Development of new cyclic plasmin inhibitors with excellent potency and selectivity

Article abstract of DOI:10.1021/jm3012917

The trypsin-like serine protease plasmin is a target for the development of antifibrinolytic drugs for use in cardiac surgery with cardiopulmonary bypass or organ transplantations to reduce excessive blood loss. The optimization of our recently described substrate-analogue plasmin inhibitors, which were cyclized between their P3 and P2 side chains, provided a new series with improved efficacy and excellent selectivity. The most potent inhibitor 8 binds to plasmin with an inhibition constant of 0.2 nM, whereas K_i values >1 μM were determined for nearly all other tested trypsin-like serine proteases, with the exception of trypsin, which is also inhibited in the nanomolar range. Docking studies revealed a potential binding mode in the widely open active site of plasmin that explains the strong potency and selectivity profile of these inhibitors. The dialkylated piperazine-linker segment contributes to an excellent solubility of all analogues. Based on their overall profile the presented inhibitors are well suited for further development as injectable antifibrinolytic drugs.

Full text of DOI:10.1021/jm3012917

Article
pubs.acs.org/jmc
Development of New Cyclic Plasmin Inhibitors with Excellent
Potency and Selectivity
Sebastian M. Saupe, Stephanie Leubner, Michael Betz, Gerhard Klebe, and Torsten Steinmetzer*
Department of Pharmacy, Institute of Pharmaceutical Chemistry, Philipps University Marburg, Marbacher Weg 6, D-35032 Marburg,
Germany
S
* Supporting Information
ABSTRACT: The trypsin-like serine protease plasmin is a target
for the development of antifibrinolytic drugs for use in cardiac
surgery with cardiopulmonary bypass or organ transplantations to
reduce excessive blood loss. The optimization of our recently
described substrate-analogue plasmin inhibitors, which were
cyclized between their P3 and P2 side chains, provided a new
series with improved efficacy and excellent selectivity. The most
potent inhibitor 8 binds to plasmin with an inhibition constant of
0.2 nM, whereas Ki values >1 μM were determined for nearly all
other tested trypsin-like serine proteases, with the exception of
trypsin, which is also inhibited in the nanomolar range. Docking
studies revealed a potential binding mode in the widely open
active site of plasmin that explains the strong potency and
selectivity profile of these inhibitors. The dialkylated piperazine-linker segment contributes to an excellent solubility of all
analogues. Based on their overall profile the presented inhibitors are well suited for further development as injectable
antifibrinolytic drugs.
was mainly replaced by the lysine analog tranexamic acid and in
a few cases also by ε-aminocaproic acid. Both compounds
inhibit the activation of plasminogen but have no direct
inhibitory effect on active plasmin. In contrast to aprotinin,
tranexamic acid has to be administered in relatively high doses,
which has been linked to convulsive seizures.17,18 It was
proposed that high dose tranexamic acid induces hyper-
excitability and convulsions due to blocking γ-aminobutyric
acid-driven inhibition of the central nervous system.19 There-
fore, the development of new and safe plasmin inhibitors as
replacement for aprotinin, especially for use in cardiac surgery,
is of high interest.
Only a few potent plasmin inhibitors have yet been reported
in literature.20 Among them are tetrapeptide derivatives
containing a C-terminal arginal group,21 derivatives of trans-4-
aminomethylcyclohexanecarboxylic acid,22,23 and cyclic ketone-
based inhibitors.24,25
A series of highly potent plasmin inhibitors was derived from
substrate analogue structures containing a C-terminal 4-
amidinobenzylamide residue.26 One of these derivatives, CU-
2010 (structure not yet disclosed), inhibits plasmin with a Ki
value of 2.2 nM27 and has reached clinical development to
reduce blood loss in cardiac surgery. It was recently reported
that its further development was stopped in phase 2b due to
serious side effects, although no further information has been
INTRODUCTION
■
The ubiquitously expressed trypsin-like serine protease plasmin
is responsible for the dissolution of fibrin clots in blood, where
it is normally regulated by their endogenous inhibitors α2-
antiplasmin and α2-macroglobulin. Its zymogen plasminogen is
primarily released from the liver into the circulation and
becomes mainly activated by the plasminogen activator tPA,
whereas uPA is responsible for the generation of plasmin in the
intercellular space.1,2 Beyond its fibrinolytic function, plasmin
cleaves and activates many other substrates including metal-
loproteases, growth factors, and extracellular matrix proteins,
thereby mediating cell migration, which enables normal
physiological processes, like embryogenesis, wound healing,
and angiogenesis. Otherwise, plasmin also promotes tumor
progression and metastasis, as well as chronic inflammatory
processes.3,4 Inhibitors of plasmin can be clinically used for the
treatment of hyperfibrinolysis, which may occur during cardiac
surgery with cardiopulmonary bypass or organ transplanta-
tions.5,6 Additional applications could be the treatment of heavy
menstrual bleeding7 or some forms of hemophilia. A beneficial
effect of plasmin inhibitors has been also proposed for the
treatment of rare forms of hereditary angioedema8,9 and certain
lymphoid malignancies10,11 or chronic inflammatory processes.4
For many years, the 58 amino acid long Kunitz-type serine
protease inhibitor aprotinin12 was clinically used for the
reduction of perioperative bleeding.13 Due to later reported
side effects, it was withdrawn from the market in 2008,14,15 but
was reintroduced in Canada in 2011.16 Meanwhile, aprotinin
Received: September 7, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm3012917 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Scheme 1. Structures of Recently Described Cyclic Plasmin Inhibitors32
announced for the exact problems that occurred.28 It should be
noted that CU-2010 has only limited selectivity and also
inhibits plasma kallikrein, factor Xa, and factor XIa in the
nanomolar range,27 which could be a potential disadvantage, as
discussed recently.20 Although there exist many examples for
approved nonselective drugs, for example, in the field of kinase
inhibitors,29 a high target selectivity is normally a primary
objective in drug development.30 However, the design of
selective substrate analogue inhibitors for the approximately 70
different trypsin-like serine proteases is a challenging task. All of
them have a similar S1 pocket and enable recognition of
substrate analogue structures by highly conserved backbone
interactions to the P1 and P3 residues.31
Starting from linear substrate-analogue 4-amidinobenzyla-
mide derivatives, we have recently reported a structure-based
development of cyclized plasmin inhibitors with significantly
improved selectivity.32 The cyclization was performed between
the side chains of the D-configured P3 and L-configured P2
amino acids via amide bond couplings, metathesis, or click
chemistry. The most potent analog 1 (Scheme 1) of this series
inhibits plasmin and plasma kallikrein (PK) with Ki values of
0.8 and 2.4 nM, respectively, whereas it has negligible activity
against the related proteases thrombin, factor Xa (fXa), and
activated protein C (aPC). Molecular modeling revealed that
the rigid cyclic ring structure of compound 1 prevents its
binding to thrombin, factor Xa, and activated protein C due to
a steric repulsion from an active site loop next to the amino acid
residue 99 (chymotrypsinogen numbering) present in these
proteases. In contrast, the ring segment of inhibitor 1 fits well
into the widely open active site of plasmin, which is missing this
99-loop. The deletion of the segment 95−100 is a unique
structural feature of plasmin that results in a direct connection
of residue 94 to amino acid 101, which is named as 94-shunt.33
PK contains a glycine in position 99 resulting in a sterically less
crowded 99-loop, which therefore still allows an unperturbed
binding of inhibitor 1.
drug, it possesses only a limited solubility of approximately
0.2 mg/mL in 0.9% saline. In addition, its synthesis requires the
preparation of the potentially hazardous intermediate 1,3-
bis(azidomethyl)benzene, which was used for the ring closure
reaction under microwave conditions. Better soluble analogues
of this first series, like 2 and 3 containing a bis-alkylated
piperazine linker, suffered from reduced potency and
selectivity.32
Therefore, a new series of well-soluble cyclized plasmin
inhibitors was developed. All analogues were tested against a
panel of related trypsin-like serine proteases, and it was found
that some of the most potent inhibitors possess an excellent
selectivity for plasmin. The modeled binding mode of these
analogues in the active site of plasmin reveals an explanation for
their high specificity compared with other trypsin-like serine
proteases. In this paper, we describe the synthesis and
characterization of these new analogues.
