Discovery and SAR study of novel dihydroquinoline-containing glucocorticoid receptor agonists

Article abstract of DOI:10.1016/j.bmcl.2007.07.021

We have recently reported the discovery of a novel class of glucocorticoid receptor (GR) antagonists, exemplified by 3, containing a 1,2-dihydroquinoline molecular scaffold. Further SAR studies of these antagonists uncovered chemical modifications conveying agonist functional activity to this series. These agonists exhibit good GR binding affinity and are selective against other nuclear hormone receptors.

Full text of DOI:10.1016/j.bmcl.2007.07.021

Bioorganic & Medicinal Chemistry Letters 17 (2007) 5091–5095
Discovery and SAR study of novel dihydroquinoline-
containing glucocorticoid receptor agonists
Hidenori Takahashi,* Younes Bekkali, Alison J. Capolino, Thomas Gilmore,
Susan E. Goldrick, Paul V. Kaplita, Lisa Liu, Richard M. Nelson, Donna Terenzio,
Ji Wang, Ljiljana Zuvela-Jelaska, John Proudfoot, Gerald Nabozny and David Thomson
Boehringer Ingelheim Pharmaceuticals, Inc., 900 Ridgebury Road, Ridgefield, CT 06877, USA
Received 4 June 2007; revised 2 July 2007; accepted 5 July 2007
Available online 13 July 2007
Abstract—We have recently reported the discovery of a novel class of glucocorticoid receptor (GR) antagonists, exemplified by 3,
containing a 1,2-dihydroquinoline molecular scaffold. Further SAR studies of these antagonists uncovered chemical modifications
conveying agonist functional activity to this series. These agonists exhibit good GR binding affinity and are selective against other
nuclear hormone receptors.
ꢀ 2007 Elsevier Ltd. All rights reserved.
Synthetic glucocorticoids are widely used to treat many
serious inflammatory and autoimmune disorders.1 How-
ever, a major drawback in their clinical use is their asso-
ciation with a number of severe and, in some cases,
life-threatening adverse events, such as enhanced bone
resorption and muscle weakening. The discovery of glu-
cocorticoid receptor (GR) agonists that are dissociated,
that is, that exhibit a reduced incidence or a reduced
severity of side effects while maintaining potent anti-
inflammatory activity, is currently an area of intense
research activity within the medicinal chemistry commu-
nity.2 There are also additional efforts underway to iden-
tify selective GR antagonists with the expectation that
these may be useful in treating diabetes.3
onstrate dissociated GR agonist properties in both cellu-
lar and in vivo systems. In addition, we have recently
disclosed a new class of potent and selective GR ligands,
such as 3, that display antagonistic activity in a cellular
system (Fig. 1).6
Literature precedents indicate that minor structural
modifications of some GR ligands can cause a func-
tional switch from agonist to antagonist activity.7
Therefore, we decided to explore if our GR antagonists
N
OH
11
The progesterone receptor (PR) antagonist Mifepristone
(RU-486) 1 is a known GR antagonist effective at block-
ing gene-transcription mediated by endogenous gluco-
corticoids. The therapeutic utility of Mifepristone has
been demonstrated for the treatment of Cushing’s syn-
drome, diabetes, glaucoma, and depression.4 Although
structural similarity between Mifepristone and endoge-
nous glucocorticoids is noticeable, it also shares struc-
tural features with a series of compounds from
Abbott/Ligand, exemplified by 2,5 that reportedly dem-
H
H
O
1
Mifepristone (RU-486)
F
S
O
O
O
N
H
N
H
Keywords: Glucocorticoid receptor; Agonist; SAR.
*
3
2
AL-438
Corresponding author. Tel.: +1 203 798 5049; fax: +1 203 798
5297; e-mail: htakahas@rdg.boehringer-ingelheim.com
Figure 1. Steroidal and nonsteroidal nuclear receptor ligands.
0960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.07.021

5092
H. Takahashi et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5091–5095
Table 1. Effects of a-methylation at the C-4 position on GR and PR
binding affinity
a
O
R
N
X
O
N
H
O
O
5
4
O
N
H
R
OH
O
O
Compound
X
R
GRa IC50 (nM)
PRa IC50 (nM)
c
b
12
13
14
15
16
S@O
S
H
H
>2000
360
116
>2000
460
600
O
O
N
N
S
Me
H
7
O
>2000
116
>2000
480
O
O
6
O
O
Me
a Values are means of two experiments.
