2-(2-Hydroxy-3-alkoxyphenyl)-1H-benzimidazole-5-carboxamidine derivatives as potent and selective urokinase-type plasminogen activator inhibitors

Article abstract of DOI:10.1016/S0960-894X(02)00311-6

The development of potent and selective urokinase-type plasminogen activator (uPA) inhibitors based on the lead molecule 2-(2-hydroxy-3-ethoxyphenyl)-1H-benzimidazole-5-carboxamidine (3a) is described.

Full text of DOI:10.1016/S0960-894X(02)00311-6

Bioorganic & Medicinal Chemistry Letters 12 (2002) 2019–2022
2-(2-Hydroxy-3-alkoxyphenyl)-1H-benzimidazole-5-carboxamidine
Derivatives as Potent and Selective Urokinase-type
Plasminogen Activator Inhibitors
Richard L. Mackman,*,y Hon C. Hui, J. Guy Breitenbucher,{ Bradley A. Katz,
Christine Luong, Arnold Martelli, Danny McGee,x Kesavan Radika, Martin Sendzi{k,*
Jeffrey R. Spencer, Paul A. Sprengeler, James Tario,y Erik Verner and JingWang
Celera, 180 Kimball Way, South San Francisco, CA 94080, USA
Received 8 November 2001; accepted 29 March 2002
Abstract—The development of potent and selective urokinase-type plasminogen activator (uPA) inhibitors based on the lead
molecule 2-(2-hydroxy-3-ethoxyphenyl)-1H-benzimidazole-5-carboxamidine (3a) is described. # 2002 Elsevier Science Ltd. All
rights reserved.
Urokinase-type plasminogen activator (uPA) is a tryp-
sin-like serine protease that has been implicated in the
processes of tumor cell invasion and metastasis.1 Studies
In this report, we describe the potency and selectivity
optimization of lead inhibitor 3a toward uPA.
2
involvingthe use of anti-uPA antibodies, antisense oli-
Inhibitors 1 and 2 bind to the active site of uPA and
span from subsite S1 to S10.7 In a similar manner, 3a
also binds in these two subsites (Fig. 1a) placing the
5-carboxamidine substituted heterocycle in the S1 sub-
gonucleotides,3 uPA À/Àknockout mice,4 and small
molecule inhibitors,5 all support an important role
of uPA in tumor progression. Furthermore, high levels
of uPA correlate with poor prognosis in several types of
cancer.1a Potent and selective inhibitors of uPA may
therefore be therapeutically useful drugs for the treat-
ment of metastatic cancer.6 In several earlier reports
we described the discovery of a novel series of trypsin-
like serine protease inhibitors (e.g., 1–3a) that bind to
the active site of trypsin-like serine proteases through
the formation of a unique cluster of short (<2.3 A) and
ordinary range hydrogen bonds (Table 1).7 This class of
inhibitors displays broad spectrum inhibitory activity
toward the family of trypsin-like serine proteases and
affords excellent leads for the development of potent
and selective inhibitors of uPA.
0
site and directingthe 3 -ethoxy group toward S10. The
latter group provides an opportunity for exploring the
S10 subsite to optimize potency and selectivity. Several
analogues based upon 3a were synthesized accordingto
Scheme 1, and their inhibition profiles toward uPA and
several other proteases are depicted in Table 1.7 The key
precursors required to synthesize the benzimidazole
analogues are an arylaldehyde (e.g., 6), and an aryl dia-
mine (e.g., 7–9) (Scheme 1).7a A Mitsunobu reaction
was used to introduce the different 30-alkoxy groups on
salicylaldehyde 5. The arylethers, 6, were then oxida-
tively condensed with one of the aryldiamines 7–9 to
furnish the benzimidazole heterocycle.7,8 The benzyl
group was removed and where necessary, the nitrile
group was converted to an amidine using reported
procedures.7a To prepare indoles 12a–c the benzyl pro-
tected salicylaldehyde 5 was first converted to the alkyne
10,9 and then subjected to a palladium mediated cou-
pling10 with mesylaniline 1111 to yield the N-mesyl
*Correspondingauthors. Tel.: +1-650-866-6533; fax: +1-650-866-
6655; e-mail: martin.sendzik@celera.com
yCurrent address: Gilead Sciences, 333 Lakeside Dr., Foster City, CA
94404, USA.