RESULTS
■
Synthesis. Many studies with related inhibitors of various
trypsin-like serine proteases, like thrombin,34,35 factor Xa,36,37
factor VIIa,38,39 and uPA,40 emphasized the important
contribution of the P1 amidinobenzylamide and P4 benzylsul-
fonyl group for a high inhibitory potency. Therefore, both
anchor residues were maintained, and we have focused only on
the optimization of the cyclic ring structure. A bis-alkylated
piperazine derivative was generally used for connecting the P3
and P2 residues, which contributes to an excellent solubility of
all inhibitors. Replacement of the various amidomethylenephe-
nylalanines in P3 and P2 position of inhibitors 2 and 3 by the
shorter amidophenylalanine residues provided analogs 4−14,
which differ only in the position of the amide group on the P2
and P3 phenyl rings and the used bis-alkylated piperazine
segment (Table 1).
Scheme 2 summarizes the synthesis strategy for the
preparation of inhibitors 4 and 8. The synthesis started from
H-D-Phe(NO2)-OH, which was treated with benzylsulfonyl
chloride. Intermediate 18 was coupled with commercially
Despite its excellent potency against plasmin and PK,
inhibitor 1 suffers from two drawbacks. For use as an injectable
B
dx.doi.org/10.1021/jm3012917 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 1. Structures of the Synthesized Cyclic Inhibitors
a
b
These compounds have been previously published; their Ki values are provided as reference.32 Porcine trypsin; all other enzymes are from human
c
source (see Supporting Information). Due to the incorporation of the trans-2,5-dimethylpiperazine, this compound should exist in two
configuration isomers; however, only a single peak was observed for the final inhibitor in HPLC.
available H-Phe(NO2)-OMe, followed by reduction of both
nitro groups, for example, by treatment with zinc dust in 90%
acetic acid. Alternatively, this reduction could also be
performed with similar yields by hydrogenation using Pd/C
as catalyst in 90% acetic acid or by transfer hydrogenation using
ammonium formate.41 The obtained bis-amine 19 was cyclized
with the piperazine derivatives 20 or 21, followed by
saponification of the methyl ester. The obtained intermediates
22 and 23 were coupled with 4-amidinobenyzlamine·2HCl
(24)42 yielding inhibitors 4 and 8.
Inhibitors 15−17 (Table 1) were prepared as exemplarily
described for compound 15 in Scheme 3. All analogs contain a
D-4-amidomethylenephenylalanyl group in P3 position, a 4-
amidophenylalanyl as P2 residue but differ in the used
piperazine linker. The synthesis started from intermediate
25,32 which was coupled to commercially available H-
Phe(NH2)-OMe, followed by removal of the Cbz protection
C
dx.doi.org/10.1021/jm3012917 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 2. Synthesis of Inhibitors 4 and 8
a
(a) (i) 1 equiv of 18, 1 equiv of H-Phe(4-NO2)-OMe·HCl, 1 equiv of PyBOP, 2 equiv of DIPEA, DMF, 0 °C; (ii) zinc powder, 90% AcOH, room
temp, preparative HPLC, 80% yield (two steps); (b) (i) 1 equiv of 19, 1 equiv of 20 or 21, 2 equiv of PyBOP, 4 equiv of DIPEA, DMF, 2 h 0 °C,
room temp overnight; (ii) 1 N NaOH, EtOH/water, room temp, preparative HPLC, 17.9% yield (22), 12.0% yield (23) (two steps); (c) 1 equiv of
22 or 23, 3 equiv of NMM, 1 equiv of isobutyl chloroformate, DMF, −15 °C, 10 min, followed by 1.5 equiv of 24 and 1 equiv of NMM, 1 h at −15
°C, room temp overnight, preparative HPLC, 32.9% yield (4), 38.3% yield (8).
a
Scheme 3. Synthesis of Inhibitor 15
a
(a) (i) 1 equiv of 25, 1 equiv of H-Phe(4-NH2)-OMe·HCl, 1 equiv of PyBOP, 2 equiv of DIPEA, DMF, 0 °C; (ii) 32% HBr in acetic acid,
precipitation by diethyl ether, preparative HPLC, 30.5% yield (two steps); (b) (i) 1 equiv of 26, 1 equiv of 20, 2 equiv of PyBOP, 5 equiv of DIPEA,
DMF, 0 °C; (ii) 1 N NaOH, EtOH/water, room temp, preparative HPLC, 29.6% yield (two steps); (c) 1 equiv of 27, 3 equiv of NMM, 1 equiv of
isobutyl chloroformate, DMF, −15 °C, 10 min, followed by 1.5 equiv of 24 and 1 equiv of NMM, 1 h at −15 °C, room temp overnight, preparative
HPLC 32.8% yield.
with 32% HBr in acetic acid. The following steps were
performed as described above in Scheme 2. The intermediate
26 was cyclized with the appropriate piperazine derivatives,
followed by saponification and coupling of 4-amidinobenzyla-
mine·2HCl.
The piperazine dicarbonic acids 20 and 21, and the trans-2,5-
dimethylpiperazine diacetic acid (see Supporting Information)
were obtained by alkylation of the appropriate piperazine
derivatives with bromoacetic or 3-bromoproponic acid under
basic conditions. After acidification with concentrated HCl, the
products precipitated and could be isolated by filtration, as
described previously.43,44 Surprisingly, no precipitation oc-
curred after the analogous alkylation of the homopiperazine
and cis-2,5-dimethylpiperazine45 with bromoacetic acid, fol-
lowed by acidification. This prohibited a simple purification of
these very hydrophilic intermediates and required a more
D
dx.doi.org/10.1021/jm3012917 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
laborious synthesis of these dialkylated piperazine analogs. In
that case, benzyl or tert-butyl bromoacetate was used for the
alkylation of these piperazine derivatives resulting in more
hydrophobic intermediates, which could be purified by reversed
phase HPLC. Hydrogenation of the benzyl ester or acidic
cleavage of the tert-butyl ester provided the appropriate
products, which were isolated by precipitation in diethyl
ether (see Supporting Information).
Kinetic Measurements. The inhibition constants of all
newly synthesized inhibitors 4−17 are summarized in Table 1.
The Ki values for compounds 1−3 from our first series32 are
provided as reference (Table 1).
Table 2. Extended Selectivity Studies with Inhibitors 4 and 8
a
Ki (nM)
protease
inhibitor 4
inhibitor 8
urinary kallikrein
human C1s
>50000
>100000
>50000
4700
ca. 10000
>35000
>35000
3370
human C1r
human FXIa
human FXIIa
human FVIIa/TF
human FIXa
human tPA
>100000
>30000
>100000
>100000
8900
>20000
>50000
>5000
>40000
ca. 6500
8.4
human uPA
Inhibitor 4, a direct analog of compound 2 missing two
methylene spacer groups in the ring structure, was found to be
an excellent plasmin inhibitor with a Ki of 0.68 nM. It exhibits
further improved selectivity compared with compound 1, which
was also a strong inhibitor of PK and retained some affinity for
fXa. In contrast, all of these newly prepared compounds inhibit
trypsin more strongly than analog 1. However, although trypsin
was used as an additional reference enzyme in our study, it is
not a relevant protease in blood. The analogous inhibitors 5−7
contain the identical diacetyl-piperazine linker but differ in the
position of the amide substitution on both phenyl groups in the
P2 and P3 residues. These three analogs had reduced selectivity
and potency against plasmin suggesting that we should
maintain the amide substitution in para position at both
phenyl rings.
The incorporation of the longer dipropionyl-piperazine in
inhibitor 8 further improved the affinity against plasmin and
retained the excellent selectivity. Also in that case the para
amide substitution on the P2 and P3 phenyl rings is preferred,
both analogues 9 and 10 have reduced selectivity and potency
as plasmin inhibitors. The diacetyl-homopiperazine inhibitor 11
has a very similar inhibition profile, as found for its direct
analogue 4, whereas the dipropionyl-homopiperazine 12 has
reduced potency compared with the piperazine derivative 8. No
benefit was observed after incorporation of the trans- and cis-
2,5-dimethyl-piperazine linker in inhibitors 13 and 14. Analogs
15−17 contain an amidomethylene substitution in para
position of the P3 phenyl ring. Inhibitors 15 and 16 have a
comparable inhibitory potency against plasmin as found for
their direct analogs 4 and 11, but a marginally reduced
selectivity. The incorporation of the cis-2,5-dimethyl-piperazine
in inhibitor 17 provided no advantage, similar to what was
already observed for compound 14.
human trypsin
44
a
Data provided by the Medicines Company Leipzig GmbH.
