Synthesis and structure-activity relationships of a series of 3-aryl-4-isoxazolecarboxamides as a new class of TGR5 agonists

Article abstract of DOI:10.1016/j.bmcl.2010.01.003

A series of 3-aryl-4-isoxazolecarboxamides identified from a high-throughput screening campaign as novel, potent agonists of the human TGR5 G-protein-coupled receptor is described. Many analogues were readily accessible via solution-phase synthesis which

Full text of DOI:10.1016/j.bmcl.2010.01.003

Bioorganic & Medicinal Chemistry Letters 20 (2010) 1363–1367
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and structure–activity relationships of a series
of 3-aryl-4-isoxazolecarboxamides as a new class of TGR5 agonists
a
a
a
b
Brian W. Budzik a, Karen A. Evans a, , David D. Wisnoski , Jian Jin , Ralph A. Rivero , George R. Szewczyk ,
*
Channa Jayawickreme c, David L. Moncol c, Hongshi Yu d
a Discovery Medicinal Chemistry, Discovery Research, GlaxoSmithKline Pharmaceuticals, 1250 South Collegeville Road, PO Box 5089, Collegeville, PA 19426-0989, United States
b Metabolic Diseases Center of Excellence for Drug Discovery, GlaxoSmithKline Pharmaceuticals, Five Moore Drive, PO Box 13398, Research Triangle Park, NC 27709-3398,
United States
c Discovery Research, GlaxoSmithKline Pharmaceuticals, Five Moore Drive, PO Box 13398, Research Triangle Park, NC 27709-3398, United States
d Discovery Research, GlaxoSmithKline Pharmaceuticals, 709 Swedeland Road, King of Prussia, PA 19406, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of 3-aryl-4-isoxazolecarboxamides identified from a high-throughput screening campaign as
novel, potent agonists of the human TGR5 G-protein-coupled receptor is described. Many analogues were
readily accessible via solution-phase synthesis which resulted in the rapid identification of key structure–
activity relationships (SAR), and the discovery of potent exemplars (up to pEC50 = 9). Details of the SAR
and optimization of this series are presented herein.
Received 20 November 2009
Revised 24 December 2009
Accepted 4 January 2010
Available online 7 January 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
SAR
TGR5
TGR5 agonists
GPCR
Solution-phase synthesis
Diabetes mellitus is an ever-increasing threat to human health.
According to the International Diabetes Federation (IDF), in 2010
diabetes is expected to cause close to four million deaths in the
20–79 age group, accounting for 6.8% of global mortality for this
age range. The estimated diabetes prevalence for 2010 is 285 mil-
lion, and within 20 years this number is expected to rise to 438
million.1 Those that suffer from type II diabetes (T2D) have too lit-
tle insulin or cannot use insulin effectively. As a result, glucose lev-
els build up in the blood and urine, and if left untreated, can cause
life-threatening complications, including blindness, kidney failure,
and heart disease. Despite the use of various hypoglycemic agents
with or without co-therapy with insulin, current treatments often
fail to achieve significant lowering of serum glucose to a level
where the severity of diabetic complications is effectively reduced.
Thus, there is a clear need for novel therapeutics targeting the
underlying mechanism(s) of elevated serum glucose.
GLP-1 are GLP-1 receptor agonist peptides (GLP-1 mimetics) and
DPP-IV inhibitors, which enhance insulin action by prolonging
the half-life of endogenous GLP-1.5 These approaches are effective
because while type II diabetics show impaired GLP-1 secretion,
they maintain normal responsiveness to GLP-1. However, the use
of a peptide in clinical treatment is severely limited because of
the requirement for parenteral administration and low in vivo sta-
bility. Therefore, a small molecule agonist that increases GLP-1
secretion from the gut during a meal is a potentially useful thera-
peutic for metabolic disorders such as type II diabetes and its asso-
ciated complications.
