Structure and property based design of factor Xa inhibitors: Pyrrolidin-2-ones with biaryl P4 motifs

Article abstract of DOI:10.1016/j.bmcl.2007.11.023

Structure and property based drug design was exploited in the synthesis of sulfonamidopyrrolidin-2-one-based factor Xa (fXa) inhibitors, incorporating biaryl P4 groups, producing highly potent inhibitors with encouraging oral pharmacokinetic profiles and significant but sub-optimal anticoagulant activities.

Full text of DOI:10.1016/j.bmcl.2007.11.023

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 23–27
Structure and property based design of factor Xa inhibitors:
Pyrrolidin-2-ones with biaryl P4 motifs
Robert J. Young,a,* Alan D. Borthwick,a David Brown,a Cynthia L. Burns-Kurtis,b
Matthew Campbell,a Chuen Chan,a Marie Charbaut,a Chun-wa Chung,a
a
a
a
a
´
Maire A. Convery, Henry A. Kelly, N. Paul King, Savvas Kleanthous,
Andrew M. Mason,a Anthony J. Pateman,a Angela N. Patikis,a Ivan L. Pinto,a
Derek R. Pollard,a Stefan Senger,a Gita P. Shah,a John R. Toomey,b
Nigel S. Watsona and Helen E. Westona
aGlaxoSmithKline, Medicines Research Centre, Gunnels Wood Road, Stevenage, Hertfordshire SG1 2NY, UK
bGlaxoSmithKline, 709 Swedeland Road, King of Prussia, PA 19406, USA
Received 5 September 2007; revised 6 November 2007; accepted 8 November 2007
Available online 17 November 2007
Abstract—Structure and property based drug design was exploited in the synthesis of sulfonamidopyrrolidin-2-one-based factor Xa
(fXa) inhibitors, incorporating biaryl P4 groups, producing highly potent inhibitors with encouraging oral pharmacokinetic profiles
and significant but sub-optimal anticoagulant activities.
ꢀ 2007 Elsevier Ltd. All rights reserved.
The quest for anticoagulant therapies with improved
efficacy/safety profiles has included considerable efforts
to discover small molecule inhibitors of factor Xa
(fXa),1 which plays a pivotal role in the coagulation cas-
cade. Clinical data are emerging to support this ap-
proach,2 although the pursuit of molecules suitable for
chronic oral administration has presented significant
medicinal chemistry challenges, including identifying
compounds with both acceptable pharmacokinetic pro-
files and good anticoagulant activities.
affinity, whilst expanding our understanding of structure
property relationships to promote good pharmacoki-
netic and anticoagulant profiles. In this communication,
the design, synthesis and evaluation of a series of sulfo-
namidopyrrolidin-2-one-based inhibitors 3 featuring
substituted biaryl P4 ligands is described.
H
H
N
N
O
Me
N
Me
N
N
O
O
S
R1
R1
O
S
O
O
O
N
O
Cl
R2
O
1
2
Classes of molecules that meet these challenges are
beginning to emerge, including neutral or weakly basic
molecules as represented by our sulfonamidopyrroli-
din-2-one-based inhibitors 1,3 which include examples
displaying attractive anticoagulant properties as well
as good oral exposure. Effective utilisation of X-ray
structural information led to a series of highly potent
acyclic analogues 2,4 and as part of a programme of
work to build on these latter findings, we sought to ex-
ploit structural information to increase intrinsic fXa
H
N
S
X
N
R1
O
Y
O
O
3
A FLO docking study5 using our available structural
information provided a strong impetus to expand our
synthetic programme to include biaryl motifs as P4 li-
gands.6 This identified the opportunity to secure good
hydrophobic contacts between the distal ring and the
aromatic residues in the S4 pocket.
Keywords: Factor Xa; Oral inhibitors; QSAR analysis.
