The design and synthesis of thrombin inhibitors: The introduction of in vivo efficacy and oral bioavailability into benzthiazolylalanine inhibitors

Article abstract of DOI:10.1016/S0960-894X(00)00283-3

The further optimisation of the novel lead compound CGH752 (Fig. 1) is described. By introducing various substituents into the 6-position of the 3,3-dimethyltetrahydroquinoline (DMTHQS) ring we have been able to favourably affect the in vitro and in vivo activity, and the pharmacokinetics of such compounds. One of the inhibitors synthesised (CGH1484) is bioavailable and shows efficacy in animal models of thrombosis. (C) 2000 Elsevier Science Ltd. All rights reserved.

Full text of DOI:10.1016/S0960-894X(00)00283-3

Bioorganic & Medicinal Chemistry Letters 10 (2000) 1567±1570
The Design and Synthesis of Thrombin Inhibitors:
The Introduction of In Vivo Ecacy and Oral Bioavailability
into Benzthiazolylalanine Inhibitors
Judy Hayler, Peter D. Kane,* Darren LeGrand, Florence Lugrin, Keith Menear,
Richard Price, Mark Allen, Xiaoling Cockcroft, John Ambler, Keith Butler,
Karren Dunnet, Andrew Mitchelson, Mark Talbot, Morris Tweed and Nicholas Wills
Novartis Horsham Research Centre, Wimblehurst Road, Horsharn, West Sussex RH12 4AB, UK
Received 21 October 1999; accepted 9 March 2000
AbstractÐThe further optimisation of the novel lead compound CGH752 (Fig. 1) is described. By introducing various substituents
into the 6-position of the 3,3-dimethyltetrahydroquinoline (DMTHQS) ring we have been able to favourably a€ect the in vitro and
in vivo activity, and the pharmacokinetics of such compounds. One of the inhibitors synthesised (CGH1484) is bioavailable and
shows ecacy in animal models of thrombosis. # 2000 Elsevier Science Ltd. All rights reserved.
The development of orally active compounds to speci®cally
inhibit thrombin is a signi®cant therapeutic goal.1 In the
previous letter we have detailed our ®ndings with amino
acid replacements for arginine at the P1 position of MD805
derived inhibitors, that culminated in the discovery of
CGH752 (Fig. 1) as a novel, potent and selective inhibitor
of thrombin.
mM) (Fig. 1), and can probably be contributed to the
increased protein binding of the more lipophilic com-
pound. However, in this new series the possibility of re-
introducing some polar groups to tune the physichochem-
ical characteristics, and hence the pharmacokinetics and
in vivo activity, was now a possibility. This is somewhat
di€erent to the strategies pursued by some other
research groups, where the P1 charge interaction has
been preserved by introducing a basic amidine group or
by modulating the basicity in order to confer favourable
properties.2
Although having potentially eliminated some of the
poor characteristics of a guanidine based inhibitor i.e.,
removal of the cationic group that prevents satisfactory
absorption, there remained a number of properties to
improve before realising our goal of an orally e€ective
drug candidate.
Con®dent that the presence of Bta in our inhibitors would
ensure selectivity for thrombin, we sought to ®nd a site on
The removal of a charged guanidine species and the intro-
duction of a lipophilic Bta group (CGH752 ClogP=5.38)
has serious consequences on the aqueous solubility of
these compounds, hence CGH752 is largely insoluble and
is not orally bioavailable. The lipophilic pro®le of
CGH752 also alters the e€ectiveness of the compound to
inhibit coagulation in plasma, as measured by the more
physiological relevant activated partial thromboplastin
time (APTT). This is apparent from a comparison of the
APTTs of an arginine compound CGH728 (Ki=6 nM,
APTT=2.1 mM) with CGH752 (Ki=26 nM, APTT=26.5
*Corresponding author and current address. Tripos Receptor Research,
Bude, Cornwall EX23 8LY. Tel.: +44-288-359-359; fax: +44-288-359-
222; e-mail: pkane@tripos.com
Figure 1.
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00283-3

