Highlights of the Structure-Activity Relationships of Benzimidazole Linked Pyrrolidines Leading to the Discovery of the Hepatitis C Virus NS5A Inhibitor Pibrentasvir (ABT-530)

Article abstract of DOI:10.1021/acs.jmedchem.8b00082

Curative interferon and ribavirin sparing treatments for hepatitis C virus (HCV)-infected patients require a combination of mechanistically orthogonal direct acting antivirals. A shared component of these treatments is usually an HCV NS5A inhibitor. First generation FDA approved treatments, including the component NS5A inhibitors, do not exhibit equivalent efficacy against HCV virus genotypes 1-6. In particular, these first generation NS5A inhibitors tend to select for viral drug resistance. Ombitasvir is a first generation HCV NS5A inhibitor included as a key component of Viekira Pak for the treatment of patients with HCV genotype 1 infection. Since the launch of next generation HCV treatments, functional cure for genotype 1-6 HCV infections has been achieved, as well as shortened treatment duration across a wider spectrum of genotypes. In this paper, we show how we have modified the anchor, linker, and end-cap architecture of our NS5A inhibitor design template to discover a next generation NS5A inhibitor pibrentasvir (ABT-530), which exhibits potent inhibition of the replication of wild-type genotype 1-6 HCV replicons, as well as improved activity against replicon variants demonstrating resistance against first generation NS5A inhibitors.

Full text of DOI:10.1021/acs.jmedchem.8b00082

Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE
Highlights of the Structure Activity Relationships of
Benzimidazole Linked Pyrrolidines Leading to the
Discovery of the HCV NS5A Inhibitor Pibrentasvir (ABT-530)
Rolf Wagner, John T. Randolph, Sachin V. Patel, Lissa T. Nelson, Mark A. Matulenko, Ryan Keddy,
John K. Pratt, Dachun Liu, A. Chris Krueger, Pamela L. Donner, Douglas K. Hutchinson, Charles
A. Flentge, David Betebenner, Todd Rockway, Clarence J. Maring, Teresa I. Ng, Preethi Krishnan,
Tami Pilot-Matias, Christine A. Collins, Neeta S. Panchal, Thomas J. Reisch, Tatyana Dekhtyar,
Rubina Mondal, DeAnne F. Stolarik, Yi Gao, Wenqing Gao, David A. Beno, and Warren M Kati
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b00082 • Publication Date (Web): 13 Apr 2018
Downloaded from http://pubs.acs.org on April 14, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Highlights of the Structure Activity Relationships of
Benzimidazole Linked Pyrrolidines Leading to the
Discovery of the HCV NS5A Inhibitor Pibrentasvir
(ABTꢀ530)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Rolf Wagner*, John T. Randolph*, Sachin V. Patel †,Lissa Nelson, Mark A. Matulenko, Ryan Keddy ‡,
John K. Pratt, Dachun Liu, A. Chris Krueger, Pamela L. Donner, Douglas K. Hutchinson ₴, Charles
Flentge║, David Betebenner ┴, Todd Rockway±, Clarence J. Maring±, Teresa I. Ng, Preethi Krishnan,
Tami PilotꢀMatias, Christine Collins, Neeta Panchal, Thomas Reisch, Tatyana Dekhtyar, Rubina
Mondal#, DeAnne F. Stolarik, Yi Gao, Wenqing Gao, David A. Beno¤, Warren M. Kati
Global Pharmaceutical Research and Development, AbbVie, 1 North Waukegan Road, North Chicago,
Illinois, 60064.
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 47
ABSTRACT
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Curative interferon and ribavirin sparing treatments for HCVꢀinfected patients require a
combination of mechanistically orthogonal direct acting antivirals. A shared component of these
treatments is usually an HCV NS5A inhibitor. First generation FDA approved treatments, including the
component NS5A inhibitors, do not exhibit equivalent efficacy against HCV virus genotypes 1ꢀ6. In
particular, these first generation NS5A inhibitors tend to select for viral drug resistance. Ombitasvir is a
first generation HCV NS5A inhibitor included as a key component of Viekira Pak for the treatment of
patients with HCV genotype 1 infection. Since the launch of next generation HCV treatments, functional
cure for genotype 1ꢀ6 HCV infections has been achieved, as well as shortened treatment duration across
a wider spectrum of genotypes. In this paper we show how we have modified the Anchor, Linker and
Endcap architecture of our NS5A inhibitor design template to discover a next generation NS5A
inhibitor, pibrentasvir (ABTꢀ530), which exhibits potent inhibition of the replication of wildꢀtype
genotype 1ꢀ6 HCV replicons, as well as improved activity against replicon variants demonstrating
resistance against first generation NS5A inhibitors.
2
ACS Paragon Plus Environment

