Novel selective androgen receptor modulators: SAR studies on 6-bisalkylamino-2-quinolinones

Article abstract of DOI:10.1016/j.bmcl.2007.01.001

A series of selective androgen receptor modulators (SARMs) with a wide spectrum of receptor modulating activities was developed based on optimization of the 4-substituted 6-bisalkylamino-2-quinolinones (3). Significance of the trifluoromethyl group on the side chains and its interactions with amino acid residues within the androgen receptor (AR) ligand binding domain are discussed. A representative analog (9) was tested orally in a rodent model of hypogonadism and demonstrated desirable tissue selectivity.

Full text of DOI:10.1016/j.bmcl.2007.01.001

Bioorganic & Medicinal Chemistry Letters 17 (2007) 1527–1531
Novel selective androgen receptor modulators: SAR studies
on 6-bisalkylamino-2-quinolinones
Arjan van Oeveren, Mehrnoush Motamedi, Esther Martinborough, Shuo Zhao,
Yixing Shen, Sarah West, William Chang, Adam Kallel, Keith B. Marschke,
*
´
´
Francisco J. Lopez, Andres Negro-Vilar and Lin Zhi
Discovery Research, Ligand Pharmaceuticals, 10275 Science Center Drive, San Diego, CA 92121, USA
Received 29 November 2006; revised 23 December 2006; accepted 2 January 2007
Available online 8 January 2007
Abstract—A series of selective androgen receptor modulators (SARMs) with a wide spectrum of receptor modulating activities was
developed based on optimization of the 4-substituted 6-bisalkylamino-2-quinolinones (3). Significance of the trifluoromethyl group
on the side chains and its interactions with amino acid residues within the androgen receptor (AR) ligand binding domain are dis-
cussed. A representative analog (9) was tested orally in a rodent model of hypogonadism and demonstrated desirable tissue
selectivity.
Ó 2007 Elsevier Ltd. All rights reserved.
Testosterone, as a male hormone, contributes mainly to
the development and function of males, and has been
used in multiple preparations for the treatment of many
forms of androgen deficiencies.1 In addition to the geno-
mic functions of testosterone through binding and acti-
vation of the androgen receptor (AR), the biological
activity of the hormone is directly related to its two ma-
jor metabolites. Testosterone is reduced to a more po-
tent androgen, dihydrotestosterone (DHT, 1, Fig. 1), by
5-alpha reductases distributed primarily in human pros-
tate and skin tissues. In addition, testosterone is aroma-
tized to a female hormone, estradiol, which plays an
important role in bone and CNS homeostasis.2,3 Reduc-
tase inhibitors, such as finasteride, have been therapeu-
tically used for the treatment of benign prostatic
hypertrophy and androgenetic alopecia. These drugs
reduce androgen effects on prostate and hair follicles
without significantly altering androgen action on other
tissues.4 In principle, androgens that exhibit affinity sim-
ilar to testosterone for the androgen receptor which are
not modified by the reductase enzyme should behave in
a tissue selective fashion. Recently, many nonsteroidal
androgens have been discovered that have demonstrated
OH
R1
R2
N
CF3
CF3
CF3
R3
N
O
N
H
O
N
H
O
3
2
1
Figure 1. AR modulators: (1) DHT; (2) LGD2226; (3) new SARM
series.
R1
R1
NH2
a), b)
O
N
H
O
N
H
5
4
R2
N
R1
c) or d)
R3
O
N
H
3
Scheme 1. Reagents and conditions: (a) HNO3, H2SO4, 0 °C, 20 min
(65–90%); (b) H2, Pd/C, EtOAc or DMF, rt, 6–15 h (80–95%); (c) an
aldehyde or a ketone, NaB(CN)H3, AcOH, rt, 15 h (50–95%); (d)
TFA, NaBH4, rt to 50 °C (40–90%).
Keywords: Androgen receptor; SARM; 2-Quinolinone; AR agonist.
*
Corresponding author. Tel.: +1 858 550 7663; fax: +1 858 550
7249; e-mail: lzhi@ligand.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.01.001

1528
A. van Oeveren et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1527–1531
tissue selective AR agonist activity.5–7 The nonreducible
properties of the molecules may partially contribute to
the observed tissue selectivity, but potentially differential
cofactor recruitment may be also be responsible for the
observed effects. Recently, we reported the discovery of
a SARM, LGD2226 (2), and the development of the
2-quinolinone AR modulator pharmacophore.8–10 Here
we report additional SAR studies on the 2-quinolinone
pharmacophore (3) and the importance of the trifluoro-
methyl group to enhance interaction of the modulators
with specific amino acid residues within the AR ligand
binding pocket.
