Nonsteroidal progestins and antiprogestins related to flutamide

Steroids 65 (2000) 725–731

Nonsteroidal progestins and antiprogestins related to flutamide Michael Dukes*, Barrington J.A. Furr, Leslie R. Hughes, Howard Tucker, James R. Woodburn AstraZeneca, Alderley Park, Macclesfield, Cheshire SK10 4TG, UK

Abstract From the dual progestin/antiandrogenic properties of certain synthetic steroids (e.g. cyproterone acetate), it was apparent that the progesterone (P) and androgen (A) receptors must have some common ligand binding features. The nonsteroidal antiandrogen (aA) hydroxyflutamide was therefore considered a possible starting point for medicinal chemistry aimed at antiprogestin (aP) activity. Various modifications to the side chain and aryl ring substituents of flutamide yielded both P and aP activity, but always coupled with varying degrees of A or aA activity. Mineralocorticoid activity was present in some structures, but glucocorticoid and antiglucorticoid activities were not detected. Species (rat, rabbit and monkey) and chiral differences presented formidable difficulties in developing simple structure activity patterns, and low ( ⬍ 1%) in vitro uterine receptor binding belied in vivo potency of some aPs. One of the most active aPs, ZM172406, the R enantiomer of ZM150271, N-(3-chloro-4-cyanophenyl)-3,3,3-trifluoro-2-hydroxy-2-methylpropanamide, had comparable oral potency to mifepristone in rats and monkeys. The racemate ZM150271 was an effective abortifacient during early pregnancy in pigtailed monkeys (3 ⫻ 10 mg/kg) but less effective in cynomolgus monkeys. One of the most active progestins (Pn), ZM182345, N-(4-nitro-3trifluoromethylphenyl)-4-phenyl-2-hydroxy-2-trifluoromethyl-pentanamide, was at least as potent as P in rats and rabbits but also possessed A activity. © 2000 Published by Elsevier Science Inc. Keywords: Antiprogestins; Progestins; Antiandrogens; Androgens; Flutamide-derivatives

1. Introduction The research described in this paper was carried out between 1978 and 1984 with the aim of developing an orally active antiprogestin (aP) that might be used to induce menstruation and provide an ‘end-of-the-month’ mode of contraception, offering regular menstrual periods with only a few days of treatment each month and fewer or none of the side effects associated with established oral contraceptives. Most of the biologic testing was done in vivo in rats, rabbits, and monkeys, using procedures that accorded with contemporary legislative and corporate requirements. Macaque monkeys were used because of the closer similarity of their ovarian cycle to humans and in particular because, like humans, they menstruate. Since many steroids are known to have overlapping androgen (A) and progesterone (P) receptor-mediated activities, such as the antiandrogen/progestin (aA/Pn) cyprot* Corresponding author. Tel.: ⫹44-625-515126; fax: ⫹44-625516033. E-mail address: [email protected] or Mike.Dukes@ ALDERLEY.ZENECA.com (M. Dukes). 0039-128X/00/$ – see front matter © 2000 Published by Elsevier Science Inc. PII: P I I S 0 0 3 9 - 1 2 8 X ( 0 0 ) 0 0 1 7 9 - 3

erone acetate [1], we chose to explore the nonsteroidal aA, flutamide [2], as a potential source of novel P receptor ligands. It indeed proved to be an abundant source of such activities, but development of the area was confounded by species differences and chiral complexities. These, coupled with findings indicating that an end-of-cycle menstrual inducer would, in practice, be neither simple to use nor provide regular menstrual periods, led to the abandonment of the project.

