Pyrrolidine-5,5-trans-lactams. 5. Pharmacokinetic Optimization of Inhibitors of Hepatitis C Virus NS3/4A Protease

Article abstract of DOI:10.1021/ol035827n

(Matrix presented) In this, the second of two Letters, the optimization of the pyrrolidine-5,5-trans-lactam template (exemplified by 1a) as a mechanism-based inhibitor of hepatitis C NS3/4A protease is described. Right Boxn analysis of cassette dosing screening pharmacokinetic data was used to rapidly categorize the compounds. GW0014 (compound 4d) emerged as the compound displaying an optimal balance of biochemical and replicon potency, along with low i.v. clearance in the dog.

Full text of DOI:10.1021/ol035827n

ORGANIC
LETTERS
2003
Vol. 5, No. 24
4631-4634
Pyrrolidine-5,5-trans-lactams. 5.
Pharmacokinetic Optimization of
Inhibitors of Hepatitis C Virus NS3/4A
Protease
,†
David M. Andrews,* Michael C. Barnes, Mike D. Dowle, S. Lucy Hind,
Martin R. Johnson, Paul S. Jones, Gail Mills, Angela Patikis, Tony J. Pateman,
Tracy J. Redfern, J. Ed Robinson, Martin J. Slater, and Naimisha Trivedi
GlaxoSmithKline Medicines Research Centre, Gunnels Wood Road,
SteVenage, SG1 2NY, U.K.
daVid.andrews@astrazeneca.com
Received September 22, 2003
ABSTRACT
In this, the second of two Letters, the optimization of the pyrrolidine-5,5-trans-lactam template (exemplified by 1a) as a mechanism-based
inhibitor of hepatitis C NS3/4A protease is described. “Right Box” analysis of cassette dosing screening pharmacokinetic data was used to
rapidly categorize the compounds. GW0014 (compound 4d) emerged as the compound displaying an optimal balance of biochemical and
replicon potency, along with low i.v. clearance in the dog.
Hepatitis C virus (HCV) infects chronically an estimated 3%
of the global human population,1 and the development of
new therapies to treat HCV infection effectively is thus of
considerable importance.2 The viral genome encodes for a
single polyprotein of 3010-3030 amino acids from which
mature nonstructural proteins are released by the action of
the viral proteases NS2 and NS3. It has been demonstrated
that NS3 protease is an essential viral function and should
prove to be an excellent target for the development of novel
anti-HCV agents.3
of two letters, we describe how an optimal combination of
pyrrolidine and lactam ring substituents was achieved and
how the critical parameter of in vivo pharmacokinetics,
namely, clearance, was rapidly evaluated using cassette
dosing in dog.
Although compounds 1a-e (Table 1) are among the most
potent HCV protease inhibitors hitherto described, they
contain structural features that are regarded as potential
liabilities. From a reactivity viewpoint, although the cyclo-
propylcarbonyl lactams 1a-c offer excellent potency, the
hydrolytic stability of this family is inferior to that of their
isopropylcarbonyl counterparts 1d and 1e. Additionally, some
Earlier communications have described the synthesis of
disubstituted4 core trans-lactam templates. In this, the second
† Current address: AstraZeneca, Alderley Park, Macclesfield, Cheshire,
SK10 4TG, U.K.
(1) World Health Organisation Weekly Epidemiological Record 1997,
72, 65.
(2) Tan, S.-L, Pause, A.; Shi, Y.; Sonenberg, N. Nat. ReV. Drug DiscoVery
2002, 11, 867.
(3) Kolykhalov, A. A.; Mihalik, K.; Feinstone, S. M.; Rice, C. M. J.
Virol. 2000, 74, 2046.
(4) (a) Andrews, D. M.; Carey, S. J.; Chaignot, H.; Coomber, B. A.;
Gray, N. M.; Hind, S. L.; Jones, P. S.; Mills, G.; Robinson, J. E.; Slater.
M. J. Org. Lett. 2002, 4, 4475. (b) Andrews, D. M.; Chaignot, H.; Coomber,
B. A.; Good, A. C.; Hind, S. L.; Johnson, M. R.; Jones, P. S.; Mills, G.;
Robinson, J. E.; Skarzynski, T.; Slater, M. J.; Somers, D. O’N. Org. Lett.
2002, 4, 4479. (c) Andrews, D. M.; Jones, P. S.; Mills, G.; Hind, S. L.;
Slater, M. J.; Trivedi, N.; Wareing, K. J. Bio. Med. Chem. Lett. 2003, 13,
1657.
10.1021/ol035827n CCC: $25.00 © 2003 American Chemical Society
Published on Web 11/06/2003