Page 3 of 47
Journal of Medicinal Chemistry
INTRODUCTION
1
2
3
4
5
6
7
8
9
Hepatitis C virus (HCV) was unambiguously characterized in 1989 as the agent responsible for
most cases of nonꢀA and nonꢀB hepatitis infection.1 With current estimates ranging from 71ꢀ80 million
people worldwide chronically infected with HCV, this viral infection represents a global public health
problem.2 HCV genotype 1, predominant in North America, Europe, and Japan, accounts for over half
of the global infections. Genotype 2 infections are most prevalent in North America, Europe and Japan,
while genotype 3 and 6 infections are predominant within various parts of Southeast Asia. In Egypt,
HCV infections are almost exclusively genotype 4, while genotype 5 is common in South Africa.
Following initial acquisition of the virus, HCVꢀinfected individuals undergo an acute phase,
followed in most cases by development of chronic infection.3 The majority of infected individuals
remain asymptomatic for many years after the acute phase, with hepatic fibrosis progressing as the main
complication of chronic HCV infection. While there is a tremendous variation in the rate of progression
from fibrosis to cirrhosis, with no treatment, the median progression to cirrhosis is 30 years. Of this
population, a substantial fraction will continue to experience progression of disease into hepatocellular
carcinoma.4 Complications due to chronic HCV infection have now become the predominant driver for
liver transplantation.5 Additionally, the leading cause of death upon kidney transplantation, which is the
most commonly performed organ allograft, is due to chronic liver disease, most commonly precipitated
by HCV infection.6 In 2004, 2.44 deaths per 100,000 people could be attributed to complications due to
chronic HCV infection in the U.S.7 In 2015, it has been reported that 19629 deaths have been attributed
to complications related to HCV infection.8 According to one predictive model, this number is expected
to peak in 2030 at ~39,800 deaths per year,9 though a more conservative prediction places the peak
mortality level in 2030 at 12,900 deaths.10 The CDC In the U.S., of those who do not receive treatment
for HCV infection, it is estimated 1 million will die as a consequence of this disease.10
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Although population genetics studies of HCV indicate an extremely diverse collection of
genotypes (1ꢀ6) worldwide, HCV genotype 1 comprises ~75% of HCV infections in the U.S., Western
Europe, and Japan, and ~60% of HCV infection in Latin America.11 While treatment of HCV has been
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 47
revolutionized with the approval of direct acting antivirals (DAAs), pegylatedꢀinterferon/ribavirin
therapy was the preferred treatment for genotype 1a/b until recently. Usual therapy required 48 weeks of
treatment that was poorly tolerated and provided limited functional cure rates.12 This functional cure is
considered to be achieved when virus RNA levels in blood remain undetectable for 12 weeks after
completion of treatment, commonly noted as a sustained viralogic response at 12 weeks (SVR12). It
should be noted that while undetectable, sufficient viable virus may still be present to allow for relapse
in rare instances. For HCV genotype 1ꢀinfected individuals with additional risk factors, such as
cirrhosis, HIV coꢀinfection, dialysis treatment, or organ transplants, functional cure are more difficult to
achieve.13 Several teams have successfully conducted clinical studies of multiple DAAꢀonly cocktails
for well tolerated curative treatments of this disease, subsequently bringing them to market for genotype
1 HCV infected patients.14,15,16 Additionally, two of these earlier generation treatments have each been
approved for treating HCV patients beyond genotypeꢀ1. The combination sofosbuvir/ledipasvir gained
additional approval for genotypes 4ꢀ6, and ombitasvir/paritaprevir/ritonavir as well as
elbasvir/grazoprevir, were approved for genotype 4.17 Next generation treatment required the ability to
achieve SVR12 in HCV genotypes 1ꢀ6 infections, as well as the ability to inhibit viral replication of
those strains that demonstrated resistance against first generation treatments. Since that time, two next
generation treatments, sofosbuvir/velpatasvir and sofosbuvir/velpatasvir/voxilaprevir, have been
approved by the FDA for the treatment of HCV infections encompassing genotypes 1ꢀ6.18
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The RNA genome of HCV encodes several nonꢀstructural (NS) enzymatic and nonꢀenzymatic
proteins which enable formation of the polyꢀprotein replication complex prior to the production of new
HCV virions. One of these replication complex proteins is NS5A, and although no enzyme activity is
associated with this protein it is recognized as essential for viral replication.19 As already described,
several NS5A inhibitors have received FDA marketing approval for use in combination therapy of HCV
in recent years, since these inhibitors demonstrate limited utility as single agents due to the emergence of
drug resistant HCV variants encountered with monoꢀtherapy.20 It should also be noted that, although
nearly every approved IFNꢀsparing HCV therapy shares the distinction of including an NS5A inhibitor
4
ACS Paragon Plus Environment