2. Experimental 2.1. Uterine P receptor binding [3] Uteri from rabbits (n ⫽ 18) pretreated for 3 days with oestradiol benzoate (E2B; 40 ␮g sc daily) were homogenized at 0°C in a Waring blender and then centrifuged at 1800 g for 15 min at 4°C. The supernatant was further centrifuged at 105 000 g for 90 min at 4°C, and the resulting supernatant (cytosol) was separated, diluted with 120 ml of buffer comprising Tris (1.212 g), sodium azide (1 g), EDTA (3.72 g), sucrose (85.6 g), 5N hydrochloric acid (1.4 ml) in

726

M. Dukes et al. / Steroids 65 (2000) 725–731

deionized water (1-l), adjusted to pH 8.0 with 1N sodium hydroxide. Diluted cytosol was stored at ⫺20°C for up to 1 month. Binding assays were performed in duplicate in 7.5 ⫻ 0.95-cm Pyrex glass tubes. Test compounds were usually assayed at an initial concentration of 20 ␮g/tube; P was assayed at 0, 4, 8, 20, 40, and 2000 ng/tube; all were dissolved in DMSO (10 ␮l). [6,7-3H]-P (10 nCi; Amersham International, UK) in DMSO (50 ␮l) was then added to all tubes, followed by diluted cytosol (500 ␮l). After vortex mixing (10 s), the tubes were incubated at 4°C for at least 2 h. 500 ␮l of a suspension of dextran-coated charcoal (Norit A (10 g), dextran T70 (1 g) in buffer (1-l)), was then added to each tube, and the contents were thoroughly mixed. After 15 min at 0°C the tubes were centrifuged at 1800 g for 15 min, and the supernatant was decanted into 5 ml PCS scintillation fluid; the radioactivity present was counted on a liquid scintillation counter. From the amount of labelled P bound in the absence of unlabelled P, the percent of labelled P bound in the presence of unlabelled P and test compounds were calculated. Percent of crossreaction was calculated as (weight (ng) of P causing equivalent displacement of labelled P to test compound) x 100/(weight (ng) of test compound). Rat and monkey uterine preparations were made and used similarly with corresponding weights of freshly excised uterus. 2.2. Pregnancy termination in rats Pregnant rats (Alderley Park, Wistar-derived) were dosed orally in the morning and afternoon of day 9 (day of sperm positive smear ⫽ day 1) and again in the morning of day 10. The number of viable conceptuses was then counted post mortem on day 16. In a variant of this protocol, rats were ovariectomized under fluothane anesthesia on day 8 and given exogenous P (3 mg/rat sc) and estrone (E1; 1 ␮g/rat sc) daily on days 8 –15; test compounds were administered on days 9 and 10, and viable conceptuses counted on day 16, as described above. Test compounds were administered as ball-milled suspensions in 0.5% polysorbate 80. 2.3. Endometrial transformation test in rabbits 2.3.1. Assessing Pn activity Immature female rabbits (Dutch, body weight ca 1 kg) were given 6 daily doses of E1 (2.5 ␮g/rabbit, sc) and then 6 daily doses of either P (0.1 mg/rabbit, sc) or test compound (also sc). The rabbits were then killed, the uterus dissected, fixed, and blocked, and histologic sections were stained with hemotoxylin-eosin (H & E). Six sections per uterus were examined, and the extent of endometrial glandular development scored using the McPhail criteria.

2.3.2. Assessing aP activity The procedure was as above except that 6 daily doses of P and test compound were given concomitantly. E1 and P were administered in arachis oil, and test compounds were administered in tricaprolin oil. 2.4. Menstrual induction in monkeys Sexually mature female pigtailed macaques (Macaca nemestrina; body weight 5– 6 kg) with regular menstrual cycles (assessed by vaginal swabbing) were used. During treatment cycles, daily blood samples were collected from the saphenous vein. Plasma E2 was measured by radioimmunoassay (E001 antibody; Steranti Research Ltd, St Albans, UK) on days near the expected time of ovulation in order to identify the day of peak preovulatory concentrations. The day following the day of peak estradiol was designated as day 0, and the next day as day 1 of the luteal phase. Test compounds were administered orally on each of days 5, 6, and 7 of the luteal phase. Vaginal swabs were taken during and after treatment to detect the onset of menstruation. In a variant of this protocol, monkeys were ovariectomized and then given cycles of estradiol benzoate (E2B) and P sc in arachis oil in order to restore ‘normal’ patterns of menstruation. The dosage of P was adjusted to produce plasma levels of ca 3 ng/ml. In monkeys that showed predictable withdrawal bleeds, test compounds were given on 3 consecutive days during continued dosing with P to assess aP activity (induction of menstruation), or were given for 3–5 days beginning at the time of P withdrawal to assess Pn activity (prevention of P withdrawal bleeding). 2.5. Pregnancy termination in monkeys In the absence of an assay for monkey pituitary and gonadotrophic hormones, radioimmunoassays of P (P001 antibody; Steranti Research Ltd, St Albans, UK) were relied upon for diagnosing pregnancy. Comparison of hormone profiles in non-pregnant and pregnant pigtailed monkeys showed that pregnancy could be identified around 16 days post-ovulation (the day following peak plasma E2 concentrations being designated the day of ovulation); in normal cycling monkeys, plasma P at this time was 1 ng/ml or less, whereas in pregnant monkeys it was 5 ng/ml or higher. On this basis, pregnant monkeys were given 3 consecutive daily oral doses of ZM150271 beginning on days 17–20 (n ⫽ 7), days 29 –34 (n ⫽ 2), days 45–56 (n ⫽ 6), or on day 107 (n ⫽ 1) post-ovulation. Vaginal swabs were taken during and after treatment, and the animals and cages were carefully examined for signs of abortion. Daily blood samples for hormone assays were collected during and for at least 7 days after treatment, and, in monkeys not showing signs of pregnancy termination, at intervals of 2–3 days thereafter.