Capitalizing on the earlier observation that a urea linker
appended to the amino acid moiety (Figure 1) bestows
improved potency upon the series,5 an array of 15 compounds
was synthesized (Scheme 1) for comparison with the
reference compounds 1a-e.
Table 1. trans-Lactam Array; Compounds Synthesized and
obs/I Determinations (in italics, M-1 s-1
K
)
Scheme 1. Synthesis of 2a-4ea
a Reagents and conditions: (a) p-nitrophenyl chloroformate,
DCM, Et3N, L-valine dimethylamide; (b) p-nitrophenyl chloro-
formate, DCM, Et3N, L-valine tert-butyl ester; (c) trifluoroacetic
acid, DCM; (d) HATU, DIPEA, MeCN or DMF, 2-(1-piperazinyl)-
pyrimidine.
of the previously described inhibitors, e.g., 3a, have molec-
ular weight over 600 Da. These concerns identified three
main goals. First, improved biochemical potency was needed
such that the compounds should reliably demonstrate activity
in a replicon cell-based assay. Second, the molecular weight
of the compounds should be reduced, preferably to below
500 (a threshold more consistent with membrane permeability
and hence improved oral bioavailability). Third, clearance
needed to be reduced, since initial studies had demonstrated
that clearance (Clp) in the dog approximated to liver blood
flow (38 mL/min/kg, too high to be useful for a systemically
administered antiviral drug). Ultimately however, optimiza-
tion of clearance (Clp), plasma half-life (t1/2), and oral
bioavailability (%F) is necessary to ensure that the required
trough concentrations during dosing are obtained.
The previously reported amines 5a-e4 were reacted with
p-nitrophenyl chloroformate to form activated carbamate
species that were either quenched with L-valine dimethyl-
amide (a) or L-valine tert-butyl ester (b) to give the valyl
ureas 2a-e and 6a-e. Further deprotection and coupling of
6a-e generated the piperidino analogues 3a-e. The ureas
4a-e were prepared straightforwardly in a single step by
reaction of cyclopentyl isocyanate with 5a-e. Biochemical
potency data, as well as the structure of all the compounds
evaluated, is summarized in Table 1.
Compounds were scheduled for parallel evaluation in
multiple assays to facilitate rapid identification of compounds
exemplifying both good potency and pharmacokinetics.
Screening relatively large numbers of compounds in the
biochemical proteinase and cellular replicon assays was
anticipated to be straightforward, but evaluating clearance
using traditional pharmacokinetics assays would clearly not
offer the throughput required. We therefore chose to extend
previously developed in-house dog cassette methodology6
to assist with the rapid evaluation of analogues with superior
pharmacokinetics to 1a.
(5) Slater, M. J.; Amphlett, E. M.; Andrews, D. M.; Bravi, G.; Carey, S.
J.; Johnson, M. R.; Jones, P. S.; Mills, G.; Parry, N. R.; Somers, D. O’N.;
Stewart A. J. Org. Lett. 2003, 5, 4627.
(6) (a) Macdonald, S. J. F.; Dowle, M. D.; Harrison, L. A.; Clarke, G.
D. E.; Inglis, G. G. A.; Johnson, M. R.; Shah, P.; Smith, R. A.; Amour, A.;
Fleetwood, G.; Humphreys, D. C.; Molloy, C. R.; Dixon, M.; Godward, R.
E.; Wonacott, A. J.; Singh, O. M. P.; Hodgson, S. T.; Hardy, G. W. J.
Med. Chem. 2002, 45, 3878. (b) Frick, L. W.; Adkison, K. K.; Wells-Knecht,
K. J.; Woollard, P.; Higton, D. M. Pharm. Sci. Technol. Today 1998, 1,
12.
Figure 1. Urethane- and urea-substituted pyrrolidine-5,5-trans-
lactams 1a and 4d; cf. cassette standard GW311616.
4632
Org. Lett., Vol. 5, No. 24, 2003

