxy-4-androstene-6,17-dione (3-OHAT) which is a competitive and irreversible inhibitor of aromatase and that both AT and 3-OHAT are irreversibly bound to aromatase in a time-dependent manner.

INTRODUCTION The biosynthesis of estrogens involves three sequential hydroxylations of the androgen precursors which are mediated by an enzyme complex referred as aromatase [l-3]. Compounds affecting the aromatase inhibition have potential applications in the treatment of advanced estrogen-dependent mammary carcinoma and in the modulation of reproductive process [4]. A large number of steroids having 4-en-3-one system are known to be substrates or active-site directed inhibitors for the aromatase enzyme system [5-191. 4-Androstene-3,6,17-trione (AT) has found to be a potent competitive inhibitor of aromatase with an apparent K, of 0.43pM [14] or 2.5 PM [19] and is widely used to delineate the role of estrogens in the

EXPERIMENTAL Materials

regulation of endocrine process [20]. Recent study by Covey and Hood [14] showed that AT causes a time-dependent loss of aromatase activity in the presence of NADPH in human placental microsomes suggesting that AT acts as a suicide substrate of the enzyme. On the other hand, we [21] reported that AT is chemically reactive with thiols such as IV-acetyl cysteine and cysteine under mild conditions in a 1,4-Michael addition manner to give the 4a-thio ether derivatives, and suggested that the 1,Caddition reaction at the active site of aromatase may be responsible for a decrease of the enzyme activity. The dynamic aspects of aromatase inhibition with AT thus has not

[4-‘4C]3p-Hydroxy-5-androsten-17-one [DHEA] (52 mCi/mmol) and [ 1,2-3H] 4-androstene-3,17-dione [AD] (46 Ci/mmol) were purchased from New England Nuclear (Boston, MA). NADPH was obtained from Kohjin Co. Ltd (Tokyo, Japan). N,N-diethyl 4-methyl-3-oxo-4-aza-5a-androstane-l7b-carboxamide (4-MA) was kindly provided by Dr Rasmusson of Merk Sharp & Dohme Research Laboratories (Rahway, NJ). Anti-human placental aromatase monoclonal antibody was donated by Dr Osawa of Medical Foundation of Buffalo (Buffalo, NY). AT [22], 3-OHAT [23], and 6a-bromo-4-androstene3,17-dione (6-BrAD) [24] were synthesized according to known methods. Synthesis of [4-‘“C]A T

*To whom correspondence

should be addressed.

To a solution of carrier free [4-14C]DHEA (110 pg, 0.385 pmol; 20 PCi) in 400 PL of acetone was added 337

338

MITSUTERU NUMAZAWA et al.

100 ~1 of 8N Cr& solution and the reaction mixture was stirred in ice bath for 10min. After this time, saturated NaHCO, solution (500 ~1) was added to the mixture and the product was extracted with ethyl acetate (2 ml x 2) which was washed with saturated NaHCO, and NaCl solutions and evaporated to dryness under nitrogen. The residue was subjected to thin-layer chromatography [TLC] (silica gel 60 F,,, Merk AG: solvent; hexaneeethyl acetate, 2:3, v/v) and the region (R,, 0.60) corresponding to AT was scrapped off and eluted with ethyl acetate. The product was again purified by TLC as above to give [4-r4C]AT (radiochemical yield, 50%), of which radiochemical purity was obtained to be 98% by TLC analysis and also established within the statistical limits of experiment error (2-2.5%) by reverse isotope dilution analysis (data not shown). Enzyme preparation Human term placental microsomes (particles sedimenting at 105,OOOg for 60 min) were obtained as described by Ryan [25]. They were washed twice with 0.5 mM dithiothreitol solution, lyophilized, and stored at -20°C. No loss of activity occurred over the period of the study. Standard aromatase assay procedure Aromatase activity was measured essentially based upon the original assay of Thompson and Siiteri [I]. This assay quantitates the production of ‘H,O released from [ 1,2-3H]AD after aromatization. Aromatase activity study was carried out in 0.067 M phosphate buffer, pH 7.5, at a final incubation volume of 0.5 ml. The incubation mixture contained 180 PM of NADPH, 2 PM of [1,2-3H]AD alone or with AT, 3-OHAT or other compounds in 25 ~1 of methanol, and 150 ~_lgof the lyophilized microsomal protein. Incubations were performed at 37°C for 20min in the air and terminated by the addition of 3 ml of CHCI,, followed by vortexing for 40 s. After centrifugated at 2,500 rpm for 10 min, aliquots (0.3 ml) were removed from the water phase and added to ~intillation cocktail for the dete~ination of ‘H,O production. Isolation of [4-‘4C]3-OHAT and non-labeled 3-OHAT A mixture of [4-i4C]AT (0.26 PCi, 5 nmol) or AT (5 nmol), NADPH (0.95 pmol), and human placental microsomes (1.5 mg protein) in 1.0 ml of 0.067 M phosphate buffer, pH 7.5, was incubated at 37°C for 20 min under aerobic conditions. After this time, the mixture was extracted with CHCl, (2 ml x 3) which was dried over Na,SO, and then evaporated under Nz gas. The residues obtained from 10 incubations were combined and subjected to TLC (hexane-ethyl acetate = 2: 3, v/v) and the region (&, 0.30) corresponding to 3-OHAT was scrapped off and eluted with ethyl acetate. The crude “C-labeled 3-OHAT or non-labeled 3-OHAT was further purified by highperformance liquid chromatography (HPLC) [a

