Identification of orally bioavailable, non-amidine inhibitors of Urokinase Plasminogen Activator (uPA)

Article abstract of DOI:10.1016/j.bmcl.2009.08.008

In this Letter we report the synthesis and evaluation of a series of non-amidine inhibitors of Urokinase Plasminogen Activator (uPA). Starting from compound 1, a significant change provided compounds in which the amidine, binding in the S1 pocket, was rep

Full text of DOI:10.1016/j.bmcl.2009.08.008

Bioorganic & Medicinal Chemistry Letters 19 (2009) 5712–5715
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Identification of orally bioavailable, non-amidine inhibitors of Urokinase
Plasminogen Activator (uPA)
b
a
a
c
a
Christopher W. West a, , Marc Adler , Danny Arnaiz , Deborah Chen , Kieu Chu , Giovanna Gualtieri ,
Elena Ho d, Christoph Huwe a, David Light e, Gary Phillips a, Rebecca Pulk a, Drew Sukovich c, Marc Whitlow b,
Shendong Yuan a, Judi Bryant a
*
a Department of Medicinal Chemistry, Berlex Biosciences, 2600 Hilltop Drive, Richmond, CA 94804, United States
b Department of Biophysics, Berlex Biosciences, 2600 Hilltop Drive, Richmond, CA 94804, United States
c Department of Molecular Pharmacology, Berlex Biosciences, 2600 Hilltop Drive, Richmond, CA 94804, United States
d Department of Animal Pharmacology, Berlex Biosciences, 2600 Hilltop Drive, Richmond, CA 94804, United States
e Department of Antibody Technology, Berlex Biosciences, 2600 Hilltop Drive, Richmond, CA 94804, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
In this Letter we report the synthesis and evaluation of a series of non-amidine inhibitors of Urokinase
Plasminogen Activator (uPA). Starting from compound 1, a significant change provided compounds in
which the amidine, binding in the S1 pocket, was replaced with a primary amine. Further modifications
led to the identification of potent, selective, and orally bioavailable uPA inhibitors.
Ó 2009 Elsevier Ltd. All rights reserved.
Received 2 April 2009
Revised 1 August 2009
Accepted 4 August 2009
Available online 7 August 2009
Keywords:
uPA
Urokinase
Serine protease
Urokinase Plasminogen Activator (uPA) is a serine protease that
is involved in various biological processes and functions as a key
initiator in extracellular proteolytic cascades.1 Urokinase activates
plasminogen to plasmin which in turn initiates the activation of
many matrix metalloproteases (MMPs). Plasmin and activated
MMPs are capable of degrading collagen, elastin and many other
extracellular matrix (ECM) and basement membrane (BM) proteins
which can lead to an increase in tissue destruction.2 Recent studies
have associated the tissue remodeling activity of uPA, plasmin, and
MMPs in several diseases and disorders, including aneurysm for-
mation and other acute coronary syndromes. Additionally, stroke,
multiple sclerosis, rheumatoid arthritis, cancer, and AIDS have also
been implicated.3 Therefore, inhibition of uPA may be useful in the
treatment of diseases and disorders mediated by the degradation
of ECM and BM proteins.
A primary amine, discovered by X-ray fragment screening at Astex,
has been shown to be able to fit in the S1 pocket.5
Our starting point for development of an orally bioavailable uPA
inhibitor began with amidine 1 (Fig. 1), a proprietary molecule de-
rived from our earlier work on serine protease inhibitors and iden-
tified through a focused screening effort. Compound 1 is potent
against uPA (IC50 = 0.098
two related trypsin family serine proteases, tPA and plasmin
(IC50 = 3.9 and 2.1 M, respectively). However, compounds of this
lM) and shows some selectivity against
l
class were found to generally have low oral bioavailability, a result
mainly attributed to the presence of an amidine moiety.
Following analysis of the X-ray crystal structure of 1 bound to
the protease domain of uPA (not shown), we found that the ami-
dine does indeed form a salt bridge to Asp189 in the S1 pocket of
uPA. Though we needed to keep this important interaction, early
In light of the potential benefits of uPA inhibition, many articles
have described the discovery of uPA inhibitors. However, almost all
of the non-peptidic small molecule inhibitors described to date
feature either a guanidine or amidine forming a key interaction
with the Asp189 carboxylate in the S1 pocket of the active site of
Urokinase.4 So far, only one uPA inhibitor has been described that
does not use an amidine or guanidine to make this key interaction.
* Corresponding author. Tel.: +1 919 941 5206; fax: +1 919 941 0813.
E-mail address: cwest@icagen.com (C.W. West).
Figure 1. Focused screening lead.
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.08.008