{Current address: RW Johnson PRI, 3210 Merryfield Row, San Diego,
CA 92121, USA.
0
indole. A Mitsunobu couplingintroduced the 3 -alkoxy
xCurrent address: Corixa, 600 Gateway Blvd., South San Francisco,
CA 94080, USA.
group, which was then followed by removal of the mesyl
and benzyl groups and conversion of the nitrile to an
amidine to give the indoles 12a–c.7a
{Current address: Exelexis, 170 Harbor Way, South San Francisco,
CA 94080, USA.
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00311-6

2020
R.L.Mackman et al./ Bioorg.Med.Chem.Lett.12 (2002) 2019–2022
0
Inhibitors 1 and 2 bearinga 3 -phenyl ringreveal that a
single atom change (benzimidazole to indole) can result in
up to 50-fold greater affinity toward some enzymes.7 In
contrast, the indole analogues 12a–c, which have 30-alkoxy
groups, only display up to 3-fold greater affinity than
their benzimidazole counterparts, 3a–c. The 30-alkoxy
groups are situated in close proximity to the cluster of
hydrogen bonds and presumably introduce different
effects to the phenyl groups found in 1 and 2. These
effects may be both electronic, exerted through the
Scheme 1. (a) BnBr, Cs2CO3; (b) R2OH, DEAD, PPh3; (c) EtOH, Na2S2O3, reflux; (d) H2, 10% Pd/C, MeOH; (e) NH2OH, EtOH, reflux; (f) Ac2O,
AcOH; (g) H2, 10% Pd/C, MeOH; (h) K2CO3, Bestmann’s reagent;9 (i) Pd(PPh3)2Cl2, CuII, DMF, TEA; (j) R2OH, DEAD, PPh3; (k) KOH, MeOH;
(l) H2, 10% Pd/C, MeOH; (m) NH2OH, EtOH, reflux; (n) Zn, AcOH.
Table 1. Inhibition profiles of the inhibitors toward 5 trypsin-like serine proteases
Compd
R1
X
OR2
Ki (mM)
uPA
tPA
1.0
0.035
3.2
1.0
0.69
1.7
0.88
0.59
>75
16
29
>75
61
33
45
275
263
225
44
58
29
425
150
71
Factor Xa
Thrombin
Plasmin
1
2
3a
3b
3c
H
H
H
H
H
H
H
H
Cl
Cl
Cl
F
F
F
F
F
F
F
F
F
N
CH
N
N
N
CH
CH
CH
N
N
N
N
N
N
N
N
N
N
N
Ph
Ph
OEt
0.40
0.008
0.70
0.25
0.22
0.45
0.18
0.19
0.95
0.55
0.55
0.29
0.19
0.22
0.08
0.70
0.46
1.2
2.1
0.078
6.8
2.4
1.5
2.1
1.7
0.99
>75
34
15
0.32
31
15
15
9.0
7.5
7.5
>75
90
165
>75
23
32
20
70
50
50
20
16
11
5.0
0.10
1.6
1.3
0.60
0.65
0.65
0.31
1.6
2.2
1.0
1.8
2.7
1.6
4.2
27
OiBu
O(C5H9)
OEt
12a
12b
12c
14a
14b
14c
15a
15b
15c
15d
15e
15f
15g
15h
15i
15j
15k
15l
15m
OiBu
O(C5H9)
OEt
OiBu
O(C5H9)
OEt
43
41
31
32
OiBu
O(C5H9)
a
13
a
a
a
a
a
a
a
a
a
167
125
94
34
65
34
83
38
38
21
47
0.032
0.032
0.011
0.11
0.18
0.10
0.85
2.6
1.1
3.1
3.0
1.7
N
N
N
N
F
F
F
F
14
22
10
N
aOR2 groups for 15d–m

R.L.Mackman et al./ Bioorg.Med.Chem.Lett.12 (2002) 2019–2022
2021
covalent bonds, or alternatively, spatial effects, such as
hydrogen bonding interactions between the 30-O atom
and the hydroxyl group. For example, it has been pro-
posed that the preferred bindingconformer of the tau-
tomeric benzimdazole, in the short hydrogen bond
cluster, is that in which the nitrogen proximal to the
hydroxyl group is protonated.7b Therefore, one reason-
able possibility, is that the 30-O in this series of inhibi-
tors may well promote the more preferred tautomeric
form of the benzimidazole, compared to the 30-phenyl
series, through the formation of hydrogen bonds
between the protonated nitrogen, the hydroxyl group
and the 30-O. Since the indole heterocycle offered no
significant advantage over the benzimidazole in this
series of 30-alkoxy analogues, the more easily synthesized
benzimidazoles were the focus of future efforts toward
inhibitors displayinggreater potency and selectivity.
