`
`J. Med. Chem. 1995,38, 2463-2471
`
`d AA
`
`2463
`
`Novel Steroidal Inhibitors of Human Cytochrome P45017a. (17a.-Hydroxylase·
`C17,20-lyase): Potential Agents for the Treatment of Prostatic Cancer
`
`Gerard A. Potter,t S. Elaine Barrie, Michael Jarman,* and Martin G. Rowlands
`
`Cancer Research Campaign Centre for Cancer Therapeutics at the Institute of Cancer Research,
`Cancer Research Campaign Laboratory, 15 Cotswold Road, Sutton, Surrey 8M2 5NG, U.K.
`
`Received January 19, 1995*
`
`Steroidal compounds having a 17-(3-pyridyl) substituent together with a 16,17-double bond
`have been synthesized, using a palladium-catalyzed cross-coupling reaction of a 17 -enol triflate
`with diethyl(3-pyridyl)borane, which are potent inhibitors of human testicular 17a-hydroxylase(cid:173)
`Cl7,2o-lyase. The requirement for these structural features is stringent: compounds having
`2-pyridyl (9), 4-pyridyl (10), or 2-pyridylmethyl (11) substituents instead of the 3-pyridyl
`substituent were either poor inhibitors or noninhibitory. Reduction of the 16,17-double bond
`to give 17P-pyridyl derivatives diminished potency with 3-pyridyl substitution (3- 27; IC5o
`for lyase, 2.9-23 nM) but increased it with a 4-pyridyl substituent present (10- 28; ICso 1
`.uM - 53 nM). In contrast, a variety of substitution patterns in rings A -C of the steroid skeleton
`afforded inhibitors having potencies similar to those most closely related structurally to the
`natural substrates pregnenolone and progesterone, respectively 17-(3-pyridyl)androsta-5,16-
`dien-3P-ol (3, Ki. < 1 nM; IC50 for lyase, 2.9 nM) and 17-(3-pyridyl)androsta-4,16-dien-3-one
`(15; IC50 for lyas~: 2.1 nM). Thus compounds having variously aromatic ring A (18), saturated
`rings AlB (21, 22), and oxygenated ring C (26) exhibited ICso values for lyase (1.8-3.0 nM)
`falling within a 2-fold range. The most potent compounds are candidates for development as
`drugs for the treatment of hormone-dependent prostatic carcinoma.
`
`Carcinoma of the prostate is now the most prevalent
`cancer in men in the USA. In 1993, 165 000 new cases
`were expected to be diagnosed, of which 35 000 will die
`of metastatic prostatic cancer.1 The most widely ac(cid:173)
`cepted drug treatment is the use of GnRH agonists,
`which act by interfering with the production of test(cid:173)
`osterone by the testes and represent a medical alterna(cid:173)
`tive to orchiectomy.2 However neither GnRH agonists
`nor orchiectomy deplete the synthesis of androgens
`through the adrenal route, and levels of testosterone and
`dihydrotestosterone in the prostate are respectively still
`25% and 10% of pretreatment levels even after 3 months
`treatment with a GnRH agonist. 3 The importance of
`androgen synthesis by the adrenal route in maintaining
`tumor growth is suggested by the improved therapeutic
`benefit, both in terms of increase in progression-free
`survival time and survival advantage, seen in patients
`treated with the combination of GnRH agonist or
`orchiectomy with an antiandrogen, compared with those
`It is
`given GnRH agonist or orchiectomy alone.4·5
`proposed that the role of the antiandrogen is to coun(cid:173)
`teract the stimulant action of residual androgens,
`synthesized through the adrenal route, on androgen
`receptors in the prostate cancer cells.
`In principle, the effects of the combined therapy could
`be realized by a single drug which inhibits the enzyme
`steroidal17a.-hydroxylase-Cl7,2o-lyase. This enzyme is
`responsible for androgenic hormone biosynthesis which
`produces dehydroepiandrosterone and androstenedione,
`immediate precursors of testosterone, from their respec(cid:173)
`tive precursors pregnenolone and progesterone, in both
`testes and adrenals. The imidazole antifungal agent
`ketoconazole inhibits this enzyme when given in high
`
`• To whom enquiries should be addressed.
`' Present address: Chiroscience Ltd., Milton Rd., Cambridge CB4
`4WE, U.K
`~Abstract published in Advance ACS Abstracts, May 15, 1995.
`
`doses to male patients and produces the symptoms of
`androgen suppression. This drug has been used to treat
`prostate cancer,6 and although success has been re(cid:173)
`ported in some studies,7•8 it proved less promising in
`others.9·10 The undesirable side effects, coupled with
`the inconvenience of the three times daily schedule
`which is dictated by its short half-life, limit its potential
`clinical usefulness. Nevertheless the clinical results
`obtained, coupled with a very recent report that careful
`scheduling of ketoconazole can produce prolonged re(cid:173)
`sponses in previously hormone-refractory prostate can(cid:173)
`cer,11lend credence to the selection of this enzyme target
`and impetus to the design and development of a more
`enzyme-selective, less toxic, and less metabolically labile
`inhibitor.
