Synthesis and SAR of 2-carboxylic acid indoles as inhibitors of plasminogen activator inhibitor-1

Article abstract of DOI:10.1016/j.bmcl.2005.05.095

We synthesized and evaluated a novel series of 2-carboxylic acid indole-based inhibitors of plasminogen activator inhibitor-1 (PAI-1). Systematic modification of the N-1 position and the 5-position of the indole scaffold resulted in the identification of several compounds that showed good potency against PAI-1 in the spectrophotometric assay. This potency did not always translate to the antibody assay. Solubility and serum protein binding studies on selected analogs revealed that protein binding might be a factor in the poor correlation between the two assays.

Full text of DOI:10.1016/j.bmcl.2005.05.095

Bioorganic & Medicinal Chemistry Letters 15 (2005) 3514–3518
Synthesis and SAR of 2-carboxylic acid indoles as inhibitors
of plasminogen activator inhibitor-1
Baihua Hu,a,* James W. Jetter,a Jay E. Wrobel,a Thomas M. Antrilli,b Jean S. Bauer,b
Li Di,c Sergiusz Polakowski,c Uday Jainc and David L. Crandallb
aChemical and Screening Sciences, Wyeth Research, Collegeville, PA 19426, USA
bCardiovascular and Metabolic Diseases Research, Wyeth Research, Collegeville, PA 19426, USA
cChemical and Screening Sciences, Wyeth Research, Princeton, NJ 08543, USA
Received 21 April 2005; revised 19 May 2005; accepted 25 May 2005
Abstract—We synthesized and evaluated a novel series of 2-carboxylic acid indole-based inhibitors of plasminogen activator
inhibitor-1 (PAI-1). Systematic modification of the N-1 position and the 5-position of the indole scaffold resulted in the identifica-
tion of several compounds that showed good potency against PAI-1 in the spectrophotometric assay. This potency did not always
translate to the antibody assay. Solubility and serum protein binding studies on selected analogs revealed that protein binding might
be a factor in the poor correlation between the two assays.
Ó 2005 Elsevier Ltd. All rights reserved.
Plasminogen activator inhibitor-1 (PAI-1) is a member
of the serine protease inhibitor (serpin) gene family
and is the principal inhibitor of tissue plasminogen acti-
vator (tPA) and urokinase plasminogen activator (uPA)
in vivo.1 These serine proteinases convert plasminogen,
an inactive zymogen, to the active enzyme plasmin,
which digests fibrin clots by degrading insoluble fibrin
molecules to small soluble fragments.2 In the acute set-
ting, PAI-1 stored in platelet a-granules can be released
upon platelet activation, resulting in significant local
concentrations and the resistance of platelet-rich thro-
mbi to lysis. In healthy individuals, PAI-1 expression
is low, but it is elevated significantly in a number of dis-
eases, including deep vein thrombosis,3 atherosclerosis,4
and type 2 diabetes.5 Plasma PAI-1 is also elevated in
postmenopausal women and has been proposed to con-
tribute to the increased incidence of cardiovascular dis-
ease in this population.6
more, these null mice are viable and protected from the
development of atherosclerosis, due in part to its role
in regulating tissue proteolysis. Several patients with
PAI-1 deficiency have been reported and observed to
have normal coagulation.8 Most abnormal bleedings
were found only after trauma or surgery. No other
physical anomalies were reported in individuals with
complete PAI-1 deficiency. These results suggest that
modulation of PAI-1 activity offers a beneficial therapy
in treating a variety of cancers and cardiovascular dis-
eases originating from tissue remodeling and fibrinoly-
tic disorders.
The importance of PAI-1, as a major regulatory pro-
tein in a variety of biological processes, has been sup-
ported by
a number of studies in experimental
animals.7 PAI-1 null mice have normal coagulation,
but unstable thrombi that lyse spontaneously. Further-
Keyword: PAI-1 inhibitors.
*
Corresponding author. Tel.: +1 484 865 7867; fax: +1 484 865
9399; e-mail: hub@wyeth.com
Figure 1. Structures of known PAI-1 small molecule inhibitors.
