Design and synthesis of potent and selective macrocyclic thrombin inhibitors

Article abstract of DOI:10.1016/S0960-894X(03)00506-7

A series of potent and selective proline- and pyrazinone-based macrocyclic thrombin inhibitors is described. Detailed SAR studies led to the incorporation of specific functional groups in the tether that enhanced functional activity against thrombin and provided exquisite selectivity against trypsin and tPA. X-ray crystallography and molecular modeling studies revealed the inhibitor-enzyme interactions responsible for this selectivity.

Full text of DOI:10.1016/S0960-894X(03)00506-7

Bioorganic & Medicinal Chemistry Letters 13 (2003) 2781–2784
Design and Synthesis of Potent and Selective Macrocyclic
Thrombin Inhibitors
Philippe G. Nantermet,a,* James C. Barrow,a Christina L. Newton,a Janetta M. Pellicore,a
MaryBeth Young,a S. Dale Lewis,b Bobby J. Lucas,b Julie A. Krueger,b
Daniel R. McMasters,c Youwei Yan,d Lawrence C. Kuo,d Joseph P. Vaccaa
and Harold G. Selnicka
aDepartment of Medicinal Chemistry, Merck Research Laboratories, West Point, PA 19486, USA
bDepartment of Biological Chemistry, Merck Research Laboratories, West Point, PA 19486, USA
cDepartment of Molecular Systems, Merck Research Laboratories, West Point, PA 19486, USA
dDepartment of Structural Biology, Merck Research Laboratories, West Point, PA 19486, USA
Received 12 March 2003; accepted 28 April 2003
This manuscript is dedicated to the memory of our colleague and friend Christie Newton
Abstract—A series of potent and selective proline- and pyrazinone-based macrocyclic thrombin inhibitors is described. Detailed
SAR studies led to the incorporation of specific functional groups in the tether that enhanced functional activity against thrombin
and provided exquisite selectivity against trypsin and tPA. X-ray crystallography and molecular modeling studies revealed the
inhibitor-enzyme interactions responsible for this selectivity.
# 2003 Elsevier Ltd. All rights reserved.
Thrombosis-related disorders such as deep vein throm-
bosis, pulmonary embolism and thromboembolic stroke
remain a major cause of morbidity worldwide.1 The
limitations associated with the current therapies2 have
driven the recent search for small-molecule direct inhib-
itors of specific enzymes involved in the coagulation
cascade.3 Inhibitors of both thrombin and Factor Xa
have attracted considerable attention.4 The search for
orally bioavailable direct inhibitors of thrombin has led
our laboratories to the evaluation of proline5 and pyr-
azinone6 based small molecules which inhibit thrombin
with a high degree of potency and selectivity. In this
letter we report on the elaboration of such inhibitors
into highly potent and selective macrocycles.
N,N-diethylacetamide or benzenesulfonamide (2, 5)
provided a 10- to 15-fold enhancement in potency while
substitution of the P1 phenol with N-ethylacetamide (3)
resulted in a comparable 8-fold boost in potency.
Examination of X-ray crystallographic data5 revealed
the close spatial proximity of lipophilic P4 groups
attached to either P1 or P3. This analysis, along with the
Earlier studies with proline-based inhibitors5 have
shown that a significant enhancement in inhibitory
potency is obtained by accessing the S4 shelf of the
enzyme from either the P1 or P3 group7 (Fig. 1). Sub-
stitution of the N-terminal amino group (P3) with either
Figure 1. aKi values are the average of at least two determination,
standard error of the mean <10%.
*Corresponding author. Tel.: +1-215-652-0945; fax: +1-215-652-
3971; e-mail: philippe_nantermet@merck.com
0960-894X/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00506-7

2782
P. G. Nantermet et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2781–2784
possibility of enhancing potency by conformational
preorganization directed us toward the design of mac-
rocycles8 which would link the P1 and P3 groups of a
proline-based inhibitor. Ultimately, extension of this
concept to pyrazinone-based inhibitors was found to be
more fruitful.