catalytically inactive mutant of plasmin indicates that
there are some differences within the S10 region com-
pared to uPA.13 In plasmin, residue 41 is a phenylala-
nine whereas in uPA it is a smaller valine residue. The
peptide backbone in the region of Phe41 is shifted by
ꢀ2 A inwards and the disulfide bridge between Cys42
and Cys58 is also moved inwards, leadingto a smaller
S10 subsite in plasmin compared to uPA. Efforts to
increase the steric bulk of the 30-isobutyl group on the
inhibitor, 15b, in an attempt to exploit the larger S10
subsite of uPA compared to plasmin were successful
and yielded a series of inhibitors, 15d–m, with improved
plasmin selectivity. Analogues 15d–g, indicated that
secondary nitrogen atoms tended to give relatively weak
uPA inhibitors. For example, the tetrahydrofurfuryl
analogue 15d was ten times more potent toward uPA
than the correspondingpyrrolidine analouge
15e.
Bulky, hydrophobic, five- and six-membered rings (e.g.,
15h–m) yielded the most potent and selective analogues
(e.g., 15h–j). Multiple methyl groups on the cyclohexyl
ring(e.g., 15k) or movingthe methyl group around the
ring, analogues 15l and 15m, reduced affinity toward
uPA and also, to a lesser extent, plasmin. The most
selective inhibitor was trans-2-methylcyclohexanol 15j,
which had a Ki=11 nM toward uPA and greater than
100-fold selectivity against plasmin and the other anti-
targets. It was interesting to note that the cis-isomer
(15i) displayed a similar inhibition profile to the trans-
isomer (15j).
Selectivity toward uPA and against tissue-type plasmi-
nogen activator (tPA) can be significantly enhanced
by the introduction of a chlorine or fluorine atom at
the 6-position adjacent to the amidine.10,12 The halogen
substitution displaces a water molecule that is co-bound
with the inhibitor in the vicinity of residue 190. In the
enzymes that possess an Ala190 residue (tPA, factor Xa,
and thrombin) the water molecule is an essential
hydrogen bond partner for the amidine group and its
displacement leads to a marked reduction in affinity
toward these enzymes. The serine hydroxyl group in the
Ser190 enzymes (e.g., uPA and plasmin) can compen-
sate the loss of the water molecule, thereby maintaining
affinity. Comparingthe 6-chlorine and 6-fluorine sub-
stituted analogues (14a–c and 15a–c, respectively) with
their unsubstituted counterparts (3a–c) reveals that
selectivity toward uPA and against the other enzymes
was improved without a loss of affinity toward uPA.
Despite the improved selectivity of 15a–c compared to
3a–c, the selectivity against plasmin was not improved
since plasmin, like uPA, has a Ser190 residue.