Waters Associates ALC/GPC 244 liquid chromatograph equipped with a U6K injector and a JASCO UVIDEC-loo-III U.V. detector (at 250 nm); column, Radial Pak C,, (10 cm x 0.8 i.d. cm); solvent, methanol-water, 3 : 2, v/v (1 .Oml/min)]. The peak corresponding to 3-OHAT (retention time, 9.5 min, where that of AT was 1l.Omin) was collected and used for the incubation experiment or gas chromatography-mass spectrometric analysis. Gas chromatography-mass

spectrometry (GC-MS)

The product 3-OHAT was derivatized to its trimethylsilyl ether with bis-t~methylsilylt~fluoroacetoamide in acetonitrile and then subjected to GC-MS analysis [a Shimadzu LKB 9000B gas ~hromatograph-mass spectrometer; column, 2% SE30 (22O”Q 1.Om; carrier gas, He (30 mljmin)]. Covalent labeling of aromatase by [4-‘“C]AT and [4-‘4C]3-OHAT A typical experiment was carried out as follows: [4-14C]AT or [4-‘4C]3-OHAT (0.054 PCi, 1.Onmol) alone or with other compounds was added to microfuge tubes, the solvent was evaporated under nitrogen, and the residue was redissolved in 25 ~1 of methanol. Labeling was performed by the addition of human placental microsomes (1 .Omg of protein) and NADPH (0.60 mM) in 1.0 ml of 0.067 M phosphate buffer, pH7.5. The mixture was incubated at 37°C for an appropriate time in the air. After 50 ~1 aliquots of the mixture were removed for the determination of aromatase activity, 3 ml of CHCI, was added to the residue to stop the incubation. Free steroid fraction was obtained by the CHCI, extraction (3 ml x 3) and to the remaining aqueous phase was added 4ml of methanol and allowed to stand at 4°C for 2 h. After this time, the mixture was centrifugated at 2,500 rpm for 20 min to give the pellets, which was then washed with methanol (4 ml x 4) and CHCI, (2 ml x 2) and then desiccated under vacuum. The dried pellets (the protein-bound metabolite fraction) were dissolved in 400 ~1 of 0.2 N NaOH solution and heated at 80°C for 30 min. The solution was neutralized by 1 N acetic acid and aliquots (400~1) were transferred to scintillation cocktail for determination of the protein binding of 14C-labeled AT and 3-OHAT. Quantitative determination 3-OHAT from AT

of

the formation

of

A typical incubation was carried out in 0.067 M phosphate buffer, pH 7.5, at a final volume of 1.Oml. The mixture consisting of [4-14C]AT (0.054 PCi, 1.Onmol) in 25 ~1 of methanol, NADPH (0.6 mM), and 1.Omg protein of the placental microsomes was incubated for an appropriate time in the presence of other steroids or anti-aromatase antibody. The incubation was te~inated by the addition of 3 ml of CHCl, and then free steroids were extracted with the same solvent (3 ml x 3). The combined organic layer was evaporated under nitrogen and the residue