Page 5 of 47
Journal of Medicinal Chemistry
as part of its drug combination cocktail, clinical studies need to be conducted in order to assess
compatibility between any two agents, as unforeseen drugꢀdrug interactions may either decrease or
increase the combination’s effectiveness and tolerability. Additional limitations of several current
regimens include significant drugꢀdrug interactions, especially in HCVꢀHIV coinfected patients being
treated for HIV, 21 limited options for subjects with renal insufficiency,22 reduced efficacy in patients
with baseline amino acid polymorphisms associated with reduced susceptibility to the HCV NS5A
inhibitors or NS3/4A protease inhibitors (PIs).23
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Chemical Structure of Pibrentasvir and Glecaprevir
This paper will describe the discovery of pibrentasvir (ABTꢀ530) which suppresses replication of wild
type HCV genotype 1ꢀ6 replicons, with a much reduced propensity to elicit drug resistance when
compared to other first generation NS5A inhibitors. Pibrentasvir, when combined with glecaprevir
(ABTꢀ493) resulted in AbbVie’s next generation IFNꢀ and RBVꢀsparing HCV treatment for the curative
treatment of HCV genotypes 1ꢀ6 infection.24
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 47
RATIONALE
1
2
3
4
5
6
7
8
9
Our prior efforts resulted in the discovery of ombitasvir (OMB, Figure 2),25 the NS5A inhibitor
included as a component of the DAA cocktail in AbbVie’s first generation IFNꢀsparing HCV treatment.
First generation NS5A inhibitor ombitasvir has been shown to exhibit potent 50% effective
concentrations inhibiting replication (EC50s) (ranging 0.82ꢀ19.3 pM) against HCV genotypes 1ꢀ5 and an
EC50 of 366 pM against genotype 6a in the replicon assays. In vitro resistance selection experiments in
genotype 1ꢀ6 replicons selected variants which demonstrated reduced susceptibility to the actions of
ombitasvir, by factors often greater than 1000ꢀfold.26 The medicinal chemistry strategy for the discovery
of a next generation NS5A inhibitor was to retain the potent antiviral properties of OMB, while
introducing significant improvements to genotype coverage, and most significantly, to improve the
genotype 1 resistance profile relative to OMB. Replacement of OMB with a compound possessing an
improved resistance profile could potentially translate into requiring fewer DAAs in the curative
combination.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Chemical Structure of ombitasvir and Structural Template Modifications for Next Generation
NS5A Inhibitor Design
This report will provide an overview of the effects that replacements of the “End Caps,” “Linker,” and
“Anchor Groups,” as well as changing the pyrrolidine core stereochemistry (See Figure 2) had on the
6
ACS Paragon Plus Environment