The complex of 15j in uPA was solved (Fig. 1b and c)
and compared to that of 15b. The inhibitor 15j binds
more deeply in the S1 pocket than 15b, leadingto the
displacement of two water molecules co-bound in S1.
The short hydrogen bond cluster involving Ser195 is
broken and new short hydrogen bonds form between
the inhibitor oxygens and His57. The markedly different
bindingmodes of 15b compared to 15j is surprising, and
underscores how changes in inhibitor structure can have
significant affects upon the mode of binding.
The 30-alkoxy substituent provides a convenient entry
point into the S10 subsite (Fig. 1a) to try and improve
selectivity against plasmin. The X-ray structure of a
By takingadvantage of small differences between uPA
and other related proteases a series of potent and
Figure 1. (a) uPA–15b complex. The plasmin surface is in red overlaid onto the uPA surface of S10. The isopropyl group extends into S10; (b) uPA–
15j complex; (c) overlay of uPA–15b and uPA–15j complexes. In uPA–15j, the inhibitor has sunk deeper into S1 and displaced two water molecules
that can be seen in the uPA–15b complex.

2022
R.L.Mackman et al./ Bioorg.Med.Chem.Lett.12 (2002) 2019–2022
broadly selective uPA inhibitors have been generated.
Two significant advances that improved potency and
selectivity were the introduction of a fluorine atom
adjacent to the amidine and also the substitution of the
30-ethoxy group for a sterically larger alkyl ether. The
inhibitors produced will likely provide unique opportu-
nities for studyingthe role of selective uPA inhibitors in
vitro and in vivo environments.
Research, New Orleans, LA; Mar 24–28, 2001; Abstract 5074.
(e) Xing, R. H.; Mazar, A.; Henkin, J.; Rabbani, S. A. Cancer
Res. 1997, 57, 3585. (f) Evans, D. M.; Sloan-Stakleff, K. D.
Invasion Metasasis 1998/1999, 18, 252.
6. Magill, C.; Katz, B. A.; Mackman, R. L. Emerg.Ther.
Targets 1999, 3, 109.
7. (a) Verner, E.; Katz, B. A.; Spencer, J. R.; Allen, D.;
Hataye, J.; Hruzewicz, W.; Hui, H. C.; Kolesnikov, A.; Li, Y.;
Luong, C.; Martelli, A.; Radika, K.; Rai, R.; She, M.; Shra-
der, W.; Sprengeler, P. A.; Trapp, S.; Wang, J.; Young, W. B.;
Mackman, R. L. J.Med.Chem.
2001, 44, 2753. (b) Katz,
Acknowledgements
B. A.; Elrod, K. C.; Luong, C.; Rice, M.; Mackman, R. L.;
Sprengeler, P. A.; Spencer, J. R.; Hataye, J.; Janc, J.; Link, J.;
Litvak, J.; Rai, R.; Rice, K.; Sideris, S.; Verner, E.; Young,
W. B. J.Mol.Biol. 2001, 307, 1451.
The authors would like to acknowledge the support of
Dr. J. Knolle and Dr. M. Green.
8. 3,4-Diaminobenzamidine 7 was prepared as described in ref
7a. Preparation of 2-chloro-3,4-diaminobenzonitrile 8: 2-
chloro-4-fluorobenzonitrile was treated with potassium nitrate
in 18 N sulfuric acid. The nitrated product was then treated
with ammonium hydroxide to yield 2-chloro-4-amino-5-nitro-
benzonitrile. Reduction over Pd/C catalyst under 1 atm hydro-
gen gave the desired diamine. Preparation of 2-fluoro-3,4-
diaminobenzamidine 9: 2,4-difluorobenzamidine was nitrated
usingpotassium nitrate in 18 N sulfuric acid. The product was
treated with ammonia to yield 2-fluoro-4-amino-5-nitrobenzo-
nitrile and then reduced over Pd/C catalyst under 1 atm
hydrogen to yield 2-fluoro-4,5-diaminobenzonitrile. Treatment
with di-tert-butyldicarbonate (2 equiv) was followed by con-
version of the nitrile to the amidine usingstandard proce-
dures.10 Removal of the BOC groups using HCl generated the
2-fluoro-3,4-diaminobenzamidine.
