Steroids 78 (2013) 662–669

Contents lists available at SciVerse ScienceDirect

Steroids
journal homepage: www.elsevier.com/locate/steroids

Design, synthesis and biochemical studies of new 7a-allylandrostanes as

aromatase inhibitors
Carla L. Varela a, Cristina Amaral b,c, Georgina Correia-da-Silva b,c, Rui A. Carvalho d, Natércia A. Teixeira b,c,
Saul C. Costa a, Fernanda M.F. Roleira a,⇑, Elisiário J. Tavares-da-Silva a,⇑
a
CEF, Center for Pharmaceutical Studies, Pharmaceutical Chemistry Group, Faculty of Pharmacy, University of Coimbra, 3000-548 Coimbra, Portugal
b
Laboratory of Biochemistry, Department of Biological Sciences, Faculty of Pharmacy, University of Porto, Rua Jorge Viterbo Ferreira, 228, 4050-313 Porto, Portugal
c
Institute for Molecular and Cellular Biology (IBMC), University of Porto, Rua do Campo Alegre, 823, 4150-180 Porto, Portugal
d
Department of Life Sciences, Faculty of Science and Technology, University of Coimbra, Portugal

a r t i c l e i n f o a b s t r a c t

Article history: Two series of derivatives of 7a-allylandrostenedione, namely its 3-deoxo and 1-ene analogs, were
Received 14 December 2012 designed and synthesised and their biochemical activity towards aromatase evaluated. In each of these
Received in revised form 17 February 2013 series, the C-17 carbonyl group was further replaced by the hydroxyl and acetoxyl groups. The attained
Accepted 26 February 2013
data pointed out that the absence of the C-3 carbonyl group led to a slightly decrease in the inhibitory
Available online 13 March 2013
activity and the introduction of an additional double bond in C-1 revealed to be a very beneficial struc-
tural change in the studied compounds (compound 12, IC50 = 0.47 lM, Ki = 45.00 nM). Furthermore, the
Keywords:
relevance of the C-17 carbonyl group in the D-ring as a structural feature required to achieve maximum
Breast cancer
Synthetic steroids
aromatase inhibitory activity is also observed for this set of derivatives.
7a-Allylandrostanes Ó 2013 Elsevier Inc. All rights reserved.
Aromatase inhibitors
Structure-activity relationships

1. Introduction quently the growth effects of estrogens on primary and metastatic

hormone responsive breast cancers [4].
Many breast tumors are dependent upon estrogens for their The third-generation aromatase inhibitors (AIs) in clinical use,
development and continued growth [1]. As a consequence, it might such as the steroidal exemestane along with the non-steroidal
be expected that estrogen deprivation will both prevent the anastrazole and letrozole (Fig. 1), appear to be very potent with
appearance of these tumors and cause regression of the established greater target specificity, which allow circulating estrogens to fall
ones [2]. This has been the rationale that supports the hormone below detectable levels [5]. Despite their success, they still have
therapy of breast tumors. some disadvantages since they significantly increase musculoskel-
The primary site of estrogen biosynthesis varies with the men- etal symptoms, osteopenia, osteoporosis and fracture rate [3]. In
opausal status. Thus, in premenopausal women the ovaries are the addition, after some years of usage they can induce tumor resis-
major source of estrogen production whereas peripheral tissues tance. For this reason, it is important to seek for other potent and
such as adipose tissue, brain, blood vessels, skin, bone, endome- specific molecules in order to increase the number of AIs available
trium, and breast tissue are more important in postmenopausal pa- to allow surpassing the side effects and the resistance phenomena.
tients [3]. Estrogens are the end-products in the sequence of In the present study, our focus is the design, synthesis and bio-
steroid biosynthesis being the last step catalyzed by the cyto- chemical evaluation of a new series of steroids, based in the 7a-
chrome P-450 enzyme aromatase (CYP19). Aromatase is responsi- allylandrostenedione framework, containing extra modifications
ble for the conversion of androgen precursors (testosterone and in the A- and D-rings, in order to further study the structure–activ-
androstenedione) into the respective estrogens (estradiol and ity relationships (SAR) on AIs.
estrone). For this reason, aromatase inhibition has been a useful Brueggemeier and co-workers [6–9], have reported the synthe-
approach to selectively block the estrogen production and conse- sis of many C-7-substituted androstanes as AIs, having the most
part of the substituents a phenyl moiety. Also, Labrie et al. [10],
have described a rather active AI with an allyl functional group
⇑ Corresponding authors. Tel.: +351 239 488 400; fax: +351 239 488 503. in the C-7a position being this reference the only one found in
E-mail addresses: [email protected] (F.M.F. Roleira), [email protected] (E.J. Ta- the literature that describes this kind of substituent. In this man-
vares-da-Silva). ner, the allyl group was not particularly studied. Recently, Ghosh

0039-128X/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.steroids.2013.02.016
 C.L. Varela et al. / Steroids 78 (2013) 662–669 663

