Synthesis and biological evaluation of piperazine-based derivatives as inhibitors of plasminogen activator inhibitor-1 (PAI-1)

Article abstract of DOI:10.1016/j.bmcl.2003.11.035

Compound 2 was identified by high throughput screening as a novel PAI-1 inhibitor. Systematic optimization of the A, B, and C segments of 2 resulted in the identification of a more potent compound 39 with good oral bioavailability. The synthesis and SAR d

Full text of DOI:10.1016/j.bmcl.2003.11.035

Bioorganic & Medicinal Chemistry Letters 14 (2004) 761–765
Synthesis and biological evaluation of piperazine-based derivatives
as inhibitors of plasminogen activator inhibitor-1 (PAI-1)
Bin Ye,* Yuo-Ling Chou, Rushad Karanjawala, Wheeseong Lee, Shou-Fu Lu,
Kenneth J. Shaw, Steven Jones, Dao Lentz, Amy Liang, Jih-Lie Tseng,
Qingyu Wu and Zuchun Zhao
Discovery Research, Berlex Biosciences, 2600 Hilltop Drive, PO Box 4099, Richmond, CA 94804-0099, USA
Received 9 September 2003; accepted 10 November 2003
Abstract—Compound 2 was identified by high throughput screening as a novel PAI-1 inhibitor. Systematic optimization of the A,
B, and C segments of 2 resulted in the identification of a more potent compound 39 with good oral bioavailability. The synthesis
and SAR data are presented in this report.
# 2003 Elsevier Ltd. All rights reserved.
A key step in the regulation of thrombus formation and
clearance is the generation of plasmin through proteo-
lytic cleavage of plasminogen by tissue plasminogen
activator (tPA) and urokinase (uPA).1 Plasmin dissolves
fibrin clots by degrading insoluble fibrin molecules to
small soluble fragments.2 Plasminogen activator inhi-
bitor-1 (PAI-1), a member of the serine protease
inhibitor (serpin) family, is the major negative regulator
of tPA and uPA.3 PAI-1 is unique in the serpin family
and its active conformation spontaneously rearranges
into the inactive form through the insertion of a larger
portion of the reactive center loop of the active form
into b-sheet A in the middle of the molecule.4 Recently,
mechanisms by which PAI-1 interacts with small mole-
cules5 and its targeted protease such as tPA6 have been
reported. These studies provide important information
on the three dimensional structure of PAI-1, which will
help to design PAI-1 inhibitors.
angiogenesis,11 cancer invasion,11 and metastasis.12
Thus, inhibition of PAI-1 would represent a useful
strategy in treating a variety of cardiovascular and can-
cer diseases. To date, several PAI-1 inhibitors including
antibodies,13 peptides,14 and small molecules3,15 have
been reported.
We have previously reported a menthol-based inhibitor
1, which demonstrates excellent potency in the primary
enzymatic assay (IC50=0.38 mM) as well as the func-
tional clot lysis assay (IC50=0.01 mM).16 There are,
however, some limitations such as poor solubility and
low oral bioavailability associated with this series of
compounds. In searching for an alternate template, high
throughput screening of our compound library was
performed, leading to the discovery of the structurally
different compound 2 with an IC50 value of 1.8 mM in
our primary assay.16 Compound 2 has moderate
potency, but poor oral bioavailability which is pre-
sumably at least partially attributed to the existence of
the phosphonic acid moiety. Thus, an effort was made
to replace the phosphonic acid group to improve the
pharmacokinetic profile as well as increase potency.
Structurally, compound 2 can be divided into three seg-
ments designated as A, B, and C (Fig. 1). Systematic
modification of these segments is detailed in this paper.
