New aromatase inhibitors from the 3-pyridyl arylether and 1-aryl pyrrolo[2,3-c]pyridine series

Article abstract of DOI:10.1016/j.bmcl.2012.01.076

Aromatase inhibition is the new standard of care for estrogen receptor positive breast cancer and has also potential for treatment of other diseases such as endometriosis. Simple and readily available 3-pyridyl arylethers and 1-aryl pyrrolo[2,3-c]pyridines recapitulating the key pharmacophore elements of Letrozole (1) are described and their structure-activity relationships are discussed. Potent and ligand efficient leads such as compound 23 (IC ₅₀ = 59 nM on aromatase) have been identified.

Full text of DOI:10.1016/j.bmcl.2012.01.076

Bioorganic & Medicinal Chemistry Letters 22 (2012) 1860–1863
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
New aromatase inhibitors from the 3-pyridyl arylether and 1-aryl
pyrrolo[2,3-c]pyridine series
⇑
Frédéric Stauffer , Pascal Furet, Andreas Floersheimer, Marc Lang
Novartis Institutes for BioMedical Research, Global Discovery Chemistry Oncology, Postfach, CH-4002 Basel, Switzerland
a r t i c l e i n f o
a b s t r a c t
Article history:
Aromatase inhibition is the new standard of care for estrogen receptor positive breast cancer and has also
potential for treatment of other diseases such as endometriosis. Simple and readily available 3-pyridyl
arylethers and 1-aryl pyrrolo[2,3-c]pyridines recapitulating the key pharmacophore elements of
Letrozole (1) are described and their structure–activity relationships are discussed. Potent and ligand
efficient leads such as compound 23 (IC50 = 59 nM on aromatase) have been identified.
Ó 2012 Elsevier Ltd. All rights reserved.
Received 8 December 2011
Revised 20 January 2012
Accepted 21 January 2012
Available online 28 January 2012
Keywords:
Aromatase
CYP19
Aromatase inhibition
All the endogenous estrogens (estradiol and estrone) are origi-
nating from the conversion of androgens (testosterone and andro-
stenedione, respectively) by the aromatase enzyme complex, a
cytochrome P450 dependant enzyme. The aromatase enzyme com-
plex is comprised of two polypeptides. The first of these is a spe-
cific cytochrome P450, named aromatase cytochrome P450
(CYP450arom) and is the product of a single gene, the CYP19 gene.
The second is a flavoprotein, NADPH-cytochrome P450 reductase
and is ubiquitously distributed in most cells. Ovary, testis, adipose
tissue, skin, hypothalamus and placenta express aromatase
normally, whereas breast and endometrial cancers, endometriosis,
and uterine fibroids overexpress aromatase and produce local
estrogens that exert paracrine and intracrine effects. Three mecha-
nisms are responsible for aromatase overexpression in a patholog-
ical tissue versus its normal counterpart: a cellular component is
altered to increase aromatase-expressing cell types; molecular
alterations in stromal cells favor binding of transcriptional enhanc-
ers versus inhibitors to a normally quiescent aromatase promoter
and initiate transcription; or heterozygous mutations, which cause
the aromatase coding region to lie adjacent to constitutively active
cryptic promoters that normally transcribe other genes, result in
excessive estrogen formation.1
X
CN
CN
N
CN
8.5 Å
O
Fe......
N
N
N
N
N
N
N
L shape of
the minimal
CN
1
2
3
pharmacophore
Figure 1. Schematic analysis of the minimum requirement for the aromatase
pharmacophore as identified in compounds 1 (box), 2 and 3.
of the ligand as compared to the former strategy using the antago-
nistic effect of Tamoxifen on the receptor.
Letrozole (1) is a potent non-steroidal aromatase inhibitor
registered for clinical use. Its reported synthesis is only two steps
away from commercially available starting material.2 It has been
recognized from analysis of the SAR of known non-steroidal
aromatase inhibitors that the key pharmacophore necessary for
activity has a L shape displaying at one end a nitrogen electron
donor atom to interact with the iron of the heme and at the other
end a hydrogen bond acceptor or a halogen distant of around 8.5 Å.