Proline-based macrocyclic inhibitors 7–10 were eval-
uated for their thrombin (IIa), trypsin and tPA inhibi-
tory potencies and their ability to double the activated
partial thromboplastin time (2ÂAPTT) in human
plasma.12 As illustrated in Table 1, we found that
N-acetyl-4-aminobutyric acid provided the optimal tether
between the P3 amino group and the P1 phenoxy group (9,
thrombin Ki=0.4 nM). Crystallographic analysis of 8
complexed to thrombin (Fig. 2) confirmed the same
general binding mode as observed for 2.5 The macrocyclic
tether is located on the S4 shelf. Intrinsic potency against
trypsin and tPA and selectivity versus thrombin were
deemed acceptable (>2000-fold). While there seemed to
be a clear relationship between tether length and inhibi-
tory potency against thrombin, potency was not sig-
nificantly improved by converting acyclic inhibitor 2 to
macrocycle 9. A more elaborate optimization study was
undertaken within the pyrazinone series.
The synthesis of proline-derived macrocycles is illu-
strated in Scheme 1. Coupling of various aminoesters to
carboxylic acid 65a followed by deprotection and
macrolactamization provided macrocycles 7–10.
The preparation of pyrazinone-based macrocycles fol-
lowed a similar route. Aminoester 13 was prepared from
azide 119 and bromopyrazinone 126 according to
Scheme 2. Alkylation with bromides 14a or reductive
amination with aldehydes 14b, followed by deprotection
and macrolactamization led to compounds 17–23 and
27–31 of general structure 15. An alternative macro-
cyclization strategy10 based on Grubbs olefin metath-
esis11 was implemented for the synthesis of compounds
24–26. Acyclic inhibitors 16a,b were prepared from 12
by standard protecting group manipulations and amide
coupling.
Table 2 documents an extensive study of pyrazinone-
based macrocycles. Aliphatic linkers between the P3
amino group and the P1 phenoxy group were first eval-
uated (17–22). As observed in the proline series, the
inhibitory potency of the compounds is directly related
to tether length. A pentyl-derived tether appeared opti-
mal, leading to compound 20, a 50 pM thrombin inhib-
itor, which represents a 40-fold boost in potency when
compared to acyclic inhibitors 16a,b (thrombin Ki: 2
nM and 2.2 nM, respectively). Methylation of the P3
amino group (R=Me) provided an additional 2-fold
increase in potency resulting in the identification of one
of the most potent thrombin inhibitors in this series (22,
thrombin Ki=20 pM). The functional activity assay (2
APTT) is, however, a better indicator of potential
antithrombotic activity as it reflects both inherent
enzyme inhibitory potency and the impact of inhibitor
lipophilicity presumably via plasma protein binding.13
Butyl derivative 19, when compared to its pentyl analo-
gue 20, demonstrates that an improvement in thrombin
Ki does not necessarily translate into enhanced func-
tional activity, and that a better balance between
potency and physical properties must be achieved. In
addition, the selectivity of this set of compounds for
inhibition of thrombin over trypsin and tPA was insuf-
ficient. While relative selectivity appeared acceptable
(>100-fold), micromolar or higher potencies against
serine proteases other than thrombin would be more
desirable to avoid potential side effects in a clinical set-
ting. With these two goals in mind, polar functionality
was introduced into the macrocycle tether.
Scheme 1. (a) tBuO2C(CH2)nNH2, EDC, HOAt, DMF; (b) TFA; (c)
EDC, HOAt, DMF.
Replacement of the tether central methylene with oxy-
gen (23) had little effect on intrinsic potency but
improved functional activity (2ÂAPTT=0.56 mM),
presumably as a result of increased polarity14 (23 vs
20). Selectivity was also improved, especially against
tPA. Carbamate derivatives 24 and 25 followed a
similar trend with the exception of the saturated deri-
vative 26.
Scheme 2. (a) Et3N, 110 ꢀC; (b) NCS, DCE, 80 ꢀC; (c) SnCl2, MeOH;
(d) 14a, Et3N, DMF or 14b, NaBH(OAc)3, DCE; (e) HCl(g), DCM;
(f) 1 N LiOH, THF; (g) EDC, HOAt, Et3N, DMF, 50 ꢀC; (h) for
R=Me, HCHO, NaBH(OAc)3; (i) Boc2O, CH2Cl2; (j) 1 N LiOH,
THF; (k) 1-(5-chloro-2-propoxyphenyl)methanamine, EDC, HOAt,
Et3N, DMF, microwave 80 ꢀC, 10 min, then HCl; (g) for 16a; 1-(5-
chloro-2-butoxyphenyl)methanamine, same conditions, for 16b.