Clinically, PAI-1 is considered to be a thrombotic risk
factor.7 Elevated levels of PAI-1 antigen and activity
have been described to correlate with an increased risk
of deep vein thrombosis,8 atherosclerosis,9 unstable
angina and myocardial infarction.3,7 In addition, ele-
vated levels of PAI-1 are associated with a poor prog-
nosis in cancer patients,10 and PAI-1 may play a role in
We started the optimization with the replacement of the
C segment (Table 1). The initial effort focused on direct
replacement of the aminomethyl phosphonic acid with
* Corresponding author. Tel.: +1-510-669-4211; fax: +1-510-669-
4310; e-mail: bin_ye@berlex.com
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.11.035

762
B. Ye et al. / Bioorg. Med. Chem. Lett. 14 (2004) 761–765
carboxylic acids. The isonipecotic acid adduct, com-
pound 3 (Fig. 2) is slightly more potent than 2. Switch-
ing the nitro and trifluoromethyl groups on the B
segment gave compound 4, which is equally potent as
compound 3. This is advantageous from a synthetic
aspect, because a large number of analogues of 4 can be
efficiently prepared via a common intermediate 7
through the displacement of the 2-chloro group with
different nucleophiles (Scheme 1). Compound 7 was
obtained as a sole product by the displacement of the
4-chloro group of compound
5
with 1-(3-tri-
fluoromethylphenyl)piperazine at room temperature.
This is because the reactivity of 4-chloride group of
compound 7 is much higher than 2-chloride group
based on the similar results reported by Welmaker et
al.17 Further displacement of the 2-chloro group with
amino acids and hydroxyphenyl acids yielded analogues
of compound 4. It is observed that the distance between
the carboxylic acid group of C segment to the central B-
ring has an effect on potency. The closer the carboxylic
acid group of C segment to the central B-ring, the less
potent the compound is. For the amino-acid adduct
series, 4 and 12 are equally potent since the distance of
the carboxylic acid group of C segment in these two
compounds to the central ring is almost identical.
However, the distance for the carboxylic acid in 10 and
11 is shorter, and these compounds have lower potency
accordingly. This is also true for the phenyl carboxylic
acid adducts (13, 14, 15). The reason for this remains
unclear. In general, compared with 4, no significant
potency improvement was observed for this series with
most of the compounds having similar (12, 13) or lower
activities (10, 11, 15).
Figure 1. Menthol-based inhibitor 1 and new lead compound 2 from
library.
Table 1. Replacement of phosphonic acid
Number
R1
IC50 (mM)a
4
0.91
10
11
12
13
14
15
2.8
7.9
0.9
1.5
2.1
3.7
Scheme 1. Conditions: (a) DIEA, 6, CH3CN, rt, overnight; (b)
DMSO, 8, DIEA, 110 ꢁC, 10 h; (c) LiOH, THF/H2O, rt, 10 h; (d),
NaH, DMSO, 9, 110 ꢁC, 4 h.
a IC50 values are averaged from multiple determinations (nꢀ2), and
the standard deviations are <30% of the mean.
Figure 2. Initial replacement of phosphonic acid 2.

B. Ye et al. / Bioorg. Med. Chem. Lett. 14 (2004) 761–765
763
We next explored the optimization of the A segment
while keeping isonipecotic acid as the C segment. The
synthesis of the compounds in Table 2 is outlined in
Scheme 2. Displacement of the 4-chloro group with N-
Boc-piperazine 16 followed by addition of isonipecotic
acid ethyl ester 8 and deprotection with TFA afforded
17. Coupling of 17 to 2-chloro-5-(trifluoro-
methyl)pyridine 18 followed by saponification afforded
22. Reductive-amination of 17 with 3-(trifluoro-
methyl)benzaldehyde 19 followed by saponification
afforded 24. Acylation of 17 with 3-(trifluoro-
methyl)benzoyl chloride 20 followed by saponification
afforded 25. The remaining compounds in Table 2 were
prepared in
a similar manner. Shifting the tri-
fluoromethyl group from the meta to para position has
no effect on potency (4 vs 21), whereas introduction of
pyridine (22) decreases potency 3-fold. Insertion of
methylene (24), carbonyl (25), or urea (26) spacers are
also detrimental, and incorporation of lipophilic groups
causes dramatic loss of potency (23, 27).