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.05.095

B. Hu et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3514–3518
3515
To date, several PAI-1 inhibitors including antibodies,
peptides, and small molecules have been reported.1b,9
Some of the recent small molecule inhibitors (Fig. 1)
are flufenamic acid-derived AR-H029953XX (1),10 the
diketopiperazine analogues,11 benzothiophene deriva-
tives,12 menthol,13 and piperazine analogues.14 In addi-
tion, Wyeth has previously reported a number of
potent PAI-1 inhibitors,15 such as the naphthyl benzofu-
ran 2,15a the 3-indole oxoacetic acid 3 (Tiplaxtinin),15b
and oxadiazolidinediones 4.15c The key structural fea-
tures present in all inhibitors reported so far include
either a carboxylic acid or an acid bioisostere attached
to a lipophilic aromatic ring scaffold. As part of our
ongoing research to find additional novel, orally active
PAI-1 inhibitors, we designed a group of 2-indole car-
boxylic acid-based compounds. In our design, the oxo-
acetic acid of Tiplaxtinin was replaced by a carboxylic
acid and the position of the carboxylic acid was switched
from 3-indole to 2-indole. The relocation of the acid
moiety allowed easier access to indole substituent varia-
tion, which could provide an avenue for the exploration
of additional chemical space.
Scheme 2. Reagents: (a) hexane-2,5-dione, THF, reflux; (b) LiOH;
(c) 2,5-dimethoxytetrahydrofuran, THF, reflux; (d) R3B(OH)2/2,6-
lutidine/Cu(OAc)2.
reaction of aniline 9 with a large excess of 2,5-dim-
ethoxytetrahydrofuran, followed by a basic hydrolysis,
gave pyrrole 13. The aniline 9 can also be converted to
a diaryl amine with a boronic acid in the presence of
Cu(OAc)217 and the final acid 12 was obtained via basic
hydrolysis.
2-Carboxylic acid indoles were conveniently prepared by
following synthetic Schemes 1 and 2. 5-Nitro indole 516
was first reduced to amine 6 upon treatment with
RaneyÒ nickel in a mixture of hydrazine and ethanol
(Scheme 1). Sulfonylation of the aniline with sulfonyl
chlorides, followed by basic hydrolysis of ethyl esters
in LiOH, gave the desired carboxylic acids 7. The 1H-in-
dole 8 was alkylated with either iodomethane or benzyl
bromides in the presence of Cs2CO3. The resulting nitro
intermediate was then converted to acid 10 upon
RaneyÒ nickel reduction, sulfonylation/acylation, and
LiOH hydrolysis.
The inhibition18 of PAI-1 by 2-carboxylic acid indoles is
presented in Tables 1–3, together with comparative data
for AR-H029953XX (1) and Tiplaxtinin (3).15b As an
initial in vitro screen, preincubation of recombinant
human PAI-1 with tPA resulted in complete inhibition
of the enzymatic action of tPA with a chromogenic sub-
strate, as determined spectrophotometrically. Addition
of 2-carboxylic acid indoles to human PAI-1 restored
the proteolytic activity of tPA, indicating successful
inhibition of PAI-1. For the most potent inhibitors,
the IC50 values were generated using a secondary assay
that quantified residual active PAI-1 by antibody bind-
ing following incubation with various concentrations
of the compound.
Condensation of aniline 9 with excess hexane-2,5-dione
utilizing a Dean-Stark trap for azeotropic water removal
led to a 2,5-dimethylpyrrolylindole, which was convert-
ed to 11 upon saponification (Scheme 2). Similarly,
We began our SAR studies with the modification of
N-benzyl-C3-H indoles (Table 1). In general, lipophilic
sulfonamides were preferred. Thus, the presence of a
bisphenyl (16) or 3,4-dichlorophenyl (17) increased
PAI-1 % inhibition at an initial screening concentration
(100 lM) over n-butyl- (14) or phenyl (15)-substituted
compounds. The bisphenyl sulfonamide 16 also showed
49% inhibition at 25 lM; however, it was found to be a
weak PAI-1 inhibitor, with an IC50 value of 63.1 lM in
the secondary antibody assay.
Next, we examined various N-benzyl-C3-phenyl-substi-
tuted indoles (Table 2). The sulfonyl chlorides chosen
for use in the synthesis of sulfonamides 18–26 were
selected based on an earlier SAR study using a different
scaffold.19 Since the C3–phenyl sulfonamides are much
more potent than the C3–H sulfonamides, only the inhi-
bitions at lower concentrations (10 and 25 lM) are
included in Table 2. Looking across the subseries
(R1 = Bn and Me), the less lipophilic 8-quinoline substi-
tuent at the R2 position (19 and 24) was the least potent,
and highly lipophilic R2 groups [e.g., 4-tBu phenyl (20
and 25), 4-phenyl-phenyl (21 and 26)] were the most
Scheme 1. Reagents: (a) RaneyÒ nickel/EtOH; (b) R1SO2Cl/NEt3;
(c) LiOH; (d) R2Br/Cs2CO3; (e) R1COCl or R1-NCO.

