Potent noncovalent thrombin inhibitors that utilize the unique amino acid D-dicyclohexylalanine in the P3 position. Implications on oral bioavailability and antithrombotic efficacy

Article abstract of DOI:10.1021/jm970140s

In an effort to prepare orally bioavailable analogs of our previously reported thrombin inhibitor 1, we have synthesized a series of compounds that utilize the unique amino acid D-dicyclohexylalanine as a P3 ligand. The resulting compounds are extremely p

Full text of DOI:10.1021/jm970140s

J . Med. Chem. 1997, 40, 1565-1569
1565
Expedited Articles
P oten t Non cova len t Th r om bin In h ibitor s Th a t Utilize th e Un iqu e Am in o Acid
D-Dicycloh exyla la n in e in th e P 3 P osition . Im p lica tion s on Or a l Bioa va ila bility
a n d An tith r om botic Effica cy
Thomas J . Tucker,*,† William C. Lumma,† S. Dale Lewis,‡ Stephen J . Gardell,‡ Bobby J . Lucas,‡
Elizabeth P. Baskin,§ Richard Woltmann,§ J oseph J . Lynch,§ Elizabeth A. Lyle,§ Sandra D. Appleby,§
I-Wu Chen,| Kimberley B. Dancheck,| and J oseph P. Vacca†
Departments of Medicinal Chemistry, Biological Chemistry, General Pharmacology, and Drug Metabolism,
Merck Research Laboratories, West Point, Pennsylvania 19486
Received March 11, 1997X
In an effort to prepare orally bioavailable analogs of our previously reported thrombin inhibitor
1, we have synthesized a series of compounds that utilize the unique amino acid D-
dicyclohexylalanine as a P3 ligand. The resulting compounds are extremely potent and selective
thrombin inhibitors, and the N-terminal Boc derivative 8 exhibited excellent oral bioavailability
and pharmacokinetics in both rats and dogs. The des-Boc analog 6 was not orally bioavailable
in rats. The high level of oral bioavailability observed with 8 appears to be a direct function
of its increased lipophilicity versus other close analogs. Although increased lipophilicity may
serve to increase the oral absorption of tripeptide thrombin inhibitors, it also appears to have
detrimental effects on the antithrombotic properties observed with the compounds. Compound
6 performed extremely well in our in vivo antithrombotic assay, while the much more lipophilic
but essentially equipotent analog 8 performed poorly. We have found that in general with
this series of thrombin inhibitors as well as with other unreported series, increased lipophilicity
and the associated increases in plasma protein binding have detrimental effects on 2X APTT
values and subsequent performance in in vivo antithrombotic models.
In tr od u ction
Thrombin is a trypsin-like serine protease that plays
a pivotal role in the blood coagulation cascade. In an
attempt to develop orally bioavailable antithrombotic
agents, we have focused on the design and synthesis of
novel thrombin inhibitors. Recently, we reported a
series of potent thrombin inhibitors including compound
11 (Figure 1). While these compounds were extremely
potent and selective inhibitors of thrombin, they dis-
played poor oral bioavailability. Compound 1 is an
extremely polar molecule, possessing a log P value of
-0.539. The highly polar nature of 1 appears to be a
major obstacle to absorption of the compound after oral
dosing. In an effort to synthesize orally bioavailable
derivatives of these earlier inhibitors by increasing their
log P values, we have utilized the unique amino acid
D-dicyclohexylalanine as a novel P3 group (Figure 1).2
In this paper, we detail the results of these investiga-
tions and report a series of potent and orally bioavail-
able tripeptide thrombin inhibitors.
F igu r e 1.
lipophilicity of this compound could possibly enhance
its absorption from the gut after oral dosing. The
starting amino acid dicyclohexylalanine was prepared
initially in its racemic form via catalytic reduction of
racemic diphenylalanine3 (Scheme 1). We found iridium
black to be the most efficient catalyst for effecting this
reduction at normal Parr apparatus pressures (ap-
proximately 60 psi). The CBZ-protected racemic amino
acid 2a was coupled to proline methyl ester using
standard coupling conditions. Saponification, followed
by reversed phase preparatory HPLC, provided the
clean diastereomers 3a and 4a . Compounds 3a and 4a
were each coupled to CBZ-protected trans-4-(amino-
methyl)cyclohexylamine,4 and subsequent deprotection
provided the pure product diastereomers 5 and 6.