M. Dukes et al. / Steroids 65 (2000) 725–731

727

Table 1 Structural details of flutamide analogues described in tables 2– 4 Compound

Fig. 1. Hydroxyflutamide and general structure of analogues.

2.6. Other hormone activities Androgen activity was assessed by dosing castrated immature male rats for 7 consecutive days with test compounds at 25 mg/kg/day and weighing the seminal vesicles and prostate glands. Compounds that caused a significant increase in these weights, relative to those in vehicle-treated controls, were deemed to have A activity. aA activity was assessed by dosing castrate immature male rats with testosterone propionate (TP; 0.2 mg/kg) and test compound (various doses po) for 7 days and comparing prostate and seminal vesicle weights with those of rats given TP and oral dose vehicle only. Mineralocorticoid activity was assessed by measuring sodium and potassium concentrations in urine collected over 5 h from adrenalectomized, saline-loaded (2.5 ml physiological saline sc) male rats. Test compounds were administered orally or sc at the time of saline loading. Glucocorticoid and antiglucocorticoid activity were assessed by measuring thymus weights in female rats (ca 200 g body weight) after 7 days of dosing with test compound alone or dexamethasone (20 ␮g/rat/day) and test compound concomitantly. Thymus weights were compared of rats given dexamethasone alone with those given dosing vehicles alone.

3. Results The structure of hydroxyflutamide and the general structure of the analogues to be described are shown in Fig. 1. Modification of the side chain (R1 and/or R2) had great impact on pharmacological properties, and, for convenience, the side chain modification provides a simple descriptor for different groups of compounds. Modification of the aniline ring substituents, R3 and R4, often enhanced activity within the different groups of side chain modifications. The particular compounds described have been selected to illustrate the range of biologic properties encountered, which varied substantially and among species. In the tables of data, the figures relating to activity refer to the minimum dose that causes total loss of viable conceptuses in rat, that completely blocks or emulates the uterotrophic effect of P in rabbit, or that induces menstruation in monkey.

Haloalky-series ZM130352 ZM150561 ZM150271 Aryl series ZM147376 ZM159724 Haloalkyl ⫹ Arylalkyl-series ZM154389 ZM156854 ZM164503 ZM165574 ZM182345

