Optimization and metabolic stabilization of a class of nonsteroidal glucocorticoid modulators

Article abstract of DOI:10.1016/j.bmcl.2004.06.023

The optimization of a series of nonsteroidal glucocorticoid modulators is reported. Potent selective GR ligands that have improved metabolic stability were discovered typified by the subnanomolar acid 12 (GR binding IC ₅₀=0.6nM).

Full text of DOI:10.1016/j.bmcl.2004.06.023

Bioorganic & Medicinal Chemistry Letters 14 (2004) 4169–4172
Optimization and metabolic stabilization of a class of nonsteroidal
glucocorticoid modulators
J. T. Link,a,* Bryan K. Sorensen,a Jyoti Patel,b David Arendsen,a Gaoquan Li,b
Susan Swanson,a Bach Nguyen,a Maurice Emery,a, Marlena Grynfarbc and
Annika Goos-Nilssonc
aMetabolic Disease Research, Abbott Laboratories, Dept 4CB, Bldg. AP-10, Rm. L-14, 100 Abbott Park Road,
Abott Park, IL 60064-6098, USA
bMedicinal Chemistry Technologies, Abbott Laboratories, Dept 4CP, 100 Abbott Park Road, Abbott Park, IL 60064-6098, USA
cKaro Bio AB, Novum, SE-141 57 Huddinge, Sweden
Received 9 February 2004; revised 7 June 2004; accepted 9 June 2004
Available online 2 July 2004
Abstract—The optimization of a series of nonsteroidal glucocorticoid modulators is reported. Potent selective GR ligands that have
improved metabolic stability were discovered typified by the subnanomolar acid 12 (GR binding IC50 ¼ 0.6 nM).
ꢀ 2004 Elsevier Ltd. All rights reserved.
Nonsteroidal glucocorticoid receptor (GR) modulators
have been explored as potential therapeutics for multiple
diseases.1 We recently reported a class of modulators
that has potent GR affinity, moderate nuclear hormone
selectivity, but poor metabolic stability.2 A number of
approaches were explored to improve the metabolic
stability while optimizing their potency and selectivity.
Two approaches for modification of the potential anti-
diabetic GR modulators are detailed herein.
Dibenzyl aniline sulfonamides, like 1 and dihalide 2,
served as leads for the series (Fig. 1). During optimiza-
tion, substitution at the para position of the benzyl
Figure 1. GR modulator core structures.
groups improved binding affinity and two of the most
tent agonist [3H]-dexamethasone. Functional activity
potent cores discovered were diphenyl ketone 3 and di-
was assessed in a cellular assay (GRAF). In the GRAF
aryl ether 4. Rapid microsomal metabolism for these
assay, genetically engineered cells expressing a gluco-
and related compounds was observed. Metabolite
corticoid hormone response element and core promoter
identification revealed that N-debenzylation was the
sequences linked to a secreted alkaline phosphatase
primary metabolic pathway.
reporter were employed to measure blockade of an
agonist (dexamethasone) response. A second functional
assay measuring the blockade of dexamethasone
induced activity of tyrosine amino transferase (TAT) in
Human GR binding activity was determined in a radio-
ligand competition binding assay using the labeled po-
freshly isolated rat hepatocytes was also employed.
Nuclear hormone receptor selectivity was determined
using binding assays for the human androgen receptor
Keywords: Glucocorticoid modulator; Antidiabetic.
(h-AR), progesterone receptor (h-PR), mineralocorti-
coid receptor (h-MR), and thyroid-hormone receptor
(h-TRa and ).
* Corresponding author. Tel.: +1-847-935-6806; fax: +1-847-938-1674;
e-mail: james.link@abbott.com
Present address: Amgen, Thousand Oaks, CA 91320, USA.
b
0960-894X/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.06.023

