Synthesis and biological evaluation of novel selective androgen receptor modulators (SARMs). Part II: Optimization of 4-(pyrrolidin-1-yl)benzonitrile derivatives

Article abstract of DOI:10.1016/j.bmcl.2017.03.038

We recently reported a class of novel tissue-selective androgen receptor modulators (SARMs), represented by a naphthalene derivative A. However, their pharmacokinetic (PK) profiles were poor due to low metabolic stability. To improve the PK profiles, we m

Full text of DOI:10.1016/j.bmcl.2017.03.038

Accepted Manuscript
Synthesis and biological evaluation of novel selective androgen receptor mod-
ulators (SARMs). Part II: Optimization of 4-(pyrrolidin-1-yl)benzonitrile de-
rivatives
Moriteru Asano, Takenori Hitaka, Takashi Imada, Masami Yamada, Megumi
Morimoto, Hiromi Shinohara, Takahito Hara, Masuo Yamaoka, Takashi
Santou, Masaharu Nakayama, Yumi Imai, Noriyuki Habuka, Jason Yano, Keith
Wilson, Hisashi Fujita, Atsushi Hasuoka
PII:
S0960-894X(17)30276-7
DOI:
http://dx.doi.org/10.1016/j.bmcl.2017.03.038
BMCL 24790
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
14 February 2017
14 March 2017
16 March 2017
Please cite this article as: Asano, M., Hitaka, T., Imada, T., Yamada, M., Morimoto, M., Shinohara, H., Hara, T.,
Yamaoka, M., Santou, T., Nakayama, M., Imai, Y., Habuka, N., Yano, J., Wilson, K., Fujita, H., Hasuoka, A.,
Synthesis and biological evaluation of novel selective androgen receptor modulators (SARMs). Part II: Optimization
of 4-(pyrrolidin-1-yl)benzonitrile derivatives, Bioorganic & Medicinal Chemistry Letters (2017), doi: http://
dx.doi.org/10.1016/j.bmcl.2017.03.038
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Synthesis and biological evaluation of novel selective androgen recep-
tor modulators (SARMs). Part II: Optimization of 4-(pyrrolidin-1-
yl)benzonitrile derivatives
Moriteru Asano*, a, Takenori Hitakaa, Takashi Imadaa, Masami Yamadaa, Megumi Morimotoa, Hiromi
Shinoharaa, Takahito Haraa, Masuo Yamaokaa, Takashi Santoua, Masaharu Nakayamaa, Yumi Imaia,
Noriyuki Habukaa, Jason Yano#,b, Keith Wilsonb, Hisashi Fujitaa, Atsushi Hasuoka*, a
aPharmaceutical Research Division, Takeda Pharmaceutical Company Ltd, 26-1 Muraoka-higashi 2-chome, Fujisawa,
Kanagawa 251-8555, Japan
bStructural Biology, Takeda California Inc., 10410 Science Center Drive, San Diego, California 92121, USA
Keywords: Androgen receptor, Selective androgen receptor modulators (SARMs), 4-(Pyrrolidin-1-yl)benzonitrile,
Pharmacokinetics (PK), Liver microsomal metabolic stability.
Corresponding Author
* (M.A.) Tel: +81-466-32-1152. Fax: +81-466-29-4449. E-Mail: mo.masayume@gmail.com
(A.H.) Tel: +81-466-32-2913. E-mail: atsushi.hasuoka@takeda.com
Present Address
# (J.Y.) Beryllium Discovery Corp., 3 Preston Ct Bedford, MA 01730, USA
Abbreviations: PK, pharmacokinetics; CL, clearance; F, bioavailability; AUC, area under the blood concentration
time curve; MS, liver microsomal metabolic stability; sc, subcutaneous; qd, quaque die; CNS, central nervous system;
bid, bis in die.
ABSTRACT: We recently reported a class of novel tissue-selective androgen receptor modulators (SARMs), repre-
sented by a naphthalene derivative A. However, their pharmacokinetic (PK) profiles were poor due to low metabolic
stability. To improve the PK profiles, we modified the hydroxypyrrolidine and benzonitrile substituents of 4-
(pyrrolidin-1-yl)benzonitrile derivative B, which had a comparable potency as that of compound A. This optimization
led us to further modifications, which improved metabolic stability while maintaining potent androgen agonistic activi-
ty. Among the synthesized compounds, (2S,3S)-2,3-dimethyl-3-hydroxylpyrrolidine derivative 1c exhibited a suitable
PK profile and improved metabolic stability. Compound 1c demonstrated significant efficacy in levator ani muscle
without increasing the weight of the prostate in an in vivo study. In addition, compound 1c showed agonistic activity in
the CNS, which was detected using sexual behavior induction assay.
As men age, serum testosterone levels gradually decline due to testicular malfunction and hypothalamic
dysregulation.1 Reduced testosterone levels cause a variety of symptoms, including depression, loss of libido,2 and de-
creased bone and lean body mass. This syndrome is called late-onset hypogonadism (LOH),3 and affects 20% of men
aged 60 and older.4 Currently, several strategies for testosterone replacement therapy (TRT) use testosterone and/or its
ester to treat hypogonadism.5,6 These treatments effectively alleviate symptoms and provide additional health benefits.
However, widespread use of testosterone/testosterone ester-based TRT is limited by potential risk of cardiovascular
problem, prostate cancer and erythrocytosis.7 In addition, administration routes are limited to inconvenient intramuscu-
lar injection, surgical implantation, or transdermal delivery using gels or patches.
Lack of convenient administration routes has stimulated a growing interest in orally available and nonsteroidal tis-
sue-selective androgen receptor modulators (SARMs).8-11 The concept of SARMs emerged from the clinical success of
selective estrogen receptor modulators (SERMs), such as raloxifene, which acts as an estrogen receptor agonist in bone
but an antagonist in the breast and uterus.12 SARMs are expected to have desired anabolic functions, and lack unwant-