4-Androstene-3,6,17-trione

and aromatase inhibition

339

was subjected to TLC (hexane+thyl acetate, 2:3, v/v). The region (R,, 0.30) corresponding to 3-OHAT was scrapped off and its radioactivity was measured by liquid scintillation counter. The recovery rate of 3-OHAT during the isolation procedure was 91 * 2%(n = 4). Treatment of the protein -bound metabolite Ni

with Raney

The protein-bound metabolites (1.25 x lo4 and 1.02 x lo4 dpm) obtained from “C-labeled AT and

3-OHAT were dissolved and neutralized as above. To the solutions (550~1) was added a large excess of deactivated Raney Ni (suspended in water) and the mixtures were heated under reflux with stirring for 6 h and then extracted with ethyl acetate (2ml x 4), respectively. Portions of the organic phases were subjected to the radioactivity counting. RESULTS

Initially, ability of AT to cause a time-dependent decrease in aromatase activity in the presence of NADPH in human placental microsomes was examined according to Covey and Hood [14], and the essentially similar results as the previous ones were also obtained in our hands (see Fig. 9). To further explore dynamic aspects of the aromatase inhibition with AT especially in relation to its metabolism and irreversible binding to aromatase, we synthesized carrier-free [4-14C]AT by treatment of [4-14C]DHEA with 8N CrO, in 50% yield principally according to Bellino et al. [ 131.The i4C-labeled AT was incubated with the placental microsomes with and without NADPH. An irreversible binding of the radioactive substrate to the microsomal protein and aromatase activity were simultaneously determined. The results with NADPH show that a time-dependent increase of the production of the protein-bound metabolite involves a time-dependent first-order decrease of the activity as given in Fig. 2. On the other hand, no significant time-dependent change of the enzyme activity and a slight increase of the protein binding of the radioactivity were observed without NADPH. An employment of a prolonged incubation time (60 min) increased the formation of the protein-bound metabolite up to be approx. 2-fold compared to that with a IO-min incubation in the presence of NADPH, while no significant increase of the protein-binding of radioactive AT was observed without the cofactor.

01 0’ Incubation

AT

0

SOHAT

Fig. 1. Structures of AT and 3-OHAT.

1

10

time (min)

Fig. 2. The inactivation of aromatase (A) and the formation of the protein-bound metabolite (B) by incubation of [W]AT with human placental microsomes. [14c]AT (0.054 PCi, 1.0 nmol) was incubated with 1.0 mg protein of the microsomes in the presence or absence of NADPH in the air. A portion of the incubation mixture was taken for the assay of aromatase activity and the remaining was submitted to the determination of the protein-bound metabolite as described in Experimental. Each point represents mean &-SD (n = 4). l with NADPH, 0 without NADPH.

As shown in Table 1, the irreversible binding was markedly prevented by the addition of reversible natural substrate AD and anti-human placental aromatase monoclonal antibody equivalent to 150 p g of IgG which caused 75% inhibition of aromatase activTable 1. Protection effect of steroids, anti-aromatase antibody and bromoacetic acid on the irreversible binding of [?Z]AT to human nlacental microsomes

Treatment None (native microsomes) without NADPH anaerobic conditions* boiled microsomes, 95”C, 10 min AD (3.5 yM) anti-aromatase antibody (lSO/cg of IgG) bromoacetic acid (72 IBM) (144uMl

oG@ Ho& 0

I

5

Protein-bound metabolite, %

Aromatase activity, %

100

100

22 20 35 36

1 1 1

25

25

83 80

98 99

Values are expressed as relative ones and means of three determinations. [?JAT (O.O54pCi, l.OpM) was incubated with 1.Omg protein of human placental microsomes in the presence of NADPH at 37°C for 10 min with AD, anti-aromatase antibody and bromoacetic acid. The protein-bound metabolite was isolated and its radioactivity was determined as described in Experimental. Aromatase activity was determined by standard assay procedure. *The incubation was carried out in Thurnberg tube under nitrogen.

MIT~UTERU NU~~AZAWA et al.

340

Table 2. Reverse isotope dilution analysis of [“C]3-OHAT formed from [“C]AT with human placental microsomes Crystallization

Crystalline

No.