1568
J. Hayler et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1567±1570
Scheme 1. Reaction conditions: (i) (a) dimethylmalonic acid, SOCl2, , THF, re¯ux; (b) 4-sub-aniline (2 equiv), 20 ꢀC; (ii) P2O5, CH3SO3H, 70 ꢀC; (iii)
Pd/C, AcOH, H2; (iv) (a) P2S5 or Lawesson's reagent, dioxan; (v) MeI, KOtBu, THF; (vi) Na CNBH3; (vii) Pyr.SO3 Pyr, re¯ux; (viii) Pyr, POCl3,
CH3CN, ultrasound 20 ꢀC; (ix) H-Bta-FEP, DIPEA, DCM; (x) AlCl3, LAH, THF/Et2O, 20 ꢀC to re¯ux; (xi) t-BuPh2SiCl, imidazole; (xii) TBAF,
THF; (xiii) 1M NaOH, H2O/MeOH.
Scheme 2. Reaction conditions: (i) ClCO2Et, DMF, TEA; (ii) NaBH4, EtOH; (iii) meNCO, DCM, AcOH, 50 ꢀC; (iv) DIPEA, DCM, RCl; (v)
COCl2, DMSO, TEA; (vi) Ph3P=CHCO2Et, DCM, rt; (vii) NaBH4, CuCl, MeOH; (viii) NaOH, H2O, meOH, then d.HCl; (ix) PyBOP, DIPEA,
DMF, rt; (x) TBTU, DIPEA, DCM, piperazine; (xi) (PhO)2P(O)N2, TEA, Toluene, 100 ꢀC; (xii) NH3, DCM.

J. Hayler et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1567±1570
1569
the CGH752 structure where property-modifying groups
could be introduced without impairing the good binding
anity. Modelling ®rst suggested the P2 piperidine
moiety had potential for such a campaign; there appears to
be considerable space within the enzyme to accommodate
structural changes to the substituent group on piperidine.
However, from previous studies in the arginine series we
knew that it was not easy to make changes at this position
and maintain good Ki and APTT values.3 This was
con®rmed by the synthesis of a number of P2 variants in
the Bta series, ®ndings that largely paralleled those of
the equivalent arginine compounds. The other possible
site was the aromatic ring of the dimethyltetrahydro-
quinoline (DMTHQS) P3, where a small amount of direct
evidence already existed that this could be substituted in
the 6-position without detriment to potency.3 The selection
of the DMTHQS P3 6-position was also born out by a
systematic modelling study, and the synthesis and eva-
luation of compounds with various substitutions at this
position was therefore undertaken (Scheme 1).
rated while retaining reasonable activity, a ®nding that
was in contrast to our experience in the P2 area. The
various compounds shown in Table 1 all have sub-
micromolar activity and the majority have improved
APTTs when compared to the original lead. As was pre-
viously stated there appears to be a relationship between
logP (and logD) and APTT at a gross level,6 but the
e€ectiveness of an inhibitor in plasma must be governed
by a multiplicity of factors making rationalisation and
prediction dicult at the present time. Using 12 (bioa-
vailability=50% id) as a template a variety of amides
were prepared, with piperazine amides in particular dis-
playing good properties. The compound 23 (CGH1484)
(Table 2), which contains the solubilising elements of a
basic nitrogen (piperazine ring) and morpholine, had
good Ki and APTT; moreover this compound showed
reasonable bioavailablity (28%) when dosed to rats by per
oral administration. The e€ects of CGH1484A (HCl salt)
on arterial thrombosis in the rat was determined using an
acute model of injury-induced thrombus formation in the
rate dorsal aorta.7 CGH 1484A of a 10 or 20 mg/kg (po)
Chemistry
Table 1.
Key 6-substituted DMTHQS intermediates were prepared
by a route adapted from that originally used for the
unsubstituted version.3 The respective para-substituted
aniline was coupled to in situ prepared dimethylmalonyl
chloride and the resultant amide carboxylic acid eciently
cyclised to the ketoamides 2a±c in Eaton's reagent.4 These
were then converted in a stepwise fashion, ®rst by reduc-
tion of the benzylic ketone (and double bond for 3c) by
catalytic hydrogenation, and then by reduction of the
amide via the intermediary S-methyl thioimidate com-
pounds 4a±c, thus preserving the ester functionality in the
6-positive. Alternatively a total reduction was carried out
with LAH in the presence of aluminium chloride leading
to the 6-hydroxymethyl and ethyl compounds 6a and
6b respectively. Regioselective sulphonation, chlorination
and coupling with the free base of Bta-FEP completed the
synthesis. The preparation of Bta-FEP is detailed in the
preceding letter. The fully coupled compounds could be
conveniently transformed to more complex structures.
Thus the carboxylic acid 10 was reduced via the mixed
anhydride (Scheme 2) and the alcohol (14) reacted with
methyl isocyanate to form 15 or alkylated to 16. Swern
oxidation of 8 followed by Wittig reaction led to alkene
17, which could then be reduced with copper (I) chloride/
sodium borohydride,5 with no interference from Bta,
and hydrolysed to the acid 18. The carboxylic acid 12
was coupled to a variety of amines typically using
PyBOP or TBTU (Scheme 2). A Curtius rearrangement
was accomplished with diphenyl-phosphoryl azide and
the isocyanate trapped with ammonia to give the urea 24,
a number of substituted ureas and urethanes could be
prepared in this way.
-R
Ki (nM)
APTT (mM)
ClogP
752
14
26
64
26.5
16.8
5.38
4.57
8
39
13.7
19
4.58
5.1
19
119
16
125
18
9
723
359
105.4
13
6.45
5.46
10
11
12
18
15
24
20
500
82
94
77
47
34
14.9
10.7
16.6
20.7
10.2
9.2
4.25
5.1
5.63
4.85
4.03
4.54
Results and Discussion
As new compounds bearing substituents at the 6-position
were synthesised and assayed, it became apparent that a
range of chain length and functionality could be incorpo-

