Factor VIIa inhibitors: Improved pharmacokinetic parameters

Article abstract of DOI:10.1016/j.bmcl.2006.01.037

Efforts to improve the potency and pharmacokinetic properties of small molecule factor VIIa inhibitors are described. Small structural modifications to existing leads allow the modulation of half-life and clearance, potentially making these compounds suitable candidates for drug development.

Full text of DOI:10.1016/j.bmcl.2006.01.037

Bioorganic & Medicinal Chemistry Letters 16 (2006) 2243–2246
Factor VIIa inhibitors: Improved pharmacokinetic parameters
Aleksandr Kolesnikov,* Roopa Rai, Wendy B. Young, Joyce Mordenti, Liang Liu,
Steven Torkelson, William D. Shrader, Ellen M. Leahy, Huiyong Hu, Erik Gjerstad,
James Janc, Bradley A. Katz and Paul A. Sprengeler
Celera Genomics, 180 Kimball Way, South San Francisco, CA 94080, USA
Received 18 November 2005; revised 6 January 2006; accepted 9 January 2006
Available online 7 February 2006
Abstract—Efforts to improve the potency and pharmacokinetic properties of small molecule factor VIIa inhibitors are described.
Small structural modifications to existing leads allow the modulation of half-life and clearance, potentially making these compounds
suitable candidates for drug development.
Ó 2006 Elsevier Ltd. All rights reserved.
We have previously identified small molecule serine pro-
tease inhibitors of which analogs 1–3 are exemplary of
this series.1,2 As described earlier2–4 our goal was to
develop potent and selective small molecule inhibitors
of the factor VIIa/Tissue Factor (fVIIa/TF) complex
for the treatment of thromboembolic disease. We were
particularly interested in achieving >1000-fold selectivi-
ty for the fVIIa/TF complex versus factor Xa (fXa) and
thrombin as it has been postulated that such selectivity
may be important for improving the safety profile of
anticoagulants.5 Additionally, we were interested in gen-
erating analogs that would be amenable to once daily
subcutaneous dosing in human.
led to a dramatic increase in selectivity for fVIIa over
fXa as shown in Table 1.3 Compounds 1–3, however,
were cleared rapidly following intravenous (iv) adminis-
tration to rats, and the MRT was less than 1 h.6 It was
found that glucuronidation is the major mode of metab-
olism.7 In the present paper, we describe our efforts to
improve the potency, selectivity, and pharmacokinetics
of our series to achieve analogs that would be suitable
for further development.
X-ray crystallography data indicate that the acetic acid
moiety off the 50 position in structures 1–3 forms a favor-
able interaction with the Lys192 of fVIIa.1,2 The same car-
boxylate anion repels Glu192 of thrombin contributing to
the enhanced selectivity against thrombin. As Lys60A of
fVIIa is also near this region, we decided to build in anoth-
er acidic moiety here to take advantage of this potential
interaction. Additionally, this 50-vector also accesses
solvent and could simultaneously be used as a handle to
manipulate physiochemical properties which in turn
may alter the pharmacokinetic parameters. We were
CO2H
CO2H
5'
5'
OH
F
X
N
H2N
HN
H2N
HN
NH OH
NH OH
NO2
1: X = N; 2: X = CH
3
The early leads 1 and 2 lacked the desired selectivity
against fXa and contained an arylnitro-group, which
has the potential to form toxic metabolites. In response
to these concerns, the nitrophenyl group was replaced
with a 200-hydroxy-500-fluorophenyl group, as in 3, which
Table 1. SAR and pharmacokinetic parameters of early analogs in rat
following IV administration
Compound fVIIa/TF
Ki (lM)
Selectivity
fXa fIIa
CLa MRTb
(mL/min/kg) (h)
1
2
3
0.015
0.003
0.002
20 >10,000
20 960
2600 >75,000
9.2
10.5
8.1
<1
<1
<1
Keywords: Factor VIIa; Tissue factor; Serine protease inhibitor; Ant-
icoagulants; Amidines; Dicarboxylic acids.
a CL, plasma clearance.
b MRT, mean residence time.
*
Corresponding author. Tel.: +1 4158105495; e-mail: aleks_
kolesnikov@yahoo.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.01.037