4170
J. T. Link et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4169–4172
Table 1. GR binding, GR cellular functional inhibition, and rat
microsomal stability for compounds 1–4
Table 2. GR binding, GR cellular functional inhibition and rat
microsomal stability for compounds 13–15
Compds
GR binding
IC50, nMa
GRAF
IC50, nMa
Rat microsomal
t1=2 (min)b
1
2
3
4
28
250
210
340
240
7.5
nd
20
5.7
13
2.7
4
a Values are geometric means of two experiments (nd ¼ not deter-
Compds R1
R2
GR binding GRAF
IC50, nMa
Rat
IC50, nMa microsomal
mined).
b Microsomal incubations conducted at 10 lM with 0.25 mg/mL of
microsomal protein.
t1=2 (min)b
13
14
15
Me
CH2OH
Et
Me
H
71
56
12
290
840
270
3.5
4
H
9
The modulator cores 1–4 have potent GR binding
activity, but modest cellular/functional activity and poor
microsomal metabolic stability (Table 1). Diphenyl
ether 4 is the most active binder with an IC50 of 2.7 nM.
Compounds from this series were full antagonists in the
GRAF assay. All four compounds are greater than eight
times less potent in the GRAF assay relative to the
binding assay. For comparison, the potent steroidal
antagonist mifepristone (RU-486) has roughly equal
IC50’s in these assays. Rat microsomal metabolism for
sulfonamides 1, 3, and 4 is rapid with the most stable
compound being diphenyl ketone 3. None have suitable
stability for use as an orally dosed agent with the liver as
a potential target organ.
a Values are geometric means of two experiments (nd ¼ not deter-
mined).
b Microsomal incubations conducted at 10 lM with 0.25 mg/mL of
microsomal protein.
individual protocols that work efficiently on small and
multigram scale.
In order to improve metabolic stability multiple strate-
gies have been explored. First, the steric hindrance
around the dibenzylaniline nitrogen was increased. The
polarity of the compounds was then increased and the
metabolism tested. Other modifications designed to liver
target the compounds were also evaluated for their
effects on metabolic stability. Bile acid conjugation3 and
statin modulator hybrid formation4 were explored and
C-glucuronide formation5 was considered.
A representative synthetic sequence is described in
Scheme 1. Resorcinol monoacetate 5 was allylated, de-
protected, and the resulting phenol reacted with 4-flu-
orobenzaldehyde. The product, aldehyde 6, was then
subjected to reductive amination reaction conditions in
the presence of 2-methyl-3-nitroaniline and the product
benzylated to provide nitroarene 7. Deallylation with
tetrakis(triphenylphosphine)palladium in the presence
of phenylsilane provided the corresponding phenol 8.
The phenol could then be converted into a wide range of
analogs by alkylation or Mitsunobu reaction. Alkyl-
ation with a halobutyric ester yielded nitroarene 9. Nitro
reduction and mesylation then gave the methylsulfon-
amide 10 that was hydrolyzed to acid 11. Amide for-
mation followed by hydrolysis yielded the modulator 12.
Although step intensive, the routes consist of simple
The steric bulk around the aniline nitrogen could be
modestly increased without a dramatic loss of potency
(Table 2). For example, addition of a methyl group to
the R2 benzylic position as in 13 decreases the binding
affinity by a factor of 2.5 without much impact on
functional potency. Similarly, increasing the size of the
R1 substituent from Me 1 to CH2OH 14 or ethyl 15 does
not radically affect binding potency. Unfortunately,
none of the substitutions improved metabolic stability.
Larger substituents weakened GR affinity considerably
(data not shown).
Scheme 1. Representative synthesis of a GR modulator: (a) allyl bromide, K2CO3, DMF, 80 ꢁC, 12 h, 57%; (b) NaOH, THF, MeOH, H2O, rt, 1 h;
(c) 4-fluorobenzaldehyde, K2CO3, DMF, 100 ꢁC, 12 h, 60% (two steps); (d) 2-methyl-3-nitroaniline, AcOH, DCE, rt, 4 h; Na(OAc)3BH, 12 h, 82%;
(e) benzylbromide, i-Pr2NEt, DMF, 90 ꢁC, 12 h, 84%; (f) Pd(PPh3)4, PhSiH3, CH2Cl2, rt, 1 h, 96%; (g) NaH, 4-bromo-butyric acid ethyl ester,
DMF, 0 ꢁC fi rt, 12 h, 69%; (h) Fe, NH4Cl, EtOH, H2O, 80 ꢁC, 1 h; (i) MsCl, py, rt, 1 h; (j) NaOH, THF, EtOH, rt, 2.5 h, 72%; (k) EDCI, HOBt,
Et3N, DMF, rt, 12 h, 93%; (l) NaOH, THF, EtOH, rt, 2.5 h, 74%.

