Design, synthesis, and SAR of a series of activated protein C (APC) inhibitors with selectivity against thrombin for the treatment of haemophilia

Article abstract of DOI:10.1016/j.bmcl.2013.12.094

A design strategy was used to identify inhibitors of activated protein C with selectivity over thrombin featured by a basic and/or aromatic functionality for binding to the S2 pocket. Our strongest inhibitor showed an IC ₅₀-material value and selectivity for APC vs thrombin similar to a compound previously reported in the literature. However, in contrast to the reference compound, our compound showed a retained coagulant effect of thrombin with increasing substrate concentration in a modified Calibrated Automated Thrombogram (CAT) method. This was likely related to our compound being inactive against FVIIa, while the reference compound showed an IC₅₀ of 8.9 μM. Thus, the higher selectivity of our compound against all relevant coagulation factors likely explained its higher therapeutic potential in comparison to the reference compound. The data indicate that at least a 100-fold selectivity over other serine proteases in the coagulation cascade will be required for an effective APC inhibitor.

Full text of DOI:10.1016/j.bmcl.2013.12.094

Bioorganic & Medicinal Chemistry Letters 24 (2014) 821–827
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design, synthesis, and SAR of a series of activated protein C (APC)
inhibitors with selectivity against thrombin for the treatment of
haemophilia
a,
a
b
b
b
Peter Bach a, , Laurent Knerr , Ola Fjellström , Kenny Hansson , Christer Mattsson , David Gustafsson
⇑
⇑
a Department of Medicinal Chemistry, AstraZeneca R&D, CVMD iMed, Mölndal, Pepparedsleden 1, SE-431 83 Mölndal, Sweden
b Department of Bioscience, AstraZeneca R&D, CVMD iMed, Mölndal, Pepparedsleden 1, SE-431 83 Mölndal, Sweden
a r t i c l e i n f o
a b s t r a c t
Article history:
A design strategy was used to identify inhibitors of activated protein C with selectivity over thrombin fea-
tured by a basic and/or aromatic functionality for binding to the S2 pocket. Our strongest inhibitor
showed an IC50-material value and selectivity for APC vs thrombin similar to a compound previously
reported in the literature. However, in contrast to the reference compound, our compound showed a
retained coagulant effect of thrombin with increasing substrate concentration in a modified Calibrated
Automated Thrombogram (CAT) method. This was likely related to our compound being inactive against
Received 9 November 2013
Revised 20 December 2013
Accepted 22 December 2013
Available online 29 December 2013
Keywords:
Inhibitor
Activated Protein C (APC)
Thrombin
Haemophilia
FVIIa, while the reference compound showed an IC50 of 8.9 lM. Thus, the higher selectivity of our com-
pound against all relevant coagulation factors likely explained its higher therapeutic potential in compar-
ison to the reference compound. The data indicate that at least a 100-fold selectivity over other serine
proteases in the coagulation cascade will be required for an effective APC inhibitor.
Ó 2014 Elsevier Ltd. All rights reserved.
Solid-phase synthesis
S2 pocket
Haemophilia (deficiency of coagulation factors VIII or IX, FVIII or
FIX) is treated with intravenous replacement of the deficient factor
or with factor concentrates.1 Clinical studies and experience show
that prophylactic administration of the deficient factor is highly
preferable with regard to outcome.2 Prophylaxis is cumbersome
due to the necessity of intravenous self-injections every second
to third day, the cost of coagulation factors, and requirements on
hygiene and refrigerators. A serious clinical complication in 5–
25% of the patients is the development of inhibitory antibodies to-
wards the administered coagulation factor. Today only 20–30% of
the estimated ‘haemophilia population’ in the world receives ade-
quate treatment, and only a fraction of those are given the coagu-
lation factors as prophylaxis.3 Treatment with small molecules is
limited to desmopressin (intranasally or subcutaneously) for tran-
sient release of FVIII from endothelial cells, which is effective in
mild haemophilia,4 and stabilisation of the clot with the oral fibri-
nolysis inhibitor tranexamic acid, with clinically meaningful effects
on top of added coagulation factor in dental extractions.5 Thus,
prophylactic oral prevention of bleeding in haemophilia represents
a large medical need.