In order to further optimize functional activity and
selectivity we then considered the removal of the lipo-
philic P1 chloro substituent (Y=H vs Cl). As expec-

P. G. Nantermet et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2781–2784
2783
Table 1. Proline based macrocycles
Compd
n
Thrombin Ki(nM)a
2ÂAPTT (mM)a,b
Trypsin Ki(mM)a
tPA Ki(mM)a
7
8
9
10
0
1
2
3
2.9
1.3
0.4
1.3
0.69
0.36
0.31
0.36
9.3
21.5
5
7.7
5.6
8.9
4.3
48
aKi values are the average of at least two determination, standard error of the mean <10%.
bThe 2ÂAPTT value is the concentration of inhibitor in plasma required to double the activated partial thromboplastin time.
Table 2. Pyrazinone-based macrocycles of general structure 15
Compd
X1
Y
R
Thrombin Ki(nM)a 2ÂAPTT (mM)a,b Trypsin Ki(mM)a (selectivity)c tPA Ki(mM)a (selectivity)c
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
(CH2)2-O
(CH2)3-O
(CH2)4-O
(CH2)5-O
(CH2)6-O
Cl
Cl
Cl
Cl
Cl
H
H
H
H
H
560
26
0.2
ND
ND
0.74
1.22
3.17
1.89
0.56
0.78
0.87
1.97
0.6
ND
ND
ND
ND
0.02 (100)
0.006 (120)
ND
0.003 (150)
0.15 (3000)
0.006 (100)
0.48 (2800)
ND
0.36 (150)
3.2 (1500)
3.9 (1300)
165 (323000)
0.64 (7100)
0.09 (450)
0.08 (1600)
0.24 (1850)
0.04 (1800)
0.11 (2200)
0.11 (1850)
0.58 (3400)
4 (6300)
6.7 (2700)
4.6 (2200)
17 (5700)
41 (82000)
2.1 (23000)
0.05
0.13
0.02
0.05
0.06
0.17
0.63
2.5
2.1
3
0.51
0.09
(CH2)5-O
(CH2)2-O-(CH2)2-O
CO2CH2CH=CHCH2-O (Z) Cl
CO2CH2CH=CHCH2-O (E) Cl
CO2(CH2)4-O
(CH2)5-O
(CH2)2-O-(CH2)2-O
COCH2-NMe-(CH2)2-O
(CH2)4-NH-CH2
(CH2)3-NH-(CH2)2
Cl Me
Cl
H
H
H
H
H
H
H
H
H
Cl
H
H
H
H
H
0.39
0.5
0.43
0.47
aKi values are the average of at least two determination, standard error of the mean <10%.
bThe 2ÂAPTT value is the concentration of inhibitor in plasma required to double the activated partial thromboplastin time.
cThe selectivity value represents the approximate ratio of the trypsin/tPA Ki to the thrombin Ki.
Figure 2. X-ray structure of proline derived macrocycle 8 bound in the
active site of thrombin.
Figure 3. X-ray structure of 31 bound in the active site of thrombin.
Residue Glu192 is labelled.
ted,5a, 15 this resulted in a ca. 50 fold loss in potency (27
vs 20) but with an improvement in functional activity.
Similarly, removal of the P1 chloro group from 23 fur-
ther improved functional activity and selectivity leading
to macrocycle 28 with a 0.39 mM 2ÂAPTT value and
micromolar potencies against trypsin and tPA. Other
modifications of the tether proved fruitful as well.
Incorporation of sarcosine into the tether provided
inhibitor 29 with a similar profile to 28. In the light of
the previous result, introduction of amino functionality
in the linker was studied more extensively. Substitution
of one of the tether methylenes with an amino group led
to the identification of inhibitors 30 and 31 with accep-
table 2ÂAPTT values and exquisite selectivity. In an
effort to understand the nature of this newly acquired
selectivity, we turned to X-ray crystallography and
molecular modeling.