`We report here on the synthesis and inhibitory
`activity toward the individual 17a.-hydroxylase and
`c17,2o-lyase components of the target enzyme, obtained
`from human testis, of a variety of steroidal compounds
`having as their common structural feature a 17 -(3-
`pyridyl) substituent together with a 16,17-double bond
`in the steroidal skeleton. We have previously explored
`nonsteroidal inhibitors containing a pyridyl residue,
`starting from the serendipitous discovery that certain
`esters of 4-pyridylacetic acid were effective inhibitors
`of the hydroxylase-lyase enzyme from rat testis, 12 find(cid:173)
`ings which have in part been rationalized by crystal(cid:173)
`lographic and molecular modeling studies. 13 More
`recently, esters of 3-pyridylacetic acid have been evalu(cid:173)
`ated, using enzyme from human testis. 14
`The design concept used here was to consider how a
`pyridyl substituent could be incorporated into the actual
`steroid skeleton such that the pyridyl nitrogen lone pair
`would coordinate to the iron atom of the heme cofactor
`in the active site of the enzyme. The initial step of the
`de novo mechanism-based design approach was to
`postulate a complete catalytic cycle for the enzyme
`
`0022-2623/95/1838-2463$09.00/0
`
`© 1995 American Chemical Society
`
`WCK1035
`Page 1
`
`
`
`2464 Journal of Medicinal Chemistry, 1995, Vol. 38, No. 13
`
`:bo
`
`Potter et al.
`
`,I 'yo
`
`JJOH
`
`HO-Fe
`
`~ ~oto-o.-Fe
`)J0 .-H
`
`O=Fe
`
`HO-Fe•
`
`e·
`[!£] __ ) ____ .....:.;_· ___ )e·
`ET • electron transport system
`
`Figure 1. Postulated complete catalytic cycle for the 17a.-hydroxylase-Cl7.2o-lyase enzyme. For clarity, only the steroid D-ring
`and the cofactor iron atom are shown.
`
`(Figure 1) and then to consider the juxtaposition be(cid:173)
`tween the steroid D-ring and the heme cofactor from
`the putative transition state geometry. For this pur(cid:173)
`pose, three-dimensional molecular models were con(cid:173)
`structed of the putative transition states using the
`Cochrane orbit molecular modeling system. From this
`analysis, it was postulated that a steroid incorporating
`a 16,17-double bond with the 17-position substituted by
`a 2-pyridyl group may inhibit the hydroxylase step and
`a 3-pyridyl derivative may inhibit the lyase step, while
`a 4-pyridyl analog should not inhibit either step. How(cid:173)
`ever, the enzyme may not tolerate an aromatic ring
`attached to the 17 -position, and all three compounds
`may be inactive, even if the coordination geometry is
`correct.
`The steroidal skeleton chosen for the first compound
`which was synthesized on the basis of this concept,
`namely the novel steroid 3, was that of pregnenolone,
`which appears to be the preferred substrate for the
`hydroxylase activity of the human enzyme in the
`testis. 15 Alternative orientations of the pyridyl ring
`relative to the steroidal framework were explored by
`synthesizing the 2- (9) and 4- (10) pyridyl analogs, as
`was the effect of a spacer group between a 2-pyridyl
`residue and a C-17 (compound 11). The second 17-(3-
`pyridyl) derivative synthesized was 15, analogously
`related to progesterone, the alternative substrate for the
`hydroxylase activity of the target enzyme. Further
`molecules synthesized retained the ring D substitution
`pattern of 3 and 15 while further exemplifying the effect
`on enzyme inhibition of structural variations in rings
`
`Scheme 1a
`
`dal"ydroepiol"l drosterone
`
`~
`HO~
`
`a (a) TfzO, base; (b) 3-PyBEt2, Pd(PPhahCh, THF, H20, NazC03;
`(c) NaOH, H20, MeOH.
`
`A, B, and C. Finally, the effect of reducing the 16,17-
`double bond in 3 and 10 was explored.
`
`Results
`Chemistry. A general method for introducing the
`required 17-pyridyl16,17-ene functionality into ring D
`was by palladium-catalyzed cross-coupling of steroidal
`17 -enol triflates with suitable pyridyl-containing nu(cid:173)
`cleophilic coupling partners. For the synthesis of 3
`(Scheme 1), dehydroepiandrosterone 3-acetate was con(cid:173)
`verted into its 17-enol triflate 1 by base-catalyzed
`reaction with triflic anhydride in the presence of the
`hindered base 2,6-di-tert-butyl-4-methylpyridine. This
`reaction also produced the 3,5-diene 4 in 10% yield. The
`3-pyridyl group was then introduced into the 17 -position
`by reacting 1 with diethyl(3-pyridyl)borane in THF,
`
`WCK1035
`Page 2
`
`
`
`Novel Steroidal Inhibitors of Human Cytochrome P450na
`
`Journal of Medicinal Chemistry, 1995, Vol. 38, No. 13 2465
`
`Chart 1
`
`Scheme~
`
`R =IV:.