Stereochemical assignments were based on the results
of the thrombin inhibition assay, in which compound 6
was shown to be the more potent inhibitor of the two
and was thus assigned the (R) stereochemistry at the
P3 asymmetric center. When D-diphenylalanine became
commercially available,3 the sequence was successfully
Ch em istr y
Since D-diphenylalanine is well tolerated by the S3
lipophilic binding pocket of the thrombin active site, we
theorized that the fully reduced version of this amino
acid should also be well tolerated. The increased
† Department of Medicinal Chemistry.
‡ Department of Biological Chemistry.
§ Department of General Pharmacology.
| Department of Drug Metabolism.
X Abstract published in Advance ACS Abstracts, May 15, 1997.
S0022-2623(97)00140-4 CCC: $14.00 © 1997 American Chemical Society

1566 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 11
Tucker et al.
Ta ble 2. Inhibition of Various Human Serine Proteases4 by
Sch em e 1
Compound 6
enzyme
Ki (µM)
selectivity ratio
thrombin
trypsin
TPA
plasmin
plasma Kallekrein
factor Xa
0.000056
0.12
137
165
5
95
2143
2.5 × 106
3.0 × 106
89000
1.7 × 106
act protein C
417
7.5 × 106
might be expected (Table 3). Compounds 6 and 8 were
initially evaluated for oral bioavailabily in rats, and the
results are summarized in Table 3. Compound 6
exhibited only 2% oral bioavailability in the rat, whereas
the much more lipophilic Boc derivative 8 showed
excellent oral bioavailabilty and good pharmacokinetic
behavior. On the basis of these results, compound 8 was
also evaluated for oral bioavailability in dogs, where it
exhibited an excellent pharmacokinetic profile and oral
bioavailability of 90% (Table 3).
Compounds 6 and 8 were also evaluated for their
antithrombotic activity. We have extensively used 2X
APTT (activated partial thromboplastin time) values6
as a qualitative in vitro indicator of potential anti-
thrombotic activity. The 2X APTT value is defined as
the the inibitor concentration (in µM) required to double
the APTT in plasma. The values determined for com-
pounds 6 and 8 (Table 4) were high, which did not bode
well for potential in vivo antithrombotic efficacy. Evalu-
ation of compounds 6 and 8 in the rat carotid artery
FeCl3-induced thrombosis model5 (Table 4) indicated
that compound 6 possessed very good antithrombotic
activity (1/5 occlusions dosed iv at 6 µg/kg/min), while
the highly bioavailable compound 8 showed only mar-
ginal antithrombotic activity (4/6 occlusions, same dos-
ing). Plasma concentrations for compounds 6 and 8
were 1.9 and 1.1 µM, respectively, at the end of the
experiment. In the case of compound 6, the plasma
concentration was well in excess of the 2X APTT value
determined for the compound in rat plasma. This data
suggested that antithrombotic activity might be ob-
served in vivo with this inhibitor. However, in the case
of compound 8 the concentration of the inhibitor was
well below the 2X APTT value determined for the
compound, suggesting that the concentration of com-
pound necessary to achieve good antithrombotic activity
was never achieved. The rapid IV clearance of com-
pound 8 from the rat, along with its high 2X APTT value
in rat plasma appear to combine to compromise the
antithrombotic activity observed in vivo.