R
O
could also be converted to agonists via minor structural
modifications. In addition, we attempted to improve the
drug-like properties of our GR ligands represented by 3.
d, e
O
A
primary concern was that thioether-containing
N
H
compounds are likely to be extensively metabolized by
flavin-containing monooxygenase (FMO) and cyto-
chrome P450 enzymes.8 These enzymes typically oxidize
the thioether-containing xenobiotics to sulfoxide and/or
sulfone metabolites, which often display high clearance
in in vivo or in vitro systems. Efficacy was expected to
be concomitantly lost since compound 12, a sulfoxide
analog of one of our antagonists, did not demonstrate
appreciable GR binding affinity when tested at concen-
trations up 2000 nM (Table 1). We thus anticipated that
improving the compounds metabolic stability would be
beneficial for both their pharmacokinetic and pharma-
codynamic profiles. This paper discloses our initial at-
tempts at identifying a novel series of dissociated GR
agonists that are structurally derived from our original
series of GR antagonists exemplified by 3.
8
Scheme 1. Reagents and conditions: (a) Boc2O, 1.6 M n-BuLi, THF,
ꢀ78 ꢁC ! 0 ꢁC ! rt, 6 h, 91%; (b) selenium dioxide, 1,4-dioxane,
reflux, 3 h, 94%; (c) Grignard reagent, THF, ꢀ78 ꢁC, 30 min, 47–83%;
(d) allyl iodide, 1.0 M sodium bis(trimethylsilyl)amide, DMSO, rt,
10 min, 51–88%; (e) TFA, dichloromethane, rt, 2 h, 60–90%.
PR, estrogen receptor (ER), and mineralo-corticoid
receptor (MR) using a fluorescence polarization compet-
itive binding assay.9
Introduction of a methyl group at the a-position of the
1,2-dihydroquinoline C4 substituent resulted in a sig-
nificant increase in GR binding affinity for both the
thioether- and the ether-containing analogs. For the
thioether analogs, introduction of the methyl group
at the a-position resulted in an approximately threefold
increase in GR binding affinity, whereas more than a
15-fold increase in GR binding affinity was observed
for the ether analogs. Reduction of the size of the het-
eroatom from sulfur to oxygen was not beneficial for
GR binding affinity in the absence of the a-methyl sub-
stitution. However, increasing the size of C4 substitu-
ents in the presence of the a-methyl group restored
GR binding affinity (Table 1). These data suggest that
the size of the substituent at the C4 position is impor-
tant to achieve potent GR binding affinity. This discov-
ery encouraged us to prioritize the ether series for
further exploration over the thioether analogs. We also
felt that moving away from the thioether derivatives
would also alleviate the potential metabolic liability
associated with sulfur oxidation.
For SAR study of the a-position on C4 substituent, we
synthesized various derivatives, starting from compound
4.6 Protection of the secondary amine in 4 with a tBoc
group followed by allylic oxidation with selenium diox-
ide yielded aldehyde 6 in good yield. Reactions of 6 with
a variety of Grignard reagents at low temperature affor-
ded the secondary alcohols 7 in modest to good yield.
Further O-alkylation of 7 with allyl iodide, followed
by removal of the tBoc group with trifluoroacetic acid,
provided the desired allyl ether analogs 8, in acceptable
overall yield (Scheme 1).
To study the effects of oxygen substitution, we synthe-
sized several ether analogs starting from compound 6.
The aldehyde 6 was treated with methyl magnesium bro-
mide to afford the secondary alcohol 9 in good yield.
This alcohol was then activated as a mesylate and trea-
ted with various alkoxides to afford the corresponding
ether derivatives. These compounds were eventually
deprotected under standard conditions to generate the
analogs 11 (Scheme 2).
We then investigated the replacement of the a-methyl
group in compound 16 with other substituents, and ana-
lyzed the impact of this modification on the resulting
GR binding affinity. Increasing the size of the a-substi-
tuent from methyl (16; IC50 = 116 nM) to ethyl (17;
IC50 = 355 nM) or i-propyl (18; IC50 = 710 nM) was
All the compounds were evaluated for binding in a panel
of human nuclear hormone receptors, including GR,

H. Takahashi et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5091–5095
OH
5093
16, resulted in about a ninefold increase in GR binding
affinity. However, increasing the substituent size further
resulted in a loss of GR binding affinity, as illustrated
with the 2-butenyl, compound 25 (IC50 = 210 nM), the
2-pentenyl compound 27 (IC50 = 565 nM), and the
3-phenylallyl compound 28 (IC50 = 980 nM). Reduction
of the double bond in compound 28 led to a threefold
enhancement in GR binding affinity, while the phenethyl
ether analogs (30–33) displayed GR binding affinities
similar to those of compound 29 (IC50 = 185–400 nM).