N N pounds was determined from the area peaks in the chromatogram

O
N N of the sample solution.
N N

2.2. Chemistry
O NC CN
NC CN 2.2.1. 3-Oxoandrost-4-en-17b-yl acetate (2)
To a solution of 1 (2.0 g, 6.93 mmol) in dry pyridine (48 mL),
Exemestane Anastrozole Letrozole
acetic anhydride (7.9 mL, 83.85 mmol) was added and the reaction
Fig. 1. The aromatase inhibitors in clinical use. was stirred for 21 h 30 min, at room temperature, until all the
starting material was consumed (TLC control). Dichloromethane
(250 mL) was added and the organic layer was washed with 10%
and collaborators [11], have designed novel C-6-substituted 2- NaHCO3 (3  150 mL), 10% HCl (3  150 mL) and water
alkynyloxy androstanes as potent AIs, where it was found that (3  150 mL), dried over anhydrous MgSO4, filtered and concen-
the C-6 linear side chains protrude into the access channel cavity trated to dryness. Crystallization of the obtained residue from ethyl
immobilizing the catalytic residues. Based on this, the structure– acetate gave the pure compound 2 (1.92 g, 84%) as translucent
activity relationships of steroids containing C-6 or C-7 linear chains prism crystals: mp 141–142 °C (ethyl acetate); IR (NaCl plates,
deserve to be revisited and further studied. Further, evidence that CHCl3) 3018 (=CH), 1736 (C@O), 1675 (C@C), 1248 (C–O); 1H
the aromatase inhibitory activity is similar when the same kind of NMR (600 MHz, CDCl3) d 0.82 (3H, s, 18-H3), 1.18 (3H, s, 19-H3),
substituent is in position C-6 or C-7 of a given steroidal framework, 2.03 (3H, s, CH3COO), 4.58 (1H, dd, J17a,16a = 9.0, J17a,16b = 8.0,
comes from previous reported works [9,12]. Keeping this in mind, 17a-H), 5.71 (1H, s, 4-H); 13C NMR (150 MHz, CDCl3) d 12.0 (C-
we are interested in performing the synthesis of 7a-allylandros- 18), 17.4 (C-19), 20.5, 21.1, 23.4, 27.4, 31.4, 32.7, 33.9, 35.4, 35.7,
tenedione derivatives where two main structural features were 36.6, 38.6, 42.4, 50.2, 53.7, 82.4 (C-17), 123.9 (C-4), 170.9 (C-5),
considered: the introduction of a double bond in C-1 and the 171.1 (OC@O), 199.4 (C-3).
abstraction of the carbonyl group at C-3. The introduction of a dou-
ble bond in C-1 is based on the rationale that the A-ring planarity 2.2.2. 3-Oxoandrosta-4,6-dien-17b-yl acetate (3)
in steroidal AIs is important for aromatase inhibition, as reported To a solution of 2 (3.8 g, 11.50 mmol) in acetic acid (50 mL) and
in previous works [13–17], and also because it has been estab- toluene (10 mL), under nitrogen, chloranil (2.5 g, 10.17 mmol) was
lished that the introduction of a double bond at C-1/C-2 of andro- added and the reaction mixture was refluxed for 3 h (UV spectro-
stenedione affords AIs that causes mechanism-based (suicide) photometric control). After cooling, the mixture was extracted
inactivation of aromatase [18–20]. In fact, the aromatization pro- with dichloromethane (3  200 mL) and the organic layer was
cess of androstenedione encompasses the C-1 and C-2 hydrogen sequentially washed with 2.5 N NaOH, 10% NaHCO3 (3  150 mL)
elimination [21]. There are also several studies revealing that, in and water (3  100 mL). After drying over anhydrous MgSO4, the
some AIs, the C-3 carbonyl group, present in the natural substrate solvent was removed under reduced pressure yielding a crude
of the enzyme, is not essential for aromatase inhibition [13,22,23]. product which was purified by silica gel chromatography (petro-
For this reason, 3-deoxy derivatives of 7a-allylandrostenedione leum ether 40–60 °C/diethyl ether) giving the pure compound 3
were synthesized and evaluated. Further, for each set of derivatives (2.2 g, 58%), as a yellow crystalline solid: mp 141–143 °C (ethyl
it was studied the influence of the functional group at C-17 in aro- acetate/n-hexane); IR (NaCl plates, CHCl3) 3024 (=CH), 1734
matase inhibitory activity. (C@O), 1658 (C@C), 1617 (C@C), 1030 (C–O); 1H NMR (600 MHz,
CDCl3) d 0.88 (3H, s, 18-H3), 1.11 (3H, s, 19-H3), 2.05 (3H, s, CH3-
COO), 4.63 (1H, dd, J17a,16a = J17a,16b = 9.0, 17a-H), 5.67 (1H, s, 4-
2. Experimental section H), 6.09 (1H, d, J6,7 = 10.4, 6-H), 6.12 (1H, dd, J7,6 = 10.4, J7,8 = 2.4,
7-H); 13C NMR (150 MHz, CDCl3) d 11.9 (C-18), 16.3 (C-19), 20.2,
2.1. General 21.1, 23.1, 27.4, 33.8, 33.9, 36.0, 36.5, 37.3, 43.4, 48.0, 50.6, 82.1
(C-17), 123.8 (C-7), 128.2 (C-6), 139.9 (C-4), 163.4 (C-5), 171.1
Reactions were controlled by thin layer chromatography (TLC) (OC@O), 199.4 (C-3).
using silica gel 60 F254 plates and the mobile phases used were
mixtures of petroleum ether 40–60 °C/ethyl acetate in different 2.2.3. 7a-Allyl-3-oxoandrost-4-en-17b-yl acetate (4)
proportions. Column chromatography was performed using silica To a solution of 3 (0.55 g, 1.67 mmol) in dry dichloromethane
gel 60 (0.063–0.200 mm). Melting points (MPs) were determined (50 mL), under magnetic stirring, nitrogen atmosphere and cooled
on a Reichert Thermopan hot block apparatus and were not cor- at 78 °C, titanium tetrachloride (3.0 mL, 27.29 mmol) was added.
rected. IR spectra were recorded on a Jasco 420 FT/IR spectrometer. Five min later, a solution of allyltrimethylsilane (3.0 mL,
The 1H NMR and 13C NMR spectra were recorded at 600 MHz and 18.6 mmol) in dry dichloromethane (5.0 mL) was carefully added
150 MHz, respectively on a Varian Unity 600. Chemical shifts were to the previous solution. The resulting mixture was stirred for
recorded in d values in parts per million (ppm) downfield from tet- 40 min and then allowed to warm to 30 °C. After that, water
ramethylsilane as an internal standard. All J values are given in Hz. was added, the mixture was extracted with dichloromethane
Mass spectra ESI and LC–MS were obtained with a mass spectrom- (3  100 mL) and the organic layer was washed with 10% NaHCO3
eter QIT-MS Thermo Finningan, model LCQ Advantage MAX cou- (2  100 mL) and water (3  100 mL), dried over anhydrous Na2-
pled to a Liquid Chromatography Equipment of High SO4, filtered and concentrated under reduced pressure yielding
Performance Thermo Finningan. Testosterone (1) was purchased an oily product (0.674 g). Purification by silica gel flash column
from Pharmacia & Upjahn Company, Kalamazoo, Michigan (USA). chromatography (petroleum ether 40–60 °C/diethyl ether) affor-
Reagents and solvents were used as obtained from suppliers with- ded the pure compound 4 (0.35 g, 56%): mp 150–152 °C (petro-
out further purification, with exception to dichloromethane, which leum ether 40–60 °C/diethyl ether); IR (NaCl plates, CHCl3) 3012
was dried through reflux and distilled from CaH2 [24]. (=CH), 1745 (C@O ester), 1722 (C@O), 1663 (C@C), 1041 (C–O);
1
All SAR compounds possess a purity superior to 98%. The purity H NMR (600 MHz, CDCl3) d 0.84 (3H, s, 18-H3), 1.20 (3H, s, 19-
was checked by HPLC with a C18-reversed phase column and H3), 2.04 (3H, s, CH3COO), 4.61 (1H, dd, J17a,16a = 8.9, J17a,16b = 8.2,
water/acetonitrile 40:60 as solvent. The purity of individual com- 17a-H), 4.98 (1H, ddd, CH@CH2), 5.02 (1H, ddd, CH@CH2), 5.64
664 C.L. Varela et al. / Steroids 78 (2013) 662–669

[AI] = 0 µM 2200

[AI] = 0.1 µM 2000

[AI] = 0.4 µM 1800

1/V (mol/min/ug protein)-1

[AI] = 0.8 µM 1600
1400
1200
1000
800
600
400
200
0
-0.2 -0.15 -0.1 -0.05 0 0.05 0.1
1/[S] (nM)

Fig. 2. Lineweaver–Burk plot for inhibitor 10 at different concentrations (0, 0.1, 0.4 and 0.8 lM), with androstenedione as substrate at 10, 20 and 30 nM. Each point
represents the mean of three independent experiments, performed in triplicate.