The balance between haemostasis and bleeding is skewed in the
sense that there is no need to fully restore the haemostasis. In hae-
mophilia, less than 1%, 1–5%, and 5–40% of the normal level of FVIII
or FIX are the definitions for severe, moderate and mild haemo-
philia, respectively. Persons with mild haemophilia usually do
not require coagulation factor prophylaxis.
Activation of Protein C, a serine protease in the coagulation cas-
cade, by thrombin forms activated protein C (APC) which functions
as a negative feed-back inhibitor of plasma coagulation by degrad-
ing activated factor V (FVa) and activated factor VIII (FVIIIa) to
inactive factors (FVai, FVIIIai). FVa and FVIIIa strongly reinforce
plasma coagulation by assembling the prothrombinase and tenase
complexes. Reduced effectiveness of the APC pathway via the FV
Leiden mutation is a thrombosis risk in persons without bleeding
disorders but in persons with bleeding disorders it may instead
prevent bleeding.6 Stimulation of plasma coagulation by inhibition
of the anticoagulant action of APC has been described for peptides7
and aptamers.8 The only low molecular weight APC inhibitors re-
ported to date is a series of benzamidine derivatives from Servier,9
exemplified by compound 110 (selected as a reference compound,
Fig. 1) that showed an IC50 of 0.82 lM (1.7 lM in our assay) and
the highest selectivity for APC over thrombin (57-fold reported,
21-fold in our assay) in the Servier series. Calculated inhibition
constants (Ki-values) from all IC50 values are given for all com-
pounds in the Supplementary Information.
⇑
Corresponding authors. Tel.: +46 73 8208612 (P.B.); tel.: +46 31 7065788; fax:
+46 31 7763700 (L.K.).
E-mail addresses: bach@chalmers.se (P. Bach), laurent.knerr@astrazeneca.com
(L. Knerr).
0960-894X/$ - see front matter Ó 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2013.12.094

822
P. Bach et al. / Bioorg. Med. Chem. Lett. 24 (2014) 821–827
The compounds were synthesized by applying the modular so-
NH
lid-phase synthesis shown in Scheme 115 which is based on the
anchoring of the benzamidine moiety on a carbamate resin.16 Thus,
Wang carbamate resin 2 was treated with three equivalents of ben-
zamidine derivative 3 in the presence of DIPEA to give 4. Removal
of the Teoc protective group was achieved by treatment with an
excess of TBAF. Coupling of a diverse set of Fmoc protected amino
acid was effected smoothly using TBTU/HOBT. Initially, the subse-
quent Fmoc deprotection step was performed using classical con-
ditions (20% piperidine in DMF), however this approach gave
irreproducible and/or overall poor yields for the sequence. We rea-
soned that the carbamate-amidine linkage could potentially be
reactive in the presence of a high concentration of a nucleophilic
amine.17 We therefore tested the use of milder reaction conditions
by using TBAF as the Fmoc cleaving reagent,18 and this improved
both the efficiency and reliability of the sequence. Subsequent cou-
pling with carboxylic acid 13 (Scheme 2), followed by cleavage/
deprotection under acidic conditions yielded smoothly the tar-
geted amidines in 8–34% overall yield, isolated after reverse-phase
chromatography. Not all pairs of diastereomers could be separated
in this manner, thus some pairs of diastereomers were screened19
as such after chromatographic purification.
H2N
O
O
H
N
H
N
N
H
OH
O
1
Figure 1. The compound (1, reference compound), with highest selectivity for APC
over thrombin in the Servier series.
NH
H2N
R1
O
H
N
H
N
N
H
R3
O
R2
Figure 2. Benzamidine scaffold used in the present study.
A comparison of the crystal structures11 of the two serine pro-
teases APC and thrombin pointed to a difference in the preferred
amino acid for the natural ligands binding to the S2 pocket. While
the prototypic substrate in P1–P2–P3 for thrombin is D-phe-pro-
arg12 the analogous ligand for APC is D-phe-arg-arg. Given that
the guanidino functionality of arginine has a typical pKa values of
13,13 compared to the neutral side chain of proline, we hypothe-
sized that ligands with a basic functionality aimed for binding in
the S2 pocket of APC would lead to selectivity for APC over throm-
bin. In addition, for this structural class it was known that attempts
to introduce large substituents into the S2 pocket resulted in a re-
duced affinity for thrombin, since S2 in APC has a more open con-
formation.14 Basic and/or sterically large functional groups aimed
for binding interactions in the S2 pocket should thus be introduced
to attain selectivity for APC over thrombin. Based on the precedent
set by 1, we tested this hypothesis using the general benzamidine
scaffold depicted in Figure 2. Below we describe our synthetic ap-
proach and successful identification of ligands that display coagu-
lant activity as well as selectivity for APC over thrombin.