Table 1. Cotransfection and competitive binding data for the new analogs and reference compoundsa
R2
CH3
R2
N
R2
N
R1
CH3
CF3
CF3
N
CF3
N
CF3
CF3
R3
O
N
H
O
N
H
O
N
H
O
N
H
6
7
8
9
Compound
R1
R2
R3
hAR
agonist
EC50 (nM)
hAR
agonistb
Eff (%)
hAR
antagonist
IC50 (nM)
hAR
antagonistc
(Eff. %)
hAR
binding
Ki (nM)
2 (LGD2226)
6a
0.20 0.02
2.1 0.3
1.5 0.2
0.4 0.1
95 2
94 19
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
19
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
1.5
4.4
64
CH2CH3
CH2CH3
6b
6c
CH2CH2CH3
CH2CHF2
CH2CClF2
CH3
CH2CH2CH3
CH2CHF2
CH2CHF2
CH2CF3
CH2CF3
CH2CF3
CH2CF3
CH2CF3
CH2CF3
CH2CF3
CH2CF3
CH2CF3
CH2CF3
CH2CF3
CH2CF3
CH2CF3
CH2CF3
CH2CF3
CH2CF3
CH2CF3
CH2CF3
CH2CH3
CH2CH3
72
74
5
7
0.5
2.3
9.0
7.1
12
6d
6e
0.4
0
110
59 12
7
9.8 3.6
0.4 0.1
1.3 0.4
6.8 0.8
0.9 0.3
1.8 0.5
81 32
5.9 4.1
0.2 0.1
0.5 0.1
1.6 0.8
46 15
2.3 0.5
7.9 5.3
27
6f
6g
CH2CH3
CH2CH2CH3
107
88
5
4
6
6h
6i
CH2CH2CH2CH3
CH(CH3)2
CH2C(CH3)3
CH2CH2C(CH3)3
CH(CH3)CH2CH3
CH2CHF2
CH2CH2Cl
CH2CHCl2
CH2CH2CF3
CH2CF2CF3
CH2CH2OH
CH2Ph
69
33
107 17
81 10
69 17
86 16
2.9
12
6j
6k
243
29
6l ( )
6m
84
88 14
7
1.7
2.6
6.0 1.1
116
31
6n
6o
67
7
6
6p
6q
102
94
6
d
6r
125 27
71
6.6 2.8
79
6s
6t
6u
CH2(2-thienyl)
CH2CO2CH2CH3
COCF3
12
20
2
4
85
59
38
49
69
68
93
89
—
26
97
9
5
4
61
41
19d
1000
14
6v
6w
22 13
2.8
4.1
—
—
1.0
—
—
0
0
56
—
—
39
—
—
83
41
—
—
—
3
9
1
CH2(3-furyl)
CH(CH2)COCH3
CH2CF3
6x ( )
7a
5.8 1.3
57 15
0.8 0.3
5
8
5
6
3.1
24
H
CH3
7b
7c
CH2CF3
CH2CF3
1.4
5.4
42d
21
CH2CH3
12
—
9.1
6
7d
7e
CH2CH2CH3
CH(CH3)2
CHF2
Cl
CH2CF3
CH2CF3
36 15
2.0
—
7f
7g
CH2CF3
CH2CF3
0.5 0.1
0.89 0.58
1.0 0.0
35 16
4
0.4
0.5
0.5d
23
115 13
125 22
—
—
7h
7i
Br
OCH3
OH
CH2CF3
CH2CF3
34
43 12
4
15
7.0
—
20
—
—
—
—
0
24 11
43 12
—
7j
CH2CF3
CF3
83 34
2.3 0.7
0
1000
9.2
112
5.4
9.8
12
8a (R)
8b (S)
8c (R)
8d (S)
9
86
27
5
4
CF3
CH2Cl
CH2Cl
50
0
0
43
—
—
—
—
85
3
3.7 0.8
3.6 0.1
3.2 1.3
5.1 0.1
—
87 26
87 23
71
5
DHT
Bicalutamide
100
—
0.20 0.02
151 36
162 99
9
a Values with standard errors represent the mean value of at least 2 separate experiments with triplicate determinations. ‘—’ indicates not active
(an efficacy of <20% and/or a potency of >10,000 nM).
b Agonist efficacies were determined relative to DHT (100%).
c Antagonist efficacies (%) were determined as a function of maximal inhibition of DHT at the EC50 value.
d Baculovirus binding assay result that is consistent with most agonist binding data in whole-cell assays.