(1H, m, CH@CH2), 5.71 (1H, s, 4-H); 13C NMR (150 MHz, CDCl3) d (C@O), 1669 (C@C); 1H NMR (600 MHz, CDCl3) d 0.92 (3H, s, 18-
11.9 (C-18), 17.9 (C-19), 20.7, 21.1, 22.8, 27.3, 30.2, 33.9, 35.9, H3), 1.22 (3H, s, 19-H3), 5.01 (1H, ddd, CH@CH2), 5.05 (1H, ddd,
36.0, 36.1, 36.4, 38.3, 38.6, 42.5, 45.9, 46.9, 82.4 (C-17), 116.8 (C- CH@CH2), 5.67 (1H, m, CH@CH2), 5.73 (1H, s, 4-H); 13C NMR
24), 126.2 (C-4), 136.8 (C-23), 169.2 (C-5), 171.1 (C@O), 199.1 (C- (150 MHz, CDCl3) d 13.5 (C-18), 17.9 (C-19), 20.5, 21.3, 30.2, 31.1,
3); MS (ESI) m/z 369.53 ([M H]+, 100%). 33.9, 35.4, 35.6, 35.9, 36.0, 38.1, 38.7, 46.7, 47.1, 47.5, 117.1 (C-
22), 126.4 (C-4), 136.4 (C-21), 168.5 (C-5), 198.9 (C-3), 220.0 (C-
2.2.4. 7a-Allyl-3-oxoandrost-4-en-17b-ol (5) 17); MS (ESI) m/z 325.13 ([M H]+, 100%).
To a mixture of dioxane/water (85:15) (50 mL) with 2% NaOH
(10 mL), compound 4 (0.62 g, 1.67 mmol) was added and the 2.2.6. 7a-Allylandrost-4-en-17-one (9)
resulting solution was stirred overnight, at room temperature, Jones Reagent (1.1 mL); compound 8 (134.7 mg, 0.43 mmol);
after which chilled water was added leading to the formation of acetone/dioxane (60:10) (23 mL); the solid residue obtained
a precipitate. This precipitate was separate by filtration, dissolved (129.3 mg) was purified by a silica gel column chromatography
in ethyl acetate/dichloromethane (4:1), and the resulting solution (n-hexane/ethyl acetate (99:01)) affording the pure compound 9
was dried over anhydrous MgSO4 and concentrated under reduced (105.6 mg, 79%) as a white crystalline solid: mp 139–142 °C (n-
pressure yielding the pure compound 5 (0.57 g, 96%) as a white so- hexane/ethyl acetate); IR (NaCl plates, CHCl3) 3082 (=CH), 1734
lid: mp 216–218 °C (ethyl acetate/dichloromethane); IR (NaCl (C@O), 1640 (C@C); 1H NMR (600 MHz, CDCl3) d 0.89 (3H, s, 18-
plates, CHCl3) 3421 (OH), 1722 (C@O), 1645 (C@C), 1061 (C–O); H3), 1.05 (3H, s, 19-H3), 2.44 (1H, ddd, J16b,16a = 18.9, J16b,15b = 8.5,
1
H NMR (600 MHz, DMSO-d6) d 0.69 (3H, s, 18-H3), 1.16 (3H, s, J16b,15a = 1.3, 16b-H), 4.99 (2H, m, CH@CH2), 5.28 (1H, dd,
19-H3), 3.46 (1H, ddd, J17a,16a = 8.5, J17a,16b = 8.5, J17a,17b-OH = 4.8, J4,3a = 2.2, J4,3b = 2.2, 4-H), 5.72 (1H, m, CH@CH2); 13C NMR
17a-H), 4.47 (1H, d, J17b-OH,17a = 4.8, 17b-OH), 4.94 (1H, ddd, (150 MHz, CDCl3) d 13.5 (C-18), 19.2, 19.9 (C-19), 20.7, 21.4, 25.7,
CH@CH2), 5.00 (1H, m, CH@CH2), 5.56 (1H, s, 4-H), 5.72 (1H, m, 29.5, 31.4, 35.2, 35.7, 35.7, 37.2, 37.9, 38.3, 46.9, 47.3, 47.7, 115.8
CH@CH2); 13C NMR (150 MHz, DMSO-d6) d 11.0 (C-18), 17.4 (C- (C-22), 122.2 (C-4), 138.0 (C-21), 140.5 (C-5), 221.1 (C@O); MS
19), 20.4, 22.3, 29.6, 29.8, 33.6, 35.2, 35.5, 35.7, 36.0, 37.8, 38.2, (ESI) m/z 311.39 ([M H]+, 100%).
42.4, 45.5, 46.4, 79.8 (C-17), 116.1 (C-22), 125.2 (C-4), 137.4 (C-
21), 169.2 (C-5), 197.5 (C-3); MS (ESI) m/z 327.31 ([M H]+, 100%). 2.2.7. 7a-Allylandrosta-1,4-diene-3,17-dione (12)
Jones Reagent (0.3 mL); compound 10 (25.1 mg, 0.077 mmol);
2.2.4.1. General procedure to obtain 7a-allylandrost-4-ene-3,17-dione acetone/dioxane (60:10) (10 mL); the solid residue obtained
(6), 7a-allylandrost-4-en-17-one (9) and 7a-allylandrosta-1,4-diene- (27.5 mg) was purified by a silica gel column chromatography (n-
3,17-dione (12). Jones Reagent was added dropwise to solutions of hexane/ethyl acetate (70:30)) affording the pure compound 12
5, 8 or 10 in acetone/dioxane (60:10), at 0 °C, with magnetic stir- (22.1 mg, 88%) as a white crystalline solid: mp 205–207 °C (n-hex-
ring, until persistent brown coloration was obtained. The excess ane/ethyl acetate); IR (NaCl plates, CHCl3) 3076 (=CH), 1737 (C@O),
of oxidant agent was destroyed by addition of 2-propanol until a 1661 (C@C), 1617 (C@C); 1H NMR (600 MHz, CDCl3) d 0.95 (3H, s,
greenish colour was obtained and the reaction mixture was poured 18-H3), 1.26 (3H, s, 19-H3), 5.01 (1H, ddd, CH@CH2), 5.08 (1H,
in water (150 mL), extracted with ethyl acetate (3  100 mL) and ddd, CH@CH2), 5.67 (1H, m, CH@CH2), 6.06 (1H, dd, J4,2 = 1.8,
the organic layer washed with 10% NaHCO3 (2  100 mL) and J4,6a = 1.7, 4-H), 6.24 (1H, dd, J2,1 = 10.1, J2,4 = 1.8, 2-H), 7.05 (1H,
water (2  100 mL), dried over anhydrous MgSO4, filtered and con- d, J1,2 = 10.1, 1-H); 13C NMR (150 MHz, CDCl3) d 13.5 (C-18), 19.0
centrated to dryness. (C-19), 21.4, 21.7, 29.4, 30.9, 35.2, 35.3, 36.7, 37.7, 43.4, 44.9,
46.3, 47.6, 117.2 (C-22), 126.5 (C-4), 127.6 (C-2), 136.1 (C-21),
2.2.5. 7a-Allylandrost-4-ene-3,17-dione (6) 154.9 (C-1), 165.2 (C-5), 185.6 (C-3), 219.4 (C-17); MS (ESI) m/z
Jones Reagent (2.6 mL); compound 5 (0.53 g, 1.61 mmol); ace- 323.39 ([M H]+, 100%).
tone/dioxane (60:10) (70 mL); the residue obtained was insolubi-
lized in diisopropylether giving the pure compound 6 (0.386 g, 2.2.8. 7a-Allylandrost-4-en-17b-yl acetate (7)
73%) as a white crystalline compound: mp 218–221 °C (n-hex- Sodium borohydride (202.0 mg, 5.33 mmol) was added in small
ane/ethyl acetate); IR (NaCl plates, CHCl3) 3076 (=CH), 1734 proportions to a stirred and cooled mixture of trifluoracetic acid
 C.L. Varela et al. / Steroids 78 (2013) 662–669 665