The cyclohexyl alanine derivative 13 (used in step e, Scheme 1)
was prepared on a multigram scale as outlined in Scheme 2. Thus,
protection of compound 9 as a methyl ester to give 10 followed by
consecutive mono-alkylation and Boc-protection of the amine
functionality provided 12. Hydrolysis of the methyl ester under
alkaline conditions provided the carboxylic acid 13. The analogous
building blocks for the syntheses of 23, 30 (R2 = cyclohexyl) and 24
(R3 = (CH2)2CO2H) were prepared by similar procedures.
Compound 22 (Table 1) was featured by the basic side chain of
arginine in combination with a CH2-cyclohexyl R2 substituent and
a CH2COOH R3 substituent and showed a comparably high affinity
for APC with an IC50 value of 0.47 lM, but only a 2.8-fold selectiv-
ity against thrombin. Replacement of the R2 substituent with
cyclohexyl (to give 23) or of the R3 substituent with (CH2)2CO2H
(to give 24) led to lower affinity, thus the R2 and R3 substituents
of compound 22 were retained in the further SAR investigations.
Amino acid moieties with an (S)-stereochemistry were preferred
O
N
NH2
O
NH2
O
O
HN
+
N
H
N
a
H
N
O
O
O
Si
Si
O
O
4
2
3
O
NH2
O
NH2
N
b
c, d
N
R1
NH2
H
N
NH2
5
6
O
NH
O
NH2
H2N
R1
O
O
O
O
H
N
H
N
R1
O
O
R
N
H
e
N
H
OH
R
N
N
f
N
H
O
O
O
8
7
Scheme 1. Solid-phase method for modular synthesis of APC inhibitors. Reagents and conditions: (a) DIPEA, 9.0 equiv, DMF, rt, 20 h; (b) TBAF (1M in THF), 10 equiv, DMF, rt;
(c) Fmoc-NHC(R1)COOH, 1.5 equiv, TBTU, 1.5 equiv, DIPEA, 3.0 equiv, DMF, rt, 3.5 h; (d) TBAF (1M in THF), 3.0 equiv, DMF, rt, 3 Â 2 min.; (e) 13, 1.5 equiv, TBTU, 1.5 equiv,
HOBT, 1.5 equiv, DIPEA, 3.0 equiv, DMF, rt, 4 h; (f) TFA/DCM/H2O 49:49:2; rt, 1 + 1 h.

P. Bach et al. / Bioorg. Med. Chem. Lett. 24 (2014) 821–827
823
b
a
HO
O
O
O
NH2
N
H
NH3Cl
O
O
O
O
O
10
11
9
c
d
HO
O
O
O
N
N
O
O
O
O
O
O
O
12
13
Scheme 2. Preparation of the cyclohexyl alanine derivative 13. Reagents and conditions: (a) SOCl2, MeOH, À20 °C then reflux, 18 h, (77%); (b) BrCH2COO-t-Bu, 1.0 equiv,
adjust to pH = 9 with DIPEA, then stir at rt, 36 h (58%); (c) Boc2O, 1.1 equiv, DIPEA, 1.0 equiv, rt, 16 h, add Boc2O, 0.1 equiv, rt, 21 h (100%); (d) NaOH, 2.0 equiv; dioxane/water,
rt, 16 h (99%).