A. van Oeveren et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1527–1531
1529
The synthesis of compounds of structure 3 is described
in Scheme 1. Nitration of the 4-substituted 2-quinoli-
nones (4) followed by hydrogenation selectively provid-
ed the 6-amino-2-quinolinones of structure 5. The
bisalkyl groups were introduced by a typical reductive
alkylation with an aldehyde, a ketone, or TFA in case
of 2,2,2-trifluoroethyl. The analogs were characterized
by a competitive whole cell binding assay and the hu-
man AR cotransfection assay and the testing results
are summarized in Table 1.11 We used the functional
assay results to drive the SAR studies and the binding
affinity to assure the interaction with AR. As previously
described,9 small bisalkyl groups (i.e., ethyl and propyl)
or bistrifluoroethyl group on the 6-amino group of the
2-quinolinones led to potent AR agonists (6a and 6b)
and the fluorine atoms on the bisalkyl groups signifi-
cantly enhanced the activity (comparing 2 with 6a).
Other fluoro or chlorofluoroalkyl substitution provided
similarly potent agonists (6c and 6d). To systematically
investigate SAR of the side chains, we prepared the
4-trifluoromethyl analogs with one of the side chains
fixed as a 2,2,2-trifluoroethyl group (6e–6v). Apparently,
the size of the alkyl group had significant impact on the
AR modulating activities. The optimal size for a full
agonist is about 2–3 carbon units (6f, 6g, and 6i), methyl
(6e) and 4–5 carbon units (6h, 6j, 6l) are slightly off the
optimal size and generated potent partial agonists, and
the 6-carbon unit analog (6k) had about 10-fold weaker
activity. It can be observed that changing the size and
electronic properties of the other alkyl group generated
a wide spectrum of potent AR modulating activities
from full agonists (6f, 6g, 6i), partial agonists (6e, 6h,
6o) to partial agonist/antagonists (6u and 6v).
Fluorine atoms play an important role in molecules with
biological activity. The fluorine atoms on the second
carbon of the alkyl groups of the 6-amino analogs signif-
icantly enhanced the AR modulating activity (compar-
ing 2, 6a, and 6f). Interestingly, fluorine atoms on the
third carbon of the propyl group have opposite impact
on activities of the analogs (comparing 6g and 6p). How-
ever, additional fluorine atoms on the second carbon
regained the lost activity (comparing 6p and 6q). To
investigate the special effect of the fluorine atoms, we
prepared the 6-(2,2,2-trifluoroisopropyl)amino and
6-(2-chloroisopropyl)amino analogs (8) and separated
the two pairs of enantiomers by chiral HPLC. The two
trifluoroisopropyl enantiomers (8a and 8b) had very dif-
ferent activity but the two chloroisopropyl enantiomers
(8c and 8d) had similar activity. Compound 8a has the
agonist activity similar to that of its parent compound
6i, while the enantiomer 8b has much weaker partial
activity. Both of the chloro substituted enantiomers 8c
and 8d showed activity similar to that of their similar
sized racemic analog 6l. It is obvious that the fluorine
atoms on the second carbon provide unique interactions
in the AR binding pocket.
We have reported that the bistrifluoroethyl group of
LGD2226 (2) occupies the same space as the C and D
rings of DHT in AR binding pocket but has a hydrogen
bonding with residues of AR-lbd weaker than that of the
17-hydroxy group of DHT based on the cocrystal X-ray
structure.9,12 Modeling results based on the LGD2226-
AR structure indicated that the fluorine atoms on the
isopropyl rather than the ethyl group of 8a form a fluo-
˚
rine-to-hydroxyl hydrogen bond (2.4 A) with amino acid
THR711 that could neither be replaced by a methyl
˚
group nor be reached (3.0 A as the closest) by either
Various 4-substituted analogs with 6-bistrifluoroethyla-
mino group (7a–7j) were prepared and tested. Size and
shape changes of the 4-substituents gradually modulated
the activities of the compounds from partial agonist/an-
tagonist (7a, R1 = H) to full agonist (7b, R1 = methyl),
then partial agonist/antagonist again (7e, R1 = isopro-
pyl) to almost full antagonist (7d, R1 = propyl). The
4-fluoromethyl, 4-chloro, and 4-bromo analogs showed
similar activity, which suggests the size, rather than
the electronic property, plays a role. The oxygen con-
taining groups appear to give partial agonist/antagonists
(7i–7j).