2200 temperature, overnight. The reaction mixture was then cooled at

[S] = 10 nM
2000 10 °C and 3 N HCl was added until pH 5–7. Ethanol was evapo-
[S] = 20 nM
1800
rated under vacuum and the oily residue obtained was poured in
1/V (mol/min/ug protein)-1

[S] = 30 nM water (50 mL) and extracted with ethyl acetate (3  50 mL). The
1600
organic layer was washed with water (3  100 mL), dried over
1400 anhydrous MgSO4, filtered and concentrated to dryness giving a so-
1200 lid residue (18.2 mg). This residue was washed with chilled n-hex-
1000 ane leading to the pure compound 8 (4.4 mg, 29%) as a white solid:
800 mp 140–143 °C (ethyl acetate); IR (NaCl plates, CHCl3) 3297 (OH),
3076 (=CH), 1640 (C@C), 1059 (C–O); 1H NMR (600 MHz, CDCl3) d
600
0.77 (3H, s, 18-H3), 1.04 (3H, s, 19-H3), 3.64 (1H, dd, J17a,16a = 8.6,
400 J17a,16b = 8.6, 17a-H), 4.96 (2H, m, CH@CH2), 5.25 (1H, dd,
200 J4,3a = 2.3, J4,3b = 2.3, 4-H), 5.70 (1H, m, CH@CH2); 13C NMR
0 (150 MHz, CDCl3) d 10.9 (C-18), 19.3, 19.9 (C-19), 21.1, 22.8, 25.8,
-700 -600 -500 -400 -300 -200 -100 0 100 200 300 400 500 600 700 800 900 29.5, 30.4, 35.3, 36.2, 36.6, 37.2, 37.9, 38.8, 42.9, 46.3, 47.3, 82.0
[AI] (nM) (C-17), 115.5 (C-22), 121.8 (C-4), 138.5 (C-21), 141.1 (C-5); MS
(ESI) m/z 313.49 ([M H]+, 100%).
Fig. 3. Dixon plot for inhibitor 10 at different concentrations (0, 0.1, 0.4 and
0.8 lM), with androstenedione at 10, 20 and 30 nM, to determine the apparent
inhibition constant (Ki). Each point represents the mean of three independent
2.2.10. 7a-Allyl-3-oxoandrosta-1,4-dien-17b-ol (10)
experiments, performed in triplicate.
To a stirred solution of 5 (538.7 mg, 1.64 mmol) in toluene
(55 mL), DDQ (645.0 mg, 2.84 mmol) and benzoic acid (215.0 mg,
(1.24 mL), glacial acetic acid (1.24 mL) and acetonitrile (1.24 mL) 1.76 mmol) were added. After 18 h at reflux, an additional amount
followed by a solution of 4 (400.2 mg, 1.08 mmol) in dry dichloro- of DDQ (250 mg, 1.10 mmol) and benzoic acid (150 mg, 1.23 mmol)
methane (20 mL). The temperature was raised up at room temper- was added and the reaction proceeded for more 16 h. The reaction
ature and the reaction was stirred under nitrogen, until all the mixture was cooled at room temperature, filtered and the resulting
starting material had been consumed (1 h 30 min, TLC control). filtrate was evaporated giving an oily residue, which was dissolved
The reaction mixture was neutralized with a solution of 10% with dichloromethane and then mixed with silica gel. This mixture
NaHCO3, extracted with dichloromethane (4  100 mL) and the or- was filtered through basic alumina and washed with petroleum
ganic layer was washed with water (3  100 mL), dried over anhy- ether 40–60 °C/ethyl acetate (8:2) (250 mL). Evaporation of the fil-
drous MgSO4, filtered and concentrated to dryness giving a white trate afforded a crystalline residue which was purified by silica gel
solid residue (380.8 mg). This residue was separated by silica gel column chromatography (toluene/diethyl ether (65:35)) affording
column chromatography (n-hexane/ethyl acetate (99:1)) affording the pure compound 10 (268 mg, 50%) as a pure solid: mp 175–
293.8 mg of a fraction containing compound 7, which was further 178 °C (petroleum ether 40–60 °C/toluene); IR (NaCl plates, CHCl3)
purified by crystallization from ethanol/water giving the pure 7 3378 (OH), 3018 (=CH), 1652 (C@C), 1053 (C–O); 1H NMR
(26.9 mg) as clear and bright needle-like crystals: mp 129–130 °C (600 MHz, CDCl3) d 0.83 (3H, s, 18-H3), 1.24 (3H, s, 19-H3), 3.65
(ethanol/water); IR (NaCl plates, CHCl3) 3076 (=CH), 1728 (C@O), (1H, dd, J17a,16a = 8.6, J17a,16b = 8.6, 17a-H), 5.00 (1H, ddd, CH@CH2),
1041 (C–O); 1H NMR (600 MHz, CDCl3) d 0.81 (3H, s, 18-H3), 1.03 5.05 (1H, ddd, CH@CH2), 5.64 (1H, m, CH@CH2), 6.06 (1H, dd,
(3H, s, 19-H3), 2.04 (3H, s, CH3COO), 4.59 (1H, dd, J17a,16a = 8.5, J4,2 = 1.8, J4,6a = 1.7, 4-H), 6.23 (1H, dd, J2,1 = 10.1, J2,4 = 1.8, 2-H),
J17a,16b = 8.5, 17a-H), 4.96 (2H, m, CH@CH2), 5.25 (1H, dd, 7.06 (1H, d, J1,2 = 10.1, 1-H); 13C NMR (150 MHz, CDCl3) d 10.9,
J4,3a = 2.3, J4,3b = 2.3, 4-H), 5.69 (1H, m, CH@CH2); 13C NMR 19.1, 22.2, 22.9, 29.4, 30.2, 35.4, 36.1, 37.5, 38.2, 43.1, 43.6, 45.2,
(150 MHz, CDCl3) d 11.8 (C-18), 19.2, 19.9 (C-19), 20.9, 21.2, 22.9, 45.9, 81.4 (C-17), 116.9 (C-22), 126.3 (C-4), 127.5 (C-2), 136.7 (C-
25.8, 27.4, 29.5, 35.3, 36.2, 36.7, 37.1, 37.9, 38.5, 42.6, 46.1, 47.2, 21), 155.6 (C-1), 166.1 (C-5), 185.9 (C-3); MS (ESI) m/z 325.39
82.8 (C-17), 115.9 (C-24), 121.9 (C-4), 138.4 (C-23), 140.9 (C-5), ([M H]+, 100%).
171.2 (C@O); MS (ESI) m/z 355.49 ([M H]+, 100%).