Table 1
Effect on inhibition of APC and thrombin, respectively, by introduction of R1 groups with basic, aliphatic functionalities. The inhibition of activated coagulation factors X (FXa) and
XI (FXIa) was determined for some compounds
NH
H2N
R1
O
H
N
H
N
N
H
R3
O
R2
Compd
No
R1
stereo
a
-C
R1
ACDa pKa
of R1
R2
stereo
a
-C
R2
R3
APC,
IC50
Thrombin,
IC50 M)
APC/
thrombin
ratio
FXa,
IC50
FXIa,
IC50
(lM)
ACDa
logD
pH7.4
logDb,
pH7.4
(lM)
(
l
(lM)
1
—
—
R
cy-hex
CH2CO2H
CH2CO2H
1.7
36
21
48
—
>44
0.52
0.93
*
H
N
H2N
H2N
H2N
NH
22
S
S
S
13.3
R
R
R
CH2Àcy-hex
0.47
1.3
1.9
5.2
2.8
—
À4.51
—
*
*
*
H
N
NH
23
24
13.3
13.3
cy-hex
H
1.6
2.8
1.2
1.8
—
—
—
À9.33
À5.1
—
—
H
N
NH
CH2-cy-hex
(CH2)2CO2H
>44
NH
25
26
S
10.5
10.5
R
R
CH2-cy-hex
CH2-cy-hex
CH2CO2H
CH2CO2H
5.6
7.0
1.3
1
—
—
—
—
À2.79
À2.79
—
—
(rac)
(rac)
*
NH
R
>44
>44
*
(continued on next page)

824
P. Bach et al. / Bioorg. Med. Chem. Lett. 24 (2014) 821–827
Table 1 (continued)
Compd
No
R1
stereo
a
-C
R1
ACDa pKa
of R1
R2
stereo
a
-C
R2
R3
APC,
IC50
Thrombin,
IC50 M)
APC/
thrombin
ratio
FXa,
IC50
FXIa,
IC50
(lM)
ACDa
logD
pH7.4
logDb,
pH7.4
(lM)
(
l
(lM)
H
N
27
28
S
10.2
10.5
R
CH2-cy-hex
CH2-cy-hex
CH2CO2H
CH2CO2H
3.1
1.3
26
8.4
9.1
—
—
>44
À2.53
À3.07
—
—
*
HN
rac
R
12
—
*
29
30
31
rac
rac
rac
10.6
10.6
10.6
R
R
R
CH2-cy-hex
cy-hex
CH2CO2H
CH2CO2H
H
0.92
>44
3.1
22
23
1
>130
—
>44
—
À2.67
À3.38
À5.78
<0
—
HN
HN
HN
*
*
*
>44
28
CH2-cy-hex
8.9
—
—
—
a
The pKa values of the R1 functionalities in the corresponding protonated forms and the lipophilicity of the compounds, expressed by logD at pH 7.4, were calculated using
the software package from Advanced Chemical Design, Inc.
b
logD at pH 7.4 was determined by a chromatographic method. cy-hex = cyclohexyl.
for binding to the S2 pocket of APC, and all synthesized compounds
with (R)-amino acids in this position were inactive, as exemplified
by the pair of diastereomers 25/26. Further investigations of basic
side-chains with amine functionalities presented compounds 27–
was added to reach approximately 5% FVIII level. Ellagic acid, a
known FXII activator (up-stream to FVIII and FIX), was also added
to amplify the thrombin signal from the internal pathway. The re-
sults from this method are shown in Figure 3 and Table 3. The area
under the curve represents the total thrombin activity (endoge-
nous thrombin potential, ETP) in the sample and was shown to de-
crease when thrombomodulin was added, due to APC-induced
inhibition of thrombin formation. When an APC inhibiting anti-
body was added as positive control, the ETP was restored to the le-
vel when no thrombomodulin was present.
29, of which 29 showed an affinity of 0.91 lM for APC and a
remarkable 23-fold selectivity vs thrombin. Replacement of the
CH2-cyclohexyl R2 substituent of 29 with a cyclohexyl gave 30,
which was inactive both vs APC and thrombin. In the Servier series,
replacement of the cyclohexyl R2 substituent of compound 1
(Fig. 1) with a CH2-cyclohexyl substituent led to an 8-fold increase
in inhibition of APC (from IC50 = 0.82
l
M to 0.10
l
M), but also to a
Both compound 1 and compound 29 showed activity starting at
their IC50 for APC. However, while compound 1 (Fig. 4, Table 3)
showed a decreasing ETP with increasing concentration of 1, com-
pound 29 (Fig. 5, Table 3) showed an increase in ETP towards nor-
malisation with increasing concentration. Compounds 1 and 29
both shifted lagtime and time to peak to the right (Figs. 4 and 5, Ta-
ble 3). The tendency with right shifted time to peak is often seen
when more thrombin is formed in the system due to that the peak
in thrombin formation then occurs later. The delay in lagtime is
probably due to coagulation factor inhibition since it is most pro-
nounced at higher concentrations where the coagulation inhibition
is substantial.