of the fluorine atoms on the isopropyl or ethyl groups
of 8b (Fig. 2).13 In addition, the bisfluoroalkylamino
group of 8b orientes quite differently in the binding
pocket from that of 8a, which resulted in space shifting
of helix 12 backbone from the agonist position of 8a
structure that is similar to that of LGD2226.
SAR studies indicated that AR modulating activity could
be tuned accordingly by variation of the R1, R2, and R3
groups of structure 3. Analogs 6w, 6x, and 9 are typical
partial AR agonists resulting from modification of these
Figure 2. Interactions of analogs 8a and 8b with the amino acid residue THR877 (helix 11) of AR ligand binding domain in a docking model.

1530
A. van Oeveren et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1527–1531
Figure 3. Effects of compound 9 and testosterone treatment on levator ani muscle and ventral prostate weights, and osteocalcin levels in a 2-week
orchidectomized rat maintenance assay. 100% identifies intact animals, whereas 0% denotes levels for vehicle-treated animals.
groups. Analog 9 was profiled in a two-week mature
orchidectomized rat assay to assess its tissue selectivi-
ty.14,15 In the study, compound 9 was dosed orally,
whereas testosterone propionate was dosed subcutane-
ously once daily for 14 days. Weights of ventral prostate
and levator ani muscle were used to measure androgenic
and anabolic effects, respectively, and osteocalcin levels
were used as a surrogate bone turnover biomarker. As
illustrated in Figure 3, testosterone equally stimulated
all three endpoints to intact-equivalent levels at approxi-
mately 0.5 mg/kg dose. Since no differences were ob-
served in the dose of testosterone needed to achieve
eugonadal levels for each of the endpoints it was inferred
that testosterone depicts no tissue selectivity in this mod-
el. Dose of compound 9 to achieve intact-equivalent lev-
els in levator ani weights was approximately 40 mg/kg. At
this dose, compound 9 was almost inactive in stimulating
ventral prostate and showed better efficacy in osteocalcin
levels than that observed for the highest dose of testoster-
one. In other words, compound 9 had the same potency
as and better efficacy than testosterone in the osteocalcin
endpoint, was about 80-fold less potent in stimulating
levator ani, and was at least 200-fold less potent in the
ventral prostate. Within the dose range tested, compound
9 never stimulated the ventral prostate to intact-equiva-
lent levels. Overall, compound 9 demonstrated excellent
selectivity on levator ani muscle over prostate and even
better selectivity on bone marker over prostate.
analogs and amino acid residues within the AR binding
pocket provided additional insight into the SAR of the
series. More SAR studies on the series to define the scope
and limitation of the series will be published later.
Acknowledgments
We thank Mr. Mark Cummings for chiral HPLC sepa-
ration of the enantiomers, the Department of New
Leads for performing the cotransfection and binding
assays, and Department of Pharmacology for perform-
ing the in vivo assay.
References and notes
1. Ohl, D. A.; Quallich, S. A. Urol. Nurs. 2006, 26, 253.
2. Andriole, G.; Bruchovsky, N.; Chung, L. W.; Matrumoto,
A. M.; Pittmaster, R.; Roehrborn, C.; Russell, D.; Tindall,
D. J. Urol. 2004, 172, 1399.
3. Gennari, L.; Nuti, R.; Bilezikian, J. P. J. Clin. Endocrinol.
Metab. 2004, 89, 5898.
4. Libecco, J. F.; Bergfeld, W. F. Expert. Opin. Pharmac-
other. 2004, 5, 933.
5. Marhefka, C. A.; Gao, W.; Chung, K.; Kim, J.; He, Y.;
Yin, D.; Bohl, C.; Dalton, J. T.; Miller, D. D. J. Med.