C. W. West et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5712–5715
5713
Table 2
optimization efforts focused on significantly modifying the ami-
dine. Most modifications were unsuccessful, but replacement of
the amidine with a benzyl amine provided 2 (IC50 = 0.10 M) (Ta-
ble 1), a potent compound that showed an improvement in selec-
tivity against tPA and plasmin (IC50 = 18 and 6.5 M, respectively).
Additionally, this change led to a compound that showed some oral
bioavailability (%F = 9) but with low exposure levels after oral
administration to rats (Table 2).
PK results for selected compounds
l
Compd
PK data
Cmax (ng/mL)
T1/2 (h)
Cl (mL/min/kg)
Ld (L/kg)
%F
l
2
11
20
26
33
0
220
59
0.5
4.6
0.5
26
5
1.15
1
9
0
54
257
0.7
4.6
37
18
0.26
An X-ray structure of 2 bound to uPA (not shown) provided evi-
dence that the benzyl amine had indeed replaced the amidine
binding in the S1 pocket. Modifications of the primary amine by
substitution of the amine with one or two methyl groups (3 and
4, respectively), or moving the benzyl amine to the 4-position (5)
of the phenyl ring provided only inactive compounds (Table 1).
In general, modification of this portion of the molecule led to inac-
tive compounds, highlighting just how specific the S1 pocket is.
Table 1
Ring variations
Compd
R1
R2
R3
R4
R5
uPA
IC50
tPA
IC50
Plasmin
IC50 (lM)
a
a
a
(l
M)
(lM)
1
2
3
4
5
6
7
8
9
3-C(NH)NH2
3-CH2NH2
3-CH2NHMe
3-CH2NMe2
4-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
H
H
H
H
H
H
H
H
H
H
Me
H
H
H
2-CO2H
2-CO2H
2-CO2H
2-CO2H
2-CO2H
2-CO2H
2-CO2H
2-CO2H
2-CO2H
2-CO2H
5-NMe2
5-NMe2
5-NMe2
5-NMe2
5-NMe2
5-NMe2
5-NMe2
5-NMe2
5-NMe2
5-NMe2
0.098
0.10
>30
>30
>30
0.13
0.53
0.90
0.065
0.10
3.9
18
NA
NA
NA
16
>30
10
11
2.1
6.5
NA
NA
NA
3.6
2.9
2.1
4.4
6.3
H
NHMe
NMe2
Morpholine
Piperazine
10
28
N
N
11
3-CH2NH2
H
2-CO2H
5-NMe2
0.025
5.2
1.9
N
N
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
3-CH2NH2
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
2-CO2H
3-CO2H
4-CO2H
2-CH2CO2H
2-(CH2)2CO2H
2-tetrazole
2-CO2H
2-CO2H
2-CO2H
2-CO2H
2-CO2H
2-CO2H
2-CO2H
2-CO2H
2-CO2H
2-CO2H
2-CO2H
2-CO2H
2-CO2H
H
H
H
H
0.84
20
>30
8.1
>30
13
26
>30
NA
NA
NA
NA
NA
NA
NA
22
NA
NA
>30
NA
15
21
NA
NA
NA
NA
NA
NA
NA
5.0
8.6
2.8
9.4
NA
11
H
5-Me
3-Me
4-Me
5-Me
6-Me
5-iPr
5-OMe
5-NO2
5-Me
5-Me
5-Me
5-Me
5-Me
5-Me
3.5
0.24
0.54
0.68
0.22
1.3
0.089
0.039
>30
2.7
H
2-NH2
3-NH2
4-NH2
3-CH2NH2
3-OH
3-NH2
>30
NA
NA
14
5.0
NA
2.7 0.3
5.5
0.023
0.015
8.5
3.0
N
N
a
Values are the mean of P2 experiments. For assay descriptions see Ref. 6. NA = not available.