3516
B. Hu et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3514–3518
Table 1. PAI-1 inhibitory activity of C3–H indoles
tration of 10 lM. Several N-benzyl-3-phenyl indoles
exhibited high inhibition (>25%) at a lower concentra-
tion (10 lM) in the initial spectrophotometric assay.
For example, the 4-phenyl benzyl sulfonamide 22
showed 79% inhibition. However, this high percent inhi-
bition did not get transferred to our antibody assay
(IC50 = 28 lM for 22, vide infra). The most potent sul-
fonamide in the antibody assay was found to be the
4-CF3O-phenyl sulfonamide 18 with an IC50 value of
8.3 lM, which is 2-fold more potent than the literature
standard AR-H029953XX (1).
Compound R1
% Inhibitiona % Inhibitiona IC50
at 100 lM
at 25 lM
(lM)b
1
3
74
83
14
32
64
51
47
17.8
2.7
14
15
16
17
nBu
Ph
To examine the effect of sulfonamide pharmacophore on
the inhibition of PAI-1, the sulfonamide functionality
was replaced by a number of other groups, as shown
in Table 3. Compared to the sulfonamides (18–21), an
amide derivative (27) and a urea analog (28) showed
similar potency. To explore a non-amide lipophilic
piece, a series of biphenyl anilines (29–32) was synthe-
sized and all anilines showed similar potency as well.
For example, an N-(4-phenyl)benzyl aniline analog
4-Ph-Ph
49
63.1
3,4-diCl-Ph 62
a Initial spectrophotometric assay.
b Antibody assay.
potent. Comparing N-methyl with N-benzyl pairs (18
and 23, 19 and 24, 20 and 25, and 21 and 26), the N-ben-
zyl analogs always showed greater potency at a concen-
Table 2. PAI-1 inhibitory activity of C3–Ph indoles
Compound
R1
R2
% Inhibitiona at 25 lM
% Inhibitiona at 10 lM
IC50 (lM)b
1
3
3
0
18
19
20
21
22
23
24
25
26
Bn
Bn
4-CF3O-Ph
8-Quinoline
(4-tBu)-Ph
4-Ph-Ph
81
43
82
61
87
19
1
33
8
8.3
Bn
Bn
41
44
79
0
28.7
23.0
28.0
4-Ph-Bn
Me
Me
Me
Me
Ph
4-CF3O-Ph
8-Quinoline
(4-tBu)-Ph
4-Ph-Ph
0
46
68
5
11
a Initial spectrophotometric assay.
b Antibody assay.
Table 3. PAI-1 inhibitory activity of non-sulfonamide-based indoles
Compound
R1
R2
% Inhibitiona at 25 lM
% Inhibitiona at 10 lM
IC50 (lM)b
27
28
29
30
31
32
33
34
Bn
Bn
(2,4-diF)Ph-CONH
(2,4-diF)Ph-NHCONH
3-MePh-NH
43
61
69
73
71
80
69
83
24
27
24
45
36
52
12
23
Bn
Bn
(4-tBu)-Ph-NH
Bn
4-Ph-Ph-NH
4-Ph-Bn
Bn
Bn
(4-tBu)-Ph-NH
1-Pyrrole
(2,5-diMe)-1-Pyrrole
12.0
a Initial spectrophotometric assay.
b Antibody assay.

B. Hu et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3514–3518
3517
(32) showed 52% inhibition at 10 lM and had an IC50
Acknowledgments
value of 12.0 lM in the antibody assay. Similarly, 5-pyr-
role derivatives 33 and 34 showed essentially the same
potency in the initial in vitro screen. This preliminary
SAR information suggested that the 5-position of the in-
dole ring was flexible and could accommodate a variety
of substitutions, although no additional increase in the
PAI-1 inhibitory activity was achieved compared to 18.
The authors thank Drs. Thomas J. Commons and
Eugene J. Trybulski for leading this project in chemistry.
We also thank the Discovery Analytical Chemistry
group at Wyeth Research, Collegeville, PA, for spectral
data.
We next focused on a preliminary investigation of the
potency correlation discrepancy between the initial spec-
trophotometric assay and the antibody assay. Since
these compounds are highly lipophilic, we initially sus-
pected that the lower than expected potency of newly
synthesized PAI-1 compounds (18, 20, 22, and 32) in
the antibody assay compared to 3 may be due to solubil-
ity. The concentrations of the five compounds (3, 18, 20,
22, and 32) were determined in the two assay buffers at a
target concentration of 50 lM by LC-MS.20 All the
compounds were soluble (>25 lM) in the antibody assay
buffer and in the initial spectrophotometric assay buffer.
Thus, solubility was probably not a factor for the dis-
crepancy between the two assays.