Solvent

Weight (ma)

I

Acetone Acetone CHCI, CHCI, Acetone

25 17 14 6 4

2 3 4 5

Sp. act. (dnm/ma) 1288 1248 1198 1231 1220

3b-OHAT (13OOdpm) isolated from the incubation mixture with [?]AT by TLC was mixed with 33 mg of non-labeled 3-OHAT and then subjected to repeated crystallization.

ity in the absence of AT, leading to the non-specific level of the radioactivity binding obtained without NADPH or O2 or with boiled microsomes. On the other hand, bromoacetic acid, a non-specific sulfhydry1 reactive reagent, lowered the binding by approx. 20%, in which any detectable changes of the enzyme activity were not observed. To further validate AT to be a suicide substrate of aromatase, we then explored its metabolism with the placental microsomes. TLC analysis of CHCI, extract obtained by incubation of [14C]AT with the placental microsomes and NADPH showed two radioactive peaks corresponding to AT and the polar 3/l-reduced product, 3-OHAT. The structure of the product was determined by reverse isotope dilution method and GC-MS analysis. Repeated crystallization of the radioactive metabolite diluted with non-labeled 3-OHAT showed a constant specific radioactivity, by which more than 94% of the metabolite was charac-

terized as 3-OHAT (Table 2). The retention time (8.0 mm) and fragmentation pattern in the MS spectrum of the trimethylsilyl derivative of the nonlabeled metabolite which was isolated by using TLC and reversed-phase HPLC were consistent with those of the authentic 3-OHAT (Fig. 3). The formation of estrogen product having an 0x0 function at C-6 was also not detected as suggested by the previous report [ 141. AT was rapidly converted into 3p-hydroxy derivative by incubation with the placental microsomes in the presence of NADPH, in which about 55% of AT was reduced to 3-OHAT by IO-min incubation (Fig. 4). This biotransformation was not affected by using anaerobic conditions instead of aerobic ones and by the addition of anti-aromatase antibody but decreased by the additions of AD and 6-BrAD, an active-site-directed irreversible inhibitor of aromatase [13], in a dose-dependent manner (Table 3). Unexpectedly, 4-MA, which is known as a potential steroidal Sa-reductase inhibitor [26,27] and has no significant inhibitory effect on aromatase activity, effectively prevented the carbonyl reduction at C-3 and 5.10 PM of it was enough for a complete inhibition of the biotransformation. To know that whether or not the 3/?-hydroxy compound 3-OHAT is involved in an apparent timedependent aromatase inactivation caused by the parent steroid AT, aromatase inhibition studies with 3-OHAT were then performed. Lineweaver-Burk and Dixon plots showed 3-OHAT to be a competitive inhibitor with an apparent K, of 6.5 PM (Fig. 5) [for

-1 6.9

0

100

Fig. 3. MS spectrum of the trimethylsilyl

300

200

derivative of the metabohte 3-OHAT incubation of AT with the microsomes.

obtained

from AT by the

4-Androstene-3,6,17-trione

0

I/

20

5 Incubation

time

(mid

Fig. 4. The formation of 3-OHAT by the incubation of AT with the microsomes. [‘%ZjAT (0.054 ptci, 1 nmol) was incubated with 1.0mg protein of the microsomes in the air. 3-OHAT formed was quantified by TLC as described in

Experimental. Each point represents the mean of duplicate determinations.

AT the apparent Ki was 1.25 p M in our system, data not shown], showing 3-OHAT to be a less potent inhibitor than AT. The reverse conversion of 3-OHAT into 3-keto compound AT was not detected during the incubation. Next, 3-OHAT was tested for its ability to cause a time-dependent decrease in aromatase activity with and without NADPH. The enzyme activity apparently decreased in a time-dependent manner that follows first-order kinetics in the presence of NADPH as shown in Fig. 6 and the cofactor was essential for the activity loss. Incubation of 3-OHAT in the presence of the substrate AD showed an increase in the t,,, of inactivation compared to that conducted in the presence of the inhibitor alone (Fig. 7); the enzyme activity was completely recovered. On the other hand, a nucleophile, L-cysteine, had no significant effect on the t,,r of inactivation caused by 3-OHAT.