R1

R2

R3

R4

CF3 CF3 CF3

CH3 C2H5 CH3

CF3 CF3 Cl

NO2 NO2 CN

p-NO2-C6H4 p-CN-C6H4

CH3 CH3

CF3 CF3

NO2 NO2

CF3 CF3 CF3 CF3 CF3

CH2C6H5 C2H5C6H5 CH2(p-F-C6H4) CH2(p-F-C6H4) CH2CH(C6H4)CH3

CF3 CF3 CF3 Cl CF3

NO2 NO2 NO2 CN NO2

3.1. ‘Haloalkyl’ series (Table 1) The simplest modification to yield aP activity was halogenation of one of the side chain methyls, e.g. the trifluoro analog ZM130352 (Table 2). The simple homolog of this compound, ZM150561, initially appeared to be more potent in rat (intact), but substantially less so in rabbit and monkey. However, testing in male rats revealed substantial A activity, and it proved to be inactive in ovariectomized rats supported with exogenous P. Activity in intact pregnant rats was thus probably due to gonadotrophin inhibition and not P antagonism. Thus, a small additional change to the sidechain of ZM130352 extinguished aP activity and produced a non-steroidal A. Retaining the side chain of ZM 130352 and modifying the aniline ring substituents yielded, in ZM150271, enhanced aP potency in rat and monkey but not in rabbit. Its efficacy in rat and monkey was comparable to that of mifepristone. Comparability in potency in vivo was, however, in marked contrast to their uterine P receptor binding (⬍ 1% and ca 100% respectively), and ZM 150271’s affinity was also lowest in the species in which it was most active. Whereas mifepristone has potent antiglucocorticoid activity, ZM150271 had no discernible effects mediated via the glucocorticoid receptor (data not shown). It did, however, possess substantial mineralocorticoid activity in rat at doses comparable to those eliciting aP activity, and was also a potent aA, hypotensive, and antidiuretic in rat. Hypotensive activity was also demonstrated in a primate (baboon). Subsequent investigations established that hypotensive activity was due to its being a K⫹ channel opener [4]. Resolution of ZM150271 into its enantiomers revealed an interesting split in properties. Both enantiomers were equally endowed with aA and aP activity, but mineralocorticoid, hypotensive, and antidiuretic activities were present only in the S(-) enantiomer. The R(⫹) enantiomer (ZM 172406) therefore represents a comparatively selective and

728

M. Dukes et al. / Steroids 65 (2000) 725–731

Table 2 Activities of selected anilides of the ‘Haloalkyl series’ in antiprogestin (aP) tests in rats, rabbits, and monkeys and in tests for other activities in rats Activity

ZM130352

Rat (pregnancy termination): Intact Ovxd ⫹ P Rabbit (Endometrial transformation test) aP Monkey Menses induction in intact Menses induction in ovxd ⫹ P Uterine P receptor binding: Rat Rabbit Monkey Other activities in rats: aA A agonist Mineralocorticoid Hypotensive

ZM150561

50 50

12.5 ⬎100

25

⬎25

50 25

⬎25

3.2 1.1 0.26

⬃0.3

⬃25 ⫹233%

1 ⫹75% 25 50

50

ZM150271 7.5 7.5

25 5 0.5–1

Mifepristone 2.5 sc; 10 po

10 5–10 1

0.7 0.2 0.06 0.5 0 5 10

For aP activity, the numbers given are the minimum doses in mg/kg that gave full activity. Figures for uterine progesterone (P) receptor binding are the % cross reactions, relative to P ⫽ 100. Figures for antiandrogen (aA) activity are the smallest doses causing statistically significant reductions in prostate and seminal vesicle weights in pubertal rats. Androgen agonist (A) activity is the % increase in prostate gland weight in immature males given 25 mg/kg ⫻ 7 days (⫹100% ⫽ doubling). Figures for mineralocorticoid, hypotensive, and antidiuretic activity are the smallest doses associated with maximal activity.

potent nonsteroidal aP/aA. Effects of ZM150271 in pregnant monkeys will be described later.

kyl series these compounds did not possess mineralocorticoid or K⫹ channel opening activities.

3.2. ‘Aryl’ series (Table 1)

3.3. ‘Haloalkyl/arylalkyl’ series

aP activity was also present when one of the side chain methyls was replaced by a substituted phenyl group. These compounds, e.g. ZM147376 and ZM159724, showed good aP activity in rabbit (comparable to mifepristone) but not in rat; only modest activity was seen in monkey (Table 3). They were potent aAs in rat. Like the haloalkyl analogues, their P receptor binding was low (⬍ 1%). Unlike the haloal-