*
Corresponding author. Tel.: +44 1438 768372; fax: +44 1438
763620; e-mail: Rob.J.Young@gsk.com
0960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.11.023

24
R. J. Young et al. / Bioorg. Med. Chem. Lett. 18 (2008) 23–27
The target molecules were accessed as outlined in
Scheme 1, utilising the reaction of the appropriate ani-
lines, activated as trimethylaluminium complexes, with
N-Boc homoserine lactone as the key step.7 The result-
ing homoserine amide was converted into the required
N-Boc aminopyrrolidin-2-one 4 using Mitsunobu condi-
tions.8 In this manner, the monoaryl bromide intermedi-
ate 4 was accessed from 4-bromo-2-fluoroaniline.
Subsequent cross couplings of 4 with aryl boronic acids9
or heterocycles gave the requisite biaryl intermediates 5
(Y = CF); however the C-3 stereocentre was racemised
under these conditions.10 The final compounds 6 were
accessed by removal of the Boc protection, followed
by sulfonylation under standard conditions. Alterna-
tively, the target compounds 6 were generated as race-
mates by conversion of 4 into the corresponding
sulfonamides 7 followed by cross-coupling with aryl
boronic acids.10 The pyridines 6 (Y = N) were accessed
using a variation on this route starting from 2-amino-
5-iodopyridine via the aryl iodide 8, which was con-
verted into the pinacol boronate ester 9. This was cross
coupled with the requisite aryl bromides,9 to yield the
biaryl intermediates 5 (Y = N) en route to the target
compounds 6 (Y = N).
An initial set of compounds explored a variety of distal
aromatic rings attached to the 4-position of a 2-fluoro
aniline.11 Gratifyingly, potent inhibition of fXa12 was
achieved with a number of these biaryl motifs, which
gave insight into the optimal requirements for activity,
with subtle changes producing significant shifts in activ-
ity (Table 1). Incorporation of an ortho substituent on
the distal ring was generally associated with increased
potency, with the 2-methylsulfonyl analogue 10 being
particularly noteworthy. In addition, various heterocy-
clic distal rings were tolerated. Interestingly, retaining
a terminal phenyl ring and incorporating a proximal 2-
pyridyl ring was well tolerated (cf. 19 and 20). However,
Table 1. fXa inhibitory activities12 anticoagulant potency13 for com-
pounds 10–20
H
H
H
N
Boc
c
N
Boc
HO
N
Boc
H
N
N
O
F
O
N
O
F
O
HN
O
S
+
a
b
O
X
O
Ar
NH2
Cl
F
4
Br
Br
Entry
X
Ar
fXa Ki (nM)
1.5· PT (lM)
Br
10
C–F
<0.1
1.1
51.7
O
SO2Me
H
H
N
Boc
R1
N
S
O
N
d, e
11
12
13
14
15
16
17
18
19
20
C–F
C–F
C–F
C–F
C–F
C–F
C–F
C–F
C–F
N
54.5
O
N
O
N
N
Y
Y
Y = CF or N
0.7
64.3
6
CN
N
5
Ar
Ar
O
H
R1
N
S
8.9
92.9
N
O
c
O
N
d, e
N
N
2.7
>100
>100
>100
>100
>100
>100
>100
F
4
7
N
Br
4.4
N
H
H
N
N
Boc
Boc
6.3
N
O
O
N
N
g
f
N
N
O
15.5
74.6
215
80
N
5
N
I
B
8
9
O
O
Scheme 1. Reagents and conditions: (a) Me3Al, DCM, 0 ꢁC to rt; (b)
DtBAD, Bu3P, THF, rt; (c) ArB(OR)2,Pd[P(Ph)3]4, DME, aq Na2SO4,
80 ꢁC or CuI, trans-1,2-diaminocyclohexane, K3PO4, DMSO, 110 ꢁC;
(d) HCl, dioxane, rt; (e) R1SO2Cl, pyridine, MeCN, rt; (f) bis(pina-
colato)diboron, PdCl2dppa, KOAc, 80 ꢁC; (g) ArBr, Pd[P(Ph)3]4,
DME, aq K2CO3, 80 ꢁC.