J. T. Link et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4169–4172
4171
Table 3. GR binding, GR cellular functional inhibition, and rat
microsomal stability for compounds 16–32 and 12
Adding polar groups around the periphery of the di-
phenyl ether core 4 proved to be a more effective strat-
egy. Large substituents were tolerated on the terminal
aryl ether ring (R1 and R2 positions––Table 3). A wide
variety of functional groups provided potent com-
pounds and most efforts focused on introducing groups
that would be charged at physiological pH. Introducing
amines that retained activity was challenging, but potent
compounds like morpholine 16 were discovered.
Unfortunately, these compounds did not show im-
proved metabolic stability. Less polar amides like 17
nearly maintained the potency of the core 4 but were
also metabolized rapidly. Ether 18 showed improved
functional potency but also did not enhance stability.
The same metabolism trend was observed in amide 19
and lactam 20.
Com- Substituents
pds
GR
binding IC50
IC50
nMa
GRAF Rat
,
micro-
somal t1=2
(min)b
,
nMa
O
N
16
(R2)
19
330
280
73
6
11
13
O
Introduction of carboxylic acids substantially improved
the metabolic stability of this series. All of the acids
tested had greater than 75% parent remaining after a
30 min incubation. For instance, the propionic or bu-
tyric acid sidechain in either the R1 (21 and 23) or R2 (22
and 24) position increased the microsomal half-life sig-
nificantly. Moving the acid further from the core at the
R1 site (while including an amide in the linker) did not
diminish the improved metabolic stability. For instance,
amides 25–27 showed improved potency relative to the
core 4. The same was true at the R2 site, where amides
and ethers could also be included in the chain (28–32).
Structurally related tetrazoles and acyl sulfonamides
were also prepared (data not shown). They had mod-
erate metabolic stability that fell between amides like 17
and the acids. Functional activity in the rat hepatocyte
TAT assay for most of the metabolically stable com-
pounds was moderate to weak. For instance, acid 21 has
an IC50 ¼ 8.5 lM (RU-486 IC50 ¼ 0.27 lM).
17
(R2)
5.4
2.4
O
NHAc
O
18
(R2)
O
O
NHAc
19
(R2)
2.9
2.2
nd
30
12
O
H
( )3
20
(R1)
180
N
N
O
21
(R1)
34
500
>30
CO2H
22
(R2)
26
380
670
670
120
>30
>30
>30
>30
CO2H
One of the most potent compounds found during this
survey was modulator 12 that had subnanomolar
binding IC50 for h-GR, good binding potency against
rat GR (IC50 ¼ 1 nM, although rat hepatocyte TAT
IC50<30 lM), improved functional activity (GRAF
IC50 ¼ 86 nM), and excellent microsomal stability. The
compound also displayed excellent nuclear hormone
receptor selectivity (>200-fold selective over AR, MR,
CO2H
23
(R1)
8.8
75
24
(R2)
CO2H
O
25
(R1)
1.6
N
H
CO2H
PR, and TRa and ––data not shown).
b
O
In order to assess the in vivo antidiabetic potential of
these compounds, we performed pharmacokinetic
studies on several analogs. The systemically available
antagonist RU-486 lowers blood glucose levels partially
due to its suppression of hepatic glucose production.6
Therefore, high liver drug levels were sought. Extra-
hepatic GR activity can be detrimental, including
hypothalamic–pituitary–adrenal axis activation effects,
so low peripheral exposure was desired. Representative
profiles of the metabolically stable analogs were ob-
served during oral and iv evaluations of amide 12 and
acid 21 in Sprague–Dawley rats. Amide 12 had a
modest half-life (t1=2 ¼ 2:2 h), low bioavailability
(F ¼ 12%), and low liver levels 7 h post oral dose
(0.56 lg/g) at 5 mpk. A high iv clearance (CLp ¼ 12:7 L/
h kg) greater than liver blood flow led to a low area
CO2H
26
(R1)
1.5
nd
>30
N
H
O
27
(R1)
1.4
200
460
>30
>30
N
H
CO2H
O
O
28
(R2)
11
N
H
CO2H
29
(R2)
CO2H
6.5
nd
>30
N
H
(continued on next page)