The large difference in the 2X APTT values observed
for compounds 6 and 8 is difficult to explain if one
considers only the relative inhibitory potencies of these
two compounds. Our results suggest that factors other
than inherent potency play an important role in the
antithrombotic properties of thrombin inhibitors. It
appears that lipophilicity and the associated plasma
protein binding (Table 4) are crucial determinants of
potential antithrombotic activity. Despite the fact that
compounds 6 and 8 are both extremely potent inhibitors
of thrombin in vitro, compound 8 is far more lipophilic
and is also more highly protein bound in rat plasma
than 6. We have found that in general with these as
well as with other unreported series of compounds,
increased lipophilicity and increased plasma protein
a
Reagents: (a) H2 (60 PSI)/Ir black catalyst, AcOH/H2O, 48 h;
(b) CBZ-Cl, 2 N NaOH where X ) CBZ or di-(tert-butyldicarbonate,
1 N NaOH/H2O/dioxane, 0 °C where X ) BOC; (c) EDC/HOBT/N
(Et)3/L-proline methyl ester hydrochloride, DMF; (d) 1 M LiOH/
DME; (e) reversed phase (C18) prep HPLC separation of diaster-
eomers; (f) EDC/HOBT/N (Et)3/trans-(4-N-CBZ-aminocyclohexyl)-
methylamine, DMF; (g) H2 (60 PSI)/Pd (OH)2 catalyst, EtOH/H2O/
AcOH, 95:5:1.
Ta ble 1. Thrombin Inhibition Activity4 for Compounds and 1
and 5-8
compd no.
Ki (nM)
compd no.
Ki (nM)
1
5
6
0.10
440
0.056
7
8
3000
0.10
repeated on this material, with no racemization ob-
served throughout the sequence. Results from this work
were in agreement with the earlier assignments. The
inhibitors which employed a Boc group at the N-
terminus were prepared via an identical synthetic
sequence, in which the racemic Boc-protected amino
acid 2b was used as the starting material. Stereochem-
ical assignments were again based on the results of the
thrombin inhibition assay, with compound 8 assigned
the (R) stereochemistry at the P3 asymmetric center.
Resu lts a n d Discu ssion
Compounds containing D-dicyclohexylalanine in the
P3 position displayed inhibitory potency similar to that
of their D-diphenylalanine analog, compound 1 (Table
1).1 Compound 6 was evaluated for its selectivity
against various human serine proteases and was shown
to posess excellent selectivity versus all enzymes tested
(Table 2).5 log P values for these inhibitors were
increased substantially as anticipated, rising above zero
and into a region where reasonable oral absorption

Potent Noncovalent Thrombin Inhibitors
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 11 1567
Ta ble 3. Oral Bioavailability and Pharmacokinetic Parameters for Compounds 6 and 8 in Rat and Dog
compd no.
log P
dose (mg/kg)
route
AUC (µM h)
Rat
T1/2 (min)
CI (mL/min/kg)
F (%)
6a
8b
0.794
2
10
2
iv
po
iv
3.30
0.36
3.34
6.86
12
6
2
>3.088
249
17.9
10
po
41
Dog
8c
>3.088
1
5
iv
po
90.0
45.3
196
428
6.06
-
90
a
b
Dosed po as 0.1% methocel suspension, iv in saline. Dosed po as 0.1% methocel suspension, iv in 30% propylene glycol/saline. c Dosed
d
po as 1% methocel suspension, iv in 20% PEG/saline. log P values were determined by the standard partition method using octanol/pH
7.4 phosphate buffer.
Ta ble 4. Antithrombotic Efficacy of Compounds 6 and 8. Rat
diluted with Et2O, scraped, and sonicated to give 1.38 g (66%)
of D,L-3,3-dicyclohexylalanine hydrochloride as a tan solid, mp
) 261-264 °C. The amino acid (4.76 mmol) was dissolved in
40 mL of 2 N NaOH, and the solution was cooled to 0 °C. The
solution was treated dropwise with 1.06 g (6.19 mmol) of
benzyl chloroformate with the temperature maintained at <5
°C. After completion of the addition, the reaction mixture was
stirred at 0 °C for 15 min and then at room temperature for 1
h. The suspension was acidified to pH 2 with 2.75 M KHSO4
solution, and the suspension extracted with 3 × 50 mL of
EtOAc. The combined extracts were dried, decolorized with
activated carbon, and filtered through a Celite pad. Concen-
tration provided a peach-colored oil which partially crystallized
on pumping. The residue was triturated with hexanes, which
induced further crystallization. Filtration provided 1.00 g
(55%) of desired product as a white crystalline solid: mp )
Carotid Artery Fe Cl3-Induced Thrombosis Assay4
plasma protein
2X APTT (µM) binding (% free)
no. of
occlusions
compda
n
rat
human
rat
human
6
8
6
6
6
1/6
4/6
0/6
1.3
2.7
ND
1.4
6.0
0.28
6
<1
ND
2
ND
ND
argatroban
a
Compounds 6 and 8 were infused at 6 µg/kg/min iv for 120
min prior to the FeCl3 insult, and argatroban was infused at 10
µg/kg/min iv. The vehicle for administration of compound 6 and
argatroban was saline; vehicle for compound 8 was 30% propylene
glycol/saline. Vehicle control animals showed essentially complete
b
occlusion. The 2X APTT value is defined as the concentration of
inhibitor in plasma (µM) required to double the activated partial
thromboplastin time.