R. J. Young et al. / Bioorg. Med. Chem. Lett. 18 (2008) 23–27
25
Table 3. fXa inhibitory activities,12 anticoagulant potency,13 hydro-
phobicity calculations14a and plasma protein binding figures14c for
compounds 27 to 37
in spite of the very high fXa potency of some of these
compounds, very poor anticoagulant activity was
achieved, as assessed in the PT assay.13
H
S
Cl
N
N
To try to address this very poor conversion of excellent
potency into plasma-based activity, the effect of modu-
lating the overall hydrophobicity of the molecule
through changes to the P1 ligand was investigated—an
approach used with success in our previously reported
1-methyl-2-morpholin-4-yl-2-oxoethyl series 1.3b
X
O
S
O
N
O
Entry
X
fXa Ki
(nM)
1.5· PT clogD7.4 % HSA
(lM)
bound
27
28
29
30
31
32
33
34
35
36
37
SO2NH2
SO2NMe
SO2Me
NSO2Me
SO2NMe2
CN
<0.4
<0.3
<0.3
(14)a
0.4
17
25
1.1
1.7
1.3
1.7
2.1
2.2
0.9
1.8
3.5
4.7
3.7
98.0
Thus, retaining the highly potent 20-methylsulfonyl biar-
yl P4 group, a series of P1 sulfonamides was generated
(Table 2). Mirroring SAR established for the 1-methyl-
2-morpholin-4-yl-2-oxoethyl series 1, molecules with re-
duced hydrophobicity were identified (e.g. 21) that
showed better anticoagulant activity, even though they
had marginally lesser intrinsic potency.
38.9
58.4
65.1
73.7
78.6
>100
>100
>100
—
97.6
98.2
1.8
CONMe2
CF3
11.8
6
NMeSO2Me 9.1
OiPr
tBu
180
(2334)a
The activity of 20 (Table 1), the 2-pyridyl analogue, of-
fered a further opportunity to investigate the potential
benefits of reducing hydrophobicity, based on clogD
values.14a Thus, SAR were investigated in compounds
incorporating this 2-pyridyl proximal ring, whilst retain-
ing the preferred (5-chlorothien-2-yl)ethenyl P1 group
(Table 2) and making systematic changes to the ortho
substituent on the distal phenyl ring in the P4 group
(Table 3). A wide range of intrinsic activities was ob-
a fXa chromogenic assay data.4
served. However, the expected benefit of the reduced
hydrophobicity was not apparent in the measured values
(CHI logD values14b for 21 and 29 were 3.07 and 3.04,
respectively) contributing to the poorer translation of
high intrinsic potency into plasma-based activity. This
is evident in other fXa series incorporating biaryl S4 li-
gands and has been attributed to protein binding3b,6b
and/or substrate-dependent inhibition mechanisms.6b,15
Table 2. fXa inhibitory activities12 anticoagulant potency13 and
hydrophobicity calculations14a for compounds 10, 21–26 and human
serum albumin binding data14c for compounds 10, 21, 25
As part of the concerted process of using both property
and structure based design, a crystal structure of 21
bound into fXa (Fig. 1) revealed very clear density for
the compound in initial difference electron density
maps.16 This structure confirmed the binding mode pre-
dicted by modelling studies, including the anticipated
preference for the 3-S stereochemistry of the lactam ring
substituent. The (5-chlorothien-2-yl)ethenyl group sits in
the S1 pocket, the central pyrrolidinone makes a water
mediated interaction with the main chain carbonyl of
Ser214 and the biaryl group sits with the rings orthogo-
nal to each other in the P4 pocket. The methylsulfonyl
group sits above the S4 pocket and does not make any
direct interactions with the protein. This P4 group is sig-
nificantly longer than those for which structures have
previously been determined3,4 and this additional length
is accommodated by rearrangement of the water struc-
ture bridging Tyr99 and Phe174.
H
Me
N
N
F
R1
O
S O
O
S
O
O
Entry R1
fXa Ki 1.5· PT clogD7.4 % HSA
(nM)
(lM)
bound
S
Cl
21
22
23
10
24
25
26
0.2
0.1
<0.4
<0.1
2
21.3
28.7
30
3
97.6
S
Cl
4.5
3.1
3.7
3.1
2.5
2.5
Cl
Cl
51.7
99.5
>100
>100
98.8
99.6
This X-ray structure substantiated the modelled binding
mode and supported the idea that locking the two rings
into a preferred orthogonal conformation would confer
good levels of activity. However, this could not explain
the range of potencies observed with various ortho sub-
stituents of comparable bulk on the distal ring (Table 3).