These results are surprising as both 1 and 29 have the same
selectivity for APC towards thrombin (21 and 23-fold, respectively)
and also the selectivity ratios against FXa (immediately up-stream
to thrombin, 28 and >140, respectively) and FXIa (up-stream to FXa
and FIXa in the internal pathway, >26 and >48, respectively) seem
reasonable. Both compounds were also inactive against FIXa at
lowering of selectivity for APC over thrombin from 57-fold to 13-
fold. This shows that the SAR does not translate directly between
the two series. Replacement of the acidic R3 substituent of com-
pound 29 with hydrogen gave 31 which showed a 3.5-fold reduc-
tion in affinity (to 3.1 lM) and a reduced selectivity vs thrombin
compared to 29.
Introduction of aromatic groups as R1 substituents, exemplified
by compounds 33–35 (Table 2), led to inhibitory effects on the le-
vel of 29, however with lower selectivity for APC over thrombin. In
comparison to benzyl compound 33, substitution in the benzyl 4-
position (compounds 36–39), led to an increased selectivity over
thrombin only for the phenolic derivative 36. Interestingly, this
derivative also showed significant selectivity levels over FXa and
FXIa. The anilinic derivative 37 turned out to be the strongest
inhibitor of APC with however a decreased selectivity over throm-
bin compared to 33. Nitrile and methylamino derivatives 38 and 39
both showed decreases in inhibition of APC and in selectivity over
thrombin compared to 33.
The efficiency of compounds 1 and 29, which showed similar
inhibitory constants and selectivity for APC over thrombin, was
evaluated by a modified thrombin generation assay known as the
Calibrated Automated Thrombogram (CAT) method, using citrated
plasma from patients with severe haemophilia A.21 This method
was based on activation of coagulation and APC by addition of tis-
sue factor and thrombomodulin, respectively. To speed up throm-
bin generation, a small fraction of normal human pooled plasma
100
FVIIa on the external pathway show an interesting difference with
1 having an IC50 of 8.9 M (selectivity ratio 5.2) while compound
29 was inactive against FVIIa (IC50 > 133 M, selectivity ratio
lM (see Supplementary material). However, the analysis of
l
l
>140). Thus, most likely the high selectivity against relevant coag-
ulation factors explains the higher therapeutic potential of com-
pound 29 in comparison to compound 1 since inhibition of the
anti-coagulant APC can be achieved while retaining the pro-coagu-
lant effect in plasma.

P. Bach et al. / Bioorg. Med. Chem. Lett. 24 (2014) 821–827
825
Table 2
Effect on inhibition of APC and thrombin, respectively, by introduction of R1 groups with aromatic functionalities. The inhibition of activated coagulation factors X (FXa) and XI
(FXIa) was determined for some compounds
NH
H2N
R1
O
O
H
N
H
N
R
N
H
OH
O
Compd No
R1
stereo
a
-C
R1
ACDa pKa
of R1
APC,
IC50
Thrombin,
APC/
thrombin
ratio
FXa,
IC50
FXIa,
IC50
ACDa logD
pH7.4
logDb,
pH7.4
(
lM)
IC50
(l
M)
(
lM)
(lM)
33
34
35
S
—
0.67
1.1
6.1
9.1
11
—
—
1.75
2.96
2.95
—
*
S
S
—
—
12
—
—
—
—
1.97
—
*
0.98
5.2
5.3
*
HO
H2N
N
36
37
S
S
9.8
4.5
0.26
0.13
3.9
0.9
15
23
1.9
1.11
0.57
0.02
*
6.9
>44
0.58
—
*
38
39
rac
rac
—
3.8
1.1
3.8
3.7
5.4
3.3
—
—
1.43
—
—
*
H2N
9.1
42
43
À0.76
*
a
The pKa values of the R1 functionalities in the corresponding protonated forms and the lipophilicity of the compounds, expressed by logD at pH 7.4, were calculated using
the software package from Advanced Chemical Design, Inc.