Table 2. In Vivo Pharmacokinetic and in Vitro Potency Data for Pyrrolidine-5,5-trans-lactams
molecular
weight
compound Clp
(mL/min/kg)
standard Clp
(mL/min/kg)
Clp as ratio
t1/2
Vd
K
obs/I
IC50
compound
to Std (Std ) 1)
(h)
(L/kg)
(M-1 s-1
)
(µM)
1a
1b
1c
1d
1e
2a
2b
2c
2d
2e
3a
3b
3c
3d
3e
4a
4b
4c
4d
4e
433
447
483
435
487
503
517
553
505
555
622
636
672
624
674
444
458
494
446
496
65
18
nt
noa
no
0.3
1.4
0.3
2.2
400
36
621
166
362
4.0
ntb
86
3.1
17
<0.78c
ntd
ntd
ntd
ntd
0.45
2.2e
nt
36
63
54
54
86
44
45
45
25
77
48
60
20
39
85
33
72
no
no
12
11
28
34
34
34
28
28
11
12
43
43
43
43
43
0.2
0.4
0.1
0.1
0.5
0.2
0.8
0.6
0.3
0.2
0.1
0.5
0.7
0.9
0.5
0.2
1.3
0.6
0.5
0.8
2.0
0.9
3.0
1.2
2.1
0.6
0.6
0.9
2.4
6.6
1.4
1.0
4.5
4.9
3.1
1.3
1.3
1.3
0.9
2.8
4.4
5.0
0.5
0.9
2.0
0.77
1.7
2935
676
6112
1668
4865
7761
989
8361
3099
5704
912
1.8e
ntd
2.6
1.6e
80
122
1785
270
1.0
6.6
609
a No reference run (compound run as single discrete entity). b Not tested; biochemical activity too low to translate to cellular activity. c First dose response
performed at too high a concentration. Compound shows very high clearance; therefore IC50 determination was not repeated. d Not tested; clearance higher
than standard. e In-assay toxicity observed at concentrations <10-fold greater than the quoted IC50.
Prior to inclusion of dog cassette dosing in the screening
sequence, the technique was validated to ensure that within
this series no drug-drug interactions occurred. Five spiro-
cyclobutyl-substituted trans-lactams (previously examined
in the dog as singletons) were coadministered with the well-
characterized elastase inhibitor GW311616 (Figure 1).7 The
data confirmed that the rank order of clearance did not alter
and that, in general, the absolute values for clearance did
not vary much between the discrete and cassette doses. With
one exception, the values were within 25%. In light of the
generally self-consistent data, this variability was ascribed
to interanimal variation rather than a compound or technique
issue (see Supporting Information).
The tert-butyloxy reference compounds 1a-e had been
subject to prior examination by discrete dosing, but because
none combined biochemical potency with low clearance, they
were not subject to reevaluation in cassette. The standard
compound, GW311616 has a mean clearance of 25 mL/min/
kg, around two-thirds liver blood flow (about 38 mL/min/
kg), but as seen in Table 2, the actual value can vary between
11 and 43 mL/min/kg. This indicated a higher degree of
interanimal variability than we had observed in our initial
validation experiments. In choosing which compounds to
progress for further study, we therefore paid less attention
to the absolute clearance values, preferring to interpret
clearance data as a ratio to the standard (i.e., a “Right Box”
approach). Compounds that were cleared more rapidly than
the standard were not progressed further down the testing
cascade, and in some cases, if replicon data had not already
been obtained, cellular potency data was not collected.
Independently, White and Manitpisitkul have shown by
reanalysis of literature reports that Right Box analysis is one
of the most useful screening cassette pharmacokinetics
approaches.8
It can be appreciated readily that the valine dimethylamide
substituted compounds 2a-e, although showing good bio-
chemical potency, were all cleared rapidly and therefore not
progressed to the replicon assay in most cases. It is likely
that the high clearance could arise (in part) as a result of
rapid amide cleavage, demethylation, or N-oxidation.
The (piperazinyl)pyrimidine-substituted valineamides 3a-e
demonstrated marginally improved potency compared to that
of 2a-e, and one example (3a) combined moderate clearance
with sub-micromolar replicon potency, but the majority of
compounds (3) showed high clearance. In view of this, and
since 3a-e all had molecular weights above 600, the latter
compounds were not progressed further.
The cyclopentyl urea series 4a-e was of particular interest,
since molecular weight was in a lower range and predictive
of better membrane permeability. The dithiolanes 4c and 4e,
although biochemically highly potent, failed to show good
replicon potency and were also cleared more rapidly than
the standard. Indeed, all of the diothiolane members of the
other series were also rapidly cleared, leading to the
conclusion that this moiety is too unstable to confer druglike
properties. The poor translation of biochemical potency to
(7) MacDonald, S. J. F.; Clarke, G. D. E.; Dowle, M. D.; Harrison, L.
A.; Hodgson, S. T.; Inglis, G. G. A.; Johnson, M. R.; Shah, P.; Upton, R.
J.; Walls, S. B. J. Org. Chem. 1999, 64, 5166.
(8) (a) Christ, D. D. Drug Metab. Dispos 2001, 29, 935. (b) White, R.
E.; Minitpisitkul, P. Drug Metab Dispos 2001, 29, 957.
Org. Lett., Vol. 5, No. 24, 2003
4633