We previously reported a series of 3-aryl-4-isoxazolecarboxa-
mides exemplified by compounds 9 and 22 as novel, potent small
molecule agonists of the human TGR5 receptor which demon-
strated improved GLP-1 secretion in vivo via an intracolonic dose
(1 mg/kg) co-administered with glucose (0.125 g/kg) in conscious
dogs as compared to control.6 Herein we describe the SAR and
optimization of this isoxazolecarboxamide series from the high-
throughput screening (HTS) hit 1 to the identification of lead
compounds such as 48 and 61.
TGR5, also known as BG37, M-BAR, or hGPCR19, is a bile acid G-
protein-coupled receptor primarily expressed in monocytes and
macrophages, lung, spleen, and the intestinal tract. It has been sug-
gested that bile acids induce GLP-1 secretion from primary intesti-
nal cells by increasing intracellular cAMP levels via the TGR5
receptor.2–4 Recent FDA approved therapies for T2D based on
High-throughput screening (HTS) of the GSK proprietary com-
pound collection7 in a BacMam transduced human osteosarcoma
cell line (U2-OS) led to the identification of isoxazole 1 as a specific
agonist with a pEC50 of 5.3 and 100% maximum response.8 In addi-
tion, the compound had excellent physicochemical properties
* Corresponding author. Tel.: +1 610 917 7764; fax: +1 610 917 7391.
E-mail address: karen.a.evans@gsk.com (K.A. Evans).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.01.003

1364
B. W. Budzik et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1363–1367
which made it of particular interest as a starting point for an opti-
mization program designed to improve potency via exploration of
the SAR (Fig. 1).
R2
R2
O
O
R1
a, or
b, c
To explore this novel HTS hit, an efficient one- or two-step syn-
thesis was initially utilized (Scheme 1). Commercially available 5-
methyl-3-aryl-4-isoxazole-carbonyl chlorides 2 could be converted
to the desired amides 3 in one step from commercially available, or
prepared,9 substituted N-alkylanilines. Alternatively, the acid chlo-
ride could be coupled with a substituted primary aniline, and then
alkylated in a second step. Final compounds were typically purified
by reverse-phase HPLC to purities of >95% (LC–MS, UV 214 nm
detection) and then evaluated for pEC50 against the human TGR5
receptor in both a U2-OS and a melanophore cell line.10 The results
are detailed below. Initially the SAR of substitutions on the amide
phenyl ring (R1) was investigated (Table 1). The key finding was
that para-substitution (6, 9, 14, 17) was preferred over substitu-
tions at either the ortho- (4, 7, 12, 15) or meta-positions (5, 8, 13,
16) as well as the unsubstituted parent compound 25. Moreover,
at the para-position, the chloro group (9), as well as the methyl
substitution (14), were preferred over the strongly electron-rich
methoxy group (17). Interestingly, some of these substituents
could be combined to afford a further boost in potency, while other
effects were not additive. For example, the 4-chloro-3-methyl
disubstitution (18) lost some potency as compared to the mono-
substituted 4-chloro analogue (9), with further reductions seen
in the 4-chloro-2-methyl analogue (19). This same trend in po-
tency reduction is seen with the dichloro combinations: 4-Cl
(9) > 3,4-diCl (20) > 2,4-diCl (21). In contrast, when the 4-substitu-
tion was methyl, additional halo (22–23) and methyl substitution
(24) resulted in a boost in potency in each case.
To further probe these findings, monosubstitution at the para-
position of the aryl amide was further investigated with additional
electron-withdrawing groups as well as alkyl groups with branch-
ing, both cyclic and acyclic (Table 2). The modifications in this re-
gion clearly showed that small lipophilic groups were preferred.
Any alkyl substitutions larger than methyl, as with compounds
26–30, demonstrated reduction in potency with increasing size.
The CF3 group (31) was well tolerated, and to a lesser extent the
OCF3 (32), but all other groups prepared (nitrile 33, dimethylamino
34, ketone 35, sulfone 36 and ester 37) lost potency, likely due to a
combination of either steric and/or electronic effects.