ed androgenic properties such as prostate stimulation. Various nonsteroidal SARMs have been reported6,13 and some
compounds such as GTx-02414,15 and DT-20016-18 have reached clinical development (Figure 1).
Figure 1. Structures of selected nonsteroidal SARMs
We have previously reported that 1-(4-cyano-1-naphthyl)-2,3-disubstituted pyrrolidine derivatives, represented by
compound A, showed highly potent AR agonistic activity in vitro, and good tissue selectivity in vivo (Figure 2).19
However, they were rapidly metabolized in rat and human liver microsomes, and failed to achieve sufficient bioavaila-
bility in PK studies in rats (compound A: CL = 3081 mL/h/kg, AUCpo = 32.2 ng•h/mL, F = 9.7%). In addition, re-
placement of naphthonitrile with 2-chloro-3-methyl benzonitrile, a well-known privileged scaffold for SARMs,20,21 still
provided insufficient metabolic stability, and showed a PK profile similar to that of compound A (compound B: CL =
5166 mL/h/kg, AUCpo = 3.2 ng•h/mL, F = 1.6%).
Figure 2. 1-(4-Cyano-1-naphthyl)pyrrolidine derivative A and 2-chloro-3-methylbenzonitrile B
Therefore, we designed a compound by modifying the pyrrolidine and benzonitrile groups further to improve the
PK profile. The X-ray co-crystal structure of compound B with the AR ligand binding domain (LBD) revealed two key
interactions (Figure 3): water-bridged hydrogen bonds of the cyano group with the side chain of Arg752 and the back-
bone of Met745, and an interaction of the hydroxyl group with the side chains of Asn705 and Thr877. Since the struc-
tural data suggested that potency would be improved if the compounds possessed cyano and hydroxyl groups at the
ends of the molecule, we planned to explore substitutions of the hydroxypyrrolidine unit (X) and conduct further opti-
mization of the benzonitrile (Y). In this paper, we describe the design, synthesis, and biological characterization of 4-
(pyrrolidin-1-yl)-1-benzonitrile derivatives.
2