Combining the trifluoromethyl substituent with phenyl substitution of the remaining methyl (or ethyl homolog) afforded compounds that generally, but not invariably, displayed Pn activity (Table 4). The simplest of these, ZM154389, behaved as a Pn in rabbits and monkeys but terminated pregnancy in rats. A activity may contribute to the latter effect in intact rats, but pregnancy terminating activity in ovariectomized rats at 50 mg/kg suggests that it behaves more as an aP than Pn in rat. Its homolog, ZM156854, was a more potent Pn in rabbit, active at 1 mg/kg, and also in monkey, and this compound did not terminate pregnancy in rat. Both compounds retain aA activity. ZM156854 thus represents a nonsteroidal Pn/ aA. Compared to the previous series, compounds of this series had more than 10-fold higher P receptor binding. Substitution of the side chain phenyl of ZM154389 with p-fluoro (ZM164503) appeared to polarize species differences more markedly. While displaying clear Pn activity in rabbit and monkey, ZM164503 was a potent terminator of pregnancy in rat, both intact and ovariectomized. On resolution, the (-)-enantiomer was found to be responsible for the activity in rat and the (⫹)-enantiomer for the activity in rabbit. Unfortunately, insufficient enantiomers were available for testing in monkey or for determining absolute configurations. Retaining the side chain of ZM164503 but changing the aniline ring substituents to Cl/CN (as in ZM150271) re-

Table 3 Activities of selected anilides of the ‘Aryl series’ in antiprogestin (aP) tests in rats, rabbits, and monkeys and in tests for other activities in rats. Activity Rat (pregnancy termination): Intact Ovxd ⫹ P Rabbit (Endometrial transformation test) aP Monkey Menses induction in intact Menses induction in ovxd ⫹ P Uterine P receptor binding: Rat Rabbit Other activities in rats: aA A agonist See Table 2 for definitions of activity.

ZM147376

ZM159724

⬎100

50

10

5

25–50 0.3 0.23 0.25 ⫹108%

10 0.5 0.37 0.5 ⫹66%

M. Dukes et al. / Steroids 65 (2000) 725–731

729

Table 4 Activities of selected anilides of the ‘Haloalkyl/arylalkyl series’ in progesterone (P) and antiprogestin (aP) tests in rats, rabbits, and monkeys and in tests for other activities in rats Activity Rat (pregnancy termination): Intact Ovxd ⫹ P Rabbit (Endometrial transformation test) aP Pn Monkey Menses induction in intact Menses induction in ovxd ⫹ P Menses inhibition Uterine P receptor binding: Rat Rabbit Monkey Other activities in rats: aA A agonist

ZM154389

ZM156854

12.5 50

⬎100

⬎25 2.5

⬎25 1

⬎25 ⬍25

ZM164503

ZM165574

2.5 5

7.5 7.5

⬍5

⬎25 25 (partial)

ZM182345

0.1

10 bid 5 bid 10

24.6 7.1 2.8

14.8 15.7 5.1

1 ⫹125%

1 ⫹141%

⬃25

3.8

⬃10 ⫹118%

1 sc*

2 0.64 2.5 0

17.4 26.6

⫹⬎250%

See Table 2 for ‘other’ activities. For progestin (Pn) activity, the numbers given are the doses in mg/kg that stimulated uterine development in rabbits to an equivalent degree to progesterone (P) (0.1 mg/kg) or prevented menses on P withdrawal in monkeys, except for ZM182345 where the figure (indicated *) refers to inhibition of ovulation in cynomolgus monkeys.

duced Pn and increased aP character. In rabbit, this compound, ZM165574, elicited only partial endometrial development at 25 mg/kg and did not inhibit P. In monkey it behaved as an aP, but required twice daily dosing owing to poorer pharmacokinetics (lower peak blood levels and more rapid clearance) than, for example, ZM150271 (data not shown). In rat it was also a potent aP (and aA). The separate enantiomers of this compound were not available for testing. The final example to be described is ZM182345. This is a further homolog of ZM156854 and notable for potent Pn activity in all three species. In rabbit, it is equipotent with P and in rat (deciduoma test), even more potent (Fig. 2). Ovulation inhibiting activity in cynomolgus monkeys was