a
Table 3. Further Characterization of Inhibitors 4 and 8
inhibitor 4 inhibitor 8
thermodynamic solubility in 0.9% saline [mg/mL]
>2
>2
tPA-induced fibrinolysis in human plasma: EC200
280
180
b
for clot-to-lysis time [nM]
plasma clotting times in human plasma (EC200
)
b
[μM]
(a) PT
>100
>100
>100
7
>100
>100
>100
4
(b) aPTT
(c) TT
c
hERG [% binding at 10 μM]
c
Na channel [% binding at 10 μM]
−8
27
c
L-type Ca channels [% binding at 10 μM]
(a) benzothiazepine
(b) dihydropyridine
(c) phenylalkylamine
1
−4
−5
3
−2
13
61
0
human plasma protein binding [% free fraction]
67
2
stability in human liver microsomes [% degradation
in 60 min]
c
CYP-450 inhibition, IC50 [μM]
3A4
2D6
2C9
1A2
2C19
>22
>10
>40
>10
>100
>100
>10
>100
>100
>100
a
b
Data provided by the Medicines Company Leipzig GmbH. The
assays were described previously.46 Assays were performed at MDS
Pharma services, Taipei; the used protocols are provided in the
Supporting Information (negative values reveal no binding).
c
Further Characterization of Inhibitors 4 and 8. The
detected inhibitory profile revealed that several derivatives of
the current series are excellent and highly selective plasmin
inhibitors. Among several suitable candidates the chemically
well-accessible compounds 4 and 8 were selected for extended
selectivity studies with additional trypsin-like serine proteases
(Table 2). Both inhibitors possess a negligible affinity against
most of the other proteases confirming the hypothesis that the
rigid cyclic inhibitor structure might be repelled by their 99-
loop. However, both analogues inhibit human trypsin slightly
more strongly than the porcine enzyme used for the data
shown in Table 1.
Additional tests were performed with compounds 4 and 8
(Table 3), which possess the most suitable overall profile (data
provided by the Medicines Company Leipzig GmbH, whereby
the assays were performed as described previously).46 Due to
the piperazine linker, the inhibitors possess an excellent
solubility in 0.9% saline, which is a prerequisite for an injectable
drug. Both compounds efficiently inhibit the tPA-induced
fibrinolysis in human plasma.46 The necessary concentrations
for the doubling of the clot to lysis time (EC200)
46 are 280 nM
for inhibitor 4 and 180 nM for compound 8. Due to their
negligible inhibitory effect on the corresponding clotting
proteases, both inhibitors showed no anticoagulant activity in
plasma. EC200 values >100 μM were determined for the
doubling of clotting times in the thrombin time (TT), activated
partial thromboplastin time (aPTT), and prothrombin time
(PT) assays.46 At a concentration of 10 μM both compounds
showed no significant binding to the tested hERG, natrium, and
L-type calcium ion channels (these assays were performed at
MDS Pharma services, Taipei, Taiwan, now Ricerca Bio-
sciences; the used protocols are provided in the Supporting
Information). A 60 min incubation with human liver micro-
somes revealed negligible metabolization of both compounds.
E
dx.doi.org/10.1021/jm3012917 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 1. Stereoview of the modeled binding modes of inhibitors 4 (with green carbon atoms) and 8 (yellow carbon atoms) in plasmin (PDB code
1bui); the coordinate files are provided in pdb format in the Supporting Information.33 Both inhibitors are shown in their singly protonated form at
F
dx.doi.org/10.1021/jm3012917 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 1. continued
the piperazine moiety; the position of the protonation is indicated. (a) Overlay of both inhibitor structures in the active site of plasmin, shown with
its solvent-accessible surface color-coded by atom types (carbon in white, oxygen red, nitrogen blue, sulfur yellow). The overall binding mode reveals
that the larger macrocycle of inhibitor 8 binds closer to the plasmin surface. (b) Stereoview of inhibitor 4 in plasmin showing potential hydrogen
bonds. The residues of the catalytic triad, shown with orange carbon atoms, are not involved in any interaction with the inhibitor. (c) Analogous
stereoview of inhibitor 8 in plasmin; the residues involved in hydrogen bonds to the inhibitor are labeled.51
Figure 2. Stereoview of the modeled structure of inhibitors 4 (with green carbon atoms) and 8 (yellow carbons) in complex with plasmin (PDB
code 1bui). The protein is shown with its translucent solvent-accessible surface in gray. The plasmin backbone is also colored in gray, whereby its
segment Arg94-Lys101-Asp102 is provided as red tube in the upper right corner. These plasmin−inhibitor complexes were overlaid with the crystal
structures of aPC (1aut), fXa (2pr3), PK (2any), and thrombin (1k22), whereby of these proteases only their backbones between residues 94 and
102 are indicated as turquoise tubes. In these orientations, the loops will induce a major sterical clash with the macrocyclic ring structures of both
inhibitors and supposedly prohibit their proper binding. The analogous loop of trypsin (2ra3) is shown in magenta and located at a very similar
position.
All IC50 values for the inhibition of various cytochrome P450
enzymes were found to be >10 μM.
third case only one of them was protonated. The unlikely
double protonated form was omitted.
The docked structures shown in Figure 1 contain a singly
protonated piperazine moiety in the macrocycle and were
selected based on similar backbone interactions known from
related complexes of noncyclized substrate analogue inhibitors,
for example, with thrombin, factor Xa, factor VIIa, trypsin, or
urokinase. In addition, we have also obtained very similar
inhibitor conformations for analogs with unprotonated macro-
cycles. Identical hydrogen bonds were found in both
complexes. The P1-amidino group forms the characteristic
salt bridge to Asp189 and hydrogen bonds to the carbonyl
group of Gly219 and the side chain oxygen of Ser190, whereas
the amide NH of the P1 residue binds to the carbonyl oxygen
of Ser214. A H-bond is also suggested between the P2 carbonyl
oxygen of both inhibitors and the side chain amide of Gln192
(2.4 and 2.5 Å). In addition, the typical antiparallel β-sheet
interaction between Gly216 and the P3 backbone can be found,
and one sulfonyl oxygen is involved in a hydrogen bond to the
NH of Gly219. The comparison of both inhibitor conforma-
tions reveals that the longer and therefore more flexible cycle of
inhibitor 8 adapts better to the plasmin surface allowing for
stronger van der Waals contacts, whereby the piperazine linker
segment of analog 4 is directed more toward the solvent.
Docking of Inhibitors 4 and 8 in Plasmin. Inhibitors 4
and 8 were docked into the active site of plasmin using FlexX
considering different protonation states of the piperazine
moiety in the linker segment. In previous experimental studies
with N,N′-bisalkylated piperazine derivatives, like N,N′-
piperazine-dicarbonic acid,47 N,N′-piperazine-diacetic acid,48
and N,N′-dimethylpiperazine,49 it was found that only one of
the nitrogens is protonated at physiological pH. Additional pKa
calculations with 3,3′-(piperazine-1,4-diyl)bis(N-phenylpropa-
namide), which is identical to the linker segment of inhibitor 8,
were performed with the Marvin Suite50 and also provided a
singly protonated piperazine as major microspecies at
physiological pH 7.4 and also at pH 8.0, as used for the kinetic
measurements (calculated pKa values of 8.54 and 3.01 for the
nitrogens). In contrast, an unprotonated piperazine was
obtained as major microspecies for the shorter 2,2′-(piper-
azine-1,4-diyl)bis(N-phenylacetamide), which is part of the
linker in inhibitor 4 (calculated pKa values 6.08 and 0.58). Due
to the uncertainty of the piperazine protonation, three
possibilities were considered for docking. In the first case,
both nitrogens were unprotonated, whereas in the second and
G
dx.doi.org/10.1021/jm3012917 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
However, we could not detect any specific hydrogen bond
between the linker segment and plasmin.
the cyclization performs slightly better using the shorter
diacetyl-piperazine 20 compared with the coupling of the
longer analog 21.
The generated docking solutions of both inhibitors in
plasmin were also overlaid with crystal structures of PK,
thrombin, fXa, aPC, and trypsin. In Figure 2, plasmin is shown
with its transparent solvent-accessible surface in gray and its
specific 94-shunt as red tube. From the other proteases only
their loop segment 94−102 is shown as tube colored in
turquoise (PK, thrombin, fXa, aPC) or magenta (trypsin).
These segments protrude from the plasmin surface, because
this loop is completely missing in plasmin. Following this
model, a sterical clash between these loops and the piperazine
portion of the ligand’s macrocycle should be generated, which
serves as a reasonable explanation for the observed excellent
selectivity profile against most of the trypsin-like serine
proteases. However, an exception is trypsin; although it
contains a very similar 99 loop, it is still potently inhibited by
this type of macrocylic inhibitor.
The modeled structures of inhibitors 4 and 8 in plasmin
provide a plausible binding mode explaining their high affinity.