Cl
N
N
N
O
O
2
3
Scheme 1. Synthesis of 3-aryl-4-isoxazolecarboxamides. Reagents and conditions:
(a) various substituted N-alkylanilines, Et3N, DCM, rt, or Et3N, DME,
1 min; (b) various substituted anilines, Et3N, DCM, rt, or Et3N, DME,
1 min; (c) CH3I, powdered KOH, DMSO, 50 °C, or rt.
l
l
W 190 °C,
W 190 °C,
Table 1
pEC50 inhibition for amide phenyl substitutiona
4
O
R1
Cl
3
N
N
2
O
Compound
R1
pEC50
4
5
6
7
8
2-F
3-F
4-F
2-Cl
3-Cl
4-Cl
4-Br
4-I
5.6
6.2
6.3 ( 0.7)
5.7
6.2
7.5
7.8
6.9
6.4
5.9
6.9
5.6
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
2-Me
3-Me
4-Me
2-OMe
3-OMe
4-OMe
4-Cl, 3-Me
4-Cl, 2-Me
3,4-diCl
2,4-diCl
3-Cl, 4-Me
3,5-diBr, 4-Me
3,4-diMe
H
5.9
6.4
7.2 ( 0.9)
6.6
7.0
5.6
7.1
7.9
7.7
5.4
a
pEC50 values given are means of at least two experiments in the melanophore
Variations on the isoxazole phenyl ring were next investigated
to better understand the role of substitutions on this ring. These
compounds were prepared where possible starting from commer-
cially available 5-methyl-3-aryl-4-isoxazole-carbonyl chlorides as
described in Scheme 1. Analogues could also be prepared according
to Scheme 2. Starting from commercially available benzaldehydes,
the chloro-oxime was prepared in two steps by treatment with
hydroxylamine followed by N-chlorosuccinimide.11 A microwave-
assisted 1,3-dipolar cycloaddition of the chloro-oxime 3 with N-
(4-chlorophenyl)-N-methyl-2-butynamide 38 afforded the desired
product.12 The butynamide 38 was readily prepared in large quan-
tity from EDC/HOAt coupling of the appropriate acid with para-
chloro-N-methylaniline. It is notable that 1,3-dipolar cycloaddi-
tions of alkynamides with chloro-oximes are rare in the litera-
assay. Standard deviation is less than 0.6 except where noted in parentheses.
Table 2
pEC50 inhibition for amide phenyl para-substitutiona
R1
O
Cl
N
N
O
Compound
R1
pEC50
14
26
27
28
29
30
31
32
33
34
35
36
37
Me
Et
Pr
iPr
tBu
Cy
CF3
OCF3
CN
NMe2
COCH3
SO2Me
CO2Me
6.9
6.0
5.2
5.9
5.1
<4.6
7.7
6.5
6.5
5.4
4.9
<5.1
5.1
TGR5 pEC50 = 5.3, 100 %
U2-OS Host pEC50 < 4.6
O
cLogP = 4.95
MW = 375
Solublity = 134 uM
AMP= 310 nm/sec
Passive Papp = 448 nm/sec
Cl
N
Cl
N
O
1
a
pEC50 values given are means of at least two experiments in the melanophore
assay. Standard deviation is less than 0.6.
Figure 1. In vitro profile of HTS hit 1.

B. W. Budzik et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1363–1367
1365
ADME studies on compound 9 suggested that the developability
of this series needed to be addressed. While compound 9 had good
physicochemical properties, there was measurable activity against
two of the common cytochrome P450 (CYP450) isoforms, including
2C19 (pIC50 = 6.5) and 3A4 (vivid green) (pIC50 = 5.9). Moreover, in
rats compound 9 showed very high in vivo clearance at levels
exceeding hepatic blood flow6 as well as high intrinsic clearance
(Clint = 48 mL/min/g) in liver microsomes. In order to address these
liabilities, variations of the isoxazole 5-methyl and N-methyl sub-
stituents were studied to find more metabolically stable analogues.