Next, we turned our attention to the replacement of the
B segment. For this series, we chose 4-aminomethyl-
benzoic acid instead of isonipecotic acid as the C seg-
ment (Table 3). Replacement of trifluoromethyl with a
Scheme 2. Conditions: (a) DIEA, 16 CH3CN, rt, overnight; (b)
DMSO, 8, DIEA, 110 ꢁC, 10 h; (c), TFA, CH2Cl2; (d) DMSO, 18,
DIEA, 110 ꢁC, 6 h; (e) LiOH, THF/H2O, rt, 10 h; (f) NaBH(OAc)3,
19, ClCH2CH2Cl, rt, overnight; (g) Et3N, 20, CH2Cl2.
Table 2. Optimization of A segment
Table 3. Replacement of central ring
Number
R2
IC50 (mM)a
Number
Central ring
IC50 (mM)a
4
0.91
13
1.5
21
22
23
24
25
26
1.1
3.2
28
29
30
31
32
33
34
1.4
6.7
>15
4.3
6.1
>15
2.8
5.0
3.5
3.4
27
>15
1.7
a IC50 values are averaged from multiple determinations (nꢀ2), and
the standard deviations are <30% of the mean.
a IC50 values are averaged from multiple determinations (nꢀ2), and
the standard deviations are <30% of the mean.

764
B. Ye et al. / Bioorg. Med. Chem. Lett. 14 (2004) 761–765
Scheme 3. Conditions: (a) DIEA, 35, CH3CN, rt, overnight; (b) Pd/C, H2, EtOAc–MeOH, 2 h; (c) DIEA, 8, CH3CN, rt, overnight; (d) LiOH, THF/
H2O, rt, 2 h; (e) NaBH3CN, MeOH–CH2Cl2–HOAC, rt, 2 h.
Table 4. Optimization of C segment
Number
R3
IC50 (mM)a
Number
R3
IC50 (mM)a
34
1.7
44
1.0
39
40
41
42
43
0.5
0.9
0.8
1.0
0.9
45
46
47
48
1.7
0.7
1.4
1.7
a IC50 values are averaged from multiple determinations (nꢀ2), and the standard deviations are <30% of the mean.
nitro group retained potency (13 vs 28), but decreased
potency was observed in most of the cases such as dele-
tion of trifluoromethyl (29, 32), and incorporation of
methylcarboxylate (31). Changing the 1,3-substitution
pattern of the B segment to 1,4-substitution and the
removal of the nitro group yielded compound 34 which
has equal potency to 13 but is less lipophilic based on
the value of calculated LogP.18
Intermediate 38 was prepared by addition of iso-
nipecotic acid methyl ester 8 to 2,6-difluorobenzaldehyde
37 followed by hydrolysis. The starting aldehydes
needed for the preparation of compounds 39–48 were
either prepared in the same manner as 38 or as previously
disclosed.16 Compared with 34, all the other compounds
have similar or slightly improved potency (Table 4).
Compound 39 was pursued further as it is about 4-fold
more potent.
Further optimization of the C segment was pursued
keeping other parts of the template constant as shown
in Table 4. Compounds listed were prepared according
to Scheme 3. Addition of 1-(3-trifluoromethyl-
phenyl)piperazine 6 to 2-fluoro-5-nitrobenzotrifluoride
35 followed by hydrogenation, afforded aniline 36.
Reductive-amination of 36 with 1-(3-fluoro-2-form-
ylphenyl)-4-piperidinecarboxylic acid 38 afforded 44.