2.2.9. 7a-Allylandrost-4-en-17b-ol (8) 2.2.11. 7a-Allyl-3-oxoandrosta-1,4-dien-17b-yl acetate (11)

Compound 7 (223.2 mg, 0.63 mmol) was added to a mixture of To a solution of 10 (50.0 mg, 0.15 mmol) in dry pyridine
dioxane/water (85:15) (18 mL) with 2% NaOH (3.6 mL), at room (4.0 mL), acetic anhydride (0.2 mL, 1.79 mmol) was added and
temperature, and the reaction mixture was stirred until total trans- the reaction was stirred for 33 h 15 min at room temperature, until
formation of the starting material (48 h, TLC control) being after
this time neutralized with a solution of 5% HCl. The dioxane was Table 1
evaporated under vacuum leading to a residue that was dissolved Aromatase inhibition of A-, B- and D-ring modified steroids.a
in ethyl acetate (50 mL), diluted with water (100 mL) and extracted Compounds Aromatase inhibition (%) ± SEM
with ethyl acetate (3  100 mL). The organic layer was washed
4 10.12 ± 0.74
with water (3  100 mL), dried over anhydrous MgSO4, filtered 5 24.74 ± 1.56
and concentrated to dryness giving a white solid residue 6 83.14 ± 1.95
(209.3 mg). A portion of this residue was separated by silica gel 7 31.33 ± 2.24
8 52.44 ± 3.85
column chromatography (n-hexane/diethyl ether (80:20)) afford-
9 84.29 ± 3.23
ing compound 8 in mixture with some impurities. 10 87.94 ± 3.04
A pure sample of 8 was obtained from 9 by the following 11 13.01 ± 0.35
procedure: 12 94.49 ± 1.04
To a stirred solution of 9 (15.3 mg, 0.049 mmol) in ethanol Formestane 99.21 ± 0.04

(3.0 mL), under nitrogen, at 10 °C, sodium borohydride (0.6 mg, a

Concentrations of 40 nM [1b-3H] androstenedione, 20 lg protein from human
0.016 mmol) was added. The mixture was stirred for 1 h and after placental microsomes, 2 lM of the compounds and 15 min incubation were used.
two subsequent additions of sodium borohydride (0.6 mg, Results represent the mean ± S.E.M. of three different experiments, performed in
triplicate. Formestane at 0.5 lM was used as reference compound.
0.016 mmol; 0.7 mg, 0.018 mmol), the reaction was left, at room
666 C.L. Varela et al. / Steroids 78 (2013) 662–669

OH

O
1
i
OCOCH3

O
2
ii OCOCH3

O
3
iii
OCOCH3 OCOCH3 OCOCH3

vi

O O
11 4 7
x iv vii
OH OH OH

ix

O O
10 5 8
v O v O v O viii

O O
12 6 9

Scheme 1. Synthesis of aromatase inhibitors from testosterone. Reagents and conditions: (i) (CH3CO)2O, dry pyridine, rt, 21 h 30 min; (ii) chloranil, acetic acid, toluene,
118 °C, 3 h; (iii) TiCl4, allyltrimethylsilane, dry dichloromethane, 78 °C, 40 min; (iv) 2% NaOH, dioxane/water, rt, overnight; (v) Jones Reagent, acetone/dioxane, 0 °C; (vi)
CF3COOH, CH3COOH, CH3CN, NaBH4, dry dichloromethane, rt, 1 h 30 min; (vii) 2% NaOH, dioxane/water, rt, 48 h; (viii) NaBH4, ethanol, 10 to 5 °C, overnight; (ix) DDQ,
benzoic acid, toluene, 118 °C, 34 h; (x) (CH3CO2)2O, dry pyridine, rt, 33 h 15 min.

all the starting material had been consumed (TLC control). Dichlo- hospital, human placenta were placed in cold 67 mM potassium
romethane (100 mL) was added and the organic layer was washed phosphate buffer (pH 7.4) containing 1% KCl. The cotyledon tissue
with 10% NaHCO3 (2  100 mL), 10% HCl (2  100 mL) and water was separated and homogenized in a Polytron homogenizer with
(2  100 mL), dried over anhydrous MgSO4, filtered and concen- 67 mM potassium phosphate buffer (pH 7.4) containing 0.25 M su-
trated to dryness. The resulting residue was crystalized from ethyl crose and 0.5 mM dithiothreitol (DTT, 1:1, w/v). The homogenate
acetate/n-hexane giving the pure compound 11 (29.6 mg, 54%) as was centrifuged at 5000g for 30 min and the supernatant was cen-
white needles: mp 155–159 °C (ethyl acetate/n-hexane); IR (NaCl trifuged at 20,000g for 30 min and after, at 54,000g for 45 min to
plates, CHCl3) 3076 (=CH), 1731 (C@O), 1661 (C@C), 1617 (C@C), yield the microsomal pellet. After ultra-centrifugation, the micro-
1042 (C–O); 1H NMR (600 MHz, CDCl3) d 0.87 (3H, s, 18-H3), 1.24 somes were washed and ressuspended in 67 mM potassium phos-
(3H, s, 19-H3), 2.04 (3H, s, CH3COO), 4.59 (1H, dd, J17a,16a = 8.5, phate buffer (pH 7.4) containing 0.25 M sucrose, 20% glycerol, and
J17a,16b = 8.5, 17a-H), 5.01 (1H, ddd, CH@CH2), 5.05 (1H, ddd, 0.5 mM DTT and stored at 80 °C. All the procedures were carried
CH@CH2), 5.64 (1H, m, CH@CH2), 6.04 (1H, dd, J4,2 = 1.6, out at 0–5 °C. Using bovine serum albumin (BSA) as a standard, the
J4,6a = 1.5, 4-H), 6.23 (1H, dd, J2,1 = 10.1, J2,4 = 1.6, 2-H), 7.05 (1H, protein content was calculated by the Bio-Rad protein assay (Bio-
d, J1,2 = 10.1, 1-H); 13C NMR (150 MHz, CDCl3) d 11.9 (C-18), 19.2 Rad Labs, Munich, Germany).
(C-19), 21.1, 22.1, 23.1, 27.3, 29.4, 35.4, 36.3, 37.5, 38.0, 42.7,
43.5, 45.0, 45.7, 82.2 (C-17), 117.1 (C-24), 126.4 (C-4), 127.6 (C-
2), 136.5 (C-23), 155.4 (C-1), 165.9 (C-5), 171.0 (C@O), 185.8 (C- 2.3.2. Aromatase assay procedure
3); MS (ESI) m/z 367.49 ([M H]+, 100%). By measuring the tritiated H2O released from [1b-3H] andro-
stenedione during the aromatization process, the aromatase activ-
ity was measured in human placental microsomes according to
2.3. Biochemistry Thompson and Siiteri [21] and Heidrich et al. [25] method with
modifications [13]. The [1b-3H] androstenedione was purchased
2.3.1. Preparation of placental microsomes from PerkinElmer Life Sciences (Boston, MA, USA).
The human placental microsomes were obtained as described The steroidal AIs tested were dissolved in DMSO and diluted in
previously by our group [13,15]. After delivery from a local 67 mM potassium phosphate buffer (pH 7.4) before assay. To
 C.L. Varela et al. / Steroids 78 (2013) 662–669 667