1
150-152 °C. 400 MHz H NMR (CDCl3) 0.91-1.38 (m, 10H),
1.48 (q, 1H), 1.54-1.89 (m, 12H), 4.69 (d, 1H), 5.06 (d, 1H),
5.20 (q, 2H), 7.41 (m, 5H); high-resolution FABMS M+ theo-
retical 388.248 78, obsd 388.247 93.
binding have detrimental effects on 2X APTT values and
antithrombotic activity. While it is clearly possible to
enhance oral bioavailability of tripeptide-type thrombin
inhibitors by increasing the lipophilicity of these com-
pounds, it appears that this approach also serves to
simultaneously compromise the potential antithrom-
botic activity of the compounds. The in vitro properties
and physical parameters of a molecule must be opti-
mized simultaneously with the in vivo pharmacokinetic
profile to obtain clinically viable thrombin inhibitors.
It is likely that a much more careful and systematic
approach will be necessary to develop clinically useful
thrombin inhibitors that possess both good oral bio-
availability and good antithrombotic efficacy, and un-
derstanding the requisite physical properties of the
compounds will be crucial in this process. We are
currently making use of these principles in an effort to
develop clinically useful noncovalent thrombin inhibi-
tors, and will detail our progress in future publications.
N-CBZ-L-an d D-3,3-dicycloh exylalan yl-L-pr olin e (3a an d
4a ). A solution of 850 mg (2.19 mmol) of 2a , 363 mg (2.19
mmol) of proline methyl ester hydrochloride, 325 mg (2.41
mmol) of EDC, and 488 mg (4.82 mmol) of triethylamine in
12 mL of anhydrous DMF was treated with 462 mg (2.41
mmol) of EDC, and the resulting solution stirred at room
temperature in an N2 atmosphere for 18 h. The reaction
mixture was diluted with 3 times its volume of 10% citric acid
solution, and the suspension was extracted with 2 × 40 mL of
EtOAc. The combined EtOAc extracts were washed with
water and brine, and were dried and concentrated to provide
the crude coupling product. The crude product was purified
via column chromatography over silica gel with 2.5% MeOH/
CHCl3 to give the pure coupling product as a white foam. The
foam was dissolved in 5 mL of 2 M LiOH/5 mL of DME, and
the solution was stirred vigorously at room temperature for
18 h. The reaction mixture was acidified with 10% citric acid
solution and extracted with EtOAc. The extract was washed
with brine, dried, and concentrated to provide the crude
N-CBZ-L- and -D-3,3-dicyclohexylalanine-L-proline as a mixture
of diastereomers. The acid diastereomers were separated via
reversed phase preparatory HPLC to provide 370 mg of the
more polar diastereomer 3a as a white foam: 400 MHz 1H
NMR (CDCl3) 0.96-1.31 (m, 10H), 1.50-1.79 (m, 12H), 2.01
(m, 2H), 2.49 (m, 1H), 3.59 (m, 2H), 4.04 (t, 1H), 4.58 (m, 1H),
4.66 (t, 1H), 5.06 (d, J ) 12Hz, 1H), 5.17 (d, J ) 12Hz, 1H),
5.31 (d, J ) 10 Hz, 1H), 7.32 (m, 5H); high-resolution FABMS
Exp er im en ta l Section
Melting points were determined in open capillary tubes in
a Thomas-Hoover apparatus and are uncorrected. 1H NMR
spectra were recorded on a Varian Unity 400 spectrometer.