`
`R • H
`
`10
`
`11
`
`using bis(triphenylphosphine)palladium(II) chloride as
`catalyst (0.01 equiv) and aqueous Na2COa as nucleo(cid:173)
`philic activator. The reaction proceeded remarkably
`efficiently, without the potential side reactions oftriflate
`hydrolysis or ethyl coupling, to give the acetate 2 in 84%
`isolated yield. From 4, the 3-pyridyl derivative 5 was
`similarly obtained. The acetyl group of 2, which was
`stable to the mildly basic conditions of the coupling
`reaction, was easily removed with aqueous methanolic
`NaOH to afford the target 3-pyridyl steroid 3.
`Although these coupling reactions were developed
`independently, the palladium-catalyzed cross-coupling
`of organoboron reagents with an enol triflate has been
`reported recently by Suzuki and co-workers.16 Their
`reactions employed arylboronic acids and 9-alkyl-9-BBN
`reagents and the ~d base KaP04 as the nucleophilic
`activator under strictly anhydrous conditions. Our use
`of diethyl(3-pyridyl)borane was prompted by its com(cid:173)
`mercial availability (it is also easily synthesized17) and
`its previous use in palladium-catalyzed cross-coupling
`reactions with aryl iodides.18 Some features of our
`reaction compared with that of Suzuki are noteworthy.
`We found that the catalyst Pd(PPhahCh was superior
`to Pd(PPha)4 and consistently gave better yields of
`coupled product. The catalyst could also be used at
`much lower levels, and even at 0.001 equiv, good yields
`were obtained with prolonged reaction times. Impor(cid:173)
`tantly our reaction did not require anhydrous condi(cid:173)
`tions, and indeed an aqueous THF solvent system was
`employed. Our method of introducing the 17 -pyridyl
`16,17-ene functionality was more efficient and higher
`yielding than a previous route, 19·20 reaction of 3-pyridyl(cid:173)
`lithium with a 17-keto steroid and dehydration of the
`resulting tertiary alcohol.
`The 2-pyridyl (6), 4-pyridyl (7), and 2-picolyl (8)
`steroidal acetates (Chart 1) were synthesized similarly
`to 2 but employing different nucleophilic coupling
`partners and modifying the conditions accordingly. The
`reagents used to prepare 6 and 8 were 2~pyridyl- and
`2-picolylzinc chloride, respectively. In the latter case
`the intermediate 8 was converted without isolation
`directly into 11 in good overall yield (79%). An attempt
`to prepare the 3-picolyl analog of 11 using 3-picolylzinc
`chloride was unsuccessful due to homocoupling of this
`reagent. In the synthesis of the 2- (6) and 4- (7) pyridyl
`steroid acetates, the novel palladium catalyst bromo(cid:173)
`(isopropenyl)bis(triphenylphosphine)palladium(II) was
`employed. Its use enabled the coupling reaction to be
`carried out at ambient temperature, thereby avoiding
`side reactions, and 6 was obtained in 74% yield from
`which hydrolysis gave the required 2-pyridyl analog 9.
`The catalyst had been developed to enable low-temper(cid:173)
`ature cross-coupling reactions for the stereoselective
`synthesis of (E)-4-hydrox.ytamoxifen21,22 and was pre(cid:173)
`pared from 2-bromopropene and tetrakis(triphenylphos-
`
`and.roatenedione
`
`12
`
`13
`
`IS
`14
`"(a) C1Fs, CsF, DMF; (b) Tf20, base; (c) 3-PyBE~. Pd(PPha)2Cl2,
`THF, H20, Na2COa; (d) HCl, H20, EtOH; (e) Al(O-i-Pr)a.
`
`0
`
`Chart 2
`
`~\ .,_ R•Ac: 16
`RO~
`
`R • Ac: 19
`
`('Y\ -
`RO''''UJ
`H
`
`oDY
`
`H
`
`R • Ac: 17
`R • H: 18
`
`R • Ac: 20
`R G H: 21
`
`22
`
`phine)palladium(O) by a procedure analogous to that
`used to make benzylchlorobis(triphenylphosphine)(cid:173)
`palladium(II).23 When the coupling reaction was per(cid:173)
`formed using 4-pyridylzinc chloride, prepared from
`4-bromopyridine, only a low yield (18%) of the 4-pyridyl
`steroid acetate 7 was obtained. Instability of 4-halopy(cid:173)
`ridines can restrict the use of 4-pyridylmagnesium and
`-zinc halides in palladium cross-coupling reactions, and
`diethyl(4-pyridyl)borane has been used as an alternative
`reagent. 24 Here, lithium trimethoxy( 4-pyridyl)boronate,
`an intermediate in the synthesis25 of 4-pyridylboronic
`acid, was the organoboron reagent used, and the coupled
`product thus obtained, 7, was hydrolyzed directly to give
`the 4-pyridyl steroid 10 in 53% yield overall from 1.