The minimal efficient pharmacophore of the low nanomolar
aromatase inhibitor Letrozole (1; IC50 = 1–11 nM)3 is its synthetic
intermediate 4-([1,2,4]triazol-1-ylmethyl)-benzonitrile (2; 88%
Selective aromatase inhibition, as displayed for instance by the
third generation non-steroidal aromatase inhibitor Letrozole (1),
has become the standard of care for estrogen receptor positive
breast cancer and allows reduction of total body estrogens. The
abrogation of estrogen receptors activation is obtained by depletion
inhibition at 1 lM) (Fig. 1).
In an effort to identify new aromatase inhibitor leads, we have
investigated derivatives using 3-pyridyl as a replacement for the
5-membered ring triazole heme binding moiety. Such a modifica-
tion had already been shown to be adequate in the 4-(aminohetero-
aryl)-benzonitrile series.4 The simple commercially available
4-(3-pyridyloxy)benzonitrile (3) is fulfilling the minimal
⇑
Corresponding author. Tel.: +41 61 696 8749; fax: +41 61 696 6246.
E-mail address: frederic.stauffer@novartis.com (F. Stauffer).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2012.01.076

F. Stauffer et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1860–1863
1861
Table 1
Inhibition of aromatase at 1
compounds
lM concentration for 4-substituted benzonitrile ether
a
Compds Substituent on position 4
oxygen
Inhibition at
M (%)
Aromatase IC50
M)
1
l
(l
3
4
5
3-Pyridyl
4-Pyridyl
3-Pyridazine
70
53
1
0.289 0.010
a
Average of three measurements on aromatase from human.13
pharmacophore requirements for efficient aromatase inhibition
and shows IC50 = 0.289 lM (Table 1), corresponding to a ligand effi-
5
ciency of 0.60 kcal ꢀ molꢁ1 per heavy atom based on IC50
. The
known isomer 4-(4-pyridyloxy)-benzonitrile (4)6 shows reduced
inhibitory activity (53% inhibition at 1
pound 3. Despite being isosteric to compound 3, 4-(3-pyridazinyl-
oxy)benzonitrile (5) is inactive (1% inhibition at M).
lM) as compared to com-
1
l
Compound 5 was obtained by deprotonation of 4-hydroxybenzo-
nitrile in DMF with 55% NaH in oil, followed by adjunction of 3-
chloropyridazine and heating with microwave irradiation for 1 h
at 170 °C. (Scheme 1)
The presence of the second benzonitrile group in Letrozole (1) is
key to achieve selectivity against the other enzymes involved in
the steroidogenesis. Using a divalent oxygen linker between the
two aromatic moieties in 3 prevents to achieve selectivity follow-
ing the same vector as seen for the third substituent of the carbon
linker in Letrozole (1). In order to design new aromatase inhibitors,
we have used a ligand based pharmacophore model built by the
alignment of a diverse set of potent ligands challenging the space
Figure 2. Represented on top of the heme (beige), the volume accessible to ligands
(white) in the aromatase pocket as defined by the superimposition of diverse
sterically challenging potent aromatase inhibitors. The ligands are aligned based on
their geometry of interaction with the heme. Compound 23 docked in the
pharmacophore model appears in green. For reference, Letrozole is shown in pink
and the enzyme substrate androstenedione in yellow.
around the
L
shape minimal pharmacophore (Fig. 2).7 This
We have investigated, as an alternative vector for the elabora-
tion of compound 3, the substitution at position 2 (Table 2) or
position 3 (Table 3) of the benzonitrile of 3. The introduction of a
fluoro substituent is unfavorable; compound 6 with fluoro at posi-
tion 2 is less active than the commercially available compound 7
approach has still been considered valid despite the recent release
of the first X-ray structure of aromatase in co-crystal with andro-
stenedione.8 The enzyme pocket in this structure appears to be
not large enough to accommodate compound 1 suggesting some
level of plasticity of the active site.