Chemical shifts are reported in δ (ppm) relative to tetra-
methylsilane. All reagents used were of commercial synthetic
grade. Aldrich Sure-Seal dimethylformamide was used as
solvent in all amino acid coupling reactions. All reversed
phase preparatory HPLC purifications were performed on a
Waters Prep 4000, using a Waters C18 Prep-Pac and various
gradients.
N-CBZ-D,L-3,3-d icycloh exyla la n in e (2a ). A solution of
2.00 g (7.22 mmol) of D,L-3,3-diphenylalanine hydrochloride
in 50 mL of acetic acid/10 mL of H2O was hydrogenated at 62
psi on a Parr apparatus over 500 mg of Ir black catalyst. After
24 h, a second portion of catalyst was added and the reaction
continued for a second 24 h interval. The reaction mixture
M
+ theoretical 485.301 54, obsd 485.300 90. Also obtained was
363 mg of the less polar diastereomer 4a as a glass (69% total
yield for both diastereomers): 400 MHz 1H NMR (CDCl3)
0.98-1.24 (m, 10H), 1.39 (m, 1H), 1.51-1.78 (m, 11H), 2.04
(m, 2H), 2.39 (m, 2H), 3.61 (m, 1H), 3.99 (q, 1H), 4.64 (m, 2H),
5.04 (d, J ) 12 Hz, 1H), 5.17 (d, J ) 12 Hz, 1H), 5.29 (d, J )
10 Hz, 1H), 7.34 (m, 5H); high-resolution FABMS M+ theoreti-
cal 485.301 54, obsd 485.301 62. Each diastereomer was
completely free of the other diastereomer by analytical HPLC
and NMR.
L- a n d D-3,3-Dicycloh exyla la n yl-L-p r olin e-N-[(tr a n s-4-
a m in ocycloh exyl)m eth yl]a m id e (5 a n d 6). A solution of
361 mg (0.75 mmol) of the more polar diastereomer of N-CBZ-
3,3-dicyclohexylalanyl-L-proline, 197 mg (0.75 mmol) of trans-
was filtered through
a Celite pad, and the filtrate was
concentrated in vacuo to give a tan foam. The foam was

1568 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 11
Tucker et al.
(4-N-CBZ-aminocyclohexyl)methylamine, 112 mg (0.83 mmol)
of HOBT, and 84 mg (0.83 mmol) of triethylamine in 5 mL of
anhydrous DMF was treated with 159 mg (0.83 mmol) of EDC,
and the resulting solution was stirred at room temperature
in an N2 atmosphere for 18 h. The reaction mixture was
diluted with 3 times its volume of water, and the suspension
was stirred vigorously at room temperature for 15 min. The
suspension was filtered, and the white solid was washed with
water and dried. A 528 mg (97%) yield of the bis-CBZ-
protected coupling product was obtained, mp ) 79-82 °C. A
500 mg sample of this material was dissolved in 40 mL of 4:1
EtOH/water and was hydrogenated on a Parr apparatus at
50 psi over 150 mg of Pd (OH)2 catalyst for 18 h. The reaction
mixture was filtered through a Celite pad, and the filtrate was
concentrated to provide the crude product as a clear oil. The
oil was purified via reverse phase prep LC to provide 330 mg
(64%) of the desired product after lyophilization as an amor-
(m, 1H), 3.29 (m, 1H), 3.56 (br s, 1H), 3.88 (q, 1H), 4.59 (m,
2H), 4.99 (d, 1H), 7.09 (br s, 1H); FAB MS M+ ) 561. Anal.
(C32H56N4O4‚0.70EtOAc‚0.70 H2O) C, H, N.
In the same manner from 370 mg (0.82 mmol) of compound
4b was obtained 155 mg (34%) of desired product 8 as a clear
1
glass: 400 MHz H NMR (CDCl3) 0.94-1.35 (complex, 15H),
1.41 (s, 9H), 1.52-189 (complex, 20H), 1.86 (d, 1H), 1.99 (br s,
1H), 2.23 (m, 1H), 2.63 (br t, 1H), 3.02 (m, 2H), 3.54 (q, 1H),
3.95 (br t, 1H), 4.48 (t, 1H), 4.59 (d, 1H), 4.96 (m, 1H), 7.11
(br s, 1H); HRFAB MS M+ theoretical 561.437 98, obsd
561.437 06. Anal. (C32H56N4O4‚0.70EtOAc‚0.15 H2O) C, H, N.
