Design and synthesis of oxadiazolidinediones as inhibitors of plasminogen activator inhibitor-1

Article abstract of DOI:10.1016/j.bmcl.2004.04.058

A novel series of PAI-1 inhibitors containing an oxadiazolidinedione moiety were identified by high through-put screening. Optimization of substituents by parallel synthesis and the iterative design toward understanding structure-activity relationship to improve potency are described.

Full text of DOI:10.1016/j.bmcl.2004.04.058

Bioorganic & Medicinal Chemistry Letters 14 (2004) 3477–3480
Design and synthesis of oxadiazolidinediones as inhibitors
of plasminogen activator inhibitor-1
Ariamala Gopalsamy,a,* Scott L. Kincaid,a John W. Ellingboe,a Thomas M. Groeling,b
Thomas M. Antrilli,b Girija Krishnamurthy,a Ann Aulabaugh,a Gregory S. Friedrichsb
and David L. Crandallb
aChemical and Screening Sciences, Wyeth Research, 401 N Middletown Road, Pearl River, NY 10965, USA
bCardiovascular and Metabolic Diseases Research, Wyeth Research, Collegeville, PA 19426, USA
Received 23 March 2004; revised 15 April 2004; accepted 19 April 2004
Available online
Abstract—A novel series of PAI-1 inhibitors containing an oxadiazolidinedione moiety were identified by high through-put
screening. Optimization of substituents by parallel synthesis and the iterative design toward understanding structure–activity
relationship to improve potency are described.
Ó 2004 Elsevier Ltd. All rights reserved.
The serine protease inhibitor plasminogen activator
inhibitor-1 (PAI-1)1 is one of the primary inhibitors of
the fibrinolytic system. The fibrinolytic system includes
the proenzyme plasminogen, which is converted to the
active enzyme, plasmin, by one of the tissue type plas-
minogen activators, urokinase plasminogen activator
(uPA), and tissue plasminogen activator (tPA).2 PAI-1 is
the principal physiological inhibitor of uPA and tPA.
One of plasmin’s main responsibilities in the fibrinolytic
system is to digest fibrin at the site of vascular injury.
PAI-1 is acutely elevated in thrombotic events such as
deep vein thrombosis following post-operative recovery
from orthopedic surgery,3 but is also associated with
angiogenesis and vascular remodeling in chronic dis-
eases such as atherosclerosis and cancer.4 PAI-1 is syn-
thesized by a number of different tissues, including
hepatic, vascular, and adipose, and is stored in platelets
to provide high local concentrations for clot stabiliza-
tion.5
mice have normal coagulation, but acceleration of clot
lysis.6 PAI-1 null humans have been identified and ob-
served to have normal coagulation and abnormal
bleeding only following trauma or surgery.7 Thus,
inhibition of PAI-1 would represent a useful strategy in
treating a variety of cardiovascular diseases. A number
of small molecules have been reported8 to inhibit PAI-1
apart from antibodies9 and peptides.10 Some of the re-
cent small molecule inhibitors are diketopiperazine
analogues11 1, benzothiophene derivatives12 2, menthol
based inhibitors13 3, and piperazine analogues14
4
(Fig. 1).
Because PAI-1 regulates plasminogen activation, it
functions to regulate fibrinolysis not coagulation. Pep-
tide inhibitors of PAI-1 have been shown to lyse clots in
vitro without affecting clot formation and PAI-1 null
Keywords: PAI-1 inhibitor; Oxadiazolidine diones; Parallel synthesis.
* Corresponding author. Tel.: +1-845-602-2841; fax: +1-845-602-3045;
e-mail: gopalsa@wyeth.com
Figure 1. PAI-1 small molecule inhibitors.
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.04.058