1570
J. Hayler et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1567±1570
Table 2.
Conclusion
-R
Ki
(nM)
APTT
(mM)
Bio
Avl.
Aqu sol.
(mg/ml)
The presence of Bta in inhibitor compounds ensures good
potency and selectivity for thrombin over other serine
proteases. The introduction of a variety of groups into the
6-position of the DMTHQS aromatic ring is possible
without adversely a€ecting the good binding of Bta-FEP
based inhibitors and we believe these substituents extend
onto the surface of the enzyme. These ®ndings have
allowed us to introduce groups that markedly in¯uence
the water solubility and APTT of inhibitors, and have led
to a compound (CGH 1484) with oral ecacy in animal
models of thrombosis.
12
21
82
34
10.7
7.5
50% (id)
0% (po)
0.00515
@pH7
N.D.b
N.D.
N.D.
N.D.
22
27
57
9.7
References and Notes
1. (a) Brundish, D. E. Current Opinion in Therapeutic Patents
(Current Drugs) 1992, pp 1457±1466; (b) Ripka, W. C. Curr.
Opin. Chem. Biol. 1997, 1, 242; Wiley, M. R.; Fisher, M. J.
Expert Opin. Ther. Pat. 1997, 7, 1265. (d) Hauptmann, J.;
Sturzebecher, J.; Thromb. Res. 1999, 93, 203.
23a
7
38% (id)
28% (po)
0.13 @pH7
15 @ pH2
2. (a) Misra, R. N.; Kelly, Y. F.; Brown, R. B.; Roberts, D. G.
M.; Chong, S.; Seiler, S. M. Bioorg. Med. Chem. Lett. 1994, 4,
2165. (b) von der Saal, W.; Kucznierz, R.; Leinert, H.; Engh,
R. A.; Bioorg. Med. Chem. Lett. 1997, 7, 1283. (c) Sanderson,
P. E. J.; Naylor-Olsen, A. Curr. Med. Chem. 1998, 5, 289.
3. Brundish, D.; Bull, A.; Donovan, D.; Fullerton, J. D.;
Garman, S.; Hayler, J.; Janus, D.; Kane, P. D.; McDonnell,
M.; Smith, G. P.; Wakeford, R.; Walker, C. V.; Howarth, G.;
Hoyle, W.; Allen, M. C.; Ambler, J.; Butler, K.; Talbot, D. J.
Med. Chem. 1999, 42, 4584.
4. McGarry, L. W.; Detty, M. R. J. Org. Chem. 1990, 55, 4349.
5. Narisada, M.; Horibe, I.; Watanabe, F.; Takeda, K. J. Org.
Chem. 1989, 54, 5308.
6. Unpublished results.
7. Butler, K. D.; Ambler, J.; Dolan, S.; Giddings, J.; Talbot,
M. D.; Wallis, R. B. Blood Coag. Fibrinol. 1992, 3, 155.
aKi trypsin=153 mM, chymotrypsin=12.2 mM, plasmin=212 mM,
Xa=>200 mM, TF/VIIa=11600 mM.
bN.D.=not determined.
dose, administered 1 h prior to injury to the dorsal aorta,
signi®cantly inhibited platelet deposition by approximately
60%. The e€ect was maintained with the higher dose for
up to 3 h. The e€ect on ®brin(ogen) deposition was less
marked with a 48% and 43% inhibition being noted at
1 and 2 h following the oral administration of 20 mg/kg
CGH1484A.