This material was the more active diastereomer in a thrombin
inhibition assay and was therefore assigned the (R) configu-
ration at the dicyclohexylalanine R-position.
Con sciou s Ra t Bioa va ila bility. Male Sprague-Dawley
rats (Taconic Farms, 375-450 g) are anesthetized with Brevi-
tal (65 mg/kg). Under aseptic conditions, the carotid artery
and in some animals (those to be dosed iv) the jugular vein
are cannulated with PE 50 tubing (approximately 75 cm in
length).
The tubing is advanced into the carotid artery for a distance
of 3 cm and the jugular vein for 2 cm, and anchored securely.
The tubing is exteriorized out the back of the neck and passed
through a stainless steel spring having a button on one end.
The neck incision is closed with autoclips, and the spring’s
button is sutured to the skin, thus protecting the tubing. The
carotid catheter is loaded with approximately 0.3 mL of
heparinized saline (5 units/mL) and sealed. Rats are kept in
cages overnight where they can roam freely with access to food
and water.
The following day a control blood sample (1 mL) is with-
drawn from each rat. Collection of all blood samples are done
as follows:
1. Approximately 0.3 mL is withdrawn from the carotid
artery catheter to remove any traces of heparin.
2. A blood sample is withdrawn into a syringe containing
1 part 3.8% sodium citrate for each 9 parts of blood.
3. Once the sample is withdrawn the catheter is flushed
with 0.3 mL of heparinized saline.
4. The blood sample is placed in a 1.5 mL microtube, mixed
gently by inverting, and then centrifuged at maximum speed
for 2.5 min. The plasma is removed to another microtube and
placed on dry ice. (Any samples not centrifuged or plasma
removed are kept on ice.)
1
phous glass: 400 MHz H NMR (CDCl3) 0.98-1.57 (complex,
15H), 1.59-2.00 (complex, 16H), 2.00-2.19 (m, 4H), 2.40 (m,
1H), 3.02 (m, 2H), 3.14 (m, 1H), 3.55 (m, 1H), 3.64 (m, 1H),
4.36 (m, 1H), 4.44 (m, 1H), 7.37 (br s, 1H), 8.19 (br s, 4H). HR
FABMS M+ theoretical 461.385 55, obsd 461.386 46. Anal.
(C27H48N4O2‚2.40TFA‚0.45 H2O) C, H, N.
An identical procedure performed on 348 mg (0.72 mmol)
of the less polar diastereomer of N-CBZ-dicyclohexylalanine-
L-proline provided 275 mg (70%) of product as an amorphous
tacky solid. 400 MHz 1H NMR (CDCl3) 0.90-1.05 (m, 4H),
1.05-1.43 (complex, 11H), 1.45-1.60 (m, 4H), 1.60-1.89
(complex, 12H), 1.95-2.18 (m, 4H), 2.12 (m, 1H), 2.86 (m, 1H),
2.94 (br s, 1H), 3.16 (m, 1H), 3.46 (m, 1H), 3.87 (m, 1H), 4.40
(m, 2H), 7.61 (t, 1H), 8.11 (br s, 2H), 8.41 (br s, 2H); HR
FABMS M+ theoretical 461.386 46, obsd 461.386 64. Anal.
(C27H48N4O2‚2.50TFA‚0.50H2O) C, H, N. This material was
the more active diastereomer in a thrombin inhibition assay
and was therefore assigned the (R)- configuration at the
dicyclohexylalanine R-position.