with fluoro at position 3 (15% and 52% inhibition at 1 lM, respec-
tively). The commercially available 4-(3-pyridyloxy)-phthalonitrile
(8) confirms the negative impact of a small electron withdrawing
group with H-bonding acceptor capability at position 2 of the
benzonitrile. Compounds 6 and 7 were further processed under
microwave irradiation to displace the fluoro group with phenolates
or amines in NMP (Scheme 1). Introduction at position 2 of an
amine is unfavorable in the three cases exemplified with the intro-
duction of N-methylpiperazine (compound 9) being less preferred
than aminocyclohexyl (compound 10) and benzylamine
O
HO
a, b
N
N
CN
CN
c
5
N
N
O
F
O
d
R
N
F
F
6
CN R'
CN
CN
9 R-R' = -(CH2)2NMe(CH2)2-
(compound 11), with an IC50 = 0.989
The introduction at position 2 of a phenolate (compound 12) is
tolerated (78% inhibition at 1 M) and modulation of the activity
lM, being the best tolerated.
f
(95%)
10 R= H, R' = cyclohexyl (79%)
11 R= H, R' = benzyl (73%)
l
e
seems to be achievable with variation of the phenyl ether substitu-
tion pattern; m-chloro (compound 13) is significantly weaker than
p-methoxyphenoxy (compound 14) (25% and 84% inhibition at
1 lM, respectively). The 2,4-bis-(3-pyridyloxy)benzonitrile (15)
was obtained in one synthetic step from 2,4-difluorobenzonitrile
O
F
O
O
R''
CN
N
CN
N
12 R'' = phenyl (59%)
13 R'' = 3-chlorophenyl (77%)
14 R'' = 4-methoxyphenyl (10%)
15 R'' = 3-pyridyl (28%)
and displays a similar potency as compared to 3 (Table 2).
R'' O
F
The 3,4-substitution pattern is less easily accessible due to the
reduced reactivity of the position 3 fluoro group, as compared with
position 2, in a SNAr reaction and moderate yields of isolated
products (15% and 12% for compounds 16 and 17, respectively)
were obtained for the reaction of 7 with phenolates (Scheme 1).
Substitution at position 3 is well tolerated when introducing a
m-chlorophenoxy or a 3-pyridyloxy group (89% and 82% inhibition
at 1 lM). (Table 3)
Using 4-pyrrolo[2,3-c]pyridin-1-yl-benzonitrile (compound 18),
the required conformation to inhibit the aromatase enzyme was
g
O
O
CN
N
CN
7
N
16 R'' = 3-chlorophenyl (15%)
17 R'' = 3-pyridyl (12%)
Scheme 1. Reagents and conditions: (a) NaH, DMF, 30 min at rt. (b) 3-Chloropy-
ridazine, DMF, 1 h at 170 °C ( w) (38%). (c) 3-Hydroxypyridine, sodium salt, NMP,
2 h from 0 °C to rt (16%). d) Amine (2.5 equiv), NMP, 15–25 min at 150–170 °C ( w).
(e) Sodium phenolate, NMP, 10 min at 110–130 °C ( w). (f) 3-Hydroxypyridine,
sodium salt, NMP, 15 min at 100 °C ( w). (g) Sodium phenolate, NMP, 40-60 min at
150 °C ( w).
l
l
l
l
rigidified and led to a potent inhibitor (IC50 = 0.093 lM). Compound
l

1862
F. Stauffer et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1860–1863
Table 2
Inhibition of aromatase at 1
zonitrile compounds
X
R'
R
a
N
R'
R
lM concentration for 2-substituted 4-(pyridyloxy)ben-
+
N
N
H
N
a
Compds Substituent on
position 2
Inhibition at 1
(%)
lM
Aromatase IC50
M)
18 R= CN, R' = H, X = I (89%)
19 R= CONH2, R' = H (76%)
b
(l
20 R= Cl, R' = H, X = I (72%)
21 R= Br, R' = H, X = I (15%)
22 R= I, R' = H, X = Br (10%)
23 R= Cl, R' = Me, X = I (90%)
24 R= CN, R'= Me, X = Br (54%)
25 R= CN, R' = OPh, X = I (53%)
6
8
9
10
11
12
13
14
15
F
CN
15
27
8
33
58
78
25
84
81
5.03 0.33
4-Methypiperazinyl
Cyclohexylamino
Benzylamino
0.986 0.066
Phenyloxy
3-Chlorophenyloxy
4-Methoxyphenyloxy
3-Pyridyloxy
Scheme 2. Reagents and conditions: (a) Halogenoaryl (1.1 equiv), K3PO4
(2.1 equiv), tran-N,N0-dimethyl-1,2-cycloheanediamine (0.1 eq.), CuI (0.05 equiv),
toluene, 4–36 h at 110 °C. (b) NaOH, 1,4-dioxane/water 2:1, 40 min at 100 °C.