3478
A. Gopalsamy et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3477–3480
Figure 2.
Scheme 2. Reagents and conditions: (a) DIBAL, THF, 0–25 °C, 2 h;
(b) (1) (Boc)NHO(Boc), Ph3P, DEAD, THF, rt, 85%; (2) TFA, DCM,
rt, 82%; (c) ClCONCO, THF, rt, 95%.
Our efforts toward identifying a PAI-1 inhibitor started
with the high throughput screening of various com-
pound libraries. The effort culminated in the identifica-
tion of oxadiazolidinedione 5 (Fig. 2).
In an in vitro tPA activity assay,15 5 inhibited the PAI-1
inhibition of tPA with an IC50 of 5.29 lM, restoring
enzymatic activation by tPA of its synthetic substrate.
Using fluorescence spectroscopy,16 the direct binding of
5 to PAI-1 indicated the inhibitor bound PAI-1 and the
binding is consistent with the observed inhibition.
Oxadiazolidine 5 and analogues were prepared by fol-
lowing synthetic Schemes 1–3.17 Benzylation of the
phenol 6 with the bromide 7 gave the bromo compound
8. Lithium exchange followed by reaction with acetal-
dehyde gave the benzyl alcohol 9, which was oxidized to
ketone 10 and subjected to olefination. A mixture of cis
and trans isomers 11 and 12 were obtained in a 55:45
ratio (Scheme 1). The isomers were readily separated by
column chromatography and individual isomers char-
Scheme 3. Reagents and conditions: (a) K2CO3, acetone, R1X, reflux,
6 h; (b) NH2OH, EtOH/Py, 60 °C, 2 h; (c) BH3/Py, MeOH, 10% HCl,
rt; (d) ClCONCO, THF, rt.
1
acterized by H NMR were carried forward toward the
derivative 14, which upon treatment with chlorocar-
bonyl isocyanate gave the desired oxadiazolidine dione
derivative 5.
synthesis of oxadiazolidinediones. As shown in Scheme
2 the ester was reduced to alcohol 13 using DIBAL.
Mitsunobu reaction with bis Boc protected hydroxyl
amine followed by deprotection gave the hydroxylamine
Analogues 19 with substituents on the carbon adjacent
to the oxadiazolidinedione ring were prepared as shown
in Scheme 3. Alkylation of the a,b-unsaturated aldehyde
or ketone 15 followed by condensation with hydroxyl-
amine gave the oxime 17. Borane reduction of the oxime
and treatment of the resulting hydroxylamine 18 with
chlorocarbonyl isocyanate gave the final targets. The
synthetic sequence was found to be general and readily
amenable for parallel synthesis.18
Our effort toward optimization of the lead 5 was focused
on regions 1 and 2 (Fig. 2) while retaining the oxa-
diazolidinedione moiety intact. Since the HTS lead was
the trans isomer it was of interest to see if the cis isomer
20 possessed any activity. The cis analogue synthesized
was found to be equipotent in inhibiting the PAI-1
inhibition of tPA indicating that the stereochemistry of
the double bond did not play a role in binding. Region 1
of the lead molecule has a very lipophilic phenyl ether
tail. Switching the position of the oxygen atom and
converting the lead to benzyl ether 21 retained the
potency.
To explore the steric and lipophilic requirements for this
region several analogues were made by parallel synthe-
sis. The general trend observed is exemplified by ana-
logues 21 through 29 in Table 1. Removal of the CF3
groups to the unsubstituted benzyl ether 22 or substi-
tuting with a simple 3-methylbenzyl ether 23 lead to
Scheme 1. Reagents and conditions: (a) K2CO3, acetone, reflux, 6 h,
74%; (b) (1) n-BuLi, À78 °C, 1 h; (2) acetaldehyde, 60 °C, 2 h, 63%; (c)
Jones reagent, rt, 69%; (d) (CF3CH2O)2P(O)CH2COOMe, NaH, 0–
25 °C, 12 h, 61%.