Table 2
IC50 and kinetic studies of the most potent inhibitors.

Compounds IC50 (lM)a Type of inhibition Kinetic Studiesb Ki (nM) Real affinity
Vm (mol/min/ug prot) (Km/Ki) (nM)
6 0.59 Competitive 0.009 ± 0.001 80.00 0.458 ± 0.042
9 0.75 Competitive 0.017 ± 0.003 60.00 0.014 ± 0.003
10 0.45 Competitive 0.004 ± 0.002 65.00 0.835 ± 0.041
12 0.47 Competitive 0.006 ± 0.001 45.00 0.926 ± 0.055
Formestane 0.04 – – – –
a
Concentrations of 100 nM [1b-3H] androstenedione, 20 lg protein from human placental microsomes, different concentrations of the compounds and 15 min incubation
were used.
b
Concentrations of 10, 20 and 30 nM [1b-3H] androstenedione, 20 lg protein from human placental microsomes, different concentrations of the compounds and 5 min
incubation were used. Apparent inhibition constants (Ki) were obtained by Dixon Plot. Inhibition type was based on analysis of Lineweaver–Burk plot. The experiments were
done in triplicate in three independent experiments.

determine the percentage of aromatase inhibition as a screening submitted to base-catalysed hydrolysis giving a crude product
assay, for the reaction mixture was used 20 lg of microsomal pro- which, after column chromatography, allowed obtaining alcohol
tein, 40 nM of [1b-3H] androstenedione (1 lCi) and 2 lM of each 8, with some remaining inseparable impurities. Oxidation of the
aromatase inhibitor in a final reaction volume of 1 mL. To deter- impure compound 8 with Jones reagent gave rise, after column
mine the IC50 it was used 100 nM (1 lCi) of [1b-3H] androstenedi- chromatography, to the pure ketone 9, in 79% yield. Compound 8
one and different concentrations of the inhibitors (0.0125–5 lM). was then further reobtained in its pure form as a white solid, in
The aromatase reaction was initiated by the addition of 150 lM 29% yield, by reduction of ketone 9 with sodium borohydride in
of reduced nicotinamide adenine dinucleotide phosphate (NADPH), ethanol [29].
and incubations were performed in a shaking water bath at 37 °C Aiming to prepare the D1-derivative 10 (Scheme 1), compound
for 15 min. For the kinetic studies were used different concentra- 5 was treated with 2,3-dichloro-5,6-dicyano-benzoquinone (DDQ)
tions of [1b-3H] androstenedione (10–40 nM) and inhibitors and benzoic acid in refluxing toluene [30]. After several filtrations
(0–0.8 lM) and to minimize the time-dependent loss of the initial and purification by silica gel column chromatography, the alcohol
aromatization rate was used 5 min of incubation. The aromatase 10 was isolated in 50% yield. Compound 10 was then treated with
reactions were stopped by addition of 250 lL of 20% trichloroacetic acetic anhydride in pyridine giving a crude product which, after
acid (TCA) and the mixture was transferred to microcentrifuge ethyl acetate/n-hexane crystallization, led to acetate 11. Jones oxi-
tubes containing a charcoal–dextran pellet, vortexed and incu- dation of 10 led, after column chromatography purification, to the
bated during 1 h. After incubation and centrifugation at 14,000g pure ketone 12, in 88% yield.
for 10 min, the supernatants were transferred to new charcoal-
dextran pellets and incubated for 10 min. Following a new
3.2. Biochemistry
centrifugation cycle, the supernatant containing the tritiated water
product was mixed with a liquid scintillation cocktail from ICN
In a series of A-, B- and D-ring modified steroids it was evalu-
Radiochemicals (Irvine, CA, USA) and counted in a liquid scintilla-
ated the aromatase activity in human placental microsomes using
tion counter (LS-6500, Beckman Coulter, Inc., Fullerton, CA). All
a radiometric assay [21]. Firstly, it was done a screening assay by
experiments were carried out in triplicate. As a reference com-
determining the percentage of aromatase inhibition (%) for all the
pound it was used formestane at 0.5 lM.
studied compounds at 2 lM (Table 1). Four of the evaluated ste-
roids presented an inhibitory aromatase activity higher than 80%.
The aromatase inhibitor 4-hydroxyandrostenedione (formestane)
3. Results
at 0.5 lM was used as a reference compound. Secondly, the IC50
was determined for the four most potent steroids, 6, 9, 10 and 12
3.1. Chemistry
(Table 2). For these steroids, it was also performed kinetic studies
to characterize the type of binding to the active site of aromatase
The reaction of testosterone 1 with acetic anhydride yielded its
and the apparent inhibition constant (Ki), using different concen-
acetate 2, in 84% yield (Scheme 1). Dehydrogenation of 2 with
trations of inhibitor and substrate. As shown by the Lineweaver–
chloranil was undertaken in a dry environment, leading to a solid
Burk plot in Fig. 2, the new compounds revealed to be competitive
residue [26]. This residue was then purified by column chromatog-
inhibitors and by the Dixon plots it was characterized the enzyme
raphy affording the dienone 3, in 58% yield. Subsequently, a Saku-
affinity (Fig. 3). Our results showed that the most potent AIs are
rai reaction was performed in 3 with allyltrimethylsilane and
compounds 10 and 12, being 12 the most effective with higher
titanium tetrachloride, in order to introduce the allyl group at
affinity to the enzyme aromatase.
the C-7a position [10,27]. After work-up, an oily crude was ob-
tained which was purified over column chromatography allowing
the isolation of the 7a-allyl derivative 4, in 56% yield. Compound 4. Discussion
4 was next submitted to alkaline hydrolysis giving the alcohol 5,
in 96% yield. Compound 5 was subsequently oxidized using Jones As referred in the introduction, many studies suggest that the
Reagent leading, after insolubilization in diisopropylether, to the presence of alkyl or phenyl groups in the C-6 or C-7 positions of
ketone 6, in 73% yield. the steroidal androstane scaffold is beneficial for achieving aroma-
In order to obtain the 3-deoxo derivative 7 (Scheme 1), the reac- tase inhibitory activity [6–12,22]. This is because it was believed
tion of 4 with a mixture of sodium borohydride in trifluoracetic that a hydrophobic binding pocket of the enzyme aromatase would
acid, glacial acetic acid and acetonitrile was performed in a con- interact with these groups leading to a more effective ligand-en-
trolled environment [28] giving, after column chromatography, zyme interaction. Surprisingly, when, recently, the 3D structure
the almost pure compound 7. An analytical sample of 7 was ob- of aromatase in a complex with its substrate androstenedione
tained by ethanol/water crystallization. This compound was then was elucidated [31], no particular binding pocket near the C-6
668 C.L. Varela et al. / Steroids 78 (2013) 662–669