5714
C. W. West et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5712–5715
Fortunately, other parts of the molecule were more amenable to
served between the pyrrolidine and the enzyme. The dimethyl-
amino pyrrolidine comes in close contact with the loop from
residues 96–98. This loop contains two additional residues that
block the S4 pocket found in other proteases such as tPA and plas-
min. The dimethylamine may pick up some positive electrostatic
interactions between the basic nitrogen and the carbonyls of
Leu97 (4.0 Å) and Ala98 (4.6 Å) that account for the increase in po-
tency seen with this analog.
exploration. The 4-position on the pyridine ring (R3) was modified,
beginning with N-methyl substituted 6. In this case, a compound
equipotent with 2 was realized. Increasing the size of the R3 substi-
tuent caused a decline in potency as seen in dimethylamine 7
(IC50 = 0.53
ever, when a piperazine was placed in this position, the potency
was reclaimed as seen in 9 (IC50 = 0.065 M). The cyclic nature of
lM) and morpholine analog 8 (IC50 = 0.90 lM). How-
l
the amine was not specifically required as linear analog 10 was
equipotent with 2. The racemic dimethylamino pyrrolidine 11
Since compound 2 is zwitterionic, its overall PK was going to be
difficult to dramatically improve. As the amine is crucial, we fo-
cused on modifying the carboxylic acid group. For ease of synthe-
sis, compound 12 was made as a standard to compare the
following analogs. Attempts to either move the acid to other posi-
tions on the phenyl ring (13, 14) or change the distance between
the acid and aromatic ring (15, 16) reduced the potency. Addition-
ally, the acid isostere 17 was synthesized, but only a weakly active
compound resulted from this effort. In general, modification of the
carboxylic acid was not tolerated.
was the most potent analog in this series (IC50 = 0.025
showed selectivity against tPA and plasmin (IC50 = 5.2 and
1.9 M, respectively). Unfortunately, compound 11 had a complete
lM) and
l
lack of oral bioavailability (Table 2), a result directly attributed to
incorporating more charge in the molecule.
Throughout our optimization process, X-ray crystal structures
were a valuable tool and were used frequently in analog design.
Figure 2 shows one such structure, the complex formed between
11 and human uPA.7 The free amine binds in the bottom of the
S1 pocket and forms a salt bridge to Asp189 (2.9 Å). This amine also
has a hydrogen bond to the carbonyls of Ser190 (2.8 Å) and Gly218
(3.0 Å). The carboxylic acid forms several hydrogen bonds. One
oxygen interacts with the side chains of Ser195 (2.7 Å) and His57
(2.8 Å) of the catalytic triad. The other oxygen forms a hydrogen
bond with the backbone amide nitrogen of Gly193 (2.8 Å). These
two key interactions formed the main contact points between
our inhibitors and the enzyme. An additional interaction is ob-
Other modifications to the phenyl ring containing the carbox-
ylic acid proved more fruitful. A methyl scan of the ring (18–21)
revealed that the 5-position was favored. Compound 20
(IC50 = 0.24
ilar selectivity against tPA and plasmin (IC50 = 22.1 and 5.0
respectively). This was an interesting result especially when com-
pared to isopropyl 22 (IC50 = 0.68 M), a compound that is sixfold
l
M) was only twofold less potent than 2 and had sim-
l
M,
l
less potent than 2. Another small surprise was observed when a PK
study was performed on 20. This compound was found to have an
improved %F and Cmax when compared to 2, a result we attributed
to the reduction of overall charge of the molecule. Further work on
this ring showed electron donating groups such as methoxy 23
(IC50 = 0.22
l
M) provided compounds with greater potency than
M).
electron withdrawing groups such as nitro 24 (IC50 = 1.3
l
With the knowledge gained so far regarding our template, we
explored substitutions on R2. From our X-ray crystal structure
studies, we postulated that these substituents should access the
S1B subsite. This S1B subsite was so named by Nienaber et al.
and exploited in improving their uPA inhibitors.8 An amino scan
revealed that the 2-position 25 (IC50 = 0.089
lM) improved the po-
tency some and the 3-position 26 (IC50 = 0.039
lM) provided
nearly a sixfold improvement compared to 20. However, 4-position
27 was inactive, probably due to a change in the orientation of the
first two aromatic rings with respect to each other. Compound 26
was profiled further and showed an increase in selectivity (IC50 >30
and =5.0 lM against tPA and plasmin), some oral bioavailability,
but an extremely high clearance.
In general, we found that at R2 a variety of substituents were
tolerated and some provided an improvement in potency. How-
ever, R2 is a sensitive position as exemplified by methylene amine
Figure 2. X-ray crystal structure of compound 11 bound to human uPA.
Scheme 1. Reagents and conditions: (a) Pd(PPh3)4, IPA, 2 N Na2CO3(aq), 80 °C, 71%; (b) RaNi, IPA, MeOH, H2, 94%; (c) Boc2O, NaOH(aq), dioxane, 95%; (d) N,N,N0-
trimethylethylenediamine, Et3N, CH2Cl2, 0 °C, 100%; (e) Cs2CO3, DMF, 50 °C, 91%; (f) 32, Cs2CO3, DMSO, 80 °C, 73%; (g) (i) LiOH, H2O, MeOH, 50 °C, 24 h; (ii) TFA, CH2Cl2, 57%.