All known interactions of the P1 4-amidinobenzylamide at the
bottom of the S1 pocket and to the carbonyl oxygen of Ser214
at the entrance of this binding site can be accomplished. The
stronger affinity of inhibitor 8 compared with 4 might be
explained by more extensive van der Waals contacts of the
macrocyclic ring structure, which is oriented slightly closer
toward the plasmin surface. The modeled complexes of
inhibitors 4 and 8 in plasmin, superimposed with the X-ray
structures of the other proteases, reveal a sterical clash with
their active site loops around the amino acid 99 explaining the
observed selectivity profile. An exception is trypsin, which is
still significantly inhibited by both compounds. The digestive
protease trypsin is mainly located in the intestinal tract and
normally should not come in direct contact with an injectable
plasmin inhibitor, although minor levels of trypsinogen and
trypsin have been detected also in other organs, in body fluids,
or in cancer cells, especially in pancreatitis patients.57 Clinical
studies with the thrombin inhibitor melagatran,58 which also
inhibits trypsin with a Ki value of 4 nM,59 revealed that an
inhibitory potency against trypsin is obviously no disadvantage
for an injectable drug. It should be noted that the trypsin used
for the inhibition studies shown in Table 2 is isolated from
human pancreas (Calbiochem, CAS 9002-07-7, PN 650275-
50UG), and this material is not only one single protein. Native
trypsin consists of a mixture of approximately two-thirds
cationic trypsin (trypsin-1), one-third anionic trypsin (trypsin-
2), and a few percent of mesotrypsin (trypsin-3), including
isoforms.57 However, human trypsin-1, -2, and -3 contain a
normal loop between their residues 94−100 with leucine in
position 99 and without any amino acid deletion, which
consists of the sequences YDRKTLN, YNSRTLD, and
YNRDTLD, respectively. At present, we can only speculate
that the 99-loop in trypsin is more flexible than those in other
proteases and still enables a significant binding of inhibitors 4
and 8, although both compounds are more than 40-fold
stronger plasmin inhibitors. A higher flexibility of the 99-loop
could be also the reason that trypsin is a relatively nonselective
protease degrading most substrates as long as they contain a
basic P1 residue. However, the comparison of several crystal
structures of trypsin-like serine proteases, which were used in
our selectivity studies, did not reveal higher B-factors for the
99-loop of trypsin (see Figure S1 in the Supporting
Information). Admittedly, the analysis of the B-factors of
individual crystal structures is certainly not a sufficient criterion
to characterize the loop flexibility in proteins.
DISCUSSION
■
The results from enzyme kinetic analysis with these new
derivatives supports our previous hypothesis that it is possible
to develop selective substrate analog plasmin inhibitors by
appropriate macrocyclization between the side chains of a P2
amino acid in the L-configuration and a D-configured P3
residue. The newly developed cyclization variant, introduced
into inhibitors 4 and 8, starts directly with an amide
substitution in the para position of the P2 and P3 phenyl
rings and provides convenient chemical access to highly potent
plasmin inhibitors. They exhibit an improved selectivity profile
compared with our first series, particularly a reduced affinity
against PK.32 Although a beneficial anti-inflammatory effect of
aprotinin during cardiac surgery due to its PK inhibitory activity
(Ki 30 nM) was described in several publications,52,53 there are
also controversial reports that an inhibition of the kallikrein−
kinin system leads to a loss of the protective effects of kinins
against ischemia-reperfusion injury.54 This is considered to be
one of the reasons for the increased number of myocardial
infarction and stroke after application of aprotinin, which does
not achieve satisfactory selectivity discrimination between
plasmin and PK. Locally increased bradykinin levels, which
would be the consequence of a reduced PK inhibition, result in
an increased coronary blood flow, and have been reported to
play a beneficial role in cardiac diseases.55 Based on this
concept, inhibitors of the lysosomal serine carboxypeptidase
cathepsin A, a degrading protease of bradykinin, have been very
recently described for the potential treatment of cardiovascular
diseases.56 Despite controversial results regarding the effects of
differences in the bradykinin levels due to variations in the
enzymatic activity of bradykinin generating and degrading
proteases, a high specificity of a drug should generally minimize
undesired side effects. A molecular scaffold with convincing
discriminative power between similar proteases should be an
advantage for further development.
In conclusion, we could develop a series of new cyclic
plasmin inhibitors and have found an excellent overall profile
for inhibitors 4 and 8. In addition to their strong inhibitory
potency on plasmin and high selectivity over related trypsin-like
serine proteases, both compounds possess a strong anti-
fibrinolytic activity in plasma, excellent solubility, and a low
affinity to the tested ion channels and cytochrome P450
enzymes. These properties should justify their further
characterization in animal models for the treatment of
pathological conditions associated with increased plasmin
activities, especially for the reduction of blood loss during
cardiac surgery or as antidote during thrombolytic therapy.
Another major advantage of the new macrocyclic inhibitors is
their straightforward chemical accessibility. The synthesis of
these compounds does not require any hazardous azide
intermediates, reactions under drastic microwave conditions,
or application of special and costly catalysts, which are
necessary for click chemistry and metathesis reaction steps
during the preparation of our previous series. However, the
yield-limiting cyclization step still requires further optimization.
Several small scale reactions using intermediate 19 revealed that
H
dx.doi.org/10.1021/jm3012917 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
min, start at 10% B (purity, 88.7%). MS (ESI, positive): calcd, 570.57;
m/z, 571.23 [M + H]+.
EXPERIMENTAL SECTION
■
General. Reagents and solvents were purchased from Aldrich
Chemical, Acros Organics, Alfa Aesar, Fluka, Iris Biotech, PepTech, or
Polypeptide. The amino acid derivatives H-Phe(4-NO2)-OH, H-D-
Phe(4-NO2)-OH, and H-Phe(4-NH2)-OH were purchased from
PepTech and Bachem.
Analytical HPLC experiments were performed on a Shimadzu LC-
10A system (column, NUCLEODUR C18 ec, 5 μm, 100 Å, 4.6 mm ×
The intermediate (3.5 g, approximately 4.87 mmol, 1.0 equiv) was
dissolved in 500 mL of 90% AcOH, and after addition of zinc dust (2.6
g, 39.8 mmol, 8.0 equiv), the suspension was stirred for 5 h. After
filtration, the solvent was evaporated. The remaining residue was
suspended in 50 mL of 90% MeCN, the precipitated zinc acetate was
removed by centrifugation, and the supernatant was evaporated. The
remaining residue was purified by preparative HPLC and lyophilized
from 80% tert-butanol/water. Yield: 2.02 g (80.0%) of a yellow
250 mm, Machery-Nagel, Duren, Germany). As eluents were used
̈
1
water (A) and acetonitrile (B) both containing 0.1% TFA at a flow rate
of 1 mL/min, a linear gradient (1% B/min) and detection at 220 nm.
The indicated purity for all intermediates is based on HPLC detection
at 220 nm. The final inhibitors were purified to >95% purity
(detection at 220 nm) by preparative HPLC (pumps, Varian PrepStar
model 218 gradient system; detector, ProStar model 320; fraction
collector, Varian model 701) using a C18 column (ProntoSIL C18 SH,
5 μm, 120 Å, 32 mm × 250 mm, Bischoff Chromatography) and a
linear gradient of acetonitrile/water containing 0.1% TFA at a flow rate
of 20 mL/min. All inhibitors were finally obtained as TFA salts after
lyophilization. The molecular mass of the synthesized compounds was
determined using a QTrap 2000 ESI spectrometer (Applied
Biosystems). The 1H and 13C NMR spectra were recorded on an
ECX-400 instrument (Jeol Inc.) and are referenced to internal solvent
signals.
lyophilized solid. HPLC: 17.4 min, start at 10% B (purity, 93.5%). H
NMR (400 MHz, DMSO-d6): δ = 8.68 (d, J = 7.79 Hz, 1 H), 7.53 (d, J
= 9.16 Hz, 1 H), 7.22−7.31 (m, 4 H), 7.18 (dd, J = 3.66, 7.33 Hz, 2
H), 7.14 (d, J = 4.35 Hz, 2 H), 7.12 (s, 1 H), 6.99 (d, J = 8.47 Hz, 2
H), 6.90 (d, J = 8.24 Hz, 2 H), 4.46 (dt, J = 5.72, 8.24 Hz, 2 H), 4.09
(dt, J = 4.47, 9.45 Hz, 2 H), 3.90 (dd, J = 13.50, 53.00 Hz, 4 H), 3.53−
3.58 (m, 3 H), 2.97 (dd, J = 5.30, 13.70 Hz, 1 H), 2.81 (dd, J = 9.16,
13.74 Hz, 1 H), 2.63 (dd, J = 4.60, 13.70 Hz, 1 H), 2.54 ppm (dd, J =
9.85, 13.51 Hz, 1 H). 13C NMR (101 MHz, DMSO-d6): δ = 171.7,
171.1, 158.8, 158.5, 134.9, 134.7, 133.7, 130.7, 130.3, 130.0, 128.2,
127.9, 121.0, 120.6, 58.5, 57.8, 53.5, 52.0, 37.9, 36.2 ppm. MS (ESI,
positive): calcd, 510.61; m/z, 511.27 [M + H]+.