Figure 3. Strategies for identification of novel SARMs from compound B.
(2S,3S)-2,3-Dimethyl-3-hydroxypyrrolidine 4 was prepared using our previously reported method (Scheme 1).22
A methyl group was introduced at the C3-position of 2 using methyl magnesium bromide. In this reaction, the Gri-
gnard substitution was achieved in a highly stereo-selective manner and with an 85% yield, driven by the optically ac-
tive 2S-methyl group, but without cleavage of the Cbz group, using CeCl3 as an additive.23 The Cbz group of 3 was
removed by hydrogenation, and successive treatment with oxalic acid afforded a hemioxalate salt 4 at 88% yield.
Scheme 1. Preparation of (2S,3S)-2,3-dimethyl-3-hydroxylpyrrolidine 4a
a Reagents and conditions: (a) MeMgBr, CeCl3, THF, -78–0 oC, 1 h 85%; (b) H2 (1 atm), Pd/C, MeOH, rt, 1.5 h, then
(CO2H)2, 88%.
The coupling reactions of pyrrolidines 4 with 4-fluorobenzonitriles 5 were performed using lithium carbonate in
dimethylsulfoxide at 70–80 oC to afford 1a–c in 57–75% yield (Scheme 2). 24
Scheme 2. Synthesis of compound 1a–c and Ba
a Reagents and conditions: (a) Li2CO3, DMSO, 70–80 oC, 57–75%.
AR binding affinities were evaluated by competitive displacement of radiolabeled [3H]mibolerone from AR. The
data are reported as IC50 values. Functional activities using an AR responsive luciferase reporter were determined in
Cos-7 cells and are given as EC50 values.
To improve the PK profile of B, we first focused on modification of the pyrrolidine. The co-crystal structure of B
with AR LBD suggests that the C3-hydroxyl group on the pyrrolidine ring is essential for potency. However, the hy-
droxyl group and pyrrolidine ring are considered to be metabolized by oxidation and conjugation.25 We hypothesized
that introduction of a methyl group at the C3-position on the pyrrolidine ring would prevent both oxidative and conju-
gative metabolism due to the steric hindrance, while not influencing the potency on the basis of the co-crystal structural
information of B with AR LBD. Based on this hypothesis, (2S,3S)-2-methyl-3-hydroxylpyrrolidine derivative B was
converted into (2S,3S)-2,3-dimethyl-3-hydroxylpyrrolidine derivative 1a (Table 1). As expected, compound 1a demon-
strated better metabolic stability in human and rat assays than that of compound B, while maintaining potent AR bind-
ing affinity and agonistic activity. In addition to improved metabolic stability, compound 1a exhibited a dramatically
improved PK profile (1a: CL = 1886 mL/h/kg, AUCpo = 236.7 ng•h/mL, F = 44.5%) relative to that of compound B (B:
CL = 5166 mL/h/kg, AUCpo = 3.2 ng•h/mL, F = 1.6%).
Table 1. Biological data of (2S,3S)-2,3-dimethyl-3-hydroxylpyrrolidine 1a and (2S,3S)-2-methyl-3-
hydroxylpyrrolidine B
3

Metabolic stability
AR binding
IC50
AR reporter
EC50
(µL/min/mg)b
CL
AUCpo
F (%)d
Compd.
(mL/h/kg)c
(ng•h/mL)c
(95%CI)
(nM)a
(95%CI)
(nM)a
human
24
rat
0.49
(0.31–0.78)
0.20
(0.13–0.31)
96
1886
5166
236.7
3.2
44.5
1.6
1a
B
0.72
(0.51–1.02)
0.19
(0.12–0.31)
67
230
a IC50 (EC50) values are presented as means of duplicate experiments, with 95% confidence intervals (95%CI) in paren-
theses. b Metabolic stability was determined based on disappearance of parent compound after incubation with human
or rat liver microsomes for 20 min. c Rat dosing at 0.1 mg/kg, i.v., and 1 mg/kg, p.o. (n = 3). dF means bioavailability
With its modified hydroxypyrrolidine unit, compound 1a demonstrated an improved PK profile while maintaining
excellent potency. However, its metabolic stability in rats was still insufficient to select 1a as a suitable in vivo candi-
date. To further improve PK profiles, we optimized the substituents on the benzonitrile of compound 1a using the cal-
culated Log P (cLog P) value as an indicator of metabolic stability.26 We designed and synthesized 2-
chlorobenzonitrile derivative 1b (cLog P 2.60) and 3-fluoro-2-methylbenzonitrile derivative 1c (cLog P 2.66), which
has low lipophilicity compared to compound 1a (cLog P 3.10). As expected, both compounds demonstrated improved
metabolic stability over compound 1a in rat liver microsomes, and maintained potent AR agonistic activities. Further-
more, these compounds showed good bioavailabilities (F; 1b: 53.4% and 1c: 46.3%, respectively) and improved plas-
ma exposures (AUCpo >380 ng•h/mL) relative to those of compound 1a (AUCpo = 236.7 ng•h/mL) in rat PK tests in
accordance with the increased metabolic stabilities. From these results, we selected compound 1b and 1c for in vivo
efficacy tests.
Table 2. Biological data of 4-[(2S,3S)-2,3-dimethyl-3-hyroxylpyrrolidin-1-yl] benzonitrile derivatives 1a–c
4