The proximity of the opposite ortho hydrogen of the dis-
tal ring to the indole fragment of Trp215 suggested that
the partial charge on this hydrogen may be significant.
An electrostatic potential map of the indole ring indi-
cated a significantly enriched area of electron density
S
N
Cl
Cl
O
O
1.1
6.2
Cl

26
R. J. Young et al. / Bioorg. Med. Chem. Lett. 18 (2008) 23–27
Table 4. Pharmacokinetic parameters of compounds 10, 23 and 29 in
male Sprague–Dawley rats following intravenous and oral
administration17
Compound t1/2 (h) Clpb (mL/min/kg) Vssc (L/kg) Fd (%)
a
23
29
10
0.8
0.6
0.9
9.2
20
20
0.5
0.7
1.1
41
29
10
a t1/2, half-life of the test compound expressed in hours.
b Clp, plasma clearance of the test compound expressed as mL/min/kg.
c Vss, steady state volume of distribution of test compound expressed
as L/kg.
d F, oral bioavailability of test compound expressed as percentage.
In summary, the synthesis, activities and structure prop-
erty relationships have been explored within a rationally
designed series of sulfonamidopyrrolidin-2-one-based
fXa inhibitors incorporating non-basic biaryl P4 motifs.
This work has identified several highly potent examples
with encouraging oral pharmacokinetic profiles and sig-
nificant but sub-optimal anticoagulant activities. X-ray
structural data confirmed the binding mode anticipated
from modelling studies and quantitative analysis sug-
gested a plausible explanation of the nature of the distal
ring interaction. These findings, taken together with our
increased in-house knowledge, gave clear indications for
our ongoing programme of work; this is developed in a
second generation series, described in the accompanying
paper.
Figure 1. X-ray crystal structure of the S-enantiomer of 21 bound into
fXa, showing S4 interactions as described in the text, highlighting the
proximity of the ortho hydrogen to an enriched area of electron density
on Trp215.
H
S
Cl
N
N
X
O
S
O
N
O
H
10
9.5
9
pKi
References and notes
8.5
8
1. (a) Quan, M. L.; Smallheer, J. M. Curr. Opin. Drug
Discovery Devel. 2004, 7, 460; (b) Gould, W. R.; Leadley,
R. J. Curr. Pharm. Des. 2003, 9, 2337; (c) Hirsh, J.;
O’Donell, M.; Weitz, J. I. Blood 2005, 105, 453; (d)
Golino, P.; Loffredo, F.; Riegler, L.; Renzullo, E.;
Cocchia, R. Curr. Opin. Invest. Drugs 2005, 6, 298; (e)
Alexander, J. H.; Singh, K. P. Am. J. Cardiovasc. Drugs
2005, 5, 279.
2. (a) Saiah, E.; Soares, C. Curr. Top. Med. Chem. 2005, 5,
1677; (b) Weitz, J. I.; Linkins, L.-A. Exp. Opin. Invest.
Drugs 2007, 16, 271; (c) Spyropoulos, A. C. Exp. Opin.
Invest. Drugs 2007, 16, 431.
7.5
7
6.5
6
5.5
5
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Hammett σm constant
3. (a) Watson, N. S.; Brown, D.; Campbell, M.; Chan, C.;
Chaudry, L.; Convery, M. A.; Fenwick, R.; Hamblin, J.
N.; Haslam, C.; Kelly, H. A.; King, N. P.; Kurtis, C. L.;
Leach, A. R.; Manchee, G. R.; Mason, A. M.; Mitchell,
C.; Patel, C.; Patel, V. K.; Senger, S.; Shah, G. P.; Weston,
H. E.; Whitworth, C.; Young, R. J. Bioorg. Med. Chem.
Lett. 2006, 16, 3784; (b) Chan, C.; Borthwick, A. D.;
Brown, D.; Campbell, M.; Chaudry, L.; Chung, C.-W.;
Convery, M. A.; Hamblin, J. N.; Johnstone, L.; Kelly, H.