Regarding the PR binding affinity, no significant differ-
ence was observed between the methyl ether 24
(IC50 = 540 nM) and the allyl ether, 16 (IC50 = 480 nM).
Larger oxygen substituents were also characterized by a
significant loss of PR binding affinity. For example,
compounds 29 and 30, featuring a phenylpropyl and
O
O
a
O
O
N
N
9
O
O
6
O
OMs
O
R
b
c, d
O
O
N
N
H
10
11
O
O
phenethyl substitution, respectively, showed IC50
>
2000 nM. Although an SAR emerged through oxygen
substitution for both GR and PR binding affinity, it
remained relatively flat for the GR binding affinity when
Scheme 2. Reagents and conditions: (a) MeMgBr ether solution, THF,
0 ꢁC, 15 min, 87%; (b) MsCl, TEA, dichloromethane, 0 ꢁC, 1 h, 99%;
(c) primary alcohol, 60% NaH, DMF, 0 ꢁC ! rt, 4 h, 30–85%; (d)
TFA, dichloromethane, rt, 2 h, 60–90%.
Table 3. GR and PR binding affinity with different O-substituents
R
O
accompanied with an approximately threefold and
sixfold loss of GR binding affinity, respectively. Com-
pounds featuring a n-propyl (19) or larger substituents
at the a-position lost a significant amount of GR bind-
ing affinity. The n-propyl substituted compound 19
and the phenyl substituted analog 21 were devoid of
either GR or PR binding affinity when tested at concen-
trations up to 2000 nM. These results indicate that the
size of the a-substituent is critical for GR binding affin-
ity. Interestingly, the impact of the size of the a-substit-
uents on PR binding affinity was not as significant as it
was on GR binding affinity (Table 2).
O
N
H
Compound
R
GRa
IC50 (nM)
PRa
IC50 (nM)
24
16
Me
Allyl
1060
116
540
480
25
26
27
28
210
395
565
980
1300
980
*
Replacing the allyl group of compound 16 with other al-
kyl groups affected both the GR and the PR binding
affinities. For example, replacing the methyl substituent
in compound 24, with an allyl group, as in compound
*
*
*
1600
>2000
Table 2. GR and PR binding affinity with different a-substituents
*
29
30
31
260
405
290
>2000
>2000
>2000
R
O
*
*
O
N
H
GRa IC50 (nM)
PRa IC50 (nM)
Cl
Compound
R
15
16
17
18
19
20
21
22
23
H
>2000
116
355
>2000
480
330
32
33
34
275
185
400
1900
930
Methyl
Ethyl
Cl
Cl
*
i-Propyl
n-Propyl
Allyl
710
950
>2000
440
>2000
>2000
>2000
>2000
>2000
*
*
Phenyl
Benzyl
Phenethyl
>2000
1200
1000
>2000
a Values are means of two experiments.
a Values are means of two experiments.

5094
H. Takahashi et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5091–5095
compared to the trends observed when modifying the a-
substitution. These results suggest that a-substitution
has a more significant impact on GR binding affinity
than the oxygen substitution (Table 3).
The enantiomers of the most potent GR ligand, com-
pound 38, were separated by chiral HPLC,10 and pro-
filed for binding affinities, functional activity, and
in vitro metabolic stability. Compound 42, (+)-isomer,
and the corresponding racemate demonstrated similar
GR binding affinities, whereas compound 41, (ꢀ)-iso-
mer, was about threefold less potent. No differences
in selectivity profiles were observed between the race-
mate and either enantiomers. The in vitro metabolic
stability of these compounds was assessed by
measuring their half-lives in a human microsomal
preparation.11 Unfortunately, there was no significant
difference between the ether and thioether series, as
all compounds showed relatively short half-lives
(620 min). Finally, both enantiomers were evaluated
for functional cellular activity using an IL-1 induced
IL-6 assay in human foreskin fibroblasts.12 Although
both compounds 41 and 42 inhibited IL-6 production
with IC50’s of 260 and 170 nM, respectively, the mag-
nitude of their inhibition of IL-6 production at the
highest concentration (2000 nM) was modest. Indeed,
compounds 41 and 42 displayed only 46% and 28%
of dexamethasone’s maximum effect, respectively.
These data indicate that these ligands are partial GR
agonists (Table 5).
Structural variations of the right-hand side phenyl ring
were also examined. We have reported earlier that intro-
duction of fluorine atom at the C5 position of this phe-
nyl ring resulted in about a twofold increase in GR
binding affinity (compounds 35 and 37) in the thioether
series.6 A similar positive effect on GR binding affinity
was also observed in the ether series. Especially in the al-
lyl ether series, introduction of a fluorine atom as R2
produced a significant increase in GR binding affinity
and afforded the most potent GR ligand, compound
38 (IC50 = 34 nM), in this series (Table 4).