N-boc-D,L-3,3-d icycloh exyla la n in e (2b). To a solution of
2.00 g (6.90 mmol) of D,L-3,3-dicyclohexylalanine in in 17 mL
1 N NaOH/9 mL of water/17 mL of 1,4 dioxane cooled to 0 °C
was added 1.66 g (7.59 mmol) of di-tert-butyl dicarbonate in
portions over approximately 2 min. The solution was stirred
in the cold for 5 min and then at room temperature for an
additional 2 h. The reaction mixture was concentrated to
remove most of the dioxane and was cooled in an ice bath. The
cold reaction was acidified to pH 2 with 1 N KHSO4 and was
extracted 2 times with EtOAc. The combined extracts were
washed with water and brine, dried, and concentrated to
provide 2.43 g (90%) of the desired product as a white
crystalline solid: mp ) 188.0-189.5 °C; 400 MHz 1H NMR
(CDCl3) 0.96-1.35 (m, 10 H), 1.46 (s, 9H), 1.51-1.83 (m, 13H),
4.56 (d, J ) 8 Hz, 1H), 4.85 (d, J ) 8 Hz, 1H); high-resolution
FABMS M+ theoretical 354.264 43, obsd 354.265 88.
Rats are now dosed with drug: 2 mg/kg intravenously or
10 mg/kg orally. Blood (0.5 mL) is collected at 1, 5, 15, 30, 60,
90, 120, 150, 180, 240, and 360 min posttreatment for rats
dosed iv and (1 mL) at 15, 30, 60, 90, 120, 180, 240, and 360
min posttreatment for those dosed orally.
After the last sample is collected from the iv rats, anesthesia
is administered through the iv catheter to prove that the iv
dose was in fact received by the rat.
N-Boc-L- a n d -D-3,3-d icycloh exyla la n yl-L-p r olin e (3b
a n d 4b). In a manner identical to the N-CBZ examples 3a
and 4a , from 2.10 g (5.94 mmol) of N-BOC-D, L-3,3-dicyclo-
hexylalanine was obtained 0.82 g of the more polar acid
diastereomer 3b as a clear oil/foam: 400 MHz 1H NMR (CDCl3)
1.04-1.32 (m, 10H), 1.43 (s, 9H), 1.51-1.76 (m, 11H), 2.10 (m,
3H), 2.34 (m, 1H), 3.62 (m, 1H), 4.01 (q, 1H), 4.62 (m, 2H),
5.17 (d, J ) 9 Hz, 1H); HR FABMS M+ theoretical 451.317 20,
obsd 451.317 64. A 1.20 g quantity of the less polar diaste-
reomer 4b was also obtained HR FABMS M+ theoretical
451.317 20, obsd 451.315 87; 400 MHz 1H NMR (CDCl3) 1.02-
1.31 (m, 10H), 1.42 (s, 9H), 1.51-1.80 (m, 11H), 1.90-2.25 (m,
3H), 2.55 (m, 1H), 3.59 (q, 1H), 4.02 (t, 1H), 4.59 (m, 2H), 4.99
(d, J ) 10 Hz, 1H). An 86% yield was obtained based on the
recovery of both diastereomers. Each diastereomer was free
of any of the other diastereomer by analytical HPLC and NMR.
tr a n s-(4-Am in ocycloh exyl)m eth yl]-N-boc-L- a n d -D-3,3-
d icycloh exyla la n yl-L-p r olin a m id e (7 a n d 8). In a manner
identical to the above detailed preparation of compound 5, from
264 mg (0.59 mmol) of compound 3b (after coupling, CBZ
removal, prep LC, and neutralization) was obtained 135 mg
The collected plasma samples are analyzed using a thrombin
enzyme activity assay. Plasma samples are extracted with
acetonitrile prior to assay.
Con sciou s Dog Bioa va ila bility. Two male beagle dogs
weighing 10-12 kg were used for the absorption and kinetic
studies. After an overnight fast, the dogs received oral doses
of inhibitor at 5 mg/kg as a 1% methocal suspension in a
crossover fashion. Blood samples were collected via the
jugular vein at 0, 10, 20, 30, 40, 60, 90, 120, 180, 240, 300,
360, and 480 min after dosing. Plasma samples were kept
frozen (-20 °C) until assayed by HPLC.
The same dogs used in the above study received iv doses of
inhibitor at 1 mg/kg bolus in a 20% PEG/saline vehicle. Blood
was collected at 0, 5, 15, 30, 45, 60, 90, 120, 180, 240, 300,
360, and 480 min after dosing. Plasma samples were kept
frozen (-20 °C) until assayed by HPLC.