The isoxazole 5-alkyl analogues were synthesized as shown in
Schemes 3 and 4. Treatment of acid 56 with 2.2 equiv of butyl lith-
ium generated the dianion, which was then quenched with the
appropriate alkyl halide to afford 57.15 Preparation of the acid chlo-
ride followed by acylation with the desired aniline afforded the de-
sired 5-alkyl-substituted-4-isoxazole-carboxamides 58 (Scheme
3). The cyclopropyl analogue however, was prepared via an alter-
nate route (Scheme 4). The chloro-oxime 59 was reacted with ethyl
3-cyclopropyl-3-oxopropanoate in triethylamine and toluene at
room temperature to afford the desired cyclized 5-cyclopropyl-
isoxazole ethyl ester.16 Hydrolysis, acid chloride formation, and
coupling with 4-chloro-N-methylaniline as before afforded the de-
sired product 61.
R2
R2
a, b
N
Cl
O
O
H
OH
3
Cl
N
c
R2
38
Cl
O
N
N
O
2
Scheme 2. Synthesis of substituted 3-aryl-4-isoxazolecarboxamides. Reagents and
conditions: (a) hydroxylamine HCl, NaOAc, EtOH/H2O (5:1), W 150 °C, 1 min; (b)
NCS, DMF, 16 h, heat then rt; (c) toluene, W 160 °C, 10 min.
l
l
ture.13 Although our yields for this cyclization were low (typically
<10%), the simplicity of this convergent route was very attractive
for our purposes.14
The results for this set of compounds (Table 3) suggested that
ortho- and meta-substitutions were preferred over para-substitu-
tion for each substituent studied (39–41, 44–52), except for the
chloro substitutions in which the meta- and para-substitutions
were equipotent (9, 42–43). Some 2-chloro disubstitutions were
also prepared (54 and 55) but these compounds lost potency as
compared to the parent 9 and were not further pursued. Interest-
R2
R2
O
O
a
OH
OH
N
N
O
O
R
56
57
ingly, compound 48 with
a meta-methoxy substituent was
R2
the most potent compound prepared in the entire series
(pEC50 = 9.0). Compounds were also prepared in which the isoxaz-
ole aryl ring was replaced with unsubstituted pyridines as well as
thiophenes. All lost potency as compared to the unsubstituted
phenyl parent compound (data not shown).
O
b, c
R1
N
N
O
Table 3
58
pEC50 inhibition for isoxazole phenyl substitutiona
Scheme 3. Alkylation of isoxazole-5-methyl substituent. Reagents and conditions:
(a) n-BuLi (2.2 equiv), then RI, À78 °C; (b) oxalyl chloride, cat. DMF; (c) N-
R2
Cl
methylaniline R1, Et3N,
l
W 190 °C, 1 min.
O
N
N
O
Compound
R2
pEC50
a, b
O
Cl
39
40
41
9
2-F
3-F
4-F
2-Cl
3-Cl
4-Cl
2-Me
3-Me
4-Me
2-OMe
3-OMe
4-OMe
2-CF3
3-CF3
4-CF3
H
6.8 ( 0.7)
8.1
6.3
7.5
7.9
7.9
7.9
7.1 ( 0.9)
6.0 ( 0.9)
7.2 ( 0.6)
9.0
5.3
7.5
6.5 ( 0.9)
5.9
7.7 ( 0.7)
6.2
Cl
OH
N
O
Cl
N
O
OH
42
43
44
45
46
47
48
49
50
51
52
53
54
55
59
60
Cl
c, d
Cl
N
N
O
61
2,6-diCl
2-Cl,6-F
Scheme 4. Preparation of 5-cyclopropyl-N-methyl-4-isoxazole-carboxamide.
Reagents and conditions: (a) ethyl 3-cyclopropyl-3-oxopropanoate, Et3N, toluene,
6.2
a
pEC50 values given are means of at least two experiments in the melanophore
assay. Standard deviation is less than 0.6 except where noted in parentheses.
rt, 16 h; (b) KOH, EtOH,
methyl-4-Cl-aniline, Et3N,
l
l
W 100 °C, 2 min; (c) oxalyl chloride, cat. DMF; (d) N-
W 190 °C, 1 min.