The crystal structure of the thrombin-31 complex was
determined at 1.8-A resolution. Glu192, a flexible sur-
face residue located at the top of the S1 pocket, is not
completely resolved in the crystal structure due to its
high mobility. The most probable conformation of the
side chain, that occupying the region of highest experi-
mental density, is shown in Figure 3. In the crystal
structure, no direct salt bridge is observed, but a sol-
vent-separated salt bridge between the amino group of
the linker and glutamate side chain remains possible. A
rotamer of the Glu192 side chain generated using mole-
cular modeling techniques does allow for a direct

2784
P. G. Nantermet et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2781–2784
interaction between the phenethylamine nitrogen of 31
and the side chain and may contribute to the inhibitor’s
potency.
5. (a) Tucker, T. J.; Brady, S. F.; Lumma, W. C.; Lewis, S. D.;
Gardell, S. J.; Naylor-Olsen, A. M.; Yan, Y.; Sisko, J. T.;
Stauffer, K. J.; Lucas, B. J.; Lynch, J. J.; Cook, J. J.; Stranieri,
M. T.; Holahan, M.; Lyle, E. A.; Baskin, E. P.; Chen, I. W.;
Dancheck, K. B.; Krueger, J. A.; Cooper, C. M.; Vacca, J. P.
J. Med. Chem. 1998, 41, 3210. (b) Tucker, T. J.; Lumma,
W. C.; Mulichak, A. M.; Chen, Z.; Naylor-Olsen, A. M.;
Lewis, S. D.; Lucas, B. J.; Freidinger, R. M.; Kuo, L. C. J.
Med. Chem. 1997, 40, 830.
6. Burgey, C. S.; Robinson, K. A.; Lyle, T. A.; Sanderson,
P. E. J.; Lewis, S. D.; Lucas, B. J.; Krueger, J. A.; Lyle, E. A.;
Stranieri, M. T.; Holahan, M.; Sitko, G. R.; Cook, J. J.; Wal-
lace, A. A.; Clayton, F.; Bohn, D.; Leonard, Y.; Detwiler, T.;
Lynch, J. J.; Williams, P. D.; Coburn, C. A.; Dorsey, B. D.;
McMasters, D. R.; McDonough, C. M.; Sanders, W. M.;
Gardell, S. J.; Shafer, J. A.; Vacca, J. P. J. Med. Chem. 2003,
46, 461.
7. For simplicity we will refer to that end of the inhibitor
which binds to the distal hydrophobic binding pocket as P3.
The region of the enzyme active site overlying Gly216, Glu217,
and Gly219, which bridges the distal hydrophobic pocket and
the S1 specificity pocket, we will refer to as the S4 shelf to
distinguish it from the distal hydrophobic binding pocket.
Inhibitor moieties occupying this region of the active site we
will call P4.
8. (a) Tyndall, J. D. A.; Fairlie, D. P. Curr. Med. Chem. 2001,
8, 893. (b) Greco, M. N.; Maryanoff, B. E. Adv. Amino Acid
Mimet. Peptidomimet. 1997, 1, 41. (c) Greco, M. N.; Powell,
E. T.; Hecker, L. R.; Andrade-Gordon, P.; Kauffman, J. A.;
Lewis, J. M.; Ganesh, V.; Tulinsky, A.; Maryanoff, B. E.
Bioorg. Med. Chem. Lett. 1996, 6, 2947. (d) Maryanoff, B. E.;
Zhang, H. C.; Greco, M. N.; Glover, K. A.; Kauffman, J. A.;
Andrade-Gordon, P. Bioorg. Med. Chem. 1995, 3, 1025.
9. Horwell, D. C.; Hugues, J.; Hunter, J. C.; Pritchard, M. C.;
Richardson, R. S.; Roberts, E.; Woodruff, G. N. J. Med.
Chem. 1991, 34, 404.
In the case of the benzylamino analogue 30, on the
other hand, both direct and water-mediated salt bridges
with Glu192 are precluded by the linker. In the modeled
complex,16 the protonated amine participates in inter-
and intramolecular hydrogen bonds, to Gly216 and the
P1-P2 linker carbonyl respectively. These hydrogen
bonds, along with the overall negative charge of this
region of the active site caused by the proximity of
Glu192, offset the energy of desolvating the highly polar
protonated amine and result in a modest loss of potency
relative to 31. In contrast, the analogous residue to
thrombin’s Glu192 in both tPA and trypsin is Gln192.