4172
J. T. Link et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4169–4172
Table 3 (continued)
ments revealed the basic cores have modest pharmaco-
kinetic profiles. Related modulators with superior PK
profiles and their in vivo activity in rodent models of
diabetes will be reported shortly.
Com- Substituents
pds
GR
binding IC50
IC50
nMa
GRAF Rat
,
micro-
somal t1=2
(min)b
,
nMa
H
N
CO2H
30
(R2)
11
660
nd
O
O
Acknowledgements
31
(R2)
11
130
640
86
nd
Dr. Kennan Marsh is thanked for her assistance with
pharmacokinetic experiments. Jiahong Wang and Ste-
ven Fung are thanked for providing rat assay data.
O
O
CO2H
H
N
CO2H
32
(R1)
3.9
0.6
nd
O
H
N
References and notes
CO2H
12
(R2)
>30
O
O
1. (a) Coghlan, M. J.; Elmore, S. W.; Kym, P. R.; Kort, M. E.
Ann. Rep. Med. Chem. 2002, 37, 167–176; (b) Coghlan, M.
J.; Elmore, S. W.; Kym, P. R.; Kort, M. E. Curr. Top. Med.
Chem. 2003, 3, 1617–1635.
2. Link, J. T.; Sorensen, B. K.; Patel, J.; Emery, M.; Grynfarb,
M.; Goos-Nilsson, A. Bioorg. Med. Chem. Lett. 2004, 14,
2209–2212.
Position of substituent listed in compds column. Unlisted substituent is
hydrogen.
a Values are geometric means of two experiments (nd ¼ not deter-
mined).
b Microsomal incubations conducted at 10 lM with 0.25 mg/mL of
microsomal protein.
3. Tu, N.; Link, J. T.; Sorensen, B. K.; Emery, M.; Grynfarb,
M.; Goos-Nilsson, A.; Nguyen, B. Bioorg. Med. Chem.
Lett. 2004, 14, in press. doi:10.1016/j.bmcl.2004.06.022.
4. Link, J. T.; Sorensen, B. K.; Lai, C.; Wang, J.; Fung, S.;
Deng, D.; Emery, M.; Carroll, S.; Grynfarb, M.; Goos-
Nilsson, A.; von Geldern, T. Bioorg. Med. Chem. Lett.
2004, 14, in press. doi:10.1016/j.bmcl.2004.06.021.
5. Sorensen, B. K.; Link, J. T.; von Geldern, T.; Emery, M.;
Wang, J.; Hickman, B.; Grynfarb, M.; Goos-Nilsson, A.;
Carroll, S. Bioorg. Med. Chem. Lett. 2003, 13, 2307–
2310.
under the curve after a 5 mpk dose suggesting amide
hydrolysis may be taking place. Acid 21 had a signifi-
cantly lower iv clearance (CLp ¼ 2:8 L/h kg, 5 mpk) that
resulted in improved area under the curve. Unfortu-
nately, oral bioavailability (10 mpk) was still modest
(F ¼ 10%) leading to low drug exposure.
A series of nonsteroidal GR modulators was optimized.
Potent and selective compounds with improved meta-
bolic stability were discovered with the most active
analog being amide 12. Early pharmacokinetic experi-
6. Friedman, J. E.; Sun, Y.; Ishizuka, T.; Farrell, C. J.;
McCormack, S. E.; Herron, L. M.; Hakimi, P.;
Lechner, P.; Yun, J. S. J. Biol. Chem. 1997, 272, 31475–
31481.