C. W. West et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5712–5715
5715
compound 28 losing 10-fold in activity. Interestingly, the hydroxyl
29 (IC50 = 0.023 M) regains the potency seen in the amino series.
Combining many of the best groups from each ring provided the
potent racemic compound 30 (IC50 = 0.015 M). This analog
sign and synthesis allowed the modification of molecular proper-
ties based on specific portions of the molecule. Further work on
this template will be reported in the near future.
l
l
showed that the effects at each site were additive, which could
References and notes
be useful in further lead optimization.
1. Werb, Z.; Ashkenas, J.; MacAuley, A.; Wiesen, J. F. Braz. J. Med. Biol. Res. 1996, 29,
1087.
2. Mignatti, P.; Rifkin, D. B. Physiol. Rev. 1993, 73, 161.
The four ringed structure of compound 1 makes a modular
modification strategy ideal for SAR exploration. Numerous analogs
were synthesized in a rapid and convergent fashion simply by
varying the order of synthetic steps for incorporating an appropri-
ately functionalized ring. The synthesis of compound 10 (Scheme
1) is representative of the methodology used to prepare the com-
pounds described in this Letter.9 A Suzuki coupling of commer-
cially available 3-cyano phenyl boronic acid and 3-bromophenol
provided nitrile 31. Following reduction of the nitrile with Raney
Nickel, protection of the free amine using Boc2O provided left side
biaryl 32. The right side biaryl ether was synthesized using two
regioselective SNAr reactions. Reaction of pentafluoropyridine with
N,N,N0-trimethylethylene-diamine at lower temperature cleanly
provided the 4-position substituted pyridine 33. Upon treatment
of 33 with salicylate 34 using cesium carbonate in DMF and mild
heat, biaryl ether 35 was realized. The third SNAr reaction, between
35 and 32, required cesium carbonate in DMSO and moderate heat
to give the fully assembled ring system. Subsequent saponification
and Boc removal provided compound 10. As a side note, the sapon-
ification of the methyl ester in these compounds proved to be time
consuming and use of the tert-butyl ester in later work was
advantageous.
3. Rockway, T. W.; Nienaber, V.; Giranda, V. L. Curr. Pharm. Des. 2002, 8, 2541.
4. (a) Bruncko, M.; McClellan, W. J.; Wendt, M. D.; Sauer, D. R.; Geyer, A.; Dalton, C.
R.; Kaminski, M. A.; Weitzberg, M.; Gong, J.; Dellaria, J. F.; Mantei, R.; Zhao, X.;
Nienaber, V. L.; Stewart, K.; Klinghofer, V.; Bouska, J.; Rockway, T. W.; Giranda, V.
L. Bioorg. Med. Chem. Lett. 2005, 15, 93; (b) Fish, P. V.; Barber, C. G.; Brown, D. G.;
Butt, R.; Collis, M. G.; Dickinson, R. P.; Henry, B. T.; Horne, V. A.; Huggins, J. P.;
King, E.; O’Gara, M.; McCleverty, D.; McIntosh, F.; Phillips, C.; Webster, R. J. Med.
Chem. 2007, 50, 2341; (c) Mackman, R. L.; Hui, H. C.; Breitenbucher, J. G.; Katz, B.
A.; Luong, C.; Martelli, A.; McGee, D.; Radika, K.; Sendzik, M.; Spencer, J. R.;
Sprengeler, P. A.; Tario, J.; Verner, E.; Wang, J. Bioorg. Med. Chem. Lett. 2002, 12,
2019; (d) Joossens, J.; Ali, O. M.; El-Sayed, I.; Surpateanu, G.; Van der Veken, P.;
Lambeir, A.; Setyono-Han, B.; Foekens, J. A.; Schneider, A.; Schmalix, W.;
Haemers, A.; Augustyns, K. J. Med. Chem. 2007, 50, 6638.
5. Frederickson, M.; Callaghan, O.; Chessari, G.; Congreve, M.; Cowan, S. R.;
Matthews, J. E.; McMenamin, R.; Smith, D.-M.; Vinkovic, M.; Wallis, N. G. J. Med.
Chem. 2008, 51, 183.
6. (a) Lottenberg, R.; Christensen, U.; Jackson, C. M.; Coleman, P. L. Methods
Enzymol. 1981, 80, 341; (b) Liang, A. M.; Light, D. R.; Kochanny, M.; Rumennik,
G.; Trinh, L.; Lentz, D.; Post, J.; Morser, J.; Snider, M. Biochem. Pharmacol. 2003,
65, 1407.
7. X-ray coordinates of compound 11 bound to uPA have been deposited into the
Protein Data Bank (PDB identifier 3IG6).
8. Nienaber, V.; Davidson, D.; Edalji, R.; Giranda, V. L.; Klinghofer, V.; Henkin, J.;
Magdalinos, P.; Mantei, R.; Merrick, S.; Severin, J. M.; Smith, R. A.; Stewart, K.;
Walter, K.; Wang, J.; Wendt, M.; Weitzberg, M.; Zhao, X.; Rockway, T. Structure
2000, 8, 553.
9. All final compounds were characterized by NMR, HPLC, MS, and EA. Additional
synthetic details will be provided in future publications and patents.
In conclusion, we have described the discovery of a series of
non-amidine, multi-ring compounds that function as potent and
selective uPA inhibitors. Using a modular approach in analog de-