and aromatase

inhibition

341

Incubation experiments using [4-i4C]3-OHAT, which was obtained by the incubation of [4-r4C]AT with the placental microsomes and NADPH, produced the protein-bound metabolite. Figure 8 shows that a time-dependent increase of its formation is accompanied with an apparent time-dependent decrease of aromatase activity and its formation was markedly prevented by the addition of AD down to almost the non-specific level. TLC analysis of the free steroid fraction (the CHCI, extract) gave only one radioactive peak corresponding to the substrate 3-OHAT. To know that whether or not the metabolite 3-OHAT is involved in the apparent time-dependent loss of aromatase activity caused by the parent compound AT, it was necessary to examine a timedependency of the enzyme deactivation with AT under conditions preventing the conversion of AT into 3-OHAT during the incubation. As shown in Fig. 9, the time-dependent loss caused by 0.2 and 0.8pM of AT was not significantly changed by the addition of 2.4 PM of 4-MA. This indicates that the involvement of 3-OHAT in the apparent timedependent loss caused by AT could be neglected in the experiments with a brief incubation time (at least 10 min). Finally, to get some information about the structures of the protein-bound metabolites formed from “C-labeled AT and 3-OHAT, the radioactive Table

3. Effect of steroids and anti-aromatase antibody conversion of AT into 3fi-OHAT

Treatment

on the

3-OHAT

Control anaerobic conditions* without NADPH AD (1.96pM) (3.92 PM) 6-BrAD (0.23 pM) (0.46 PM) (1.96pM) 4-MA (5.10 PM) Anti-aromatase antibody

(150 fig of IgG)

Values are expressed as

relative

minations. ‘The incubation

100 102 3 46 37 68 50 40 3 110

ones and means

was carried out in Thumberg

0.10

of three deter-

tube under nitrogen.

q

l/V i

/

-1.0

-0.5

0

0.5

l/151.

Fig. 5. Lineweaver-Burk for 3-OHAT. Velocity

PM

0.05

I

, -3

n

-4

0

4

0

I

I

8

16

III.

15.1 3.4 5.0 PM PM JAM

JJM

plot (A) and Dixon plot to determine the apparent inhibition constant K, (B) (v) of AD aromatization is expressed as nmol/min/mg protein. Each point represents the mean of duplicate determinations.

342

MITSUTERUNIJMAZAWA ef al.

4000 2

40 I 5

0

I 20

I 10

15.8,uM I 30

Incubationtime (min)

Fig. 6. Time-course for decrease in aromatase. activity by 3-OHAT. Each point represents the mean of duplicate determinations.

B m _ 3000 u tEf. :a P 2000 I.; 5 t:

1000 -

01 protein-bound metabolites were subjected to the desulfurization reaction with Raney Ni. Approximately 80% of their radioactivities were released into the medium as free steroids by the treatment, respectively. DISCUSSION

These results demonstrate that AT is rapidly metabolized to the 3fl-hydroxy derivative 3-OHAT by the action of a non-aromatase enzyme in human placental microsomes, and that the both steroids AT and 3-OHAT act as mechanism-based irreversible inhibitors of aromatase. However, an apparent timedependent loss of aromatase activity caused by AT should be principally responsible for the metabolic

2 t_ .z i

r f ‘Z :: 91 e

50-

z E z

. 20 _: T 0

1 5

1 40

I 20

I 40

Incubation time (minf

Fig. 7. Protection of the 3-OHAT inactivation of aromatase by substrate and cysteine. 0 control, 0 controf + AD f84ApM) and 3-OHAT (15.8pM) + AD (84.4~M~, A 3-OHAT (15.8 pM), A 3-OHAT (15.8PM) + cysteine (7.9 PM).

I

I

’

0 Incubation

5 time

(min)

10

Fig, 8. The inactivation of aromatase (A) and the formation of the protein-bound metabolite (B) by incubation of [14C]3-OHAT with the microsomes. [‘%Z]3-OHAT (0.054 @.Zi, l.OpM) was incubated with l.Omg protein of the microsomes and then treated similarly as Fig. 2, Each point represents the mean of duplicate determinations. l with NADPH, 0 without NADPH, A with NADPH + AD (3.5 j.iM).