2,2′-(Piperazine-1,4-diyl)diacetic Acid (20). The synthesis was
performed as described previously.43
3,3′-(Piperazine-1,4-diyl)dipropanoic Acid (21). Piperazine·6-
H2O (5.0 g, 25.7 mmol, 1.0 equiv) was dissolved in 140 mL of 10%
NaOH, treated with 3-bromopropionic acid (8.08 g, 52.8 mmol, 2.05
equiv) and stirred overnight at room temperature. The mixture was
acidified with 37% HCl until the product started to crystallize. The
flask was kept at 4 °C overnight, and the product was obtained by
filtration, washed with a small amount of water, and dried in vacuo.
Bzls-D-Phe(4-NO2)-OH (18). H-D-Phe(4-NO2)-OH (4.93 g, 23.5
mmol, 1.0 equiv) was suspended in 56 mL of dry DCM and treated
with TMS-Cl (6.5 mL, 52.4 mmol, 2.2 equiv) and DIPEA (9.1 mL,
52.4 mmol, 2.2 equiv). The mixture was refluxed for 1 h. Then,
benzylsulfonyl chloride (5.02 g, 26.3 mmol, 1.1 equiv) was added in
several portions within 35 min at 0 °C, and the pH was maintained at
8−9 by addition of DIPEA (4.6 mL, 26.2 mmol, 1.1 equiv). The
mixture was stirred for 1 h at 0 °C and 2 h at room temperature. The
solvent was removed in vacuo, and the brown residue was dissolved in
a mixture of 5% KHSO4 and EtOAc. The solution was extracted twice
with EtOAc. The combined organic phase was washed thrice with 5%
KHSO4 and thrice with brine. The organic layer was dried over
MgSO4 and filtered, and the solvent was removed in vacuo.
The remaining residue was dissolved in 250 mL of ethyl acetate and
dropwise treated with dicyclohexylamine (7.14 mL, 35.7 mmol, 1.5
equiv). The dicyclohexylammonium salt crystallized at 4 °C and was
isolated by filtration. The crystals were washed with EtOAc and
diethylether and were dried in vacuo (yield, 7.94 g (62.1%) of beige
DCHA-salt crystals).
1
Yield: 4.58 g (77.4%) of light yellow crystals. H NMR (400 MHz,
D2O): δ = 3.71 (br. s., 8 H), 3.56 (t, J = 6.87 Hz, 4 H), 2.90 ppm (t, J
= 6.90 Hz, 4 H). 13C NMR (101 MHz, D2O): δ = 173.3, 52.2, 48.6,
28.4 ppm. MS (ESI, negative): calcd, 230.13; m/z, 229.17 [M − H]−.
Compound 22. Bzls-D-Phe(4-NH2)-Phe(4-NH2)-OMe (19) (201
mg, 0.394 mmol, 1.0 equiv) and 2,2′-(piperazine-1,4-diyl)diacetic acid
(20) (79.8 mg, 0.395 mmol, 1.0 equiv) were dissolved in 200 mL of
DMF and stirred on an ice bath. After 10 min, PyBOP (410.8 mg,
0.789 mmol, 2.0 equiv) and DIPEA (409 μL, 2.35 mmol, 6.0 equiv)
were added, and the mixture was stirred at pH 8 for 1 h at 0 °C and at
room temperature overnight. The solvent was removed in vacuo, and
the remaining residue was dissolved in a mixture of saturated aqueous
NaHCO3 and EtOAc. The organic layer was washed thrice with
saturated aqueous NaHCO3 and thrice with a small amount of brine,
dried over MgSO4, and filtered, and the solvent was evaporated. The
remaining residue was dissolved in 5.0 mL of 50% 1 N NaOH/ethanol
(v/v). The solution was stirred for 2 h at room temperature and
neutralized with TFA, and the solvent was removed in vacuo. The
residue was purified by preparative HPLC and lyophilized. Yield: 63
mg (17.9%) of a white lyophilized solid. HPLC: 22.6 min, start at 10%
The dicyclohexylammonium salt (2.70 g, 4.95 mmol) was dissolved
in a mixture of 5% KHSO4/EtOAc, and the water phase was extracted
thrice with EtOAc. The combined organic phases were washed once
with brine, dried over MgSO4, and filtered. The solvent was
evaporated. Yield: 1.80 g (99.8%) of brown oil. HPLC: 38.7 min,
1
start at 10% B (purity, 95.3%). H NMR (400 MHz, DMSO-d6): δ =
12.97 (br. s., 1 H), 8.15 (d, J = 8.70 Hz, 2 H), 7.70 (d, J = 8.70 Hz, 1
H), 7.51 (d, J = 8.70 Hz, 2 H), 7.12−7.31 (m, 5 H), 4.10−4.20 (m, 2
H), 4.04−4.10 (m, 1 H), 3.11 (dd, J = 5.50, 13.74 Hz, 1 H), 2.93 ppm
(dd, J = 9.16, 13.74 Hz, 1 H). 13C NMR (101 MHz, DMSO-d6): δ =
172.7, 146.4, 145.5, 130.9, 130.7, 130.0, 128.1, 127.9, 123.2, 58.5, 56.8,
37.5 ppm. MS (ESI, positive): calcd, 364.07; m/z, 382.17 [M + NH4]+,
387.06 [M+Na]+.
1
B (purity, >95%). H NMR (400 MHz, DMSO-d6): δ = 12.75−13.07
(m, 1 H), 9.99−10.40 (m, 2 H), 8.68 (br. s., 1 H), 6.81−7.78 (m, 13
H), 6.43−6.79 (m, 2 H), 4.12−4.59 (m, 5 H), 2.60−3.75 ppm (m, 16
H). 13C NMR (101 MHz, DMSO-d6): δ = 173.1, 170.3, 158.9, 158.6,
136.5, 131.0, 130.4, 129.5, 129.2, 128.3, 128.0, 119.6, 119.2, 118.1,
67.0, 58.5, 56.1, 54.0, 48.9, 31.3 ppm. MS (ESI, positive): calcd,
662.76; m/z, 663.35 [M + H]+.
Bzls-D-Phe(4-NH2)-Phe(4-NH2)-OMe (19). Compound 18 (1.80
g, 4.94 mmol, 1.0 equiv) and H-Phe(4-NO2)-OMe·HCl (1.29 g, 4.95
mmol, 1.0 equiv) were dissolved in 30 mL of DMF and stirred on an
ice bath. After 10 min, PyBOP (2.60 g, 5.00 mmol, 1.0 equiv) and
DIPEA (1.72 mL, 9.89 mmol, 2.0 equiv) were added, and the mixture
was stirred for 1 h at 0 °C and 2 h at room temperature, pH 7−8. The
solvent was removed in vacuo, and the remaining residue was
dissolved in a mixture of 5% KHSO4 and EtOAc. The organic phase
was washed thrice with 5% KHSO4, once with brine, thrice with
saturated aqueous NaHCO3, and thrice with brine. The organic phase
was dried over MgSO4 and filtered, and the solvent was removed in
vacuo. Yield, 3.55 g of a brown solid containing impurity. HPLC: 51.8
Compound 23. According to the synthesis of compound 22, Bzls-
D-Phe(4-NH2)-Phe(4-NH2)-OMe (19) (200 mg, 0.392 mmol, 1.0
equiv), 3,3′-(piperazine-1,4-diyl)dipropanoic acid (21) (90.1 mg,
0.391 mmol, 1.0 equiv), PyBOP (409 mg, 0.786 mmol, 2.0 equiv),
and DIPEA (273 μL, 1.57 mmol, 4.0 equiv) were cyclized in 120 mL
of DMF, followed by saponification in 5.0 mL of 50% 1 N NaOH/
ethanol (v/v) for 3 h at room temperature. The solution was
neutralized with TFA, and the solvent was removed in vacuo. The
residue was purified by preparative HPLC and lyophilized. Yield: 43
mg (12.0%) of a white lyophilized solid. HPLC: 22.5 min, start at 10%
B (purity, 90.8%). 1H NMR (400 MHz, DMSO-d6): δ = 12.81−13.17
(m, 1 H), 9.93 (br. s., 2 H), 8.60 (d, J = 6.41 Hz, 1 H), 6.62−7.68 (m,
I
dx.doi.org/10.1021/jm3012917 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
13 H), 6.48 (br. s., 2 H), 3.80−4.53 (m, 9 H), 2.55−3.57 ppm (m, 16
H). 13C NMR (101 MHz, DMSO-d6): δ = 172.7, 169.5, 169.0, 168.9,
158.6, 158.3, 137.3, 137.0, 130.8, 130.2, 130.1, 129.8, 128.3, 128.0,
119.9, 119.7, 67.0, 58.0, 56.5, 53.9, 48.9, 36.1, 31.3 ppm. MS (ESI,
positive): calcd, 690.28; m/z, 691.24 [M + H]+.