Table 4. Effects of fluorine substitution at R2 on GR and PR binding
affinity
R2
O
R2
O
S
O
R1
R1
N
N
H
H
In summary, a novel and potent series of GR partial
agonists has been identified by structural modification
of a GR antagonist lead. Structural modifications of
substituents at the C4 and the C6 position impacted
GR and PR binding affinities differently, and led to
the discovery of potent and selective binders. The most
potent GR ligand within the series (38) was resolved to
its (+)-isomer 42. Compound 42 represents a novel,
potent, and selective GR partial agonist. Unfortu-
nately, replacement of the thioether functionality
with an ether substituent did not provide compounds
metabolically more stable in human microsome
preparations.
Ethers
Thioethers
Compound Series
R1
R2 GRa PRa
IC50 (nM) IC50 (nM)
13
35
36
37
16
38
39
40
Thioether Allyl
H
F
360
190
194
84
460
1400
Thioether Phenethyl
H
F
>2000
>2000
480
Ether
Ether
Allyl
H
F
116
34
451
>2000
1300
Phenethyl
H
F
404
195
a Values are means of two experiments.
Table 5. GR, PR, ER, and MR binding affinity of compounds with the best substituents combinations at C4 and C6 position
F
F
F
O
S
S
O
O
O
N
H
N
N
H
H
16: Racemate
41: (-)-isomer
42: (+)-isomer
36
38
Compound
GRa IC50 (nM)
PRa IC50 (nM)
MRa IC50 (nM)
ERa IC50 (nM)
HLM T1/2 (min)
IL-6 IC50 (nM)
16
41
42
35
37
34
110
40
450
620
970
>2000
1100
>2000
>2000
>2000
>2000
>2000
6
5
Not tested
260 (46%)b
170 (28%)b
Not active
Not active
170
>2000
>2000
6
190
84
>2000
>2000
12
17
a Values are means of two experiments.
b Values are maximum inhibition at 2000 nM percentage to dexamethasone inhibition.

H. Takahashi et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5091–5095
5095
Baculovirus lysate containing either GR or MR was
incubated with 5 nM tetramethyl-rhodamine conjugate of
dexamethasone, and test compound dilutions in an assay
buffer containing 10 mM TES, 50 mM KCl, 20 mM
sodium molybdate, 1.5 mM EDTA, 0.04% w/v CHAPS,
10% v/v glycerol, 1 mM DTT, pH 7.4. For the PR assay,
baculovirus lysate-containing PR was incubated with
5 nM tetramethyl-rhodamine conjugate of RU486, and
the test compound dilutions. The ER binding assay was
performed using the ER Competitor Assay kit from
Panvera (Invitrogen part number P2614). This assay uses
purified baculovirus-expressed human ER and fluorescein
conjugate of a proprietary ER ligand (FluormoneTM ES2).
IC50 values shown were means of a single experiment done
in duplicate 11-point concentration–effect curves.
Bekkkali, Y.; Gilmore, T.; Spero, D. M.; Takahashi,
H.; Thomson, D. S.; Wang, J. PCT Int. Appl.,
WO2004018429.
Acknowledgment
´
We are grateful to Dr. Stephane De Lombaert for his
critique of this manuscript.
References and notes
1. Toogood, J. In Glucocorticoids; Goulding, N. J., Flowers,
R. J., Eds.; Birkhauser: Boston, 2001; pp 161–174.
2. Rosen, J.; Marschke, K.; Deepa, R. Curr. Opin. Drug
Discov. Dev. 2003, 6, 224.
3. Coghlan, M. J.; Elmore, S. W.; Kym, P. R.; Kort, M. E.
Ann. Rep. Med. Chem. 2002, 37, 167.
4. (a) Chu, J. W.; Matthias, D. F.; Belanoff, J.; Schatzberg,
A.; Hoffman, A. R.; Feldman, D. J. Clin. Endocrinol.
Metab. 2001, 86, 3568; (b) Gettys, T. W.; Watson, P. M.;
Taylor, I. L.; Collins, S. Int. J. Obes. 1997, 21, 865; (c)
Phillips, C. I.; Green, K.; Gore, S. M.; Cullen, P. M.;
Campbell, M. Lancet 1984, 1, 767; (d) Belanoff, J. K.;
Flores, B. H.; Kalezhan, M.; Sund, B.; Schatzberg, A. F.