b
logD at pH 7.4 was determined by a chromatographic method.20
The CAT assay only gives information about the thrombin gen-
eration, and, hence, plasma coagulation. However, fibrinolysis inhi-
bition is also important, as shown clinically by tranexamic acid in
some patients with bleeding disorders. Moreover incorporation of
homopiperidylalanine as P2 has recently been shown to yield
APC potency in dual plasmin and plasma kallikrein inhibitors.22
Thus, compound 1 and 29 were tested in the plasma clot-lysis as-
say using citrated plasma from healthy volunteers.23 Plasmin gen-
eration was initiated by the addition of tissue plasminogen
activator. Both compounds prolonged the clot-lysis time and it
compounds 1 and 29, respectively (n = 2). For compound 1 the
antifibrinolytic effect could be of some relevance as the inhibition
occurs in the same concentration range as the increase in thrombin
generation measured in the CAT assay.
In conclusion, by using a design strategy to introduce a basic S2
side chain we have discovered an inhibitor of activated protein C
(APC) with IC50-value and selectivity for APC vs thrombin similar
to a compound previously reported (compound 1). However, con-
trary to compound 1 (reference compound), compound 29 shows
a retained coagulant effect of thrombin with increasing substrate
concentration in a modified Calibrated Automated Thrombogram
was doubled at assay concentration of 101 and 28 lM for

826
P. Bach et al. / Bioorg. Med. Chem. Lett. 24 (2014) 821–827
Figure 3. A control experiment in the thrombin generation assay (CAT method).
The area under the curve represents the total thrombin activity in the sample. As a
validation of the method the different curves shows the different additions. The
monoclonal antibody (MAb) that inhibit APC was used as positive control.
FVIII = coagulation factor VIII, TF = tissue factor, TM = rabbit lung thrombomodulin,
Ellag = ellagic acid.
Figure 5. Thrombin generation curves showing the effects of increasing concen-
trations of compound 29 in the modified CAT method, compared to level of
thrombin generation without APC activation (FVIII 5% + TF), positive control (FVIII
5% + TF + TM + Ellag + MAb), and negative control (FVIII 5% + TF + TM + Ellag +
DMSO).
Acknowledgments
Table 3
Effect of compounds 1 and 29 on thrombin generation parameters lagtime, time to
peak, peak and endogenous thrombin potential compared to controls
We are grateful to Senior Research Scientist Erika Gyzander for
performing screening experiments and the modified Calibrated
Automated Thrombogram assay.
Conccn (
l
M)
Lagtime
(min)
Time to
Peak
(nM)
ETP
Peak (min)
(nM*min)
FVIII 5% (baseline)
Mab (positive control)
DMSO (negative control)
2.7
2.7
2.7
2.7
2.8
3
4.6
17
2.8
2.7
2.7
3
10.2
8.2
6
5.9
6.3
7
12
29
6.1
6.2
6.3
7.8
15
40
40
34
36
34
23
7
6.5
31
35
36
39
26
598
700
199
202
203
161
93
117
184
213
237
362
517
Supplementary data
Supplementary data (screening assays for enzyme inhibition
and the CAT method) associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2013.12.
094.
Compd 1, 0.0125
Compd 1, 0.125
l
M
l
M
Compd 1, 1.25
Compd 1, 12.5
Compd 1, 125
lM
l
M
lM
Compd 29, 0.0125
Compd 29, 0.125
lM
M
References and notes
l
Compd 29, 1.25
Compd 29, 12.5
Compd 29, 125
lM
1. Berntorp, E.; Shapiro, A. D. Lancet 2012, 379, 1447.
2. Berntorp, E. Haemophilia 2009, 15, 1219.
3. Skinner, M. W. Br. J. Haematol. 2011, 154, 704.
lM
lM
5
4. Mannucci, P. M. Blood 1997, 90, 2515.
5. Dunn, C. J.; Goa, K. L. Drugs 1999, 57, 1005.
6. Lee, D. H.; Walker, I. R.; Teitel, J.; Poon, M. C.; Ritchie, B.; Akabutu, J.; Sinclair, G.
D.; Pai, M.; Wu, J. W.; Reddy, S.; Carter, C.; Growe, G.; Lillicrap, D.; Lam, M.;
Blajchman, M. A. Thromb. Haemost. 2000, 83, 387.