A. Gopalsamy et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3477–3480
Table 1. PAI-1 inhibitory activity of oxadiazolidinediones
3479
Ex
R1
R2
R3
R4
Db/Sba stereo
IC50 (lM % inhibb)
5
3[(3,5-BisCF3Ph)OCH2]–
3[(3,5-BisCF3Ph)OCH2]–
3[(3,5-BisCF3Ph)CH2O]–
3(PhCH2O)–
H
Me
Me
Me
Me
Me
Me
Me
H
H
trans Db
cis Db
5.29
7.2
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
H
H
H
H
trans Db
trans Db
trans Db
trans Db
trans Db
trans Db
trans Db
trans Db
trans Db
trans Db
trans Db
trans Db
Sb
9.4
H
H
24%
37%
24.0
2.3
3[(3-MePh)CH2O]–
H
H
3[(2-Pyridyl)CH2O]–
3[(2-Biphenyl)CH2O]–
3,5[Bis(4-CF3Ph)CH2O]–
3,5[Bis(4-ClPh)CH2O]–
4[(3,4-DiClPh)CH2O]–
5[(3,4-DiClPh)CH2O]–
4[(3,5-BisCF3Ph)CH2O]–
4[(3,5-BisCF3Ph)CH2O]–
3[(3,5-DiClPh)O]–
H
H
H
H
H
H
2.5
2.1
H
H
H
3-OMe
2-Cl
H
H
H
3.22
3.8
H
H
H
H
14.9
9.0
H
Et
Bu
Bu
H
H
H
H
1.9
2.2
3[(3-MePh)O]–
H
H
4[(3-CF3Ph)CH2O]–
4[(3,4-DiClPh)CH2O]–
4[(3,5-BisCF3Ph)CH2O]–
4[(3-ClPh)CH2O]–
H
Me
Me
Ph
Ph
trans Db
trans Db
Sb
5.8
6.1
H
H
H
H
0.39
0.68
H
H
Sb
a Db: double bond; Sb: single bond.
b Inhibition @ 25 lM.
significant loss of potency. Similarly introducing a less
lipophilic heteroaromatic tail piece like a 2-pyridylm-
ethyl ether 24 showed reduced potency as well. To ex-
inhibition. Thus the SAR for region 2 seems to show
flexibility with respect to the linker, however a larger
lipophilic group is preferred at R4 or R3 irrespective of
the saturated or unsaturated nature of the linker.
plore
a nonhalogenated lipophilic tail piece, 2-
biphenylmethyl ether 25 was synthesized and was found
to show slight improvement. Similarly, bisbenzyloxy
derivatives 26 and 27 showed slight improvement in
potency although these analogues might occupy larger
volume in the binding site. Adding yet another sub-
stituent R2 on the central phenyl ring was not detri-
mental, but it did not afford any additional boost in
activity as shown by examples 28 and 29. From this
preliminary SAR information it was clear that a highly
lipophilic tail piece is preferred in region 1 and the
region was flexible to accommodate bulky substituents.
In summary, we have explored the structure–activity
relationships around the novel oxadiazolidinedione
based PAI-1 inhibitor 5 by modifying regions 1 and 2.
Initial optimization of the region 1 indicates the pref-
erence for lipophilic groups in that region and seems to
be flexible in terms of steric requirements accommo-
dating bulky substituents. Preliminary studies in region
2 indicated that the linker need not be a rigid double
bond and is flexible with respect to the stereochemistry
as well. As far as the substituents off of the linker, a
lipophilic group like phenyl is favored over smaller alkyl
substituent like methyl, resulting in submicromolar
inhibitors.
In region 2, it was of importance to determine the
contribution of the methyl substituent on the double
bond toward activity. A twofold loss of potency was
observed in the case of analogue 30, which is devoid of
the methyl group. However, increasing the size of the
methyl group to ethyl 31 or n-butyl 32 led to a slight
improvement in potency. In the case of butyl analogues
reduction of the double bond to give the saturated
analogue 33 retained the potency once again supporting
the observation that the binding of this class of com-
pounds is not affected by the rigidity provided by the
double bond or its stereochemistry.
Acknowledgements
The authors would like to thank Mr. Hassan Elokdah
for leading this project and Ms. Choi-Han Tsang for her
synthetic contributions. We also thank Discovery Ana-
lytical Chemistry group at Wyeth Research, Pearl River,
NY for spectral data.
Moving the methyl group position from the double
bond to the methylene adjacent to oxadiazolidinedione
did not affect the potency as shown by examples 34 and
35. However, an increase in activity was observed when
the R4 substituent was a phenyl group as shown in the
saturated analogues 36 and 3719 with submicromolar
References and notes
1. Lawrence, D. A. Nat. Struct. Biol. 1997, 4, 339–341.
2. Vaughan, D. E. J. Invest. Med. 1998, 46, 370–376.