and C-7 region of androstenedione was found. However, a shallow allowing the establishment of hydrogen bonds with aromatase
hydrophobic crevice was observed to accommodate the extra C-6- active site residues, and also because the A-ring planarity is in-
methylidene group of exemestane (Fig. 1), when its molecule was creased, by the presence of the D1 double bond. In fact, the A-ring
built into the active site of aromatase. In a more recent work, planarity seems to assume an important role since the same type
Ghosh and collaborators [11], determined the crystal structures of substitution in 6 to give 5 produces a dramatic decrease in the
of inhibited complexes of aromatase with exemestane and two aromatase inhibition (Table 1). The substitution of the C-17 car-
other C-6-substituted 2-alkynyloxy compounds. It was found that, bonyl group of 12 by an acetate group as in 11, decreases the aro-
the linear C-6-alkynyloxy side chains protrude into the access matase inhibitory activity (Table 1) even more than that observed
channel cavity immobilizing the catalytic residues, resulting in in the 3-deoxo derivatives series. This dramatic decrease was also
very potent inhibition. This means that the hydrophobic binding observed when substituting the C-17 carbonyl group of 6 by an
pocket of aromatase can possibly be the access channel to the en- acetate group as in 4. This can be due to stereochemical conflicts
zyme. Taking this in consideration along with the evidence that the with aromatase active site, because of the larger size of inhibitors
aromatase inhibitory activity is similar when the same substituent 4 and 11 when compared with 7.
is in position C-6 or C-7, of a given steroidal framework [9,12], we In summary, several 7a-allyl derivatives of androstenedione
have decided to synthesize C-7a-allyl derivatives of androstenedi- were prepared and found to be strong aromatase inhibitors. Con-
one by introducing an allyl group in its C-7 position (Scheme 1) and cerning its SAR, the presence of a double bond in C-1 as in 10
obtaining compound 6, which is a strong aromatase inhibitor and 12, which increases the planarity of the steroidal A-ring, is
(IC50 = 0.59 lM, Ki = 80.00 nM) (Tables 1 and 2) as before described beneficial for aromatase inhibition. The abstraction of the C-3 car-
by Labrie et al. [10]. From the 7a-allylandrostenedione scaffold, bonyl group did not particularly favour the aromatase inhibitory
two main chemical modifications were further performed. One of activity, in some of these compounds, being the introduction of a
them consisted in C-3 decarbonylation in order to obtain 7a- double bond in C-1 more beneficial for aromatase inhibition than
allyl-3-deoxo-4-androstenes. Actually, it is known from previous the C-3 carbonyl abstraction. The substitution of the C-17 carbonyl
works [13,22,23], that the C-3 carbonyl group is not essential for group by a hydroxyl group usually decreases the aromatase inhibi-
aromatase inhibition, since the C-17 carbonyl group remains in tion, except if a double bond in C-1 is present as in 10. The substi-
the molecule [14]. In fact, aromatase establishes two hydrogen tution of the C-17 carbonyl group by an acetate group dramatically
bonds with the carbonyl groups at C-3 and C-17 [31] of its sub- decreases the aromatase inhibitory activity. This effect seems to be
strate androstenedione and, at least one of them must be present lower when the length of the steroid molecule is also smaller as in
in the steroid molecule in order to bind the enzyme. 7.
Other modification was the introduction of an additional double
bond at C-1 position aiming to obtain 7a-allyl-3-oxo-1,4-androsta- Acknowledgments
dienes, since it is also known that the increase in A-ring planarity
and electronic density are very beneficial for aromatase inhibitory The authors are grateful to FCT (Fundação para a Ciência e Tecn-
activity [13–17]. In addition, for each set of derivatives it was stud- ologia) for financial support and for the PhD grants for Carla Varela
ied the influence of several functional groups at C-17. and Cristina Amaral (SFRH/BD/44872/2008 and SFRH/BD/48190/
Regarding the 7a-allyl-3-deoxo-4-androstenes 7, 8 and 9, it was 2008, respectively). We also acknowledge the ‘‘Rede Nacional de
observed that the 3-deoxo analogue 9 of compound 6 is slightly RMN’’ (REDE/1517/RMN/2005) for access to the facilities. This
less potent (IC50 = 0.75 lM) (Table 2) than 6 showing that, in this work was funded by FEDER Funds through the Operational Com-
case, the C-3 carbonyl group is important for aromatase inhibitory petitiveness Program-COMPETE and by National Funds through
activity. When one substitutes the C-17 carbonyl group of 9 by a FCT under the project FCOMP-01-0124-FEDER-020970 (PTDC/
hydroxyl group as in 8, the activity decreased, as expected (Ta- QUI-BIQ/120319/2010).
ble 1). In this case, no C-3 or C-17 carbonyl groups are present,
being therefore difficult the binding of the inhibitor to the enzyme Appendix A. Supplementary material
active site. Substituting the C-17 carbonyl group of 9 by an acetate
group as in 7, the aromatase inhibitory activity decreased even Supplementary data associated with this article can be found, in
more (Table 1). In fact, the volume of the aromatase active site is the online version, at http://dx.doi.org/10.1016/j.steroids.2013.
short (400 Å3) [31], and bulky substituents in the androstane 02.016.
framework are not allowed if one aims to obtain strong aromatase
inhibitors. References
Concerning the 7a-allyl-3-oxo-1,4-androstadienes 10, 11 and
12, it was observed that the D1 analogue 12 is a strong AI [1] Miller WR. Aromatase inhibitors. Endocr-Relat Cancer 1996;3:65–79.
(IC50 = 0.47 lM, Ki = 45.00 nM) with great affinity for the enzyme [2] Henderson IC, Canellos GP. Cancer of the breast – Past decade. New Engl J Med
1980;302:17–30.
aromatase (Tables 1 and 2). It is more potent than 6 and also more [3] Chumsri S, Howes T, Bao T, Sabnis G, Brodie A. Aromatase, aromatase
potent than the corresponding 3-deoxo compound 9 (Tables 1 and inhibitors, and breast cancer. J Steroid Biochem 2011;125:13–22.
2), proving that the introduction of a double bond in C-1, hence [4] Brodie A. Aromatase inhibitors and models for breast cancer. In: Furr BJA,
Parnham MJ, Bruinvels J, editors. Aromatase inhibitors – Milestones in drug
increasing the planarity of the A-ring, is more beneficial for aroma- therapy. Basel: Birkhäuser Verlag; 2006. p. 23–44.
tase inhibition than the C-3 carbonyl abstraction, in the studied [5] Lønning PE. Pharmacology of new aromatase inhibitors. Breast 1996;5:202–8.
compounds. In fact, compound 12 is structurally similar to [6] Snider CE, Brueggemeier RW. Potent enzyme-activated inhibition of aromatase
by a 7alpha-substituted C19 steroid. J Biol Chem 1987;262:8685–9.
exemestane (Fig. 1) differing by the longer alkenyl group (allyl) [7] Ebrahimian S, Chen HH, Brueggemeier RW. Synthesis and biochemical studies
at C-7 position, instead of the shorter one (methylene group) at of 7alpha-substituted androsta-1,4-diene-3,17-diones as enzyme-activated
C-6 position of exemestane (Fig. 1). Surprisingly, when one substi- irreversible inhibitors of aromatase. Steroids 1993;58:414–22.
[8] Darby MV, Lovett JA, Brueggemeier RW, Groziak MP, Counsell RE. 7alpha-
tutes the C-17 carbonyl group of 12 by a hydroxyl group as in 10,
Substituted derivatives of androstenedione as inhibitors of estrogen
the aromatase inhibitory activity is almost maintained biosynthesis. J Med Chem 1985;28:803–7.
(IC50 = 0.45 lM, Ki = 65.00 nM), showing 10 lower affinity to aro- [9] O’Reilly JM, Li NY, Duax WL, Brueggemeier RW. Synthesis, structure
matase (Tables 1 and 2). In this case, the decrease is much lower elucidation, and biochemical evaluation of 7alpha- and 7beta-arylaliphatic-
substituted androst-4-ene-3,17-diones as inhibitors of aromatase. J Med Chem
than that observed in the 3-deoxo derivatives (9 to 8), probably be- 1995;38:2842–50.
cause a carbonyl group at C-3 position remains in the molecule, [10] Labrie F, Merand Y. Aromatase inhibitors. WO 91/12206 1991.
 C.L. Varela et al. / Steroids 78 (2013) 662–669 669