In addition, two optional interactions were set to the side chain of
Asp189, whereby one has to be addressed and the second is optional.
The best ranked poses were subsequently minimized using the
MMFF94x force field implemented in MOE.
4-Amidinobenzylamine·2HCl (24). The synthesis was performed
ASSOCIATED CONTENT
as described previously.42
■
S
* Supporting Information
Inhibitor 4. Compound 22 (54 mg, 69.5 μmol, 1.0 equiv) and
NMM (23.0 μL, 209 μmol, 3.0 equiv) were dissolved in 1.5 mL of
DMF at −15 °C and treated with isobutyl chloroformate (9.0 μL, 69.5
μmol, 1.0 equiv). The mixture was stirred for 10 min at −15 °C and
treated with 4-amidinobenzylamine·2HCl (24 mg, 108 μmol, 1.6
equiv) and NMM (7.7 μL, 69.5 μmol, 1.0 equiv). The suspension was
stirred at −15 °C for 1 h and at room temperature overnight. The
solvent was removed in vacuo; the product was purified by preparative
HPLC and lyophilized. Yield: 26 mg (32.9%) of a white lyophilized
solid. HPLC: 18.1 min, start at 10% B (purity, > 95%).
It contains the synthesis procedures for all other inhibitors,
their analytical data, and information for enzyme kinetics and
the coordinate files of the modeled complexes of inhibitors 4
and 8 in plasmin in pdb format. This material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*Phone: +49 6421 2825900. E-mail: steinmetzer@staff.uni-
marburg.de.
1H NMR (400 MHz, DMSO-d6): δ = 10.19 (br. s, 2 H), 9.22 (br. s.,
2 H), 9.12 (br. s., 2 H), 8.73 (br. s., 2 H), 7.59−7.81 (m, 3 H), 6.74−
7.58 (m, 17 H), 3.94−4.73 (m, 10 H), 2.62−3.32 ppm (m, 12 H). MS
(ESI, positive): calcd, 793.34; m/z 794.51 [M + H]+. TLC: Rf = 0.28.
Inhibitor 8. The synthesis was performed as described for inhibitor
4. The mixed anhydride was prepared from compound 23 (35 mg,
38.1 μmol, 1.0 equiv), NMM (12.6 μL, 114 μmol, 3.0 equiv), and
isobutyl chloroformate (5.0 μL, 38.1 μmol, 1.0 equiv) in 1.5 mL of
DMF at −15 °C. After 10 min, compound 24 (13 mg, 58.5 μmol, 1.5
equiv) and NMM (4.2 μL, 38.1 μmol, 1.0 equiv) were added; the
mixture was stirred 1 h at −15 °C and at room temperature overnight.
The product was purified by preparative HPLC and lyophilized. Yield:
17 mg (38.3%) of a white lyophilized solid. HPLC: 19.8 min, start at
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We wish to thank Peter Steinmetzer and Sven Letschert from
the Medicines Company (Leipzig) GmbH for Ki determi-
nations. Additional data for inhibitors 4 and 8 (clotting assays,
ion channel binding, inhibition of tPA induced fibrinolysis,
plasma protein binding, stability measurements in human liver
microsomes and cytochrome P 450 inhibition) were provided
by The Medicines Company (Leipzig) GmbH. Financial aid by
the Medicines Company (Leipzig) GmbH is acknowledged.
1
10% B (purity, > 95%). H NMR (400 MHz, DMSO-d6): δ = 9.83−
10.01 (m, 2 H), 9.19 (s, 2 H), 8.91 (br. s., 2 H), 8.75−8.82 (m, 1 H),
8.63 (d, J = 7.33 Hz, 1 H), 7.67 (d, J = 8.24 Hz, 2 H), 7.34−7.49 (m, 4
H), 7.20−7.28 (m, 5 H), 7.08 (s, 4 H), 6.41−6.57 (m, 2 H), 6.33 (br.
s., 2 H), 4.33−4.60 (m, 8 H), 4.28 (s, 2 H), 4.18 (s, 1 H), 2.62−3.18
ppm (m, 16 H). MS (ESI, positive): calcd, 821.37; m/z, 822.33 [M +
H]+, 411.61 [M + 2H]2+. TLC: Rf = 0.10.
ABBREVIATIONS USED
■
AMe, aminomethyl; Boc, tert-butyloxycarbonyl; Bzls, benzyl-
sulfonyl; Cbz, benzyloxycarbonyl; DIPEA, N,N-diisopropyle-
thylamine; DCHA, dicyclohexylamine; DCM, dichlorome-
thane; DMF, N,N-dimethylformamide; EtOAc, ethyl acetate;
Me, methyl; MeCN, acetonitrile; MS, mass spectrometry;
NMM, 4-methylmorpholine; Phe(4-AMe), 4-aminomethylphe-
nylalanine; Phe(4-NH2), 4-aminophenylalanine; Phe(4-NO2),
4-nitrophenylalanine; PK, plasma kallikrein; PN, product
number; PyBOP, benzotriazol-1-yl-oxytripyrrolidinophospho-
nium hexafluorophosphate; tBu, tert-butyl; TFA, trifluoroacetic
acid; TMS-Cl, trimethylsilyl chloride
Enzyme Kinetics. The kinetic measurements with human plasmin
(Chromogenix; spec. activity = 13.6 casein units/mg, concentration in
assay of 0.895 nM) were performed in a microplate reader (Multiscan
Ascent, Thermo) at 405 nm using the substrate Tos-Gly-Pro-Lys-pNA
(Chromozym PL, concentrations in assay of 364, 182, and 91 μM).
The Ki values were determined from Dixon plots.60 The inhibition
constants for the reference enzymes PK, thrombin, FXa, and aPC were
performed as described previously (see Supporting Information).61
Molecular Modeling. Preparation of Plasmin. The coordinates
of the protease domain of plasmin taken from the microplasmin−
staphylokinase−microplasmin ternary complex irreversibly inhibited
by H-Glu-Gly-Arg-chloromethyl ketone (PDB code 1bui) were used
for docking. All water molecules and covalent bonds of the inhibitor to
plasmin residues His57 and Ser195 were removed. Protonation of the
plasmin’s binding site was performed by use of the receptor
preparation tool of FlexX.62
Inhibitor Preparation and Docking. Ligands 4 and 8 were
prepared using the MOE63 builder function, whereby three different
protonation states of the piperazine linker have been considered. In
the first case, both nitrogens are unprotonated, and in the second and
third cases only one of them is protonated. For the docking process,
various conformers of the macrocycle were generated using the
conformational search procedures of MOE (conformational search
using LowModeMD,64 stochastically, and systematically search; all
other standard parameters were set to default values). This search
provided a library of 87 and 275 conformers for the macrocycles of
inhibitors 4 and 8, respectively, which were used for docking with
FlexX. These pregenerated macrocycles were treated as rigid objects
during docking.
REFERENCES
■
(1) Vassalli, J. D.; Sappino, A. P.; Belin, D. The plasminogen
activator/plasmin system. J. Clin. Invest. 1991, 88, 1067−1072.
(2) Schaller, J.; Gerber, S. S. The plasmin-antiplasmin system:
Structural and functional aspects. Cell. Mol. Life Sci. 2011, 68, 785−
801.
(3) Castellino, F. J.; Ploplis, V. A. Structure and function of the
plasminogen/plasmin system. Thromb. Haemostasis 2005, 93, 647−
654.
(4) Syrovets, T.; Lunov, O.; Simmet, T. Plasmin as a proin-
flammatory cell activator. J. Leukocyte Biol. 2012, 92, 509−519.
(5) Levy, J. H. Pharmacologic methods to reduce perioperative
bleeding. Transfusion 2008, 48, 31S−38S.
(6) Johansson, P. I.; Ostrowski, S. R.; Secher, N. H. Management of
major blood loss: An update. Acta Anaesthesiol. Scand. 2010, 54, 1039−
1049.
(7) Philipp, C. S. Antifibrinolytics in women with menorrhagia.
The active site of plasmin was defined by taking all residues within a
16 Å sphere around the center of mass of the inhibitor in the crystal
structure. Two interactions between the ligand and the carbonyl
oxygens of Ser214 and Gly216 were set as essential docking restraints.
Thromb. Res. 2011, 127 (Suppl 3), S113−115.