Metabolic stability
AR binding
IC50
AR reporter
EC50
(µL/min/mg)b
CL
AUCpo
F (%)d
Compd.
(mL/h/kg)c (ng•h/mL)c
(95%CI) (nM)a (95%CI) (nM)a rat
0.49
(0.31–0.78)
0.20
(0.13–0.31)
96
39
44
1886
581
236.7
930.6
387.9
44.5
53.4
46.3
1a
1b
1c
0.52
(0.34–0.78)
0.32
(0.16–0.62)
1.0
(0.7–1.3)
0.29
(0.16–0.50)
1236
a IC50 (EC50) values are presented as means of duplicate experiments, with 95% confidence intervals (95%CI) in paren-
theses. b Metabolic stability was determined based on the disappearance of parent compound after incubation with hu-
man or rat liver microsomes for 20 min. c Rat dosing at 0.1 mg/kg, i.v., and 1 mg/kg, p.o. (n = 3). d F means bioavaila-
bility
Tissue selectivity of the selected compounds was investigated using the Hershberger assay.27 Twice daily doses
(5.0 mg/kg/day, po, bid) of compounds 1b and 1c were administered with testosterone propionate (TP) (0.5 mg/kg, sc,
qd) to 4-week-old castrated rats for 4 days (Table 3). The control group was administered TP only (0.5 mg/kg, sc, qd),
and the increases in weight of prostates and levator ani muscles were expressed relative to 100%. Compound 1c exhib-
ited tissue selectivity and strong anabolic activity in levator ani muscle (>300%), whereas compound 1b enhanced the
weight of not only levator ani muscle (>300%) but also of the prostate (>140%). Thus, we selected compound 1c as a
tissue-selective analogue for further in vivo evaluations.
Table 3. Tissue selectivity in immature rats (n = 5)a
Increase in tissue weight (%)
Compd.
Prostate
Levator ani muscle
147*
>300*
>300*
100
1b
1c
102
Controlb 100
a Test compounds were administered orally twice daily with a dose of 5 mg/kg/day with TP (0.5 mg/kg, sc, qd). b TP
only (0.5 mg/kg, sc, qd) was administered as a control. *P<0.05 versus control group by Dunnet's test.
To evaluate anabolic potency and tissue selectivity for comparison with TP (1.2 mg/kg/day, sc, qd), compound 1c
(0.15, 0.5. 1.5, 5, 15 mg/kg/day) was administrated orally twice daily with TP (0.4 mg/kg/day, sc, qd) to 8-week old
castrated rats for 7 days at each dose (Figure 4). Compound 1c exerted significant dose dependent effects (to more
than 5 mg/kg/day) on levator ani muscle that were superior to the effects of TP (1.2 mg/kg/day, sc, qd) alone. On the
other hand, 1c did not influence the weight of the prostate, at doses up to 15 mg/kg/day, while TP (1.2 mg/kg/day, sc,
qd) increased prostate weight significantly. Accordingly, compound 1c is a promising tissue selective androgen recep-
tor modulator.
5