`The preparation of 15, starting from androstenedione
`(Scheme 2), required selective protection of the 3-keto
`function, to prevent the formation of a 3-dienol triflate. 26
`Protection as the perfluorotolyl enol ether 12 by reaction
`with octafluorotoluene in the presence of cesium fluoride
`has proved to be a convenient one-step procedure.27 The
`perfluoroaryl group was stable to the subsequent steps
`needed to insert the pyridyl substituent and was then
`cleaved by acidic hydrolysis. It was later found that 15
`was more conveniently prepared directly from 3 by
`Oppenauer oxidation using cyclohexanone and alumi(cid:173)
`num isopropoxide.
`Several 3-pyridyl derivatives (18, 21, 22; Chart 2)
`exemplifying further structural variation in rings A and
`B were prepared using procedures analogous to those
`already described. Adrenosterone was the starting
`point for the synthesis of a ring-C-substituted variant,
`26 (Scheme 3), which was prepared in good overall yield
`(60%). The formation of the tert-butyldimethylsilyl
`dienol ether 23 provided an alternative protecting
`
`WCK1035
`Page 3
`
`
`
`2466 Journal of Medicinal Chemistry, 1995, Vol. 38, No. 13
`
`Scheme sa
`
`o
`
`o
`
`a
`
`0 dP
`~·~·~A ... 2
`
`0
`
`-1
`
`b
`
`-1
`
`adrenosterone
`
`23
`
`OTI
`'\
`
`24
`
`-·-
`
`tBO~SO
`
`25
`26
`a (a) t-BDMSOTf, base; (b) (i-Pr):zNLi, PhNTf2; (c) 8-PyBEt2,
`Pd(PPhal2Cb, THF, H20, Na2COa; (d) Bu.:NF, THF, H20.
`
`Scheme 4a
`
`_Lto
`
`27
`28
`a (a) N2H4, AcOH, EtOH, air; (b) Red-Al, ZnCb, THF.
`
`HO
`
`strategy for the 3-keto function. The chemical shifts for
`the two vinylic protons in this product were very similar
`to those previously reported for the silyl dienol ether
`formed from a testosterone derivative,28 and the present
`product is therefore similarly formulated as the 2,4-
`dienol ether. In the following step, N-phenyltriflimide29
`was employed to prepare the enol triflate 24 since use
`oftriflic anhydride resulted in desilylation and 3-dienol
`triflate formation. This also enabled selective formation
`of the 17-enol triflate without affecting the 11-keto
`function by preparing the intermediate lithium enolate
`under kinetic conditions at low temperature.
`Lastly, analogs containing a saturated D-ring were
`prepared from the corresponding 16,17-ene compounds.
`Reduction of 3 using diimide, generated in situ from
`hydrazine hydrate, gave the 17,S-(3-pyridyl) steroid 27
`(Scheme 4). Reduction of the 16,17 -double bond of the
`4-pyridyl steroid (10) utilized the electron-withdrawing
`influence of the 4-pyridyl substituent under electrophilic
`activation by zinc chloride to achieve direct hydride
`reduction with Red-Al to produce the 17,S-(4-pyridyl)
`steroid 28. The P-orientation of the pyridyl ring in
`compounds 27 and 28 was confirmed by 1H-NMR
`spectroscopy which showed an apparent triplet with a
`coupling constant of 10Hz for the 17a.-proton which is
`characteristic of 17 P-substituted steroids. 20·30 Attempts
`at preparing the corresponding 17a.-(4-pyridyl) analog,
`by either direct reduction of 10 or epimerization of 28,
`were unsuccessful.
`Inhibition of Human Testicular 17cx.·Hydroxyl·
`ase and Ct1,20·Lyase. Structure-Activity Relation·
`ships. We have identified as potent inhibitors of
`human testicular steroidal17a.-hydroxylase-C17,2o-lyase
`a variety of pyridyl steroids having as their common
`structural feature the 17-(3-pyridyl) 16,17-ene moiety
`(Table 1). Although it might be expected that the most
`potent compounds would be those (3, 15) with structures
`most closely related to natural substrates, there was an
`unexpected tolerance for structural variation in this
`respect. Comparing 3 and 15 with analogs (18, 21, 22,
`
`Potter et al.
`
`>50
`
`10
`
`>50
`
`>20
`
`>20
`>20
`
`1.8
`
`Table 1. Enzyme Inhibition Data
`ICso (uM)"
`IC5o (nM}a
`compound C17,2o-lyase 17a-hydroxylase aromatase Sa-reductase
`2
`17
`18
`2.9
`4
`3
`5.6
`12.5
`5
`76
`270
`9
`1000
`4000
`10
`>10 000
`>10 000
`11
`14
`>10 000
`>10 000
`15
`2.1
`2.8
`18b
`2.6
`1.8
`21
`4.3
`2.5
`22
`3
`4.7
`26
`2.9
`13
`27
`23
`47
`28
`53
`160
`ketoconazole
`26
`65
`a The standard errors were usually < 10% of the ICso value. The
`concentration of enzyme in the assays for lyase/hydroxylase
`inhibition was estimated to be about 4-5 nM, except in the assays
`of9, 11, 14, and ketoconazole for which the concentration was ca.