ETP = endogenous thrombin potential, FVIII = coagulation factor VIII, Mab = inhibi-
tory monoclonal antibody towards APC, DMSO = dimethyl sulfoxide,
Compd = compound.
7. (a) Butenas, S.; Orfeto, T.; Kalafatis, M.; Mann, K. G. J. Thromb. Haemost. 2006, 4,
2411; (b) Brummel-Ziedins, K. E.; Whelihan, M. F.; Rivard, G. E.; Butenas, S. J.
Thromb. Haemost. 2011, 9, 2262.
8. (a) Gal, S. W.; Amontov, S.; Urvil, P. T.; Vishnuvardhan, D.; Nishikawa, F.;
Kumar, P. K. R.; Nishikawa, S. Eur. J. Biochem. 1998, 252, 553; (b) Chang, J. Y.;
Chantrathammachart, P.; Monroe, D. M.; Key, N. S. Thromb. Res. 2012, 130,
e151.
9. De Nanteuil, G.; Gloanec, P.; Béguin, S.; Giesen, P. L. A.; Hemker, H. C.;
Mennecier, P.; Rupin, A.; Verbeuren, T. J. J. Med. Chem. 2006, 49, 5047.
10. Prepared according to the procedure in the patent application: De Nanteuil, G.;
Gloanec, P.; Verbeuren, T. J.; Rupin, A. WO 03/070690 A2.
11. From the protein data bank (PDB); web-page: http://www.wwpdb.org.
12. Bode, W.; Brandstetter, H.; Mather, T.; Stubbs, M. T. Thromb. Haemost. 1997, 78,
501.
13. Stryer, L. Biochemistry; W. H. Freeman and Co.: New York, 1995.
14. Mather, T.; Oganessyan, V.; Hof, P.; Huber, R.; Foundling, S.; Esmon, C.; Bode,
W. EMBO J. 1996, 15, 6822.
15. The amino acids used in step (c) of Scheme
Novabiochem.
1 were purchased from
General experimental procedure: Step a (see Scheme 1): To para-nitrophenyl
Wang carbamate resin in a plastic syringe equipped with a frit was added
amidine 3 (3.0 eq.). To the solids was added DIPEA (9.0 equiv) in DMF (10 mL/g
resin) and the resulting slurry was agitated at rt for 18 h. The resin was filtered
Figure 4. Thrombin generation curves showing the effects of increasing concen-
trations of compound 1 in the modified CAT method (0.0125 lM of compound 1
was also tested without effect (not depicted)), compared to level of thrombin
generation without APC activation (FVIII 5% + TF), positive control (FVIII
5% + TF + TM + Ellag + MAb), and negative control (FVIII5% + TF + TM + Ellag +
DMSO).
and rinsed with 6 Â DMF and then with 3 Â (1 Â DCM
+
1 Â MeOH).
A
qualitative reaction completion check was performed by treating an aliquot
of the washed resin with 1-methylpiperazine to look for any yellow coloration,
in which case the reaction was run a second time. The resin was dried under
vacuum and engaged in the next step.
Step b: The resin was washed with 3 Â DMF and swelled in DMF before TBAF
in THF (1M ; 10 equiv) was added. The resulting slurry was then agitated at rt
for 7 h. The resin was filtered and washed with 6 Â DMF and then with
3 Â (1 Â DCM + 1 Â MeOH). The resin was then dried under vacuum and
engaged in the next step.
(CAT) method most likely due to a higher selectivity against all
relevant coagulation factors. These data indicate that at least a
100-fold selectivity over other serine proteases in the coagulation
cascade will be required for an effective APC inhibitor.
Step c: To resin 4 was added Fmoc-D,L-homo-[ala-4-pip(N-Boc)] (1.5 equiv),

P. Bach et al. / Bioorg. Med. Chem. Lett. 24 (2014) 821–827
827
TBTU (1.5 equiv), HOBT (1.5 equiv), and finally DIPEA (3.0 equiv) in DMF. The
reaction slurry was shaken at rt for 3.5 h. The resin was filtered and rinsed with
6 x DMF and then with 3 Â (1 Â DCM + 1 Â MeOH). Completion of the coupling
was confirmed by Kaiser test (Kaiser, E.; Colescot, R. L.; Bossinge, C. D.; Cook, P.