9. (a) Ohira, S. Synth.Commun. 1989, 19, 561. (b) Muller, S.;
Liepold, B.; Roth, G. J.; Bestmann, H. J. Synlett 1996, 6, 521.
10. Mackman, R. L.; Katz, B. A.; Breitenbucher, J. G.; Hui,
H. C.; Verner, E.; Luong, C.; Sprengeler, P. A. J.Med.Chem.
2001, 44, 3856.
11. 4-Aminobenzonitrile was treated with N-iodosuccinimide
in acetic acid to yield 3-iodo-4-aminobenzonitrile. The aryl
halide was then treated with mesyl chloride (2 equiv). The
product was hydrolysed with potassium hydroxide to generate
the N-mesyl-3-iodobenzonitrile 11.
References and Notes
1. (a) Andreasen, P. A.; Kjøller, L.; Christensen, L.; Duffy,
M. J. Int.J.Cancer 1997, 72, 1. (b) Rabbani, S. A.; Xing, R. H.
J.Int.Oncol. 1998, 12, 911. (c) Schmitt, M.; Wilhelm, O.;
Janicke, F.; Magdolen, V.; Reuning, U.; Ohi, H.; Moniwa, N.;
Kobayashi, H.; Weidle, U.; Graeff, H. J.Obstet.Gynaecol.
1995, 21, 151.
2. (a) Ossowski, L.; Russo-Payne, H.; Wilson, E. L. Cancer Res.
1991, 51, 274. (b) Ossowski, L.; Reich, E. Cell 1983, 35, 611.
3. Wilhelm, O.; Weidle, U.; Will, C.; Hohl, S.; Rettenberger,
P.; Brunner, N.; Senekowitch, R.; Schmitt, M.; Graeff, H. In
Molecular and Cellular Biology of Plasminogen Activation;
Cold SpringHarbor, Sep 28–Oct 3, 1993; Abstract 81.
4. Frandsen, T. L.; Holst-Hansen, C.; Nielsen, B. S.; Chris-
tensen, I. J.; Nyengaard, J. R.; Carmeliet, P.; Brunner, N.
Cancer Res. 2001, 61, 532.
5. (a) Rabbani, S. A.; Harakidas, P.; Davidson, D. J.; Henkin,
J.; Mazar, A. P. Int.J.Cancer 1995, 63, 840. (b) Alonso, D. F.;
Tejera, A. M.; Farias, E. F.; Joffe, E. B.; De, K.; Gomez, D. E.
Anticancer Res. 1998, 18, 4499. (c) Magill, C.; Gosink, L.;
Biroc, S.; Tam, B.; Qian, F.; Waldeck, N.; Shrader, B.; Spen-
cer, J.; Martelli, A.; Mackman, R. Abstract of Papers, 92nd
Annual Meetingof the American Association of Cancer
Research, New Orleans, LA; Mar 24–28, 2001; Abstract 3151.
(d) Kantak, S.; Bass, K.; Waldeck, N.; Gosink, L.; Leung, L.;
Neeley, Z.; Verner, E.; Hataye, J.; Mackman, R.; Radika, K.;
Wang, J.; Rajan, S.; Magill, C. Abstract of Papers, 92nd
Annual Meetingof the American Association of Cancer
12. Katz, B. A.; Sprengeler, P. A.; Luong, C.; Verner, E.;
Elrod, K.; Kirtley, M.; Janc, J.; Spencer, J. R.; Breitenbucher,
J. G.; Hui, H. C.; McGee, D.; Allen, D.; Martelli, A.; Mack-
man, R. L. Chem.Biol. 2001, 8, 1107.
13. Wang, X.; Lin, X.; Loy, J. A.; Tang, J.; Zhang, X. C.
Science 1998, 281, 1662.