Deter m in a tion of P er cen t F r ee. The fraction of free
inhibitor in plasma is determined as follows: To 495 µL of
platelet poor plasma (PPP) or HBSP buffer (50 mM HEPES,
150 mM NaCl, 0.1% PEG8000, pH 7.5) is added 5 µL of a
concentrated inhibitor solution (water or DMSO) for a final
concentration of 1 µM. After mixing, the solutions are passed
through a Microcon 10 (Amicon No. 42406) by centrifugation
(5000 rpm for 10 min at room temperature) and ∼50 µL is
1
(41%) of the desired product 7 as a clear glass: 400 MHz H
NMR (CDCl3) 0.84-1.39 (m, 16H), 1.46 (s, 9H), 1.50-1.99
(complex, 18H), 2.11 (m, 1H), 2.22 (m, 1H), 2.75 (m, 2H), 2.81

Potent Noncovalent Thrombin Inhibitors
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 11 1569
(2) There is a previous patent that generically describes the use of
this amino acid as a P3 ligand in thrombin inhibitors: Teger-
Nilsson, A. C. E.; Bylund, R. E. WO9311152-A1, 1995.
(3) Boc and Fmoc-D-diphenylalanine are commercially available
from Synthetech, Inc., Albany, OR. Prior to this material
becoming commercially available, racemic diphenylalanine was
prepared according to Cheng et al.; Synthesis and Biological
Activity of Ketomethylene Pseudopeotide Analogs as Thrombin
Inhibitors. J . Med. Chem. 1992, 35, 3364-3370 and references
cited therein.
(4) Full experimental details for the preparation of CBZ-protected
trans-4- (aminomethyl)cyclohexylamine are provided in the
patent claiming this series of compounds: Brady, S. F.; Feng,
D. M.; Freidinger, R. M.; Lumma, W. C.; Lyle, T. A.; Sanderson,
P. E. J .; Stauffer, K. J .; Tucker, T. J .; Vacca, J . P. US 5, 510,
369, 1996.
(5) Details of these assays are provided in several previous publica-
tions from our group: (a) Lewis, S. D.; Ng, A. S.; Baldwin, J . J .;
Fusetani, N.; Naylor, A. M.; Shafer, J . S. Inhibition of Thrombin
and Other Trypsin-like Serine Proteinases by Cyclotheonamide
A. Thromb. Res. 1993, 70, 173-190. (b) Lewis, S. D.; Ng, A. S.;
Lyle, E. A.; Mellott, M. J .; Appleby, S. D.; Brady, S. F.; Stauffer,
K. J .; Sisko, J . T.; Mao, S.-S.; Veber, D. F.; Nutt, R. F.; Lynch,
J . J .; Cook, J . J .; Gardell, S. J .; Shafer, J . A. Inhibition of
Thrombin by Peptides Containing Lysyl-a Keto Carbonyl De-
rivatives. Thromb. Haemostasis 1995, 74, 1107-1112.
(6) Details of the experimental determination of this parameter are
provided in the publications cited in ref 5.
collected from each. The eluent is mixed with a known
concentration of thrombin ([thrombin] and [inhibitor] . Ki),
and after a brief incubation chromogenic substrate is added.
The inhibition generated from the eluent of the PPP sample
is compared to a standard curve prepared from the inhibitor
that has not been filtered. The nonspecific binding (generally
<10%) of the inhibitor to the membrane is corrected for by
using the eluent from the HSBP control. Eluent from filtered
PPP with no added inhibitor was shown not to interfere with
the thrombin chromogenic assay.
Ack n ow led gm en t. The authors thank Mr. Art
Coddington and Dr. Harri Ramjet for Mass Spectra, Dr.
Graham Smith, Mr. Matt Zrada, Mr. Ken Anderson, and
Ms. Patrice Ciecko for log P determinations and elemen-
tal analyses, and Ms. J ean Kaysen for manuscript
preparation. The authors also thank Prof. Barry Trost
for helpful discussions.
Refer en ces
(1) Tucker, T. J .; Lumma, W. C.; Naylor-Olsen, A. M.; Lewis, S. D.;
Lucas, R.; Freidinger, R. M.; Mulichak, A. M.; Chen, Z.; 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.
J M970140S