0.304 0.023
a
Average of three measurements.
using a non-optimized modified Ullmann copper (I) catalyzed
coupling reaction (Scheme 2).9 Compound 20 has been described
as an intermediate in the synthesis of semicarbazide-sensitive
amine oxidase inhibitors and was obtained in a similar manner.10
In summary, the simple 3-pyridyloxybenzonitrile (3) has been
found to be a submicromolar inhibitor of aromatase. We have used
a ligand based pharmacophore to guide our SAR investigation as an
alternative to the claimed X-ray structure based11 or homology
model based12 approaches. Some readily accessible derivatives of
compound 3 have been synthesized in one or two steps allowing
a preliminary assessment of the structure–activity relationship
for aromatase inhibition. The rigidified 4-pyrrolo[2,3-c]pyridin-1-
yl-benzonitrile (18) represents a new efficient aromatase inhibitor
fulfilling the pharmacophore requirements as defined for fragment
2. Derivatisation at several positions is allowed in agreement with
our pharmacophore model. In addition, compound 18 shows a
good ligand efficiency and straightforward accessibility making it
an attractive lead for further optimization towards more specific
aromatase inhibitors.
Table 3
Inhibition of aromatase at
benzonitrile compounds
1
l
M
concentration for 3-substituted 4-(pyridyloxy)
a
Compds Substituent on
position 3
Inhibition at 1
(%)
lM
Aromatase IC50
M)
(
l
7
16
17
F
52
89
82
3-Chlorophenyloxy
3-Pyridyloxy
0.229 0.045
a
Average of three measurements.
Table 4
Inhibition of aromatase at 1
c]pyridine compounds
lM concentration for substituted 1-aryl-1H-pyrrolo[2,3-
a
Compds Substituted aryl
Inhibition at 1
(%)
lM
Aromatase IC50
M)
(
l
18
19
20
21
22
23
24
25
4-Cyanophenyl
4-Aminocarbonyl
4-Chlorophenyl
4-Bromophenyl
4-Iodophenyl
4-Chloro-3-methyl
4-Cyano-3-methyl
4-Cyano-3-
94
60
85
56
26
96
>99
97
0.093 0.005
Acknowledgments
0.523 0.064
We thank Arco Jeng and Chii-Whei Hu for their technical
support in measuring aromatase inhibition.
0.059 0.011
0.114 0.032
References and notes
phenoxyphenyl
1. Bulun, S. E.; Lin, Z.; Imir, G.; Amin, S.; Demura, M.; Yilmaz, B.; Martin, R.;
Utsunomiya, H.; Thung, S.; Gurates, B.; Tamura, M.; Langoi, D.; Deb, S.
Pharmacol. Rev. 2005, 57, 359.
a
Average of three measurements.
2. Bowman, R.M.; Steele, R.E.; Browne, L. U.S. Patent 4 978 672, 1990.
3. Haynes, B. P.; Dowsett, M.; Miller, W. R.; Dixon, J. M.; Bhatnagar, A. S. J. Steroid.
Biochem. Mo.l. Biol. 2003, 87, 35.
4. Okada, M.; Yoden, T.; Kawaminami, E.; Shimada, Y.; Kudoh, M.; Isomura, Y.
Chem. Pharm. Bull. 1997, 45, 482.
5. Hopkins, A. L.; Groom, C. R.; Alex, A. Drug Discovery Today 2004, 9, 430.
6. Buchstaller, H.P.; Wiesner, M.; Schadt, O.; Amendt, C.; Zenke, F.; Sirrenberg, C.;
Grell, M.; Finsinger, D. WO 2004037789, 2004.