A.; Kleanthous, S.; Kurtis, C. L.; Patikis, A.; Patel, C.;
Pateman, A. J.; Senger, S.; Shah, G. P.; Toomey, J. R.;
Watson, N. S.; Weston, H. E.; Whitworth, C.; Young, R.
J.; Zhou, P. J. Med. Chem. 2007, 50, 1546.
4. Young, R. J.; Campbell, M.; Borthwick, A. D.; Brown,
D.; Burns-Kurtis, C. L.; Chan, C.; Convery, M. A.;
Crowe, M. C.; Dayal, S.; Diallo, H.; Kelly, H. A.; King,
N. P.; Kleanthous, S.; Mason, A. M.; Mordaunt, J. E.;
Patel, C.; Pateman, A. J.; Senger, S.; Shah, G. P.; Smith,
P. W.; Watson, N. S.; Weston, H. E.; Zhou, P. Bioorg.
Med. Chem. Lett. 2006, 16, 5953.
Figure 2. QSAR analysis of measured pKi values for all 11 compounds
in Table 3 versus the Hammett rm constant for the ortho substituents X
in close proximity to the hydrogen, the charge on which
would be influenced by the r-electron withdrawing effect
of the substituent on the opposite ortho position. This
hypothesis was substantiated by a plot of the Hammett
r
m constant17 versus fXa pKi, in which a clear trend was
observed towards increased potency with increased elec-
tron withdrawing power (Fig. 2). The trend line for this
analysis showed for this series that pKi = 5.28 ·
rm + 6.39, with an R2 value of 0.897.
Pharmacokinetic studies with selected compounds were
carried out in rats18 (Table 4) and demonstrated that
encouraging profiles with significant oral exposure were
achievable in this series.

R. J. Young et al. / Bioorg. Med. Chem. Lett. 18 (2008) 23–27
27
5. McMartin, C. FLO/QXP Molecular Modeling software.
Thistlesoft, Colebrook, CT, USA.
ray structures determined with various fXa bound inhib-
itors of this type in our work often falls within this
distance range and could be consistent with such a
fluorophilic interaction, although the occupancy of this
particular ring orientation appeared to be not quite
complete.16.
6. Other groups have reported independently on series of fXa
inhibitors incorporating biaryl P4 ligands, for example (a)
Kohrt, J. T.; Filipski, K. J.; Cody, W. L.; Bigge, C. F.; La,
F.; Welch, K.; Dahring, T.; Bryant, J. W.; Leonard, D.;
Bolton, G.; Narasimhan, L.; Zhang, E.; Peterson, J. T.;
Haarer, S.; Sahasrabudhe, V.; Janiczek, N.; Desiraju, S.;
Hena, M.; Fiakpui, C.; Saraswat, N.; Sharma, R.; Sun, S.;
Maiti, S. N.; Leadley, R.; Edmunds, J. J. Bioorg. Med.
Chem. 2006, 14, 4379; (b) Pinto, D. J. P.; Galemmo, R. A.;
Quan, M. L.; Orwat, M. J.; Clark, C.; Li, R.; Wells, B.;
Woerner, F.; Alexander, R. S.; Rossi, K. A.; Smallwood,
A.; Wong, P. C.; Luettgen, J. M.; Rendina, A. R.; Knabb,
R. M.; He, K.; Wexler, R. R.; Lam, P. Y. S. Bioorg. Med.
Chem. Lett. 2006, 16, 5584; (c) Pinto, D. J. P.; Orwat, M.
J.; Wang, S.; Fevig, J. M.; Quan, M. L.; Amparo, E.;
Cacciola, J.; Rossi, K. A.; Alexander, R. S.; Smallwood,
A. M.; Luettgen, J. M.; Liang, L.; Aunggst, B. J.; Wright,
M. R.; Knabb, R. M.; Wong, P. C.; Wexler, R. R.; Lam,
P. Y. S. J. Med. Chem. 2001, 44, 566; (d) for further
examples see Ref. 1 and cited references.
12. Factor Xa inhibitory activities were largely determined
using Rhodamine 110, bis-(CBZ-glycylglycyl)-L-arginine
amide as fluorogenic substrate, with the others using N-a-
Z-D-Arg-Gly-Arg-p-nitroanilide as chromogenic sub-
strate; details are described in Ref. 3b. For compounds
tested in both assays, comparable levels of activity were
observed.