1366
B. W. Budzik et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1363–1367
Table 5
In vitro rat DMPK results
The assay results for these compounds demonstrated a key find-
ing that while the N-methyl amide (R3) could not be varied without
some loss of potency, the 5-methyl substituent on the isoxazole
ring (R4) appeared to tolerate more variability (Table 4). Replace-
ment of either one of the methyl groups with hydrogen resulted
in significant loss of potency (62, 67) as compared to 9. In addition,
conformational constraint of the N-methyl group as seen in prepa-
ration of the 2,3-dihydro-1H-indole analogue (73) resulted in weak
potency (pEC50 = 5.5).
a
Compds
pEC50
Clint (mL/min/g)
9
68
61
74
7.5
8.4
8.4
7.9
46
45
10
6.8
a
Intrinsic clearance in rat liver microsomes.
O
Cl
O
O
Cl
Cl
Cl
N
N
N
N
N
N
N
O
N
O
73
74
75
pEC50 = 5.5
pEC50=7.9
pEC50=7.2
P450 pIC50 <5.7 (all forms)
P450 2C19 pIC50 =6.6
Increasing bulk at R3 corresponded with a trend in decreasing
potency (9, 63–66). However, increasing bulk at R4 corresponded
to a trend of increasing potency (67–70, 61) by as much as one
log unit (61, pEC50 = 8.4) as compared to 9. Interestingly, some in-
creased steric bulk at both positions concurrently (R3 = R4 = Et, 71)
was well tolerated, but further substitution (R3 = R4 = iPr, 72)
caused potency loss.
ADME studies on compounds 9, 68, and 61 confirmed the
hypothesis that the 5-methyl group was a source of metabolic lia-
bility in this series. The increasing substitution at R4 from methyl
(9) to ethyl (68) to cyclopropyl (61) (Table 5) resulted in reduced
in vitro clearance in rats, though 61 unfortunately still had very
high clearance. We therefore considered other potential sources
of metabolic instability in this series.
Figure 2. 1,2,3-Triazole as an isoxazole replacement. pEC50 values are means of at
least four experiments in the melanophore assay. Standard deviation is less than
0.6.
had an improved P450 profile (Fig. 2). The intrinsic clearance was
also further reduced as compared to compound 68 (Table 5).
In summary, SAR exploration of multiple regions of the HTS hit
1 led to a series of 3-aryl-4-isoxazolecarboxamides as novel, potent
agonists of the human TGR5 G-protein-coupled receptor. Potent
exemplars such as 9, 48 (pEC50 = 9.0), and 61 (pEC50 = 8.4) were
quickly identified. Triazole 74 (pEC50 = 7.9) showed the most
promising improvements in the in vitro metabolic stability profile.
It has been reported in the literature that other isoxazole tem-
plates also suffer from undesirable DMPK properties such as high
clearance and poor exposure.17 When we replaced the isoxazole
ring with benzene or furan, the resulting compounds had reduced
potency (by one log unit or greater) as compared to the parent
compound (data not shown). Triazole 74, however, was slightly
more potent than isoxazole 75, had equally good solubility, and
Acknowledgments
The authors thank Walter Johnson and Chad Quinn for their
assistance with the LC–MS spectra, Victoria Magaard and Carl Ben-
nett for their assistance with compound purification, and Minghui
Wang for additional NMR support.
Supplementary data
Table 4
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2010.01.003.
pEC50 Inhibition for alkyl variations at R3 and R4a
Cl
References and notes
O
Cl
1. IDF Diabetes Atlas, 4th ed. http://www.diabetesatlas.org/.
2. Maruyama, T.; Miyamoto, Y.; Nakamura, T.; Tamai, Y.; Okada, H.; Sugiyama, E.;
Nakamura, T.; Itadani, H.; Tanaka, K. Biochem. Biophys. Res. Commun. 2002, 298,
714.
N
R3
N
O
R4
3. Thomas, C.; Pellicciari, R.; Pruzanski, M.; Auwerx, J.; Schoonjans, K. Nat. Rev.
Drug Disc. 2008, 7, 678.
Compound
R3
R4
pEC50
4. Katsuma, S.; Hirasawa, A.; Tsujimoto, G. Biochem. Biophys. Res. Commun. 2005,
329, 386.