(8) Ritchie, B. C. Protease inhibitors in the treatment of hereditary
angioedema. Transfus. Apheresis Sci. 2003, 29, 259−267.
J
dx.doi.org/10.1021/jm3012917 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(9) Du-Thanh, A.; Raison-Peyron, N.; Drouet, C.; Guillot, B. Efficacy
of tranexamic acid in sporadic idiopathic bradykinin angioedema.
Allergy 2010, 65, 793−795.
(10) Okada, Y.; Tsuda, Y.; Wanaka, K.; Tada, M.; Okamoto, U.;
Okamoto, S.; Hijikata-Okunomiya, A.; Bokonyi, G.; Szende, B.; Keri,
G. Development of plasmin and plasma kallikrein selective inhibitors
and their effect on M1 (melanoma) and HT29 cell lines. Bioorg. Med.
Chem. Lett. 2000, 10, 2217−2221.
(11) Ishihara, M.; Nishida, C.; Tashiro, Y.; Gritli, I.; Rosenkvist, J.;
Koizumi, M.; Okaji, Y.; Yamamoto, R.; Yagita, H.; Okumura, K.;
Nishikori, M.; Wanaka, K.; Tsuda, Y.; Okada, Y.; Nakauchi, H.;
Heissig, B.; Hattori, K. Plasmin inhibitor reduces T-cell lymphoid
tumor growth by suppressing matrix metalloproteinase-9-dependent
CD11b(+)/F4/80(+) myeloid cell recruitment. Leukemia 2012, 26,
332−339.
(12) Ascenzi, P.; Bocedi, A.; Bolognesi, M.; Spallarossa, A.; Coletta,
M.; De Cristofaro, R.; Menegatti, E. The bovine basic pancreatic
trypsin inhibitor (Kunitz inhibitor): A milestone protein. Curr. Protein
Pept. Sci. 2003, 4, 231−251.
(13) Mannucci, P. M.; Levi, M. Prevention and treatment of major
blood loss. New Engl. J. Med. 2007, 356, 2301−2311.
(14) Mangano, D. T.; Tudor, I. C.; Dietzel, C. The risk associated
with aprotinin in cardiac surgery. New Engl. J. Med. 2006, 354, 353−
365.
(15) Schneeweiss, S.; Seeger, J. D.; Landon, J.; Walker, A. M.
Aprotinin during coronary-artery bypass grafting and risk of death.
New Engl. J. Med. 2008, 358, 771−783.
(16) Sniecinski, R. M.; Karkouti, K.; Levy, J. H. Managing clotting: A
North American perspective. Curr. Opin. Anaesthesiol. 2012, 25, 74−
79.
(27) Szabo, G.; Veres, G.; Radovits, T.; Haider, H.; Krieger, N.;
Baehrle, S.; Miesel-Groeschel, C.; Niklisch, S.; Karck, M.; van de
Locht, A. Effects of novel synthetic serine protease inhibitors on
postoperative blood loss, coagulation parameters, and vascular
relaxation after cardiac surgery. J. Thorac. Cardiovasc. Surg. 2010,
139, 181−188.
(28) Mitchell, M. The Medicines Company Discontinues Phase 2b
Trial of MDCO-2010. http://ir.themedicinescompany.com/phoenix.
zhtml?c=122204&p=irol-newsArticle&ID=1741990&highlight= (ac-
cessed 10-31-2012).
(29) Fabian, M. A.; Biggs, W. H., 3rd; Treiber, D. K.; Atteridge, C. E.;
Azimioara, M. D.; Benedetti, M. G.; Carter, T. A.; Ciceri, P.; Edeen, P.
T.; Floyd, M.; Ford, J. M.; Galvin, M.; Gerlach, J. L.; Grotzfeld, R. M.;
Herrgard, S.; Insko, D. E.; Insko, M. A.; Lai, A. G.; Lelias, J. M.; Mehta,
S. A.; Milanov, Z. V.; Velasco, A. M.; Wodicka, L. M.; Patel, H. K.;
Zarrinkar, P. P.; Lockhart, D. J. A small molecule-kinase interaction
map for clinical kinase inhibitors. Nat. Biotechnol. 2005, 23, 329−336.
(30) Huggins, D. J.; Sherman, W.; Tidor, B. Rational approaches to
improving selectivity in drug design. J. Med. Chem. 2012, 55, 1424−
1444.
(31) Perona, J. J.; Craik, C. S. Structural basis of substrate specificity
in the serine proteases. Protein Sci. 1995, 4, 337−360.
(32) Saupe, S. M.; Steinmetzer, T. A new strategy for the
development of highly potent and selective plasmin inhibitors. J.
Med. Chem. 2012, 55, 1171−1180.
(33) Parry, M. A.; Fernandez-Catalan, C.; Bergner, A.; Huber, R.;
Hopfner, K. P.; Schlott, B.; Guhrs, K. H.; Bode, W. The ternary
̈
microplasmin-staphylokinase-microplasmin complex is a proteinase-
cofactor-substrate complex in action. Nat. Struct. Biol. 1998, 5, 917−
923.
(17) Murkin, J. M.; Falter, F.; Granton, J.; Young, B.; Burt, C.; Chu,
M. High-dose tranexamic acid is associated with nonischemic clinical
seizures in cardiac surgical patients. Anesth. Analg. 2010, 110, 350−
353.
(34) Tucker, T. J.; Lumma, W. C.; Mulichak, A. M.; Chen, Z.;
Naylor-Olsen, A. M.; Lewis, S. D.; Lucas, R.; Freidinger, R. M.; Kuo, L.
C. Design of highly potent noncovalent thrombin inhibitors that utilize
a novel lipophilic binding pocket in the thrombin active site. J. Med.
Chem. 1997, 40, 830−832.
(18) Sander, M.; Spies, C. D.; Martiny, V.; Rosenthal, C.; Wernecke,
K.-D.; von Heymann, C. Mortality associated with administration of
high-dose tranexamic acid and aprotinin in primary open-heart
procedures: a retrospective analysis. Crit. Care 2010, 14, R148.
(19) Furtmueller, R.; Schlag, M. G.; Berger, M.; Hopf, R.; Huck, S.;
Sieghart, W.; Redl, H. Tranexamic acid, a widely used antifibrinolytic
agent, causes convulsions by a gamma-aminobutyric acid(A) receptor
antagonistic effect. J. Pharmacol. Exp. Ther. 2002, 301, 168−173.
(20) Swedberg, J. E.; Harris, J. M. Natural and engineered plasmin
inhibitors: Applications and design strategies. ChemBioChem 2012, 13,
336−348.
(21) Swedberg, J. E.; Harris, J. M. Plasmin substrate binding site
cooperativity guides the design of potent peptide aldehyde inhibitors.
Biochemistry 2011, 50, 8454−8462.
(22) Okada, Y.; Tsuda, Y.; Tada, M.; Wanaka, K.; Okamoto, U.;
Hijikata-Okunomiya, A.; Okamoto, S. Development of potent and
selective plasmin and plasma kallikrein inhibitors and studies on the
structure-activity relationship. Chem. Pharm. Bull. 2000, 48, 1964−
1972.
(23) Tsuda, Y.; Tada, M.; Wanaka, K.; Okamoto, U.; Hijikata-
Okunomiya, A.; Okamoto, S.; Okad, Y. Structure-inhibitory activity
relationship of plasmin and plasma kallikrein inhibitors. Chem. Pharm.
Bull. 2001, 49, 1457−1463.
(24) Xue, F.; Seto, C. T. Selective inhibitors of the serine protease
plasmin: probing the S3 and S3′ subsites using a combinatorial library.
J. Med. Chem. 2005, 48, 6908−6917.
(25) Xue, F.; Seto, C. T. Structure-activity studies of cyclic ketone
inhibitors of the serine protease plasmin: design, synthesis, and
biological activity. Bioorg. Med. Chem. 2006, 14, 8467−8487.
(26) Steinmetzer, T.; Schweinitz, A.; Stuerzebecher, J.; Steinmetzer,
P.; Soffing, A.; Van de Locht, A.; Nicklisch, S.; Reichelt, C.; Ludwig, F.-
A.; Schulze, A.; Daghisch, M.; Heinicke, J. Trypsin-like serine protease
inhibitors, and their preparation and use. U.S. Pat. Appl. 20100022781,
2010.
(35) Wiley, M. R.; Chirgadze, N. Y.; Clawson, D. K.; Craft, T. J.;
Gifford-Moore, D. S.; Jones, N. D.; Olkowski, J. L.; Weir, L. C.; Smith,
G. F. D-Phe-Pro-p-amidinobenzylamine: A potent and highly selective
thrombin inhibitor. Bioorg. Med. Chem. Lett. 1996, 6, 2387−2392.