J. Clin. Psychopharmacol. 2001, 21, 516.
5. (a) Elmore, S. W.; Coghlan, M. J.; Anderson, D. D.; Pratt,
J. K.; Green, B. E.; Wang, A. X.; Stashko, M. A.; Lin, C.
W.; Tyree, C. M.; Miner, J. N.; Jacobson, P. B.; Wilcox,
D. M.; Lane, B. C. J. Med. Chem. 2001, 44, 4481; (b)
Kym, P. R.; Kort, M. E.; Coghlan, M. J.; Moore, J. L.;
Tang, R.; Ratajczyk, J. D.; Larson, S. W.; Elmore, J. D.;
Pratt, J. K.; Stashko, M. A.; Douglass Falls, H.; Lin, C.
W.; Nakane, M.; Miller, L.; Tyree, C. M.; Miner, J. N.;
Jacobson, P. B.; Wilcox, D. M.; Nguyen, P.; Lane, B. C. J.
Med. Chem. 2003, 46, 1016; (c) Coghlan, M. J.; Jacobson,
P. B.; Lane, B.; Nakane, M.; Lin, C. W.; Elmore, S. W.;
Kym, P. R.; Luly, J. R.; Carter, G. W.; Turner, R.; Tyree,
C. M.; Hu, J.; Elgort, M.; Rosen, J.; Miner, J. N. Mol.
Endocrinol. 2003, 17, 860.
6. Takahashi, H.; Bekkali, Y.; Capolino, A. J.; Gilmore, T.;
Goldrick, S. E.; Nelson, R. M.; Trenzio, D.; Wang, J.;
Zuvela-Jelaska, L.; Proudfoot, J.; Nabozny, G.; Thomson,
D. Bioorg. Med. Chem. Lett. 2006, 16, 1549.
7. Zhi, L.; Ringgenberg, J. D.; Edwards, J. P.; Tegley, C. M.;
West, S. J.; Pio, B.; Motamedi, M.; Jones, T. K.;
Marschke, K. B.; Mais, D. E.; Schrader, W. T. Bioorg.
Med. Chem. Lett. 2003, 13, 2075.
8. (a) Furnes, B.; Schlenk, D. Drug Metab. Dispos. 2005, 33,
214; (b) Usmani, K. A.; Karoly, E. D.; Hodgson, E.; Rose,
R. L. Drug Metab. Dispos. 2004, 32, 333; (c) Furne, B.;
Schlenk, D. Toxicol. Sci. 2004, 78, 196.
9. Fluorescence polarization competitive binding assays were
performed to quantitate the ability of test compounds to
displace ligands from GR, MR, ER, and PR in solution.
Binding reactions were assembled in 96-well microplates.
10. Enantiomers of compound 38 were separated by following
conditions. Column: Chiralcel OD (25 cm · 2.5 cm ID)
Mobile phase: 0.5% isopropanol in hexanes. Flow: 10 ml/
min. Compound 41: 94.4% ee (chiral HPLC), [a]D ꢀ70ꢁ
(c = 0.05, CHCl3, 25 ꢁC). Compound 42: 94.5% ee (chiral
HPLC), [a]D +76ꢁ (c = 0.05, CHCl3, 25 ꢁC). Absolute
configurations of compounds 41 and 42 have not been
determined.
11. The assay was performed in 50 mM potassium phosphate
buffer, pH 7.4, and 2.5 mM NADPH. Test samples were
dissolved in acetonitrile for a final assay concentration of
1–10 lM. Human liver microsomes were diluted in assay
buffer to a final assay concentration of 1 mg protein/ml. A
volume of 25 ll compound solution and 50 ll microsome
suspension were added to 825 ll assay buffer. The
preparation is incubated for 5 min in a 37 ꢁC water bath.
The reaction was started by the addition of 100 ll
NADPH. Volumes of 80 ll were removed from the
incubation mix at 0, 3, 6, 10, 15, 20, 40, and 60 min after
the start of the reaction and added to 160 ll acetonitrile.
The samples were shaken for 20 s and then centrifuged for
3 min at 3000 rpm. A 200 ll volume of the supernatant is
transferred to 0.25 mm glass fiber filter plates and
centrifuged for 5 min at 3000 rpm. Injection volumes of
10 ll were typically added to Zorbax SB C8 HPLC
columns with formic acid in water or acetonitrile at a flow
rate of 1.5 ml/min. Percent loss of parent compound was
calculated from the area under each time point to
determine the half-life.
12. Human foreskin fibroblasts were stimulated with 1 ng/ml
recombinant human IL-1 in the presence of test com-
pound. After 24 h, the degree of GR agonist activity
(transrepression) was determined by measuring IL-6 in the
tissue culture media.