2244
A. Kolesnikov et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2243–2246
Table 2. Structure–activity relationship and pharmacokinetic parameters of diacid analogs (300-nitrophenyl series) in rat following IV administration
R
X
H2N
NH OH
HN
NO2
R
Compound
X
fVIIa/TF Ki (lM)
Selectivity
fIIa
CLb (mL/min/kg)
MRTc (h)
fXa
40
COOH
4a
N
0.022
0.036
0.017
0.021
0.001
0.013
9545
0.04
na
>24
na
COOH
COOH
5
N
25
40
20
50
5
5000
9000
COOH
COOH
COOH
6a
N
0.5
5.2
0.1
1.8
5
HOOC COOH
7
N
40,000
700
<1
>24
1.5
COOH
8a
CH
CH
COOH
COOH
COOH
9
3900
a Racemic mixture.
b CL, plasma clearance.
c MRT, mean residence time.
looking to identify compounds that did not suffer from
extensive glucuronidation and offered a good pharmaco-
kinetic profile to support once daily dosing in human. To-
ward this goal, a series of diacids were generated off the 50
position of scaffolds 1 and 2, as depicted in Table 2. When
we began this work, the more selective and preferred 200-
hydroxo-500-phenyl analog 3 had not yet been uncovered
so the work was initiated in the arylnitro series. We gener-
ated analogs in both the benzimidazole and indole series
in order to evaluate the SAR in parallel. Succinic (4 and
8), 3-pentanedioic (5 and 9), 2-pentanedioic (6), and meth-
ylmalonic (7) acids were chosen to explore the region near
the Lys60A and Lys192. The benzimidazole analogs (4–7)
ranged in potency from 17 to 36 nM, all showing only a
slight change in binding energy as compared to com-
pound 1. Indole derivatives (8–9), although more potent,
show a similar trend in SAR. Selectivity against fXa re-
mains low in both benzimidazole and indole series.
more methylene group (viz. 6, 9). The pharmacokinetic
properties of these pentanedioic acids depend on their
structure—the asymmetrical one being more stable. It
should be noted that compounds with asymmetrical diac-
ids were tested as a racemic mixture of the enantiomers.
These are likely to undergo metabolic transformation at
different rates, so the pharmacokinetic parameters repre-
sent a statistical average of an ever-changing ratio.
Metabolism of the malonic acid derivative 7 may in part
involve a mechanism other than glucuronidation (decar-
boxylation) resulting in a faster clearance.
100.000
10.000
1.000
Although we did not realize any gains in potency in ana-
logs 4–9 as compared to 1 and 2, we did achieve some dra-
matic alterations in pharmacokinetic properties (Table 2).
While monocarboxylic acids 1 and 2 demonstrate rela-
tively high CL, introduction of a succinic acid moiety
directly bound to the central phenol ring dramatically
decreases CL and increases MRT (Fig. 1). This improve-
ment in pharmacokinetics seems to be directly related to
the succinic acid. In vivo half-life substantially decreases
when two carboxylate groups become separated by one
0.100
0.010
0.001
0
24
48
72
Time (h)
Figure 1. Concentration versus time profile of analogs 2 (j), 4 (m),
and 8 (s) in rat following IV administration at 1 mg/kg.