Fig. 2. Decidual reaction in ovariectomized rats treated with E1 (1 ␮g/rat/ sc) and either P, norgestrel or ZM182345. Treatment was started the day following ovariectomy and continued for 7 days; one uterine horn was exposed on the fourth day of treatment and gently traumatized with a sterile cotton thread. The weight of the traumatized horn was assessed 24 h after the last dose.

demonstrated by Dr Tangpraprutgal (personal communication). In rat it also displayed substantial A activity (Fig 3). This nonsteroidal compound thus has a similar profile and potency to steroidal Pns such as norgestrel. This molecule contains two asymmetric centers; the properties of its enantiomers were not explored. 3.4. Effects of ZM150271 in pregnant monkeys In pregnant pigtailed monkeys, three consecutive daily doses of 10 mg/kg of ZM150271 terminated pregnancy up to around 34 days postovulation, but not thereafter. Monkeys whose pregnancies were terminated usually showed signs of vaginal bleeding by the third day of dosing, which continued for 7–10 days, about twice the duration of normal menstrual bleeding. All but one of the more advanced pregnancies, not interrupted by treatment, progressed to term and produced normal infants.

Fig. 3. Androgen activity of ZM182345 in immature male rats after 7 days of dosing. The points marked T indicate the organ sizes in rats given 0.1– 0.2 mg/kg/day of testosterone propionate.

730

M. Dukes et al. / Steroids 65 (2000) 725–731

By contrast to its consistent activity in early pregnancy in pigtailed monkeys, ZM150271 was much less effective in cynomolgus monkeys. Four monkeys were given 3 days of treatment, starting 18 –24 days postovulation; in one of two monkeys given 10 mg/kg there was fetal loss, but placenta and membranes were retained. In one of two monkeys given 25 mg/kg, delayed abortion occurred approximately 50 days later. In the pigtailed monkeys whose pregnancies were terminated by treatment with ZM150271, particular attention was paid to the timing of the first post-abortion menstrual bleed. This almost invariably occurred after a longer interval (mean 46 days) than the normal menstrual cycle interval in each individual (mean 32 days). This and plasma hormone measurements indicated that, following early pregnancy interruption, resumption of ovarian activity was delayed by about 2 weeks.

4. Discussion Starting from a nonsteroidal compound known only for antiandrogenic activity, structural modifications of various sorts produced compounds with a spectrum of activity ranging from aP to full Pn, but all retained prominent effects on the A receptor, either antagonist or partial or full agonist; some had mineralocorticoid and other actions as well. However, no consistent structure activity patterns in relation to the sought after aP activity could be discerned, nor could any simple explanations for the remarkable differences in activity across the three species. ZM172406, the R-enantiomer of ZM150271, was the most potent and comparatively selective aP developed in this program with oral aP potency in rat and monkey comparable to that of mifepristone. aA activity was its only other property, in contrast to its optical isomer ZM172405, which also possessed mineralocorticoid and K⫹ channel activities. ZM182345 was the most active Pn to emerge, comparable in potency to P itself; however, with significant A activity, it seemed unlikely to offer any advantages over synthetic steroids already in clinical and contraceptive use. Measurements of uterine cytosolic P receptor binding using McGuire and Bariso’s method [3] provided little assistance in this program, afforded no explanations for species differences, and did not correlate with in vivo performance. The only consistent feature to emerge – that compounds with Pn activity had notably higher receptor binding than those with aP activity – is itself almost incongruous. The reverse would have been easier to comprehend. No attempt was made to explore the way(s) in which our compounds interacted with the P receptor, but it is possible that some could be useful tools, allied with currently available techniques, such as receptor expression, X-ray structural determination of ligand-receptor complexes etc. for probing P receptor function. The observations with ZM 165403, a potent Pn in rabbits, seemingly due to its (⫹)-