`25 and 10 nM for the lyase and hydroxylase assays, respectively.
`b Other biological activity: estrogen receptor binding affinity
`(estradiol = 100), 4.9.
`26) synthesized from other naturally occurring steroid
`precursors, there was little variation (from 1.8 to 3.0
`nM) in the ICso values for inhibition of the lyase
`component. The absence of any functionality at the
`3-position in the steroid skeleton leads to a modest drop
`in potency (compound 5). The markedly lower potency
`of the acetoxy derivative 2 compared with 3 could reflect
`a limited bulk tolerance at the 3-position, as indicated
`by the total loss in activity for the much more sterically
`demanding perfluorotolyl derivative 14 of the potent
`inhibitor 15. The stringent requirement for the 17 -(3-
`pyridyl) 16,17 -ene functionality for good inhibition was
`in marked contrast to the relative flexibility in relation
`to other features discussed and is reflected in the
`marked reduction, or abolition of activity, on relocating
`the pyridyl nitrogen (compounds 9, 10) or on reducing
`the 16,17-double bond of 3 to give the 17,S-pyridyl
`derivative 27. In contrast, reduction of the 4-pyridyl
`derivative 10 gave a product, 28, with markedly im(cid:173)
`proved inhibitory potency over its parent.
`The most inhibitory compounds in the present study
`were far more potent than any inhibitor of hydroxylase/
`lyase for which comparable data have been previously
`described. The K;•PP for 3 was < 1 nM, whereas the most
`potent inhibitor, also steroidal, reported to date is 17 {3-
`(cyclopropylamino)androst-5-en-3,S-oP1 with a K;•PP of90
`nM. Another steroidal compound, 4-pregnen-3-one-20,S(cid:173)
`carboxaldehyde oxime has been developed as a com(cid:173)
`bined inhibitor of this enzyme and testosterone 5a.(cid:173)
`reductase.32 Though a potent inhibitor (K; = 16 nM) of
`the reductase, it was much less inhibitory toward the
`rat hydroxylase/lyase, being comparable to ketoconazole.
`17,S-Ureido-substituted steroids with potent activity
`toward the rat hydroxylase/lyase enzyme have been
`described. 33·34 Though the data are presented in a way
`not easily comparable with the results of the present
`study, one of these compounds, 17,S-ureido-1,4-andros(cid:173)
`tadien-3-one, markedly suppressed testosterone levels
`and ablated androgen-dependent organs in the rat.
`Liarozole is a nonsteroidal imidazole derivative having
`activity toward the rat testicular enzyme very similar35
`to that of ketoconazole. No example among our previ(cid:173)
`ously mentioned12·14 esters of 4- and 3-pyridylacetic acid
`compares in potency with the best of the present
`steroidal derivatives.
`
`WCK1035
`Page 4
`
`
`
`Novel Steroidal Inhibitors of Human Cytochrome P450na.
`
`Journal of Medicinal Chemistry, 1995, Vol. 38, No. 13 2467
`
`Other Biological Activities. While inhibition of
`other targets was not explored in detail in the present
`study, limited evaluations have been carried out (Table
`1), particularly where such activity might be antici(cid:173)
`pated, from structural analogy with compounds known
`to interact with the target in question. Thus 15,
`structurally related to androstenedione, a substrate for
`aromatase, was a moderate inhibitor of aromatase.
`Likewise the inhibition by 15 of testosterone 5o.-reduc(cid:173)
`tase might reflect its structural resemblance to the
`natural substrate testosterone, whereas 22, correspond(cid:173)
`ingly related to the product 5a.-dihydrotestosterone, was
`not an inhibitor. Notably, compound 3 inhibited neither
`aromatase nor testosterone 5a.-reductase at the highest
`concentration tested, respectively 20 and 50 ,uM. Lastly,
`the estradiol-related analog 18 had an appreciable
`binding affinity for the estrogen receptor, 5% of that of
`estradiol itself.
`
`Concluding Remarks
`Two of the compounds described here, namely 2 (as
`a prodrug for 3) and 15, have been evaluated in vivo in
`the WHT mouse. 36 Each markedly reduced the weights
`of androgen-dependent organs, and 2 depressed test(cid:173)
`osterone to undetectable levels. The adrenals were
`unaffected, implying that 3 and 15, unlike ketoconazole,
`do not inhibit enzymes in the pathway leading to
`corticosterone. This evidence for selective inhibition of
`testosterone biosynthesis, together with the further
`evidence for selectivity of action provided here for 3 in
`particular, makes 3 a strong candidate for further
`development as a potential drug for the treatment of
`prostatic carcinoma in humans.