A. Kolesnikov et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2243–2246
2245
Table 3. Structure–activity relationship and pharmacokinetic parameters of diacid analogs (200-hydroxy-500-fluorophenyl series) in rat following IV
administration
R
OH
N
H2N
HN
NH OH
F
R
Compound
fVIIa/TF Ki (lM)
Selectivity
fIIa
CLb (mL/min/kg)
MRTc (h)
fXa
COOH
COOH
10a
11a
0.004
0.004
1800
55,500
37,000
0.3
2.6
9
COOH
COOH
1250
<1
a Racemic mixture.
b CL, plasma clearance.
c MRT, mean residence time.
Our next goal was to transfer the ‘PK SAR’ that we
learned on scaffolds 1 and 2 to our more selective series
3. Toward this end, we placed a succinic acid moiety at
the 50-position giving analog 10. Similar to the aryl nitro
series, no potency improvement was obtained with this
modification, but the pharmacokinetic enhancement
did transfer (Table 3), resulting in identification of phen-
ylsuccinic acid derivative 10 with not only excellent
potency and selectivity but also a good pharmacokinetic
profile in rat. Again, deviation from phenylsuccinic acid
moiety (separation by one methylene group, as in 11)
resulted in a sharp decrease in stability in vivo.
Reaction of 4-iodoanisole and dimethylitaconate using
Heck reaction conditions8 and subsequent hydrogena-
tion led to 2-(4-methoxy-benzyl)-succinic acid dimethyl
ester (precursor to 12). The same method was applied
for the preparation of 13, although 4-iodophenol and
diethylfumarate were used as the starting materials.3
Dimethyl ester of 2-(4-methoxy-phenyl)-pentanedioic
acid (precursor to 14) was synthesized by addition of
methylacrylate to 4-methoxyphenylacetate.9 Dimethyl
R
R
X
a,b,c
d
12-16
OHC
OHC
Br
OMem
Oral bioavailability evaluations in rat, with selected ana-
logs in this series, demonstrated limited absorption (i.e.,
<1%), most likely due to the charged succinate and ami-
dino groups. Some analogs (e.g., 10) show high bioavail-
ability (i.e., >90%) after subcutaneous administration in
rats. Combined with their good pharmacokinetic pro-
files, these compounds represent promising candidates
as parenteral anticoagulants.
OMem
Y
17
18: X=H, Y=NO2
19: X=OMe, Y=F
e,h, ref. 2
e,g
R
R
OH
F
X
N
H2N
HN
H2N
HN
NH OH
NH OH
Compounds 4–11 were obtained by modifying known
synthetic procedures2,3 as depicted in Scheme 1. The
necessary 4-substituted phenol intermediates 12–16 were
each prepared in a unique manner. Methyl ether and
alkyl ester groups (where applicable) were removed by
heating the corresponding anisoles in aqueous hydrogen
bromide yielding phenolic derivatives of dicarboxylic
acids, which were then reesterified.
NO2
4-9
10,11
Scheme 1. Reagents and conditions: (a) HCHO, MgCl2, triethylamine,
MeCN, heat; (b) NBS, DMF; (c) MEM-Cl, Hunig’s base, DCM; (d)
Arylboronic acid, Pd(PPh3)4, 2 M aq Na2CO3, dimethoxyethane,
reflux; (e) 3,4-diaminobenzamidine, benzoquinone, MeOH, reflux; (h)
4 M HCl aq, 100 °C; (g) 48% HBr aq, 100 °C.
MeOOC
COOMe
COOMe
MeOOC
COOMe
COOMe
COOMe
COOMe
MeOOC
MeOOC
OH
OH
OH
14
OH
15
OH
16
12
13