13. Anticoagulant activities were determined in the prothrom-
bin time (PT) assay; see Ref. 3b.
14. (a) Hydrophobicity predictions, expressed as clogD7.4
,
were calculated using Advanced Chemistry Development
software v8.0; Hydrophobicity measurements are reported
as Chromatographic Hydrophobicity Index (CHI) logD7.4
´
values, for details see (b) Valko, K.; Du, C. M.; Bevan, C.;
Reynolds, D. P.; Abraham, M. H. Curr. Med. Chem. 2001,
8, 1137; high throughput human serum albumin binding
´
7. Bell, I. M.; Beshore, D. C.; Gallicchio, S. N.; Williams, T.
M. Tetrahedron Lett. 2000, 41, 1141.
was measured as described in (c) Valko, K.; Numhuck, S.;
Bevan, C.; Abraham, M. H.; Reynolds, D. P. J. Pharm.
Sci. 2003, 92, 2236.
8. Typically this reaction, in our hands, furnished products in
other series with ee’s comparable to those reported in Ref.
7 as determined by 19F NMR of Mosher amide derivatives
of particular deblocked amines analogous to 7 or chiral
hplc analysis of specific final compounds.
9. The majority of aryl bromides and boronic acids combi-
nations employed were available from the GSK monomer
collection; or prepared as described in Borthwick, A. D.;
Chan, C.; Kelly, H. A.; King, N. P.; Kleanthous, S.;
Mason, A. M.; Pinto, I. L.; Pollard, D. R.; Senger, S.;
Shah, G. P.; Watson, N. S.; Young, R. J. WO 03053925
A1, 2003.
15. As part of a broader investigation, a number of examples
from this and other related fXa series were subjected to an
in depth analysis of their protein binding and fXa kinetics,
using the Biacoreꢂ surface plasmon resonance technique.
These experiments indicated that biaryl-based molecules,
in comparison to other structural classes we have inves-
tigated,3,4 were generally more tightly bound to specific
human serum albumin sites and had reduced fXa on-rates,
consistent with their relatively poorer translation of
intrinsic potency into plasma-based activity. Chung,
C.W. et al. manuscript in preparation.
10. The stereochemistry of the 3-amino linkage we believe was
most probably racemised under the strongly basic cross-
coupling conditions. However, the preferred (3S)-enan-
tiomer, anticipated from docking studies, was clearly
evident in the structures bound within fXa when co-
crystallised from the racemates. A fuller exploration of the
preferred stereochemistry of similar biaryl molecules is
discussed in more detail in the accompanying paper.
11. The fluorine substituent was chosen to promote non-
planarity of the proximal aryl ring and pyrrolidin-2-one
ring. An intermolecular CAF ꢀ ꢀ ꢀ C@O interaction has
been suggested when fluorine atoms are located within the
16. The only ambiguity in the electron density was regarding
the orientation of the proximal ring. A small percentage of
molecules in the crystal contain the ring rotated through
˚
180ꢁ. The structure for 21 was refined at 1.95 A (overall
Rmerge is 0.049) in Refmac5 to a final Rfactor of 0.196 and
Rfree of 0.241, using procedures described in Ref. 3b. Co-
ordinates are deposited in the protein data bank with code
2vh6.
17. Values for rm were used from Hansch, C.; Leo, A.; Taft,
W. Chem. Rev. 1991, 91, 165.
18. The formulation used for both iv and po dosing was a
5:95% (v/v) mixture of DMSO and 50:50 PEG-200:sterile
water. Serial blood samples were collected into heparinised
containers at various time-points and blood centrifuged to
yield plasma. These studies used two animals for each (iv/
po) leg.
˚
range of 3–4 A; Olsen, J. A.; Banner, D. W.; Seiler, P.;
Obst Sander, U.; D’Arcy, A.; Stihle, M.; Muller, K.;
¨
Diederich, Angew. Chem. Int. Ed. 2003, 42, 2507. The
intramolecular CAF ꢀ ꢀ ꢀ C@O distances observed in the X-