`
`Experimental Section
`Chemical Methods. 1H-NMR spectra (250 MHz) (internal
`Me4Si = o 0) were determined in CDCh (unless otherwise
`indicated) using a Bruker AC 250 spectrometer. Infrared
`spectra were determined with a Perkin-Elmer 1720X spec(cid:173)
`trometer. Mass spectra (electron impact, 70 eV) were obtained
`by direct insertion with a VG 7070H spectrometer and VG
`2235 data system. Melting points were determined with a
`Reichert micro hot stage apparatus and are uncorrected.
`Chromatography refers to column chromatography on silica
`gel (Merck Art. 15111) with the solvent indicated applied
`under positive pressure. Light petroleum refers to the fraction
`with bp 60-80 oc. 3-Pyridyl(diethyl)borane was purchased
`from Aldrich Chemical Co., Gillingham, Dorset, U.K. Elemen(cid:173)
`tal analyses were determined by CHN Analysis Ltd., South
`Wigston, Leicester, England.
`3/J-Acetoxyandrosta-5,1&-dien-17-yl Trifluoromethane(cid:173)
`sulfonate (1) and Androsta-3,5,16-trien-17-yl Trifluo(cid:173)
`romethanesulfonate (4). To a stirred solution of dehydro(cid:173)
`epiandrosterone 3-acetate (24.8 g, 75 mmol) in dry CHzCb (500
`mL) containing 2,6-di-tert-butyl-4-methylpyridine (18.5 g, 90
`mmol) was added trifluoromethanesulfonic anhydride (12.6
`mL, 75 mmol). After 12 h the mixture was filtered, washed
`with water (50 mL), and dried (Mg804) and the solvent
`evaporated. Chromatography, on elution with CH2Ch-light
`petroleum (1:6), gave first 4 (3.02 g, 10%) as an oil: Vmax for
`C-O str absent; 1H NMR o 0.99 (s, 3, H-18), 1.02 (s, 3, H-19),
`5.39 (m, 1, H-6), 5.59 (m, 1, H-16), 5.62 (m, 1, H-3), 5.93 (dm,
`1, J = 9.4 Hz, H-4); mlz 402 (M+).
`Further elution with CH2Cb-light petroleum (1:3) afforded
`1 (20.1 g, 58%): mp 75-76 °C (from hexane); Vmaz 1734 cm-1
`(C=O str); 1H NMR o 1.00 (s, 3, H-18), 1.06 (s, 3, H-19), 2.04
`(s, CHaCO), 4.59 (m, 1, H-3a.), 5.39 (dm, 1, J = 4.9 Hz, H-6),
`5.58 (m, 1, H-16); mlz 402 (M+- AcOH). Anal. (C22H290sFaSl
`C,H,F,S.
`3/)-Acetoxy-17-(3-pyridyl)androsta-5,16-diene (2). Di(cid:173)
`ethyl(3-pyridyl)borane (3.3 g, 23 mmol) was added to a stirred
`
`solution of 1 (6.94 g, 15 mmol) in THF (75 mL) containing bis(cid:173)
`(triphenylphosphine)palladium(II) chloride (0.105 g, 0.15 mmol).
`An aqueous solution ofNa2C03 (2M, 30 mL) was then added
`and the stirred mixture heated at 80 •c for 1 h and then
`partitioned between EtzO and H20. The organic phase was
`dried (NazCOa), filtered through a short column of silica gel,
`and concentrated. Chromatography, on elution with Et20-
`light petroleum (1:2), afforded 2 (4.95 g, 84%): mp 144-145
`oc (from hexane); Vmax 1732 cm-1 (C=O str); 1H NMR 6 1.05
`(s, 3, H-19), 1.08 (s, 3, H-18), 2.04 (s, 3, CHaCO), 4.60 (m, 1,
`H-3a.), 5.42 (dm, 1, J = 4.7 Hz, H-6), 5.99 (m, 1, H-16), 7.23
`(dd, 1, Js,4 = 8.1 Hz, J 5,s = 3.9 Hz, pyridyl H-5), 7.65 (ddd, 1,
`J,,2 = 2.0 Hz, Ju = 1.6 Hz, pyridyl H-4), 8.46 (dd, 1, pyridyl
`H-6), 8.62 (d, 1, pyridyl H-2); m/z 392 (M+ +H). Anal. (CzsHaa·
`NOz) C, H, N.
`17-(3-Pyrldyl)androsta-5,16-dien·3/J·ol (3). To a solution
`of2 (4.90 g, 12.5 mmol) in methanol (50 mL) was added 2.5 M
`NaOH (10 mL), and the mixture was stirred at 80 oc for 5
`min and then allowed to cool, poured into water, neutralized
`with 1 M HCl, rebasified with saturated aqueous NaHCOa,
`and extracted with hot toluene (3 x 100 mL). The toluene
`extracts were dried (Na2C0a) and concentrated. Chromatog(cid:173)
`raphy, on elution with Et20-toluene (1:2), gave 3 (3.45 g,
`79%): mp 228-229 •c (from toluene); Vmax 3351 (OH str); 1H
`NMR t5 1.05 (s, 3, H-19), 1.07 (s, 3, H-18), 3.54 (m, 1, H-3a.),
`5.40 (dm, 1, J = 5.0 Hz, H-6), 5.99 (m, 1, H-16), 7.22 (dd, 1,
`pyridyl H-5), 7.65 (ddd, 1, pyridyl H-4), 8.46 (dd, 1, pyridyl
`H-6), 8.62 (d, 1, pyridyl H-2); mlz 349 (M+). Anal. <C24Ha1·
`NO)C,H,N.