I. Anal. Biochem. 1970, 34, 595) on a few beads. The resin was dried under
vacuum for 16 h and engaged in the next step.
Step d: The resin was washed with 3 Â DMF and subsequently treated at rt
with a solution of DMF/1M TBAF in THF) 4:1 (v/v) for 3 Â 2 minutes. The resin
was then washed thoroughly with 6 Â DMF, 2 Â (DMF/H2O 1:1 (v/v), 2 Â H2O,
2 Â (DMF/H2O 1:1 (v/v), 2 Â DMF and finally 3 Â (1 Â DCM + 1 Â MeOH). The
resin was dried under vacuum for 2 h and engaged in the next step.
Step e: To the resin was added 13 (1.5 equiv), TBTU (1.5 equiv), HOBT
(1.5 equiv), and finally DIPEA (3 equiv) in DMF. The reaction slurry was shaken
at rt for 4 h. The resin was filtered and rinsed with 6 Â DMF and then with
3 Â (1 Â DCM + 1 Â MeOH). Completion of the coupling was confirmed by
Kaiser test on a few beads. The resin was dried under vacuum for 16 h and
engaged in the next step.
in solvent B: MeCN. Relevant fraction were pooled and freeze-dried. Yield of 29
as a mixture of inseparable epimers as a monoacetate salt: 0.014 g (19% yield
from para-nitrophenyl Wang carbamate resin). 1H NMR (400 MHz, D2O) d
0.65–1.16 (m, 11H), 1.17–1.34 (m, 8H), 1.34–1.6 (m, 15H), 1.6–1.91 (m, 16H),
2.86 (t, J = 12.8 Hz, 4H), 2.96 (d, J = 16.1 Hz, 1H), 3.08 (d, J = 16.1 Hz, 1H), 3.13
(d, J = 16.4 Hz, 1H), 3.19 (d, J = 16.3 Hz, 1H), 3.30 (d, J = 12.6 Hz, 4H), 3.45–3.6
(m, 2H), 4.18 (dd, J = 5.2, 9.4 Hz, 1H), 4.27 (t, J = 7.4 Hz, 1H), 4.34 (d, J = 15.
9 Hz, 2H), 4.41 (d, J = 15.9 Hz, 1H), 4.42 (d, J = 15.9 Hz, 1H), 7.35 (d, J = 8.1 Hz,
2H), 7.38 (d, J = 8.1 Hz, 2H), 7.61 (d, J = 8.1 Hz, 2H), 7.65 (d, J = 8.1 Hz, 2H).
HRMS [M+H]+ calcd for C28H44N6O4: 529.3424, found 529.3445.
16. Roussel, P.; Bradley, M. Tetrahedron Lett. 1997, 38, 4861.
17. For a similar observation using a carbamate linker to immobilise indole NH
see: Smith, A. L.; Stevenson, G. I.; Lewis, S.; Patel, S.; Castro, J. L. Bioorg. Med.
Chem. Lett. 2000, 10, 2693.
18. Ueki, M.; Amemiya, M. Tetrahedron Lett. 1987, 28, 6617.
19. A description of the screening assay is found in the Supplementary material.
20. Valko, K. J. Chromatogr. A 2004, 1037, 299.
Step f: The resin was treated with TFA/DCM/water 49:49:2 for 1 h. The resin
was filtered and the treatment was repeated once. The resin was then rinsed
thoroughly with cycles of MeCN and DCM. The combined solvent fractions
were concentrated to give a crude which was purified by preparative HPLC
21. Hemker, H. C.; Giesen, P.; Al Dieri, R.; Regnault, V.; de Smedt, E.; Wagenvoord,
R.; Lecompte, T.; Béguin, S. Pathophysiol. Haemost. Thromb. 2003, 33, 4.
22. Herold, P., Daghish, M., Jelakovic, S., Ludwig, A., Reichelt, C., Schulze, A.,
Schweinitz, A. WO 11/094496 A2.
(Kromasil C8, 10
l
m, 250 Â 21.2 mm ID, flow 20 mL/min., detection by UV at
23. Boström, J.; Grant, J. A.; Fjellström, O.; Thelin, A.; Gustafsson, D. J. Med. Chem.
2013, 56, 3273.
254 nm) using a gradient of solvent A: 95% 0.1 M ammonium acetate/5% MeCN