activation of AT itself by aromatase and not for that of 3-OHAT produced during the incubation. Several lines of evidence have been presented here. (a) 3-OHAT, of which formation is not influenced by anti-aromatase monoclonal antibody but almost completely prevented by a potent steroid 5areductase inhibitor 4-MA (Table 3), decreases aromatase activity in a competitive manner (Fig. 5). (b) A time-dependent first-order loss of the enzyme activity is caused by 3-OHAT in the presence of NADPH (Fig. 6). (c) The time-dependent deactivation process of aromatase with AT or 3-OHAT requires the irreversible covalent binding of their metabolites to the enzyme protein (Figs. 2 and 8). (d) AD, reversible substrate for aromatase, protects the enzyme from the time-dependent inactivation with AT and 3-OHAT (Fig. 7). (e) The time-dependent process with AT is not significantly affected by the addition of 4-MA (Fig. 9). The K, value (1.25 FM) of AT obtained here is somewhat different from those (0.43 [14] and 2.5 [19] PM) reported previously. This would be due to the different method of enzyme preparation. However, AT has much higher affinity to aromatase compared to the 3-hydroxy product 3-OHAT (Ki: 1.25 vs 6.50 PM). This supports that no significant difference in the time-dependent deactivation process

4-Androstene-3,6,17-trione and aromatase inhibition

40 1 0

1 1

. I 8

I 4

1 2 Incubation

time tmin)

Fig. 9. Time-course for decrease in aromatase activity by AT in the presence or absence of 4-MA. Each point represents the mean of duplicate determinations. l control, A 4-MA (2.4pM), 0 AT (0.2rM), n AT (0.2pM)+4-MA (2.4pM), V AT (0.8pM), V AT (0.8pM) + 4-MA (2.4 PM).

of aromatase activity caused by AT is observed between conditions with and without 4-MA (Fig. 9). Activation of a suicide substrate to its reactive intermediate occurs at the same enzymatic site responsible for catalysis of the normal substrates [28,29]. Suicide inhibition requires that once the suicide substrate is activated, it immediately binds in a covalent fashion at the active site without first diffusing into the incubation medium. If the activated intermediate diffuses out of the active site before it inactivates the enzyme, nucleophiles present in the incubation could react with the electrophilic intermediate and slow the time-dependent loss of the activity. For the suicide substrates AT and 3-OHAT, the formations of their oxygenated metabolites were not detected in the free steroid fraction during the incubations, and the presence of a nucleophile cysteine in the incubation did not protect aromatase from a time-dependent loss of the activity (Fig. 7). The similar results have been obtained in aromatase inactivation with suicide substrates having a 10/Ithiol group [12]. Raney Ni desulfurizations of the protein-bound metabolites produced from AT and 3-OHAT released steroid aglycones in the medium, of which major ones are more polar than the corresponding substrates, respectively, indicating that the activated (oxygenated) intermediate probably reacts with a sulfydryl group(s) at the active site of aromatase before reversible diffusion out of the site, although some portions of the released aglycones would be responsible for the non-specific binding of the substrates. We can not yet identify the struture of the aglycones. However, reactions of the 4,5-epoxide derivatives of AT and 3-OHAT with N-acetyl cysteine in aqueous methanol gave the steroid-amino

343

acid complexes, respectively, in high yields (reported elsewhere). Furthermore, the 4,5-epoxide of AD has previously been proposed to be an intermediate of AD aromatization [30]. These facts along with the previous result suggest that the epoxides may be possible reactive intermediates of formation of the protein-bound metabolites. The See-reduced product of AT, Sa-androstane3,6,17-trione, which is an isomer of 3-OHAT, was not detected as the metabolite by the incubation experiment with the placental microsomes, and it was found that the Scr-reduced compound is not epimerized to 3-OHAT under the conditions for its derivatization to the trimethylsilyl ether and even under drastic conditions with sodium hydroxide and hydrochloric acid. These unambiguously confirm that 3-OHAT is not the epimerized product of the SE-reduced compound but the metabolite enzymatically produced by the direct 3/I-reduction of a C-3 carbonyl group of AT. Although it is not clear that whatever enzyme is involved in the 3/I-reduction, the enzyme has affinity for AD and is sensitive to 6-BrAD, a SH-reactive reagent, and interestingly, 4-MA, a Sa-reductase inhibitor, also strongly inhibits the reduction (Table 3). It should be noted that relatively high concentration (2.4pM) of 4-MA causes a slight but significant inhibition of aromatase activity (Fig. 9). Bellino and Osawa [31] have suggested the existence of at least 2 classes of SH-reactive compounds: compound such as bromoacetic acid which require relatively high concentration to inhibit aromatase and compound such as 6-BrAD to which aromatase is much more sensitive. Relatively low concentrations of bromoacetic acid selectively prevented the nonspecific protein binding of AT without a significant change of aromatase activity (Table 1). This along with our previous results that AT reacts with a thiol to give a 1,Cadduct [21] strongly suggest that the non-specific binding may in part depend on the 1,6addition reactions with a thiol of the microsomal protein not involved in the catalytic activity. Brodie’s group [5, 321 has demonstrated the effectiveness of a number of analogues of AD as inhibitors of aromatase, which have the 4-en-3-one conjugated system. It is noteworthy that 3-OHAT is a fairly potent in vitro inhibitor of aromatase despite the disruption of the 4-en-3-one system. To our knowledge, this is the first report that a steroid without 3-keto group can be an effective aromatase inhibitor. Our current results may be helpful to develop a new type of aromatase inhibitors. So far, a number of mechanism-based irreversible inhibitors of aromatase have been reported by many investigators. The present result is suggestive that since a lot of enzymes are included in the placental microsomes, the involvement of enzymes other than aromatase should not be neglected in the metabolic activation of inhibitors using the crude microsomal preparation as an enzyme source.