`17-(3-Pyridyl)androsta-3,5,16-triene (5). The method
`followed that described for 2 but used 4 (2.01 g, 5.0 mmol).
`Chromatography, on elution with CH2Ch, gave 5 (1.39 g,
`84%): mp 110-112 oc (from hexane); 1H NMR t5 1.02 (s, 3,
`H-19), 1.07 (s, 3, H-18), 5.44 (m, 1, H-6), 5.61 (m, 1, H-3), 5.95
`(dm, 1, J = 9.8 Hz, H-4), 6.01 (m, 1, H-16), 7.23 (dd, 1, pyridyl
`H-5), 7.66 (ddd, 1, pyridyl H-4), 8.46 (dd, 1, pyridyl H-6), 8.63
`(d, 1, pyridyl H-2); mlz 331 (M+). Anal. (C24H29N) H, N; C:
`calcd 86.96; found, 86.24.
`3/)-Acetory-17 ·(2-pyridyl)androsta-5, 16-diene ( 6). To
`EtzO (6 mL), at -18 °C, was added n-butyllithium (0.96 mL,
`2.5 M solution in hexanes) followed dropwise by 2-bromopy(cid:173)
`ridine (0.228 mL, 2.4 mmol) in Et20 (2 mL). The resulting
`blood-red solution of 2-pyridyllithium was added dropwise to
`a solution ofZnC}z (382 mg, 2.8 mmol) in THF, cooled to -18
`•c, and the orange-brown solution of 2-pyridylzinc chloride was
`stirred for a further 30 min. For the preparation of the
`palladium catalyst, a solution oftetrakis(triphenylphosphine)(cid:173)
`palladium(O) (1.16 g, 1 mmol) in benzene (10 mL) was treated
`with 2-bromopropene (0.18 mL, 242 mg, 2 mmol) and the
`mixture stirred for 16 h at ambient temperature, whereupon
`the initially orange suspension became a yellow solution. The
`solvent was removed under vacuum, the residue was triturated
`with Et20, and the pale yellow product (0.70 g) bromo(cid:173)
`(isopropenyl)bis(triphenylphosphine)palladium(II) was recov(cid:173)
`ered by filtration: 1H NMR o 0.81 (s, CHa), 4.6 (m, C=CH2l,
`7.2-7.8 (m, arom H).
`To a solution of 1 (926 mg, 1 mmol) in THF (10 mL)
`containing the palladium catalyst (76 mg, ca. 0.1 mmol) was
`added the solution of 2-pyridylzinc chloride, and the mixture
`was stirred at ambient temperature. After 1 h, the mixture
`was partitioned between Et20 and H 20 and the organic phase
`was dried (Na2COa) and concentrated. Chromatography, on
`elution with Et20-light petroleum (1:4), gave 6 (0.583 g,
`74%): mp 189-190 °C (from light petroleum); Vmax 1734 cm-1
`(C-Q str); 1H NMR t5 1.09 (s, 3, H-19), 1.15 (s, 3, H-18), 2.04
`(3, s, CHaCO), 4.62 (m, 1, H-3a.), 5.42 (dm, 1, H-6), 6.37 (m, 1,
`H-16), 7.09 (dd, 1, J 5,4 = 7.9 Hz, Js,s = 4.1 Hz, pyridyl H-5),
`7.38 (d, 1, J 3•4 = 7.9 Hz, pyridyl H-3), 7.59 (t, J = 7.7 Hz, 1,
`pyridyl H-4), 8.55 (d, 1, pyridyl H-6); mlz 391 (M+). Anal.
`(C2aHaaN02) C, H, N.
`3{1-Acetoxy-17-(4-pyridyl)androsta-5,16-diene (7). 4-Bro(cid:173)
`mopyridine (4.5 g) was liberated from its hydrochloride (5 g,
`26 mmol) using the procedure previously applied to 4-chloro(cid:173)
`pyridinea7 but keeping solutions below 10 oc during concentra(cid:173)
`tion to prevent polymerization. The free base was twice
`concentrated from Et20 (to remove residual CHCla), and then
`a solution of the freshly prepared 4-bromopyridine (1.58 g, 10
`
`WCK1035
`Page 5
`
`
`
`2468 Journal of Medicinal Chemistry, 1995, Vol. 38, No. 13
`
`Potter et al.