Page 7 of 47
Journal of Medicinal Chemistry
virological profile and how these observations provided key insights into crafting the final structure of
next generation NS5A inhibitor pibrentasvir.
1
2
3
4
5
6
7
8
Chemistry
9
A stereochemicallyꢀcontrolled method used to synthesize pyrrolidineꢀcore HCV inhibitors
targeting the NS5A protein is shown in Scheme 1. Condensation of 4ꢀchloroꢀ3ꢀnitroꢀphenyl methyl
ketone, 1, with alphaꢀbromo derivative 2, gave symmetrical 1,4ꢀdiketone 3.27 Chiral reduction using
(R)ꢀ(+)ꢀα,αꢀdiphenylꢀ2ꢀpyrrolidinemethanol, trimethylborate, and diethylaniline borane gave the S,Sꢀ
diol, which was converted to dimesylate 4 and then reacted with the appropriate aniline reagent to give
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2,5ꢀR,Rꢀpyrrolidine intermediates 5.28
Palladiumꢀcatalyzed Buchwald coupling of 5 with MocꢀLꢀ
valineꢀLꢀprolinamide provided symmetricallyꢀsubstituted orthoꢀnitro acylaniline 6. Reduction of the
nitro groups to give the diamine, followed by heating in acetic acid, formed bisꢀbenzimidazole 7. In an
alternative procedure, compound 5 was coupled with BocꢀLꢀprolinamide to give 8, which was converted
to the bisꢀbenzimidazole intermediate. Removal of the Boc protecting groups with HCl, followed by
peptide coupling with MocꢀLꢀvaline, gave 7. Intermediate 8 provided access to HCV inhibitors with
various peptide endꢀcapping groups, such as Oꢀmethylꢀthreonine analog 9 and 4ꢀtetrahydropyranylꢀ
glycine analog 10.29
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 47
Scheme 1.a
1
O
S
R
Cl
R
Cl
O
O
O
O
2
3
4
O
O2N
Cl
O2N
Cl
Br
O2N
Cl
O2N
Cl
A
NO2
B
NO2
O
O
O
R
R
R
S
R
O
3
4
5
1
2
6
7
8
9
a: R = H
b: R = F
C
Z
Z
X
Y
X
Y
O2N
NO2
O2N
NO2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
H
N
H
N
H
N
H
D
N
Cl
O
O
Cl
N
O
N
N
N
O
O
O
O
O
R
R
R
R
6
5
E
F
Z
Z
O
O
O
X
Y
X
Y
HN
R1
N
O
NH
O
N
O2N
NO2
NH
HN
H
N
E, G
H
N
N
R1
N
O
N
N
N
N
O
O
O
O
O
O
R
R
R
R
8
7: R1 = isopropyl
9: R1 = (R)-methoxyethyl
10: R1 = 4-tetrahydropyranyl
aKey: (A) ZnCl2, Et2NH, tꢀBuOH, benzene. (B) i. (R)ꢀdiphenyl(pyrrolidinꢀ2ꢀyl)methanol, trimethyl borate, N,Nꢀdiethylaniline borane, THF;
ii. MsCl, TEA, CH2Cl2, 0 oC. (C) ArNH2, Hunig's base, CH3CN, 60 oC. (D) MocꢀValꢀProꢀNH2, Pd2(dba)3, xantphos, Cs2CO3,
1,4ꢀdioxane, 100 oC. (E) i. PtO2 or RaꢀNi, H2, THF; ii. AcOH, 70 oC. (F) BocꢀProꢀNH2, Pd2(dba)3, xantphos, Cs2CO3, 1,4ꢀdioxane, 100 oC.
(G) i. HCl, 1,4ꢀdioxane; ii. MocꢀValꢀOH or MocꢀThr(OMe)ꢀOH or Mocꢀ2ꢀ(S)ꢀ(4ꢀThp)GlyꢀOH, HATU, Hunig's base, DMF.
The method was applied for the synthesis of pibrentasvir as shown in Scheme 2. Heating a
DMSO solution of 1,2,3ꢀtrifluoroꢀ5ꢀnitrobenzene and dioxolaneꢀprotected 4ꢀketoꢀpiperidine, followed
by ketal hydrolysis, provided Nꢀarylꢀpiperidinone 11. The ketone was converted to a vinyl triflate in
order to prepare for Suzuki coupling to add the fluorophenyl group. Hydrogenation reduced both the
nitro group and olefin to give 12, which was reacted with S,Sꢀdimesylate 4b (R = F) to give R,Rꢀ
pyrrolidine 13. Proline amide groups were installed via Buchwald coupling, providing diꢀnitro
intermediate 14, which was converted to bisꢀbenzimidazole intermediate 15 as described above.
8
ACS Paragon Plus Environment