MITSUTERU NUMAZAWAet al.

344 Acknowledgements-We

are extremely grateful to Dr G. H. Rasmusson of Merck Sharp & Dohme for the generous gift of 4-MA and his comments and to Dr Y. C&awa of Medical Foundation of Buffalo. Inc. for the generous gift of anti-monoclonal aromatase antibody and his critical discussion. We also thank Dr Y. Matsuki of Food and Drug Safety Center, Hatano Research Institute for GC-MS analysis. REFERENCES E. A. Jr and Siiteri P. K.: Utilization of oxygen and reduced nicotinamide adenine dinucleotide phosphate by human placental microsomes during aromatization of androstenedione. J. biol. Chem. 249 (1974) 53645372. 2. Osawa Y., Tochigi B., Higashiyama T., Yarborough C., Nakamura T. and Yamamoto T.: Multiple forms of aromatase and response of breast cancer aromatase to antiplacental aromatase II antibodies. Cancer Res.

16.

17.

18.

1. Thompson

SUDDI.42 (1982) 3299s-3306s. 3 Tai L. and Muto N.: Purification and reconstitution

4.

5.

6.

7.

properties of human placental aromatase, a cytochrome P-450 type monooxygenase. Eur. J. Biochem. 156 (1986) 243-250. For reviews see: Harvey H. A., Lipton A. and Santen R. J.: Aromatase: new perspectives for breast cancer. Cancer. Res., Suppl. 42 (1982) 3261s-3469s. Brodie A. M. H.: Overview of recent development of aromatase inhibitors. Cancer Res., Suppl. 42 (1982) 3312s-3314s. Covey D. F., Hood W. F. and Parikh V. D.: IO/?-Propynyl-substituted steroids. J. biol. Chem. 256 (1981) 10761079. Marcotte P. A. and Robinson C. H.: Inhibition and inactivation of estrogen synthetase (aromatase) by fluorinated substrate analogues. Biochemistry 21 (i98i)

19.

20.

4-androsten-3-one: Potential affinity labels of huma: placental aromatase. Steroicis 38 (1981) 149-159. G. and Kerb U.: Henderson D., Norbisrath I-Methyl-1,4-androstadiene-3,17-dione (SH 489): Characterization of an irreversible inhibitor of estrogen biosynthesis. J. steroid Biochem. 24 (1986) 303-306. Covey D. F., Parikh V. D. and Chien W. W.: 19-Acetylenic, 19-hydroxyandrost-4-en-3,17-diones. Potential suicide substrates of estrogen biosynthesis. Tet. Let!. 23 (1979) 21052108. Schwarzel W. C., Kruggel W. G. and Brodie H. J.: Studies on the mechanism of estrogen biosynthesis VIII. The development of inhibitors of the enzyme system in human placenta. Endocrinology 92 (1973) 866880. Tan L., Hrycay E. G. and Matsumoto K.: Synthesis and properties of the epimeric 6_hydroperoxyandrostenediones, new substrates/inhibitors of human placental aromatase. J. steroid Biochem. 19 (1983) 1329-1338. Booth J. E.: Effects of aromatization inhibitor, androst4-ene-3,6,17-trione on sexual differentiation induced by testosterone in neonatally castrated rat. J. Endocr. 79 (1978) 69-76.