7. Furet, P.; Batzl, C.; Bhatnagar, A.; Francotte, E.; Rihs, G.; Lang, M. J. Med. Chem.
1993, 36, 1393. The molecules used to construct the model are compounds 4, 7,
8, 11, 12, 13 and 14 of this publication in addition to Letrozole, androstedione,
3-cyclohexyl-3-(4-aminophenyl)piperidine-2,6-dione (an aminoglutethimide
derivative) and LY-113174 a pyrimidine inhibitor.
18 shows good selectivity against CYP 3A4, 2C9 and 2D6 (>200ꢀ).
Hydrolysis of the cyano substituent to the amide (compound 19) re-
duces the activity and replacement of the cyano with a chloro (com-
pound 20) is tolerated with half a log unit decrease in inhibitory
activity. When increasing the size of the halogeno substituent (Br
compound 21 and I compound 22) the potency of the inhibitors is
decreased. The introduction of an extra methyl positioned ortho
to the chloro substituent in compound 23 increases the potency
by one log unit as compared to compound 20. The beneficial effect
of this methyl substituent is confirmed by comparing compound 19
8. (a) Ghosh, D.; Griswold, J.; Erman, M.; Pangborn, W. J. Steroid Biochem. Mol. Biol.
2010, 118, 197; (b) Ghosh, D.; Griswold, J.; Erman, M.; Pangborn, W. Nature
2009, 457, 219.
with compound 24 (94% and >99% inhibition at 1 lM, respectively).
Similarly to what as been discussed previously comparing
compounds 3 and 12, the introduction of a phenoxy substituent
at the position ortho to the cyano group is well-tolerated giving rise
to a similar potency though the ligand efficiency is notably
decreased (0.57 and 0.40 kcal ꢀ molꢁ1 per heavy atom for
compound 18 and 25, respectively, based on IC50). (Table 4)
Taking up the challenge of competing with the synthetic
conciseness of Letrozole (1), compounds 18 to 25 were synthesized
with poor to good yields (10–90%) in one step from the
commercially available 6-azaindole and the suitable halogenoaryl
9. Antilla, J.C.; Klapars, A.; Buchwald, S.L. J. Am. Chem. Soc. 2002, 124, 11684;
Representative synthesis of compound 23: In
a Schlenk tube are added
6-azaindole (0.846 mmol), potassium phosphate (1.78 mmol) and CuI
(0.042 mmol). The tube is twice evacuated and back-filled with Ar. Toluene
(1 ml), 2-chloro-5-iodotoluene (0.846 mmol) and (trans)-N,N0-dimethyl-1,2-
cyclohexanediamine (0.085 mmol) are added. The tube is sealed and the
reaction mixture is stirred for 22 h at 110 °C. The reaction mixture is diluted
with dichloromethane and dry loaded on silica gel to be purified by flash
chromatography (hexane/EtOAc 10% to 30%) to give 1-(4-chloro-3-
methylphenyl)-1H-pyrrolo[2,3-c]pyridine as
a
white solid (184 mg).
ESI-HRMS (M+H): 243.06827, 245.06514 (cal. C14H12N235Cl : 243.06835); H
NMR (DMSO-d6, 600 MHz) d 8.92 (s, 1H), 8.24 (d, 1H), 7.72 (br s, 1H), 7.66

F. Stauffer et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1860–1863
1863
(d, 1H), 7.63 (d, 1H), 7.55 (br d, 1H), 6.79 (d), 2.44 (s, 3H); C NMR d 139.2, 137.5,
137.0, 133.7, 133.5, 132.1, 132.1, 131.5, 130.1, 126.4, 122.9, 115.4. 103.0, 19.7.
10. Evans, D.; Carley, A.; Stewart, A. et al. WO 2011113798, 2011.
11. Ghosh, D. WO 2009061859, 2009.
12. Karkola, S.; Höltje, H. D.; Wähälä, K. J. Steroid Biochem. Mol. Biol. 2007, 105, 63.
13. Human CYP19 + P450 reductase SUPERSOMES (#456260, BD Biosciences);
fluorometric assay adapted from Stresser, D.M.; Turner S.D.; McNamara, J;
Stocker, P.; Miller, V.P.; Crespi, C.L.; Patten, C. J. Anal. Biochem. 2000, 284, 427.