As a result, whereas thrombin carries a net negative
electrostatic potential in this region of the active site,
this region is nearly electroneutral in trypsin and carries
a slight positive potential in tPA due to the presence of
nearby basic residue Lys143. The change in electrostatic
potential helps rationalize the poor potency of 30 for
trypsin and tPA, which unlike thrombin cannot offset
desolvation of the amino group with favorable electro-
static interactions.17
In conclusion, we have described the design, prepara-
tion, and evaluation of a series of macrocyclic proline
and pyrazinone thrombin ihibitors. In the case of pyr-
azinone derivatives, macrocyclization resulted in sig-
nificant improvements in potency relative to acyclic
analogues. The SAR observed led us to incorporate
specific functional groups in the tether resulting in
enhanced functional activity against thrombin and
exquisite selectivity against trypsin and tPA. Crystal-
lographic and molecular modeling studies revealed the
interactions between inhibitors and residue Glu192
which allow for such high selectivity.
10. Nantermet, P. G.; Selnick, H. G. Tetrahedron Lett. 2003,
44, 2401.
11. Furstner, A.; Langemann, K. Synthesis 1997, 792.
12. Lewis, S. D.; Ng, A. S.; Lyle, E. A.; Mellott, M. J.;
Appelby, S. D.; Brady, S. F.; Stauffer, K. S.; Sisko, J. T.; Mao,
S.-S.; Veber, D. F.; Nutt, R. F.; Lynch, J. J.; Cook, J. J.;
Gardell, S. J.; Shafer, J. A. Thromb. Haemost. 1995, 74, 1107.
13. Tucker, T. J.; Lumma, W. C.; Lewis, S. D.; Gardell, S. J.;
Lucas, B. J.; Baskin, E. P.; Woltmann, R.; Lynch, J. J.; Lyle,
E. A.; Appleby, S. D.; Chen, I.-W.; Dancheck, K. B.; Vacca,
J. P. J. Med. Chem. 1997, 40, 1565.
14. Calculated LogP (ChemDraw) for 20 and 23 are 3.7 and
2.6. respectively.
15. Lumma, W. C.; Witherup, K. M.; Tucker, T. J.; Brady,
S. F.; Sisko, J. T.; Naylor-Olsen, A. M.; Lewis, S. D.; Lucas,
B. J.; Vacca, J. P. J. Med. Chem. 1998, 41, 1011.
Acknowledgements
We thank M. Bogusky, J. Murphy, S. Pitzenberger, S.
Varga, A. Coddington, C. Ross, H. Ramjit, K. Anderson,
P. Ciecko and M. Zrada for analytical support, and Kris-
ten Glass for technical assistance in the preparation of 10.
16. (a) Multiple conformations of the macrocycles were gen-
erated using the metric matrix distance geometry algorithm JG
(S. Kearsley, Merck & Co., Inc., unpublished). The generated
conformers were energy-minimized within the thrombin active
site using MacroModel (ref 17b) with the MMFF force field,
and allowing Glu192 to adjust to the inhibitor. (b) Mohamadi,
F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.; Lipton,
M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J.
Comput. Chem. 1990, 11, 440.
17. Differences in electrostatic potential caused by E/Q192
have been proposed as an important factor in the selective
binding of PAI-1 to tPA versus thrombin. Rezaie, A. R. Bio-
chemistry 1998, 37, 13138.
References and Notes
1. Thrombosis in Cardiovascular Disorders; Fuster, V., Ver-
straete, M., Eds.; W.B. Saunders: Philadelphia, 1992.
2. Parenteral administration of low molecular weight heparin
and intense patient monitoring required with warfarin limit
their chronic utility.
3. Adang, A. E. P.; Rewinkel, J. B. M. Drugs Future 2000, 25,
369.
4. Coburn, C. A. Exp. Opin. Ther. Patents 2001, 11, 1. San-
derson, P. E. J. Annu. Rep. Med. Chem. 2001, 36, 79. Vacca,
J. P. Annu. Rep. Med. Chem. 1998, 33, 81.