62
9
H
Me
Et
Me
Me
Me
Me
Me
Me
H
<5.5
7.5
7.5
7.4
6.5
5.5
5.9
8.4
7.7
8.0
8.4
8.0
6.1 ( 1)
5. Arulmozhi, D. K.; Portha, B. Eur. J. Pharm. Sci. 2006, 28, 96.
6. Evans, K. A.; Budzik, B. W.; Ross, S. A.; Wisnoski, D. D.; Jin, J.; Rivero, R. A.;
Vimal, M.; Szewczyk, G. R.; Jayawickreme, C.; Moncol, D. L.; Rimele, T. J.;
Armour, S. L.; Weaver, S. P.; Griffin, R. J.; Tadepalli, S. M.; Jeune, M. R.; Shearer,
T. W.; Chen, Z. B.; Chen, L.; Anderson, D. L.; Becherer, J. D.; De Los Frailes, M.;
Colilla, F. J. J. Med. Chem. 2009, 52, 7962.
7. The HTS used a MRE/CRE (multiple response element/cAMP response element)
directed reporter gene assay measuring luciferase production in response to
changes in cAMP via a Gs-protein coupled signaling pathway.
8. The compound had no response in the host cell line (pEC50 <4.6 in
untransfected U2-OS cells) which confirmed the specificity.
63
64
65
66
67
68
69
70
61
71
72
iPr
n-Pr
cPr
Me
Me
Me
Me
Me
Et
Et
iPr
n-Pr
cPr
Et
9. Typical reaction conditions for preparation of substituted N-alkyl anilines:
treatment of the aniline with the appropriate aldehyde or ketone (R), sodium
triacetoxyborohydride (Na(OAc)3BH), and acetic acid in dichloroethane (DCE)
for 16 h at rt afforded the desired product. See also: Thorstensson, F.;
iPr
iPr
a
pEC50 values given are means of at least two experiments in the melanophore
assay. Standard deviation is less than 0.5 except where noted in parentheses.

B. W. Budzik et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1363–1367
1367
Kvarnstrom, I.; Musil, D.; Nilsson, I.; Samuelsson, B. J. Med. Chem. 2003, 46,
1165.
14. Regiochemistry of the cycloaddition products was confirmed with 2D NMR and
NOE studies.
10. For assay protocols as well as experimental procedures and spectral data for
compounds 1, 9, and 22 see Ref. 6. Melanophore assay potencies were typically
ꢀ0.8–1 log unit more potent than the U2-OS cell data, and thus only one data
set is reported.
11. Liu, K.-C.; Shelton, B. R.; Howe, R. K. J. Org. Chem. 1980, 45, 3916.
12. Hamper, B. C.; Leschinsky, K. L.; Massey, S. S.; Bell, C. L.; Brannigan, L. H.;
Prosch, S. D. J. Agric. Food Chem. 1995, 43, 219.
15. Natale, N. R.; McKenna, J. I.; Niou, C. S.; Borth, M.; Hope, H. J. Org. Chem. 1985,
50, 5660.
16. Epple, R.; Russo, R.; Azimioara, M.; Cow, C.; Xie, Y.; Wang, X.; Wityak, J.;
Karanewsky, D.; Gerken, A.; Iskandar, M.; Saez, E.; Seidel, H. M.; Tian, S.-S.
Bioorg. Med. Chem. Lett. 2006, 16, 4376.
17. (a) Zhao, H.; Xin, Z.; Liu, G.; Schaefer, V. G.; Falls, H. D.; Kaszubska, W.; Collins,
C. A.; Sham, H. L. J. Med. Chem. 2004, 47, 6655; (b) Kanda, Y.; Kawanishi, Y.; Oda,
K.; Sakata, T.; Mihara, S.; Asakura, K.; Kanemasa, T.; Ninomiya, M.; Fujimoto,
M.; Konoike, T. Bioorg. Med. Chem. 2001, 9, 897.
13. Lee, C. K. Y.; Herlt, A. J.; Simpson, G. W.; Willis, A. C.; Easton, C. J. J. Org. Chem.
2006, 71, 3221.