3480
A. Gopalsamy et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3477–3480
3. Siemens, H.; Brueckner, S.; Hagelberg, S.; Wagner, T.;
15. Determination of PAI-1 inhibition of tPA activity: Test
compounds were dissolved in DMSO at a final concen-
tration of 10 mM, then diluted 100 Â in physiologic buffer.
The inhibitory assay was initiated by the addition of test
compound (1–100 lM final concentration, maximum
DMSO concentration of 0.2%) in a pH 6.6 buffer contain-
ing 140 nM recombinant human plasminogen activator
inhibitor-1 (PAI-1; Molecular Innovations, Royal Oak,
MI). Following a 1 h incubation at room temperature,
70 nM of recombinant human tissue plasminogen activa-
tor (tPA) was added, and the combination of test
compound, PAI-1 and tPA was incubated for an addi-
tional 30 min. Following the second incubation, Spectro-
Schmucker, P. J. Clin. Anesth. 1999, 11, 622–629.
4. (a) Schneiderman, J.; Sawdey, M.; Keeton, M.; Bordin,
G.; Bernstein, E.; Dilley, R.; Loskutoff, D. Proc. Natl.
Acad. Sci. U.S.A. 1992, 89, 6998–7002; (b) Stefansson, S.;
McMahon, G.; Petitclerc, E.; Lawrence, D. A. Curr.
Parm. Des. 2003, 9, 1545–1564.
5. (a) Crandall, D. L.; Busler, D. E.; McHenry-Rinde, B.;
Groeling, T. M.; Kral, J. G. J. Clin. Endocrinol. Metab.
2000, 85, 2609–2614; (b) Booth, N.; Simpson, A.; Croll,
A.; Bennett, B.; MacGregor, I. R. Br. J. Heamatol. 1988,
70, 327–333.
6. Carmeliet, P.; Stassen, J.; Schoonjans, L.; Ream, B.; van
den Oord, J.; De Mo, M.; Mulligan, R.; Collen, D. J. Clin.
Invest. 1993, 92, 2756–2760.
7. Fay, W. P.; Parker, A.; Condrey, L.; Shapiro, A. Blood
1997, 90, 204–208.
8. (a) Wu, Q.; Zhao, Z. Curr. Drug Targets-Cardiovas.
Haemat. Dis. 2002, 2, 27–42; (b) Crandall, D. L.; Hennan,
K. K.; Elokdah, H.; Krishnamurthy, G.; Antrilli, T. M.;
Bauer, J. S.; Morgan, G. A.; Swillo, R. E. Biochem.
Biophys. Res. Commun. 2003, 311, 904–908.
9. Levi, M.; Biemond, B. J.; Van Zonneveld, A. J.; Ten Cate,
J. W.; Pannekoek, H. Circulation 1995, 91, 1175.
10. Eitzman, D.; Fay, W. P.; Lawrence, D. A.; Francis-
Chmura, A.; Shore, J. D.; Olson, S. T.; Ginsburg, D. J.
Clin. Invest. 1995, 95, 2416–2420.
11. (a) Bryans, J.; Charlton, P.; Chicarelli-Robinson, I.;
Collins, M.; Faint, R.; Latham, C.; Shaw, I.; Trew, S. J.
Antibiot. 1996, 49, 1014; (b) Folkes, A.; Brown, S. D.;
Canne, L. E.; Chan, J.; Englehardt, E.; Epsheteyn, S.;
Faint, R.; Golec, J.; Hanel, A.; Kearney, P.; Leahy, J. W.;
Mac, M.; Matthews, D.; Prisbylla, M. P.; Sanderson, J.;
Simon, R. J.; Tesfai, Z.; Vicker, N.; Wang, S.; Webb, R.
R.; Charlton, P. Bioorg. Med. Chem. Lett. 2002, 12, 1063–
1066; (c) Wang, S.; Golec, J.; Miller, W.; Milutinovic;
Folkes, A.; Williams, S.; Brooks, T.; Hardman, K.;
Charlton, P.; Wren, S.; Spencer, J. Bioorg. Med. Chem.
Lett. 2002, 12, 2367–2370.
zyme-tPA (American Diagnostica, Greenwich, CT),
a
chromogenic substrate for tPA, was added and absor-
bance read at 405 nm at 0 and 60 min. Relative PAI-1
inhibitory activity was equal to the residual tPA activity in
the test compound/PAI-1 treatment. Control treatments
include the complete inhibition of tPA by PAI-1 at the
molar ratio employed (2:1), and the absence of any effect
of the test compound on tPA alone.
16. Direct binding to PAI-1: Fluorescence spectroscopy was
used to determine the binding of 5 to recombinant human
PAI-1. The changes in the quantum yield of the protein
were used to determine the binding affinity of 5 to PAI-1
using a Jobin Yvon fluorometer. A fixed concentration of
the protein was serially titrated with increasing concen-
trations of 5. The intrinsic fluorescence of the protein was
measured at 335 nm using an excitation wavelength of
295 nm.
17. Malamas, M. S.; Sredy, J.; Gunawan, I.; Mihan, B.;
Sawicki, D. R.; Seestaller, L.; Sullivan, D.; Flam, B. R. J.
Med. Chem. 2000, 43, 995–1010.
18. Compounds were purified by HPLC and the purity was
>90%. LC conditions: HP 1100, 23 °C, 10 lL injected;
column: YMC-ODS-A 4.6 Â 5.0 5 lm; gradient A: 0.05%
TFA/water, B: 0.05% TFA/acetonitrile; time 0 and 1 min:
98% A and 2% B; 7 min: 10% A and 90% B; 8 min: 10% A
and 90% B; 8.9 min: 98% A and 2% B; post time 1 min;
flow rate 2.5 mL/min; detection: 215 and 254 nm, DAD.
Semi-Prep HPLC: Gilson with Unipoint software; Col-
umn: Phenomenex C18 Luna 21.6 mm  60 mm, 5 lM;
solvent A: water (0.02% TFA buffer); solvent B: acetoni-
trile (0.02 % TFA buffer); solvent gradient: time 0: 5% B;
2.5 min: 5% B; 12 min: 95% B; hold 95% B 3 min; flow rate:
22.5 mL/min; detection: 215 and 254 nm.
12. De Nanteuil, G.; Lila-Ambroise, C.; Rupin, A.; Vallez,
M.; Verbeuren, T. J. J. Bioorg. Med. Chem. Lett. 2003, 13,
1705–1708.
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, 3361–3365.
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, 761–765.
19. During the reduction of the a,b-unsaturated oxime only
saturated hydroxyl amines were obtained during the
synthesis of analogues 36 and 37.