`
`mmol) in Et20 (8 mL) was added dropwise during 30 min to a
`mixture ofn-butyllithium (2.5 Min hexanes, 4 mL, 10 mmol)
`and Et20 (20 mL) at -20 °C. Of the resulting solution of
`4-pyridyllithium, 15 mL (4.5 mmol) was added to a stirred
`solution of anhydrous ZnCl2 (681 mg, 5 mmol) in dry THF (25
`mL). After 1 h at ambient temperature, 30 mL of the red
`solution of 4-pyridylzinc chloride was added to a solution of 1
`(925 mg, 2 mmol) in THF containing bromo(isopropenyl)bis(cid:173)
`(triphenylphosphine)palladium(II) (see 6 above; 75 mg, 0.1
`mmol) and the mixture stirred at ambient temperature
`overnight and then filtered through Celite. Chromatography,
`on elution with Et20-light petroleum (1:2), gave 7 (140 mg,
`18%): mp 175-177 °C (from light petroleum); Vmax 1732 cm-1
`(C-O str); 1H NMR o 1.08 (s, 3, H-19), 1.63 (s, 3, H-18), 2.05
`(s, 3, CHaCO), 4.63 (m, 1, H-3a.), 5.42 (dm, 1, H-6), 6.18 (m, 1,
`H-16), 7.26 (d, 2, J = 6.0 Hz, pyridyl H-3, H-5), 8.50 (d, 2,
`pyridyl H-2, H-6); mlz 331 (M+ - AcOH). Anal. CC2sHaaN02)
`C,H,N.
`17-(2·Pyridyl)androsta-5,16-dien·3·P·ol (9). The method
`followed that described for 3 but used 6 (392 mg, 1 mmol),
`except that on completion of the reaction the product was
`extracted with Et20 followed by benzene and crystallized
`without prior chromatography giving 9 (273 mg, 78%): mp
`206-207 oc (from benzene-light petroleum); Vmax 3390 cm-1
`(OH str); 1H NMR o 1.08 (s, 3, H-18), 1.15 (s, 3, H-19), 3.56
`(m, 1, H-3a.), 5.41 (m, 1, H-6), 6.38 (m, 1, H-16), 7.10 (m, 1,
`pyridyl H-5), 7.38 (d, 1, J = 7.8 Hz, pyridyl H-3), 7.59 (m, 1,
`pyridyl H-4), 8.55 (d, 1, J = 4.2 Hz, pyridyl H-6); mlz 349 (M+).
`Anal. (C24Ha1NO) C, H, N.
`17-(4-Pyridyl)androsta-5,16-dien-17-ol (10). A solution
`of 4-bromopyridine (from the hydrochloride; 25 g, 129 mmol;
`see 7 above) in E~O (80 mL) was added dropwise to a stirred
`mixture of n-butyllithium (51.6 mL, 2.5 M in hexanes, 129
`mmol) and Et20 (200 mL) at -76 °C. The resulting solution
`of 4-pyridyllithium was added by transfer needle to a cooled
`( -76 °C) solution of trimethyl borate (13.4 g, 14.6 mL, 129
`mmol) in E~O (75 mL); the mixture was stirred for 20 min
`and then allowed to reach ambient temperature. Water (10
`mL) was added, and the resulting light brown precipitate of
`lithium trimethoxy(4-pyridyl)boronate (22.04 g, ca. 90%) was
`collected by filtration, washed with Et20, and dried in vacuo.
`This product (2.83 g, ca. 15 mmol) was added to a solution of
`1 (1.21 g, 5 mmol) in THF (30 mL) containing bis(triphenyl(cid:173)
`phosphine)palladium(II) chloride (175 mg, 0.25 mmol), followed
`by 2 M aqueous Na2COa (12.5 mL), and the mixture heated at
`80 °C, for 6 h, and then partitioned between Et20 and H20.
`The organic phase was dried (N a2C03) and filtered through a
`short column of silica gel to give crude 7 which was used for
`the next step without further purification. The method for
`3.
`converting 7 into 10 followed that described for 2 -
`Chromatography, eluting with Et20-toluene (1:2), gave 10
`(928 mg, 53% from 1): mp 226-228 oc (from toluene); 1H NMR
`o 1.08 (s, 3, H-19), 1.62 (s, 3, H-18), 3.55 (m, 1, H-3a), 5.40
`(dm, 1, H-6), 6.18 (m, 1, H-16), 7.26 (d, 2, J = 6.1 Hz, pyridyl
`H-3, H-5), 8.51 (d, 2, pyridyl H-2, H-6). Anal. CC24Ha1NO) C,
`H,N.
`17·(2-Pyridylmethyl)androsta-5,18·dien-17-ol (11). To
`a solution of 2-picoline (7.45 g, 7.9 mL, 80 mmol) in THF (42
`mL) at -20 oc was added n-butyllithium (50 mL, 1.6 M in
`hexanes, 80 mmol) during 30 min. The red solution of
`2-picolyllithium38 (10 mL) was added with vigorous stirring
`under argon to anhydrous ZnCl2 (1.09 g, 8 mmol), followed by
`benzene (10 mL). The resulting homogeneous solution of
`2-picolylzinc chloride (15