Chem. 2004, 47, 993.
6. Ostrowski, J.; Kuhns, J. E.; Lupisella, J. A.; Manfredi, M.
C.; Beehler, B. C.; Krystek, S. R., Jr.; Bi, Y.; Sun, C.;
Seethala, R.; Golla, R.; Sleph, P. G.; Fura, A.; An, Y.;
Kish, K. F.; Sack, J. S.; Mookhtiar, K. A.; Grover, G. J.;
Hamann, L. G. Endocrinology 2007, 148(1), 4.
7. Hanada, K.; Furuya, K.; Yamamoto, N.; Nejishima, H.;
Ichikawa, K.; Nakamura, T.; Miyakawa, M.; Amano, S.;
Sumita, Y.; Oguro, N. Biol. Pharm. Bull. 2003, 26(11),
1563.
8. Miner, J. N.; Chang, W.; Chapman, M.; Finn, P.; Hong,
M. H.; Lopez, F. J.; Marschke, K. B.; Rosen, J.; Schrader,
W. T.; Turner, R.; van Oeveren, A.; Viveros, H.; Zhi, L.;
Negro-Vilar, A. Endocrinology 2007, 148(1), 363.
9. van Oeveren, A.; Motamedi, M.; Mani, N. S.; Marschke,
K. B.; Lopez, F. J.; Schrader, W. T.; Negro-Vilar, A.; Zhi,
L. J. Med. Chem. 2006, 49(21), 6143.
A group of the representative compounds (6c, 6g, 6i,
6m, 6n, 7b, 7f, and 7g) were tested in related steroidal
hormone receptor cotransfection assays in both agonist
and antagonist modes to check their receptor selectivity.
All of the compounds showed more than 1000-fold sep-
aration from the estrogen, progesterone, glucocorticoid,
and mineralocorticoid receptors, activities.
In summary, we systematically investigated the SAR of
the three substituents of 6-amino-2-quinolinone 3 and
generated a number of potent SARMs with different
receptor modulating profiles. A representative analog 9
demonstrated oral bioavailability and desirable tissue
selectivity on muscle and surrogate bone marker endpoint
over prostate in an orchidectomized rat model of hypogo-
nadism. The unique interaction of fluorine atoms of the
10. van Oeveren, A.; Pio, B.; Tegley, C. M.; Higuchi, R.
I.; Wu, M.; Jones, T. K.; Marschke, K. B.; Negro-
Vilar, A.; Zhi, L. Bioorg. Med. Chem. Lett. 2007, 17,
1523.

A. van Oeveren et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1527–1531
1531
11. The detailed assay conditions were described in our
previous reports.
12. Wang, F.; Liu, X.; Li, H.; Liang, K.; Miner, J. N.; Hong,
M.; Kallel, E. A.; Van Oeveren, A.; Zhi, L.; Jiang, T. Acta
Crysyallogr., Sect. F 2006, F62, 1067.
13. Calculations were performed using the Tripos forcefield
and Gasteiger–Huckle, and charges 8a and 8b were
docked using the X-Ray cocrystal structure of AR-lbd
14. Mature male Sprague–Dawley rats from Harlan (India-
napolis, IN) were used and the protocol was approved by
IACUC. One group of animal was sham-operated, treated
with vehicle, and served as additional control. Five
animals were used in each dosing group and animals were
treated immediately after orchiectomy.
15. Compound 9: 1,2-dihydro-6-(N-2,2,2-trifluoroethyl-N-iso-
butyl)amino-4-methyl-2-quinolinone. Rf 0.18 (MeOH/
CH2Cl2, 3:7); 1H NMR (500 MHz, CDCl3) 11.80–12.00
(br s, 1H), 7.32 (d, J = 9.3, 1H), 7.12 (dd, J = 2.4, 9.3, 1H),
7.02 (d, J = 2.9, 1H), 6.56 (s, 1H), 3.92 (q, J = 8.8, 2H),
3.23 (d, J = 7.3, 2H), 2.47 (s, 3H), 2.04 (quin, J = 6.8, 1H),
0.95 (d, J = 6.8, 6H).
and LGD2226 as
refined using
a template. The structures were
simulated annealing protocol as
a
described in the SYBYL manuals, SYBYL 7.0, Tripos
Inc., 1699 South Hanley Rd., St. Louis, MO 63144,
USA.