enantiomer, and a potent aP in rats, seemingly due to its (-)-enantiomer, make it a particularly intriguing contender in this respect. The differing abortifacient efficacy of ZM150271 in early pregnancy in pigtailed and cynomolgus monkeys was another puzzling anomaly. It could not be explained by pharmacokinetics, nor on obvious differences in the competitive P ‘challenge,’ as plasma P concentrations were also comparable in the two species (data not shown). Confounding the mystery was the fact that ZM150271 was equally effective at inducing menstruation in nonpregnant females of both species. We could only surmise that subtle differences may exist in the P dependency of key structures and functions at the local endometrial/conceptus level. Such considerations probably also underly the decline in abortifacient efficacy in later pregnancy in pigtailed macaques. Among all the species differences encountered, the differing performance of ZM150271 in pregnant pigtailed and cynomolgus monkeys was of particular concern in regard to predicting efficacy and utility in women. But in addition, the observations on timing of ovulation following early pregnancy interruption with ZM150271 in the pigtailed monkeys cast serious doubt on the feasibility of using an aP as a regular form of contraception, either on a ‘missed-period’ or regular basis, even if it were to prove an effective abortifacient in women. The approximately 2-week delay in resumption of ovarian activity following early pregnancy interruption in our studies compared closely with contemporary observations in monkeys [5,6] and in humans [7]. The inherent variability of the human cycle [8,9], primarily a consequence of inconsistency in the timing of ovulation, itself complicates consistent alignment of regularly spaced treatment with a given stage of the ovarian cycle; the additional perturbation of the timing of ovulation caused by conception and exposure to chorionic gonadotrophin, renders such alignment virtually impossible. Despite extensive computer modeling of many theoretical and actual cycle patterns and a variety of treatment regimens, we were unable to identify any simple treatment regimen that would guarantee to preclude conception and to provide regular menstrual periods of normal intervals. Accordingly, the program was abandoned, and none of the compounds was taken into clinical trial.

Acknowledgments Many colleagues contributed to this work, and we are indebted to them for their skilled technical and cerebral assistance. Alasdair T Glen synthesized a number of the compounds. Much of the biologic and associated test data were generated by: Rosemary Chester, Ann Major, John Morrell, Barbara Valcaccia, and Caroline Lunning. We are also grateful to Dr R Haywood (Huntingdon Research Centre) for the abortifacient study in cynomolgus monkeys and to Professor Tangraprutgal (Chulalongkorn University,

M. Dukes et al. / Steroids 65 (2000) 725–731

Bangkok) for examining the effects of ZM 182345 on ovulation in cynomolgus monkeys.

References [1] Neumann F, Graf K-J, Hasan SH, Schenck B, Steinbeck H. Central actions of antiandrogens. In: Martini L, Motta M, editors. Androgens and Antiandrogens. New York: Raven Press, 1977. pp. 163–77. [2] Neri Rudolph O. Studies on the biology and mechanism of action of nonsteroidal antiandrogens. In: Martini L, Motta M, editors. Androgens and Antiandrogens. New York: Raven Press, 1977. pp. 179 – 88. [3] McGuire JL, Bariso CD. Isolation and preliminary characterisation of a progestogen specific binding macromolecule from the 273,000g supernatant of rat and rabbit uteri. Endocrinology 1972;90:496 – 506.

731

[4] Grant TL, Ohnmacht CJ, Howe BB. Anilide tertiary carbinols: a novel series of K⫹ channel openers. Trends Pharm Sci 1994;15:402– 4. [5] diZerega GS, Williams RW, Morin ML, Hodgen GD. Anovulation after pregnancy termination: ovarian versus hypothalamic pituitary factors. J Clin Endocrinol Metab 1979;49:594 –9. [6] Goodman AL, Hodgen GD. Evidence for an extraluteal antifolliculogenic action of chorionic gonadotrophin (CG) in rhesus monkeys. Endocrinology 1982;110:1315–9. [7] Blazar AS, Harlin J, Zaidi AA, Diczfalusy E. Differences in hormonal patterns during the first postabortion menstrual cycle after two techniques of termination of pregnancy. Fertility and Sterility 1980;33: 493–500. [8] Treloar AE, Boynton RE, Behn BG, Brown BW. Variation of the human menstrual cycle through reproductive life. Int J Fertil 1967;12: 77–126. [9] Chiazze L, Brayer FT, Macisco JJ, Parker MP, Duffy BJ. The length and variability of the human menstrual cycle. J Am Med 1968;203: 377– 80.