References and notes
1. (a) Lawrence, D. A. Nat. Struct. Biol. 1997, 4, 339; (b) Wu,
Q.; Zhao, Z. Curr. Drug Targets Cardiovasc. Haematol.
Disord. 2002, 2, 27; (c) Juhan-Vague, I.; Alessi, M-C.;
Mavri, A.; Morange, P. E. J. Thromb. Haemost. 2003, 1,
1575.
2. Vaughan, D. E. J. Invest. Med. 1998, 46, 370.
3. Juhan-Vague, I.; Valadier, J.; Alessi, M.; Aillaud, M.;
Ansaldi, J.; Philip-Joet, C.; Holvoet, P.; Serradimigni, A.;
Collen, D. Thromb. Haemost. 1987, 57, 67.
4. Schneiderman, J.; Sawdey, M.; Keeton, M.; Bordin, G.;
Bernstein, E.; Dilley, R.; Loskutoff, D. Proc. Natl. Acad.
Sci. U.S.A. 1992, 89, 6998.
5. Juhan-Vague, I.; Alessi, M. C. Thromb. Haemost. 1997,
78, 565.
6. (a) Schneiderman, J.; Sawdey, M.; Keeton, M.; Bordin,
G.; Bernstein, E.; Dilley, R.; Loskutoff, D. Proc. Natl.
Acad. Sci. U.S.A. 1992, 89, 6998; (b) Stefansson, S.;
McMahon, G.; Petitclerc, E.; Lawrence, D. A. Curr.
Pharm. Des. 2003, 9, 1545.
7. Carmeliet, P.; Stassen, J.; Schoonjans, L.; Ream, B.; van
den Oord, J.; De Mo, M.; Mulligan, R.; Collen, D. J. Clin.
Invest. 1993, 92, 2756.
8. Fay, W. P.; Parker, A.; Condrey, L.; Shapiro, A. Blood
1997, 90, 204.
9. Eitzman, D.; Fay, W. P.; Lawrence, D. A.; Francis-
Chmura, A.; Shore, J. D.; Olson, S. T.; Ginsburg, D. J.
Clin. Invest. 1995, 95, 2416.
10. Bjoerquist, P.; Ehnebom, J.; Inghardt, T.; Hansson, L.;
Lindberg, M.; Linschoten, M.; Stroemqvist, M.; Deinum,
J. Biochemistry 1998, 37, 1227.
11. Wang, S.; Golec, J.; Miller, W.; Milutinovic, F. A.;
Williams, S.; Brooks, T.; Hardman, K.; Charlton, P.;
Wren, S.; Spencer, J. Bioorg. Med. Chem. Lett. 2002, 2367.
12. De Nanteuil, G.; Lila-Ambroise, C.; Rupin, A.; Vallez,
M.; Verbeuren, T. J. J. Bioorg. Med. Chem. Lett. 2003, 13,
1705.
13. Ye, B.; Bauer, S.; Buckman, B. O.; Ghannam, A.; Griedel,
B. D.; Khim, S.; Lee, W.; Sacchi, K. L.; Shaw, K. J.;
Liang, A.; Wu, Q.; Zhao, Z. Bioorg. Med. Chem. Lett.
2003, 13, 3361.
Since the test compounds were run in the presence of
10 mg/L bovine serum albumin (BSA) in the antibody
assay, and serum proteins were not present in the spec-
trophotometer assay, the antibody assay faced an addi-
tional hurdle of competition of the test compounds
between BSA and PAI-1. Binding characteristics of the
five compounds (3, 18, 20, 22, and 32) with BSA were
then evaluated by equilibrium dialysis21 and Biacore as-
says.22 It was found that all five compounds were highly
bound to BSA with an unbound fraction below the
detection limit (<0.05%) in the equilibrium dialysis
experiment. Biacore analysis showed a 1:1 stoichio-
metric binding and comparable binding constants
(Kd ꢀ 10 lM) for 3 and 18, while other three com-
pounds were rather non-specific (stoichiometry: 2.5–3.7
at 10 lM). Thus, the presence of BSA decreased the con-
centration of free drugs for 20, 22, and 32 in the anti-
body assay much more than it did for compounds 3
and 18. It may thus be concluded that the magnitude
of drug–BSA protein interaction influenced the biologi-
cal activity and this may partially account for the poten-
cy correlation discrepancy between the two assays.
In summary, we have explored the structure–activity
relationships for the novel 2-carboxylic acid indole-
based PAI-1 inhibitors. Introduction of a lipophilic
phenyl group at the C3 position of the indole ring signif-
icantly increased PAI-1 potency. Solubility study indi-
cated that the selected compounds were soluble in the
initial assay buffer and in the antibody assay buffer.