Most of the compounds were tested in a functional clot
lysis assay which was described in a previous paper16
(data not shown). Compound 39 was found to be the
most potent inhibitor with an IC50 of 0.01 mM. The
compound 39, however, was less potent in the primary
assay (IC50=0.5 mM). The reason for this apparent dif-
ference in potency in these assays is not clear, but par-

B. Ye et al. / Bioorg. Med. Chem. Lett. 14 (2004) 761–765
765
Table 5. Pharmacokinetic data for compound 1 and 39a
Haemost. Thromb. 1989, 9, 87. (b) Wu, Q.; Zhao, Z. Curr.
Drug Tar. 2002, 2, 27.
b
b
b
Number Cmax
(mM)
Tmax
(h)
T1/2
(h)
Clc
Vdssc
F
4. (a) Sancho, E.; Declerck, P. J.; Price, N. C.; Kelly, S. M.;
Booth, N. A. Biochemistry 1995, 34, 1064. (b) Mottonen,
J.; Strand, A.; Symersky, J.; Sweet, R. M.; Danley, D. E.;
Geoghegan, K. F.; Gerard, R. D.; Goldsmith, E. J.
Nature 1992, 355, 270.
5. Bjoerquist, P.; Ehnebom, J.; Inghardt, T.; Hasson, L.;
¨
Lindberg, M.; Linschoten, M.; Stroemqvist, M.; Deinum,
J. Biochemistry 1998, 37, 1227.
6. Blouse, G. E.; Perron, M. J.; Thompson, J. H.; Day, D. E.;
Link, C. A.; Shore, J. D. Biochemistry 2002, 41, 11997.
7. Salomaa, V.; Stinson, V.; Kark, J. D.; Folsom, A. R.;
Davis, C. E.; Wu, K. K. Circulation 1995, 91, 284.
8. (a) Erikson, L. A.; Fici, G. J.; Lund, J. E.; Bogle, T. P.;
Polites, H. G.; Marotti, K. R. Nature 1990, 346, 74. (b)
Wiman, B. Haemost. Thromb. 1995, 74, 71.
9. Schneiderman, J.; Sawdey, M. S.; Keeton, M. R.; Bordin,
G. M.; Bernstern, E. F.; Dilley, R. B.; Loskutoff, D. J.
Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 6998.
10. Frandesen, T. L.; Stephenes, R. W.; Pedersen, A. N.;
Engelholm, L. H.; Holst-Hansen, C.; Brunner, N. Drugs
Future 1998, 873.
11. Bajou, K.; Noel, A.; Gerard, R. D.; Masson, V.; Brunner,
N.; Holst-Hansen, C.; Skobe, M.; Fusening, N. E.; Car-
meliet, P.; Collen, D.; Foidart, J. M. Nat. Med. 1998, 4,
923.
12. Tsuchiya, H.; Katsuo, S.; Matsuda, E.; Sunamaya, C.;
Tomita, K.; Veda, Y.; Binder, B. Gen. Diag. Pathol. 1995,
141, 41.
(mL/min/kg) (L/kg) (%)d
1
39
0.16
0.48
3
3
3.0
2.8
4.84
10.4
0.4
1.33
23
43
a Compound administered in 94% PEG 300, 4% EtOH, and 2% water
at 0.5 mg/kg iv, 2 mg/kg po.
b Cmax, Tmax, and T1/2 for po.
c Cl and Vdss for iv.
d F% values are averaged from three rats.
tially may be attributed to the different assay systems
because we observed that in the presence of lipid-con-
taining plasma the compound was consistently more
potent. To examine the possibility that compound 39
might inhibit other enzymes of the coagulation cascade,
which could lead to enhanced clot lysis, we did test the
compound in the clot lysis assay in the absence of PAI-
1. The results showed that the compound 39 did not
alter clot lysis time (data not shown). Thus, it is unlikely
that compound 39 directly inhibits other enzymes in the
blood-clotting cascade. Further pharmacokinetic stud-
ies of 39 were performed in rats, and the results are
summarized in Table 5. Compound 39 has improved
bioavailability (43%) compared with 1, moderate clear-
ance rate (10.4 mL/min/kg), and good half-life (2.8 h).