Figure 4. Mean weight of levator ani muscle and prostate in castrated mature rats dosed with compound 1c and TP.
*P<0.025 versus control group (Williams test), ###P<0.001 versus control group (unpaired Student's t test).
Next, we performed a sexual behavior induction assay to confirm agonistic activity of 1c on the CNS (Table 4).
Selected fertile male rats were castrated to eliminate sexual behavior. Sexual behavior induction was then quantitated
on the basis of pseudopregnancy rate in female rats. After treatment with 1c (7.5 mg/kg/day, po, bid) for 3 weeks, sex-
ual behavior increased to 87.5% compared to that in the non-treated control group (10%). This result suggested that 1c
could act as an androgen agonist in the CNS.
Table 4. Sexual behavior induction assaya
Dose
Sexual behavior
(positive/total, %)
87.5*
Compd.
(mg/kg/day for 18–21 days)
7.5
1c
Control
10
a n = 8 (1c), 10 (Control). *P<0.01, versus control (Fisher's Exact test).
To confirm the binding mode of compound 1c, we solved the X-ray co-crystal structure of this compound bound
to the AR LBD (Figure 5). The co-crystal structure of compound 1c with the AR LBD, residues 671–919, was obtained
at 2.70 Å resolution. The binding mode of compound 1c was almost identical to that of compound B. Based on this
result, the 3-methyl group on the pyrrolidine ring of 1c was confirmed to be tolerated without influencing the surround-
ing amino acid residues such as Thr877 or Asn705.
6

Thr877
Asn705
Met745
Arg752
Figure 5. Superimposed X-ray co-crystal structures of 1c (green) and B (orange) bound to androgen receptor LBD
(PDB code 5T8J, 2.70 Å resolution (1c), PDB code 5T8E, 2.71 Å resolution (B)).
To improve the PK profile of the previously reported naphthalene derivative A, we modified its 4-(pyrrolidin-1-
yl)benzonitrile derivative B having comparable potency to that of compound A. To improve PK profiles, we investigat-
ed introduction of a methyl group at the C3-position of the pyrrolidine ring and identified 4-[(2S,3S)-2,3-dimethyl-3-
dhyroxylpyrrolidin-1-yl]benzonitrile derivative 1a. Further optimization of the substituents on the benzonitrile led us to
discover compound 1c, which shows a suitable agonistic activity and a promising PK profile together with sufficient
metabolic stability. In in vivo studies, compound 1c was confirmed to exhibit potent anabolic effects on levator ani
muscle in a dose-dependent manner without increasing the weight of the prostate. In addition, compound 1c induced
sexual behavior in rats. Accordingly, we concluded that compound 1c was a useful in vivo probe molecule to study the
fundamental biology and therapeutic potential of a SARM. Further pharmacological and toxicological studies of these
benzonitrile derivatives will be reported in due course.
Acknowledgments
We are grateful to Satoshi Yamamoto, Katsuji Aikawa, Douglas R. Cary and Ryo Mizojiri for helpful discussions
on chemistry and manuscript preparation. We express our gratitude to Yoichi Nagano and Mitsuyoshi Nishitani for
measurement of the small-molecule X-ray single-crystal structure and their assistance in crystallographic data deposi-
tion to the Cambridge Crystallographic Data Centre. We thank Ken Bragstad and Jasmine Nguyen for protein chemis-
try efforts, and Gyorgy Snell for X-ray data collection of protein complexes. We also thank the Berkeley Center for
Structural Biology, which is supported in part by the National Institutes of Health, National Institute of General Medi-
cal Sciences, and the Howard Hughes Medical Institute. The Advanced Light Source is supported by the Director, Of-
fice of Science, Office of Basic Energy Sciences, of the U.S. Department of Energy under Contract No. DE-AC02-
05CH11231
Supplementary material
Supplementary data (chemical synthesis procedures, compound data, in vitro and in vivo assay methods, PK anal-
ysis methods, and single and co-crystal structural data) associated with this article can be found, in the online version.
7