[11] Ghosh D, Lo J, Morton D, Valette D, Xi J, Griswold J, et al. Novel aromatase [21] Thompson Jr EA, Siiteri PK. The involvement of human placental microsomal
inhibitors by structure-guided design. J Med Chem 2012;55:8464–76. cytochrome P-450 in aromatization. J Biol Chem 1974;249:5373–8.
[12] Numazawa M, Yamaguchi S. 6-Phenylaliphatic-substituted androst-4-ene- [22] Numazawa M, Yamada K, Nitta S, Sasaki C, Kidokoro K. Role of hydrophilic
3,17-diones as aromatase inhibitors: structure-activity relationships. J Steroid interaction in binding of hydroxylated 3-deoxy C-19 steroids to the active site
Biochem 1998;67:41–8. of aromatase. J Med Chem 2001;44:4277–83.
[13] Cepa MMDS, Tavares-da-Silva EJ, Correia-da-Silva G, Roleira FMF, Teixeira [23] Nagaoka M, Watari Y, Yajima H, Tsukioka K, Muroi Y, Yamada K, et al.
NAA. Structure-activity relationships of new A, D-ring modified steroids as Structure-activity relationships of 3-deoxy androgens as aromatase inhibitors.
aromatase inhibitors: design, synthesis, and biological activity evaluation. J Synthesis and biochemical studies of 4-substituted 4-ene and 5-ene steroids.
Med Chem 2005;48:6379–85. Steroids 2003;68:533–42.
[14] Cepa MMDS, Tavares-da-Silva EJ, Correia-da-Silva G, Roleira FMF, Teixeira [24] Perrin DD, Armarego WLF. Purification of laboratory chemicals. 3rd
NAA. Synthesis and biochemical studies of 17-substituted androst-3-enes and ed. Oxford: Pergamon; 1988.
3,4-epoxyandrostanes as aromatase inhibitors. Steroids 2008;73:1409–15. [25] Heidrich D, Steckelbroeck S, Klingmuller D. Inhibiton of human cytocrome P-
[15] Varela C, Tavares-da-Silva EJ, Amaral C, Correia-da-Silva G, Baptista T, Alcaro S, 450 aromatase activity by butylins. Steroids 2001;66:763–9.
et al. New structure-activity relationships of A- and D-ring modified steroidal [26] Specht H, Jahn H, Stachowiak A. Preparation of D 4,6–3-oxo steroids. Chem
aromatase inhibitors: design, synthesis, and biochemical evaluation. J Med Abstr 1966;64:14246.
Chem 2012;55:3992–4002. [27] Sakurai H. Reactions of allylsilanes and aplication to organic synthesis. Pure
[16] Cole PA, Robinson CH. Mechanism and inhibition of cytochrome P-450 Appl Chem 1982;54:1–22.
aromatase. J Med Chem 1990;33:2933–42. [28] Hanson JR, Hitchcock PB, Liman MD, Nagaratnam S. Facial selectivity in the
[17] Numazawa M, Sohtome N, Nagaoka M. Stereochemistry of NaBH4 reduction of hydroboration of androst-4-enes. J Chem Soc, Perkin Trans 1995;1:2183–7.
a 19-carbonyl group of 3-deoxy androgens. Synthesis of [19S-3H]- and [29] Gardner JN, Anderson BA, Oliveto EP. Total synthesis of some (+-)-18-methyl-
[19R-3H]-labeled aromatase inhibitors having a 19-hydroxy groups. Chem 9beta,10alpha-androstanes and (+-)-18-methyl-9beta,10alpha-D-
Pharm Bull 2004;52:722–6. homoandrostanes. J Org Chem 1969;34:107–12.
[18] Covey DF, Hood WF. A new hypothesis based on suicide substrate inhibitor [30] Zinczuk J, Bacigaluppo JA, Colombo MI, Cravero RM, González-Sierra M,
studies for the mechanism of action of aromatase. Cancer Res Rúveda EA. An efficient and environmentally benign chemical synthesis of
1982;42:3327s–32s. testolactone. J Braz Chem Soc 2003;14:970–4.
[19] Numazawa M, Ando M, Zennyoji R. Studies on the catalytic function of [31] Ghosh D, Griswold J, Erman M, Pangborn W. Structural basis for androgen
aromatase: aromatization of 6-alkoxy-substituted androgens. J Steroid specificity and oestrogen synthesis in human aromatase. Nature
Biochem 2002;82:65–73. 2009;457:219–24.
[20] Marcos-Escribano A, Bermejo FA, Bonde-Larsen AL, Retuerto JI, Sierra IH. 1,2-
Dehydrogenation of steroidal 6-methylen derivatives. Synthesis of
exemestane. Tetrahedron 2009;65:7587–90.