The interaction of BSA with the selected compounds
was investigated by equilibrium dialysis and Biacore as-
says, and the initial results have suggested that the drug–
protein interaction might account for the biological
activity correlation discrepancy between the initial
screening assay and the antibody assay. This study has
shown that comparative selective and nonselective pro-
tein binding analysis can be used in further understand-
ing of the SAR of small molecule PAI-1 inhibitors.
14. Ye, B.; Chou, Y.; Karanjawala, R.; Lee, W.; Lu, S.; Shaw,
K. J.; Jones, S.; Lentz, D.; Liang, A.; Tseng, J.; Wu, Q.;
Zhao, Z. Bioorg. Med. Chem. Lett. 2004, 14, 761.
15. (a) Crandall, D. L.; Elokdah, H.; Di, L.; Hennan, J. K.;
Gorlatova, N. V.; Lawrence, D. A. J. Thromb. Haemost.
2004, 2, 1422; (b) Elokdah, H.; Abou-Gharbia, M.;
Herman, J. K.; McFarlane, G.; Mugford, C. P.; Krishna-
murthy, G.; Crandall, D. L. J. Med. Chem. 2004, 47, 3491;
(c) Gopalsamy, A.; Kincaid, S. L.; Ellingboe, J. W.;
Groeling, T.; Antrilli, T. M.; Krishnamurthy, G.; Aulab-
augh, A.; Friedrichs, G.; Crandall, D. L. Bioorg. Med.
Chem. Lett. 2004, 14, 3477; (d) Elokdah, H.; McFarlane,
G. R.; Krishnamurthy, G.; Crandall, D. L. 227th ACS
National Meeting, Anaheim, CA, United States, March
28–April 1, 2004; (e) Commons, T. J.; Croce, S.; Wood-
worth, R. P.; Trybulski, E. J.; Elokdah, H.; Crandall, D.
L.; Hennan, J.; Krishnamurthy, G.; Mugford, C. 227th

3518
B. Hu et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3514–3518
ACS National Meeting, Anaheim, CA, United States,
DMSO concentration in all dialysis units was 5%.
Buffer samples spiked with the test compound were used
to verify the ability of the compounds to permeate
across the dialysis membrane. Samples were dialyzed for
18 h in a 37 °C water bath. Following dialysis, samples
were collected from all dialysis compartments and
combined with an equal volume of acetonitrile. After
12 min of centrifugation at 12,000 rpm, the supernatant
of each sample was collected and analyzed by HPLC.
Based on HPLC data, percentage drug free was
calculated for the compound.
March 28–April 1, 2004; (f) Havran, L. M.; Butera, J. A.;
Jenkins, D.; Elokdah, H.; Krishnamurthy, G.; Crandall,
D. L. 227th ACS National Meeting, Anaheim, CA, United
States, March 28–April 1, 2004.
16. Murakami, Y.; Watanabe, T.; Kobayashi, A.; Yokoyama,
Y. Synthesis 1984, 738.
17. Antilla, J. C.; Buchwald, S. L. Org. Lett. 2001, 3, 2077.
18. The primary screen and the antibody assay were described
in detail in Ref. 15b.
19. Hu, B. et al, unpublished results.
20. Initial spectrophotometric assay buffer (pH 6.6): Na2HPO4
50 mM, NaCl 200 mM, and EDTA 1 mM. Antibody assay
buffer (pH 7.5): Tris 50 mM, NaCl 150 mM, BSA 10 mg/L,
and 0.01% Tween 80.
21. Equilibrium dialysis: stock solutions at concentrations
of 400 lM of compounds were prepared in 100%
DMSO, 4% BSA in phosphate buffer (67 mM, pH 7.4)
was spiked with an appropriate compound stock solu-
tion to obtain 20 lM final drug concentration. Four
percent (w/v) dextran 70,000 was used in dialysis buffer
to maintain similar osmolarity in reciprocal dialysis
chambers minimizing the volume shift phenomena.
22. Biacore assay: BSA was immobilized on the flow cell of a
CM-5 sensor chip. Stock solutions (0.8 mM) of com-
pounds were prepared in DMSO. Test solutions of
compounds were prepared from the stock solutions and
diluted to their final concentration in phosphate buffer
(67 mM, pH 7.4). Samples of all test compounds in the
concentration range of 0–10 lM were injected over the
biosensor chip surface. Dilutions for all prepared samples
were made in such a way that the final DMSO concen-
tration was 5% in samples and running buffer. Matrix
binding effect from blank runs was subtracted from dose–
response curve data points.