13. Abrahamsson, T.; Nerme, V.; Stromqvist, M.; Akerblom,
B.; Legnehed, A.; Petterson, K.; Westin-Eriksson, A.
Haemost. Thromb. 1996, 75, 118.
14. (a) Eitzman, D. T.; Fag, W. P.; Lawrence, D. A.; Francis-
Chmura, A. M.; Shore, D.; Olson, S. T.; Ginsburg, D. J.
Clin. Invest. 1995, 95, 2416. (b) Guzzo, P. R.; Trova,
M. P.; Inghardt, T.; Linschoten, M. Tetrahedron. Lett.
2002, 43, 41.
15. (a) Charlton, P. Drugs Future 1997, 22, 45. (b) Folkes, A.;
Brown, D. S.; Canne, L. E.; Chan, J.; Engelhardt, E.;
Epshteyu, S.; Faint, R.; Goldec, J.; Hanel, A.; Kearneg,
P.; Leadhy, 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. (c) Wang, S.; Golec, J.; Miller, W.;
Milutinovic, S.; Folkes, A.; Williams, S.; Brooks, T.;
Hardman, K.; Charlton, P.; Wren, S.; Spencer, J. Bioorg.
Med. Chem. Lett. 2002, 12, 2367. (d) Nanteuil, G. D.;
Lila-Ambroise, C.; Rupin, A.; Vallez, M.-O.; Verbeuren,
T. J. Bioorg. Med. Chem. Lett. 2003, 13, 1705.
In summary, we have explored the structure–activity
relationships around the novel PAI-1 inhibitor 2 by
modifying the A, B, and C segments. Initial optimiza-
tion of the C segment indicates that the aminomethyl
phosphonic acid could be replaced with amino-acids.
No improvement was seen in the modification of the A
segment, but efforts to optimize the B segment resulted
in the identification of the 2-trifluoromethylphenyl as
the optimal group. Further optimization of the C seg-
ment led to the discovery of the most potent compound
39 in both the primary and functional assays. Currently,
animal efficacy studies of 39 to validate the disease
model are in progress.
Acknowledgements
The authors would like to thank Faye Wu and Kathy
Tran for performing the clot-lysis assays, Marilyn Lam
and Jun Shen for pharmacokinetic studies, and Monica
Kochanny and Gary Phillips for manuscript preparation.
16. Ye, B.; Bauer, S.; Buckman, B. O.; Ghannam, A.; Grie-
del, B. D.; Khim, S.-K.; Lee, W.; Sacchi, K. L.; Shaw,
K. J.; Liang, A.; Wu, Q.; Zhao, Z. Bioorg. Med. Chem.
Lett. 2003, 13, 3361.
17. Welmaker, G. S.; Nelson, J. A.; Sabalski, J. E.; Sabb,
A. L.; Potoski, J. R.; Graziano, D.; Kagan, M.; Coupet,
J.; Dunlop, J.; Mazandarani, H.; Rosenzweig-Lipson, S.;
Sunkoff, S.; Zhang, Y. Bioorg. Med. Chem. Lett. 2000, 10,
1991.
18. The calculated LogP value was obtained through the cal-
culation of the compound via the software created by
Advanced Chemistry Development, Inc, Canada. The
cLog P value for 13 and 34 is 9.09Æ0.64 and 6.97Æ0.57,
respectively.
References and notes
1. (a) Collen, D.; Lijnen, H. R. Blood 1991, 78, 3114. (b)
Vassalli, J. D.; Sappino, A. P.; Belin, D. J. Clin. Invest.
1991, 88, 1067.
2. Declerk, P. J.; Juhan-Vaguo, I.; Felez, J.; Wiman, B. J.
Intern. Med. 1994, 236, 425.
3. (a) Loskutoff, D. J.; Sawdey, M.; Mimuro, J. Prog.