cellular activity is in agreement with our earlier findings and
suggests that hydrolytic instability is an inherent liability with
dithiolanes. The spirocyclopentyl lactam 4b showed reason-
ably low clearance but poor biochemical activity, and the
measured replicon potency could not be distinguished from
in-assay toxicity. The spirocyclobutyl lactams 4a and 4d
demonstrated an optimal combination of low clearance (i.e.,
less than the standard and around one-third dog liver blood
flow), good biochemical potency, and micromolar replicon
potency.
Interestingly, the isopropylcarbonyl-substituted lactam 4d,
although less potent in the biochemical assay than its
cyclopropyl-substituted counterpart 4a, demonstrated better
cellular potency. This mirrors the situation with the analogues
1a and 1d. This trend has been previously explained;4c 1d
shows considerably greater stability in aqueous media than
1a.
The favorable pharmacokinetic behavior of 4d extends to
other species: clearance was within an acceptable range in
marmoset, rat, and hamster (data not shown). Lactam 4d
(GW0014) has been progressed to pharmacodynamic evalu-
ation in GBV-B infected marmosets and shows good efficacy
(three logs lower virus titer compared to animals given
placebo) following subcutaneous administration.9
the use of rapid parallel synthesis and purification, a
substantial number of compounds were prepared for evalu-
ation, a subset of which is described herein. The use of the
dog cassette protocol permitted a thorough evaluation of the
limiting property of the series, namely, clearance, and
allowed the rapid identification of a candidate for in vivo
efficacy testing.
Acknowledgment. We thank Drs. Paul Feldman and
George Hardy for support and encouragement; Drs. Graham
Baker, Sue Bethell, and Malcolm Ellis for provision of NS3
protease protein and initial assay systems; Andy Craven,
Derek Evans, and John Leahy for provision of intermediates;
Norman M. Gray and Seb J. Carey for biochemical test data;
and Dr. Nigel Parry and Liz Amphlett for replicon test
results.
Supporting Information Available: Experimental pro-
cedures for synthesis of representative compounds, abridged
characterization data for test compounds, and experimental
procedures and validation data for dog cassette dosing. This
material is available free of charge via the Internet at
http://pubs.acs.org.
OL035827N
(9) Bright, H.; Watts, P.; Carroll, T.; Fenton R. The Validation of GBV-B
as a Surrogate Model for HCV in the Drug DiscoVery Process. Abstracts
of Papers, 16th International Conference on Antiviral Research, Savannah,
GA, 2003; Poster 151.
In conclusion, described here is the strategy for the
successful discovery of potent, low molecular weight,
druglike inhibitors of hepatitis C NS3/4A proteinase. Through
4634
Org. Lett., Vol. 5, No. 24, 2003