(36) Sturzebecher, A.; Donnecke, D.; Schweinitz, A.; Schuster, O.;
̈
̈
Steinmetzer, P.; Sturzebecher, U.; Kotthaus, J.; Clement, B.;
̈
Sturzebecher, J.; Steinmetzer, T. Highly potent and selective substrate
analogue factor Xa inhibitors containing D-homophenylalanine
̈
analogues as P3 residue: Part 2. ChemMedChem 2007, 2, 1043−1053.
(37) Donnecke, D.; Schweinitz, A.; Sturzebecher, A.; Steinmetzer, P.;
̈
̈
Schuster, M.; Sturzebecher, U.; Nicklisch, S.; Sturzebecher, J.;
̈
̈
Steinmetzer, T. From selective substrate analogue factor Xa inhibitors
to dual inhibitors of thrombin and factor Xa. Part 3. Bioorg. Med. Chem.
Lett. 2007, 17, 3322−3329.
(38) Shiraishi, T.; Kadono, S.; Haramura, M.; Kodama, H.; Ono, Y.;
Iikura, H.; Esaki, T.; Koga, T.; Hattori, K.; Watanabe, Y.; Sakamoto,
A.; Yoshihashi, K.; Kitazawa, T.; Esaki, K.; Ohta, M.; Sato, H.; Kozono,
T. Factor VIIa inhibitors: target hopping in the serine protease family
using X-ray structure determination. Bioorg. Med. Chem. Lett. 2008, 18,
4533−4537.
(39) Kadono, S.; Sakamoto, A.; Kikuchi, Y.; Oh-eda, M.; Yabuta, N.;
Yoshihashi, K.; Kitazawa, T.; Suzuki, T.; Koga, T.; Hattori, K.;
Shiraishi, T.; Haramura, M.; Kodama, H.; Ono, Y.; Esaki, T.; Sato, H.;
Watanabe, Y.; Itoh, S.; Ohta, M.; Kozono, T. Structure-based design of
P3 moieties in the peptide mimetic factor VIIa inhibitor. Biochem.
Biophys. Res. Commun. 2005, 327, 589−596.
(40) Schweinitz, A.; Steinmetzer, T.; Banke, I. J.; Arlt, M. J.;
Sturzebecher, A.; Schuster, O.; Geissler, A.; Giersiefen, H.; Zeslawska,
̈
E.; Jacob, U.; Kruger, A.; Sturzebecher, J. Design of novel and selective
̈
inhibitors of urokinase-type plasminogen activator with improved
pharmacokinetic properties for use as antimetastatic agents. J. Biol.
Chem. 2004, 279, 33613−33622.
K
dx.doi.org/10.1021/jm3012917 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(41) Anwer, M. K.; Spatola, A. F. An advantageous method for the
rapid removal of hydrogenolyzable protecting groups under ambient
conditions; synthesis of leucine-enkephalin. Synthesis 1980, 929−932.
(42) Becker, G. L.; Sielaff, F.; Than, M. E.; Lindberg, I.; Routhier, S.;
Day, R.; Lu, Y.; Garten, W.; Steinmetzer, T. Potent inhibitors of furin
and furin-like proprotein convertases containing decarboxylated P1
arginine mimetics. J. Med. Chem. 2010, 53, 1067−1075.
(43) Shen, L.; Wang, F. W.; Cheng, A. B.; Yang, S. 2,2′-(Piperazine-
1,4-diium-1,4-diyl)diacetate dihydrate. Acta Crystallogr., Sect. E: Struct.
Rep. Online 2006, E62, o2242−o2244.
thrombin inhibitors: piperazides of 3-amidinophenylalanine. J. Med.
Chem. 1997, 40, 3091−3099.
(62) LeadIT, version 2.0.2; BioSolveIT GmbH: Sankt Augustin,
Germany.
(63) MOE, version 2010; Chemical Computing Group: Montreal,
Canada.
(64) Labute, P. LowModeMDImplicit low-mode velocity filtering
applied to conformational search of macrocycles and protein loops. J.
Chem. Inf. Model. 2010, 50, 792−800.
(44) Ivanyi, T.; Lazar, I. Synthesis and in vitro enzyme hydrolysis of
trioxadiaza- and tetraoxadiaza-crown ether-based complexing agents
with disposable ester pendant arms. Synthesis 2005, 3555−3564.
(45) Tanatani, A.; Mio, M. J.; Moore, J. S. Chain length-dependent
affinity of helical foldamers for a rodlike guest. J. Am. Chem. Soc. 2001,
123, 1792−1793.
(46) Dietrich, W.; Nicklisch, S.; Koster, A.; Spannagl, M.; Giersiefen,
H.; van de Locht, A. CU-2010–a novel small molecule protease
inhibitor with antifibrinolytic and anticoagulant properties. Anesthesi-
ology 2009, 110, 123−130.
(47) Mateescu, C.; Wikete, C.; Costisor, O.; Bouet, G.; Khan, M. A.
Protonation behaviour of N,N′-piperazine-dipropionic acid. C. R.
Chim. 2005, 8, 1149−1153.
(48) Ascenso, J. R.; Santos, M. A.; Frausto da Silva, J. J. R.; Vaz, M. C.
T. A.; Drew, M. G. B. Protonation of diaza cyclic complexons: Proton
NMR, calorimetric and molecular mechanics studies. J. Chem. Soc.,
Perkin Trans. 2 1990, 2211−2218.
(49) Khalili, F.; Henni, A.; East, A. L. L. pKa Values of Some
Piperazines at (298, 303, 313, and 323) K. J. Chem. Eng. Data 2009,
54, 2914−2917.
(50) Marvin 5.11.4, 2012, ChemAxon (http://www.chemaxon.com)
(51) Figures 1 and 2 were created with the PyMOL Molecular
Graphics System, version 1.2r2, Schrodinger, LLC.
̈
(52) Lazar, H. L.; Bao, Y.; Tanzillo, L.; O’Gara, P.; Reardon, D.;
Price, D.; Crowley, R.; Cabral, H. J. Aprotinin decreases ischemic
damage during coronary revascularization. J. Card. Surg. 2005, 20,
519−523.
(53) Engles, L. Review and application of serine protease inhibition
in coronary artery bypass graft surgery. Am. J. Health-Syst. Pharm.
2005, 62, S9−14.
(54) Saxena, P.; Thompson, P.; d’Udekem, Y.; Konstantinov, I. E.
Kallikrein-kinin system: A surgical perspective in post-aprotinin era. J.
Surg. Res. 2011, 167, 70−77.
(55) Westermann, D.; Schultheiss, H.-P.; Tschope, C. New
perspective on the tissue kallikrein-kinin system in myocardial
infarction: Role of angiogenesis and cardiac regeneration. Int.
Immunopharmacol. 2008, 8, 148−154.
(56) Ruf, S.; Buning, C.; Schreuder, H.; Horstick, G.; Linz, W.; Olpp,
T.; Pernerstorfer, J.; Hiss, K.; Kroll, K.; Kannt, A.; Kohlmann, M.;
Linz, D.; Hubschle, T.; Rutten, H.; Wirth, K.; Schmidt, T.; Sadowski,
T. Novel β-amino acid derivatives as inhibitors of cathepsin A. J. Med.
Chem. 2012, 55, 7636−7649.
(57) Paju, A.; Stenman, U.-H. Biochemistry and clinical role of
trypsinogens and pancreatic secretory trypsin inhibitor. Crit. Rev. Clin.
Lab. Sci. 2006, 43, 103−142.
(58) Gustafsson, D.; Bylund, R.; Antonsson, T.; Nilsson, I.; Nystrom,
̈
J. E.; Eriksson, U.; Bredberg, U.; Teger-Nilsson, A. C. A new oral
anticoagulant: The 50-year challenge. Nat. Rev. Drug Discovery 2004, 3,
649−659.
(59) Gustafsson, D.; Antonsson, T.; Bylund, R.; Eriksson, U.;
Gyzander, E.; Nilsson, I.; Elg, M.; Mattsson, C.; Deinum, J.; Pehrsson,
S.; Karlsson, O.; Nilsson, A.; Sorensen, H. Effects of melagatran, a new
low-molecular-weight thrombin inhibitor, on thrombin and fibrinolytic
enzymes. Thromb. Haemostasis 1998, 79, 110−118.
(60) Dixon, M. The determination of enzyme inhibitor constants.
Biochem. J. 1953, 55, 170−171.
(61) Sturzebecher, J.; Prasa, D.; Hauptmann, J.; Vieweg, H.;
̈
Wikstrom, P. Synthesis and structure-activity relationships of potent
̈
L
dx.doi.org/10.1021/jm3012917 | J. Med. Chem. XXXX, XXX, XXX−XXX