2246
A. Kolesnikov et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2243–2246
Shrader, W.; Sprengeler, P. A.; Trapp, S.; Wang, J.;
Young, W. B.; Mackman, R. L. J. Med. Chem. 2001, 44,
2753.
ester of 3-(4-methoxy-phenyl)-pentanedioic acid (pre-
cursor to 15) was obtained by reaction of 4-methoxy-
benzaldehyde with ethyl acetoacetate catalyzed by
piperidine followed by alkali hydrolysis.10 Commercially
2. Young, W. B.; Kolesnikov, A.; Rai, R.; Sprengeler, P. A.;
Leahy, E. M.; Shrader, W. D.; Sangalang, J.; Burgess-
Henry, J.; Spencer, J.; Elrod, K.; Cregar, L. Bioorg. Med.
Chem. Lett. 2001, 11, 2253.
3. Shrader, W. D.; Kolesnikov, A.; Burgess-Henry, J.; Rai,
R.; Hu, Huiyong; Torkelson, S.; Young, W. B.; Spreng-
eler, P.; Katz, B.; Yu, C.; Cabuslay, R.; Sanford, E.; Janc,
J. J., Bioorg. Med. Chem. Lett., in press.
available
2-(4-hydroxy-benzylidene)-malonic
acid
dimethyl ester was directly hydrogenated to malonic es-
ter 16. Phenol derivatives 12–16 were then transformed
into the final benzimidazole and indole targets 4–11
via a series of well-precedented steps. Formylation of
each phenol 12–16 with paraformaldehyde and magne-
sium chloride, bromination with NBS, and subsequent
MEM-protection of the phenol group afforded com-
pounds of general formula 17. Each aldehyde, as in
17, was then subjected to coupling with commercially
available boronic acids under Suzuki reaction condi-
tions producing compounds of the general structure 18
and 19. Corresponding benzimidazoles 4–7 and 10, 11
were synthesized by an oxidative cyclization of diamino-
benzamidine onto the aldehyde group of 18 followed by
cleavage of the protecting groups. Synthesis of the in-
doles 8, 9 was conducted using Sonogashira coupling
route starting from aldehydes 19 as previously
described.2
4. Kolesnikov, A.; Rai, R.; Young, W. B.; Torkelson, S.;
Shrader, W. D.; Leahy, E. M.; Katz, B. A.; Sprengeler, P.
A.; Liu, L.; Mordenti, J.; Gjerstad, E.; Janc, J. Abstracts of
Papers, 229th ACS National Meeting, San Diego, CA,
United States, March 13–17, 2005, MEDI-250.;
5. Suleymanov, O. D.; Szalony, J. A.; Salyers, A. K.;
Lachance, R. M.; Parlow, J. J.; South, M. S.; Wood, R.
S.; Nicholson, N. S. J. Pharm. Exp. Ther. 2003, 306, 1115.
6. Plasma concentrations of the compounds were determined
by LC/MS/MS. The plasma sample was processed using
acetonitrile precipitation and then the supernatant was
injected onto the LC column. The limit of quantitation of
the assay was 1–6 nM. Pharmacokinetic data were ana-
lyzed by WinNonlin-Pro (Pharsight Corp.), using com-
partmental analysis. Pharmacokinetic parameters,
including the area-under-the-curve (AUC), clearance
(CL), volume of distribution (Vd), and mean residence
time (MRT), were determined. Clearance and mean
residence time were calculated as follows: CL = Dose/
AUC and MRT = Vd/CL.
In summary, a series of potent and selective dicarboxylic
acid analogs were generated as factor VIIa/TF inhibi-
tors. The pharmacokinetic properties of these analogs
after IV dosing to rat were dramatically different. The
succinic acid analog 10 shows the best pharmacokinetic
profile and will be the focus of future lead optimization
for development of a parenteral anticoagulant.
7. Rai, R.; Kolesnikov, A.; Young, W., Shrader, W.; Liu, L.;
Enriquez, P. Abstracts of Papers, 224th National Meeting
of the American Chemical Society, Boston, MA, August
18–22, 2002; MEDI-290.;
8. Cortese, N.; Ziegler, C.; Hrnjez, B.; Heck, R. J. Org.
Chem. 1978, 43, 2952.
References and notes
9. Aoki, S.; Sasaki, S.; Koga, K. Tetrahedron Lett. 1989, 30,
7229.
10. Eiher, T.; Wobido, M.; Speicher, A. J. Prakt. Chem./
Chem. Ztg 1996, 338, 706.
1. Verner, E.; Katz, B. A.; Spencer, J. R.; Allen, D.; Hataye,
J.; Hruzewicz, W.; Hui, H. C.; Kolesnikov, A.; Li, Y.;
Luong, C.; Martelli, A.; Radika, K.; Rai, R.; She, M.;