References and notes
1.
2.
Ahern T, Wu FCW. Age Ageing 2015;44:188–195.
Travison TG, Morley JE, Araujo AB, O’Donnell AB, McKinlay JB. J. Clin. Endocrinol. Metab. 2006;91:2509–
2513.
3.
4.
5.
6.
7.
8.
9.
10.
Isidori AM, Greco EA, Aversa A. BJU Int. 2005;96:212–216.
Harman SM, Metter EJ, Tobin JD, Pearson J, Blackman MR. J. Clin. Endocrinol. Metab. 2001;86:724–731.
Udedibia E, Kaminetsky J. Expert Opin. Investig. Drugs 2014;23:1605–1618.
Choi SM, Lee, B-M. Expert Opin. Drug Saf. 2015;14:1773–1785.
Grech A, Breck J, Heidelbaugh J. Ther Adv Drug Saf. 2014; 5:190-200.
Narayanan R, Mohler ML, Bohl CE, Miller DD, Dalton JT. Nucl. Recept. Signal. 2008;6:e010.
Haendler B, Cleve A. Mol. Cell. Endocrinol. 2012;352:79–91.
Mohler ML, Bohl CE, Jones A, Coss CC, Narayanan R, He Y, Hwang DJ, Dalton JT, Miller DD. J. Med. Chem.
2009;52:3597–3617.
11.
12.
13.
14.
15.
16.
Zhang X, Sui Z. Expert Opin. Drug Discov. 2013;8:191–218.
Cosman F, Lindsay R. Endocr Rev. 1999;20:418–434.
Zhang X, Lanter JC, Sui Z. Expert Opin. Ther. Pat. 2009;19:1239–1258.
Jones A, Chen J, Hwang DJ, Miller DD, Dalton JT. Endocrinology 2009;150:385–395.
Srinath R, Dobs A. Futur. Oncol. 2014;10:187–194.
Nique F, Hebbe S, Triballeau N, Peixoto C, Lefrançois J-M, Jary H, Alvey L, Manioc M, Housseman C,
Klaassen H, Beeck KV, Guédin D, Namour F, Minet D, Ellem VdA, Feyen J, Fletcher S, Blanqué R, Robin-
Jagerschmidt C, Deprez P. J. Med. Chem. 2012;55:8236–8247.
17.
18.
Cozzoli A, Capogrosso RF, Sblendorio VT, Dinardo MM, Jagerschmidt C, Namour F, Camerino GM, Luca
AD. Pharmacol. Res. 2013;72:9–24.
Nique F, Hebbe S, Peixoto C, Annoot D, Lefrançois J-M, Duval E, Michoux L, Triballeau N, Lemoullec J-M,
Mollat P, Thauvin M, Prangé T, Minet D, Clément-Lacroix P, Robin-Jagerschmidt C, Fleury D, Guédin D,
Deprez P. J. Med. Chem. 2012;55:8225–8235.
19.
20.
21.
Aikawa K, Miyawaki T, Hitaka T, Imai YN, Hara T, Miyazaki J, Yamaoka M, Kusaka M, Kanzaki N, Tasaka
A, Shiraishi M, Yamamoto S. Bioorg. Med. Chem. 2015;23:2568–2578.
Saeed A, Vaught GM, Gavardinas K, Matthews D, Green JE, Losada PG, Bullock HA, Calvert NA, Patel NJ,
Sweetana SA, Krishnan V, Henck JW, Luz JG, Wang Y, Jadhav P. J. Med. Chem. 2016;59:750–755.
Nirschl AA, Zou Y, Krystek SR, Sutton JC, Simpkins LM, Lupisella JA, Kuhns JE, Seethala R, Golla R, Sleph,
PG, Beehler BC, Grover GJ, Egan D, Fura A, Vyas VP, Li Y-X, Sack JS, Kish KF, An Y, Bryson JA,
Gougoutas JZ, DiMarco J, Zahler R, Ostrowski J, Hamann LG. J. Med. Chem. 2009;52:2794–2798.
22.
23.
Yamamoto S, Kobayashi H, Kaku T, Hara T, Ymaoka M, Kanzaki N, Hasuoka A, Baba A, Ito M. Bioorg. Med.
Chem. 2013;21:70–83.
Imamoto T, Takiyama N, Nakamura K. Tetrahedron Lett. 1985;26:4763–4766.
8

24.
The stereochemical configuration of compound 1c was confirmed by X-ray crystallographic analyasis, see sup-
plementary material.
25.
26.
27.
Bolleddula J, DeMent K, Driscoll JP, Worboys P, Brassil PJ, Bourdet DL. Drug Metab. Rev. 2014;46:379–419.
cLog P was calculated by daylight.
Hershberger LG, Shipley EG, Meyer RK. Proc. Soc Exp. Biol. Med. 1953;83:175–180.
9

10