21. Numazawa M., Tsuji M. and Osada R.: Studies directed towards a mechanistic evaluation of aromatase inhibition with androst-4-ene-3,6,17-trione; Its reactions with thiols. J. them. Res. (S) (1986) 54-55; J. them. Res. (M) (1986) 07184734. 22. Petrson D. H., Eppstein S. H., Meister P. D., Magerlein B. J., Murray H. C., Leigh H. M., Weintraub A. and Reineke L. M.: Microbiological transformations of steroids. IV. The II-epimer of compound F and other new oxygenated derivatives of Reichstein’s compound S. A new route to cortisone. J. Am. them. Sot. 75 (1953) ~ 412415. 23. Labler L., Slama K. and Sorm F.: Steroids. CXV. Toxic effect of some androstane derivatives on Pvrrhocoris apterus larvae. Collct. Czech. them. Commun:33 (1968) I

2773-2778. 8. Metcalf B. W., Wright

22262237. 24. Numazawa

9.

34 (1979) 347-360. 25. Ryan K. J.: Biological aromatization

10.

11.

12. 13.

14.

15.

C. L., Burkhart J. P. and Johnston J. 0.: Substrate-induced inactivation of aromatase by allenic and acetylenic steroids. J. Am. them. Sot. 103 (1981) 3221-3222. Bueggemeier R. W., Floyd E. E. and Counsel1 R. E.: Synthesis and biological evaluation of inhibitors of estrogen biosynthesis. J. med. Chem. 21 (1978) 1007-1011. Flynn G. A., Johnston J. O., Wright C. L. and Metcalf B. W.: The time-dependent inactivation of aromatase by 17B-hydroxy-IO-methylthioestra-1,4-dien-3-one. Biothem. biophys. Res. Commun. 103 (1981) 913-918. Lovett J. A., Darby M. V. and Counsel1 R. E.: Synthesis and evaluation of 19-aza and 19-aminoandrostenedione analogues as potential aromatase inhibitors. J. med. Chem. 27 (1984) 734740 Bednarski P. J., Porubek D. J. and Nelson S. D.: Thiol-containing androgens as suicide substrates of aromatase. J. med. Chem. 28 (1985) 175-779. Bellino F. L., Gilani S. S. H., Eng S. S., Osawa Y. and Duax W. L.: Active-site directed inactivation of aromatase from human placental microsomes by brominated androgen derivatives. Biochemistry 15 (1976) 47304736. Covey D. F. and Hood W. F.: Enzyme-generated intermediates derived from 4-androstene-3,6,17-trione and 1,4,6-androstatriene-3,17-dione cause a timedependent decrease in human placental aromatase activity. Endocrinology 108 (1981) 1597-1599. Numazawa M. and Osawa Y.: Synthesis of 16abromoacetoxy androgens and 17/3-bromoacetylamino-

M. and Osawa Y.: Synthesis and some reactions of 6-bromo-androgens: Potential affinity ligand and inactivator of estrogen synthetase. Steroids

of steroids. J. biol. Chem. 234 (1959) 2688272. 26. Liang T. and Heiss C. E.: Inhibition of 5a-reductase, receptor binding, and nuclear uptake of androgens in the prostate by a 4-methyl-4-aza-steroids. J. biol. Chem. 256 (1981) 7998-8005. 27. Liang T., Heiss C. E., Ostrove S., Rasmusson G. H. and Cheung A.: Binding of a 4-methyl-4-aza-steroid to Sa-reductase of rat liver and prostate microsomes. Endocrinology 112 (1983) 146&1268. 28. Walsh C. T.: Suicide substrates, mechanism-based enzyme inactivators: Recent developments. A. Rev. Biothem. 53 (1984) 493-535. 29. Rando R. R.: Mechanism based enzyme inactivators. Pharmac. Rev. 36i (1984) 111-142. 30. Morand P., Williamson D. G., Layne D. S., Lompa-

Krzymien L. and Salvador J.: Conversion of androgen epoxide into 17fi-estradiol by human placental microsomes. Biochemistry 14 (1975) 635438. 31. Bellio F. L. and Osawa Y.: Estrogen synthetase. Effect of sulfhydryl reactive reagents on androgen aromatization by human term placental microsomes. J. steroid Biochem. 13 (1980) 339-344. 32. Marsh D. A., Brodie H. J., Garret W., Tsai-Morris

C.-H. and Brodie A. M. H.: Aromatase inhibitors. Synthesis and biological activity of androstenedione derivatives. J. med. Chem. 28 (1985) 788-795.