Page 9 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Biology
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The primary in vitro tool used to screen analogs was the subgenomic HCV cell culture replicon system.
Although HCV has a very robust replication rate in vivo, efforts to propagate the virus in cell culture
have been largely unsuccessful for genotype 1, despite the success of a genotype 2a infectious system.30
This challenge has led to the development of subgenomic HCV replicons that can be propagated
efficiently in the human hepatoma cell line Huhꢀ7.31 These replicons contain 5’ and 3’ nonꢀtranslated
regions as well as the HCV nonꢀstructural genes. Similar to conventional tissue culture and in vivo viral
replication, HCV replicon RNA efficiently replicates by undergoing a cycle consisting of translation,
polyprotein cleavage (by the NS3/4A protease), RNA synthesis (by the NS5B polymerase), unwinding
of the doubleꢀstranded RNA (by the NS3 helicase), and assembly of replication complexes containing
the HCV nonstructural proteins within Huhꢀ7 cells. The replicon system produces 1,000ꢀ10,000 copies
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 47
of subgenomic RNA per cell, which is substantially greater than that estimated for in vivo infection (50ꢀ
100 RNA strands/cell). The replicon system lacks assembly and release of virus particles. Therefore, the
replicon system is incapable of producing infectious virus, and viral infection of new cells does not
occur. Nevertheless, the predictive value of HCV replicons to characterize HCV inhibitors has now long
been validated in the clinic, with one of the first instances using the NS3/4A protease inhibitor BILNꢀ
2061.32 Compound potency is reported as EC50 against the subgenomic replicons of different genotypes
incorporating the firefly luciferase reporter gene. Amino acid positions 24, 28, 30, 31, 32, 58, 92, 93 in
NS5A are considered as positions of interest for the NS5A inhibitor class. Of these, variants at positions
28, 30 and 93 were most commonly detected in patients experiencing virologic failure with the first
generation NS5A inhibitors, ombitasvir, daclatasvir, and ledipasvir. The resistance profile of compounds
was therefore evaluated against representative amino acid substitutions, M28T, Q30E, Q30R, Y93C,
Y93H, Y93N in genotype 1a, and Y93H, Y93N in genotype 1b replicons.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In order to establish the effect of plasma protein binding on the antiviral activity, we tested these
compounds in the presence of 5% fetal calf serum (FCS) with or without 40% human plasma. Although
it is recognized that the shift in EC50 value due to protein binding is static in vitro, versus the dynamic
conditions encountered in vivo, and that individual patient plasma protein binding shows variability, it
has nevertheless been established that evaluating protein shifting effects in vitro can help to prioritize
the advancement of compounds and deꢀrisk hydrophobic chemical compounds as they advance into
more costly experiments. This prioritization allows one to roughly estimate the effect of plasma proteins
on the potency of compounds and target therapeutic plasma concentrations in the clinic.33
10
ACS Paragon Plus Environment

Page 11 of 47
Journal of Medicinal Chemistry
RESULTS AND DISCUSSION
1
2
3
4
5
6
Table 1
7
8
Antiviral Activity (EC50, pM) of NS5A Inhibitors in HCV Stable Replicons
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Inhibition of HCV Stable Replicons Containing NS5A from Genotypes
1ꢀ6 EC50 (pM)
Chirality
cmpd Carbons
2,5
Linker
Pair
40% H.
Plasma
1a
1b
2a
2b
3a
4a
5a
6a
1a
1b
56
90
S,S
R,R
S,S
A
A
B
B
14
136
71
9
5
9
186
986
12
4.3
NT
NT
10
19
NT
NT
6
1.7
NT
NT
6
3.2
NT
NT
NT
366
NT
NT
NT
OMB
16
NT
NT
13 1980 354
13
17
R,R
148 268 152
18
H. Plasma: human plasma; NT: not tested; pM: picomolar
While seeking to improve upon the virological properties of ombitasvir, whose pyrrolidine coreꢀ
phenyl amide linker had already served as an exhaustive platform to search for optimal antiꢀHCV
activity, we entertained the notion that perturbing this core structure could open new opportunities for
improved virology, that were not otherwise attainable with the original core. An alternative inhibitor
scaffold (Table 1), investigated the impact of the “linker” moieties A and B, as well as the absolute
stereochemistry at carbons 2 and 5. While the phenyl amide linker presented the desired broad genotype
coverage with the chirality at carbon’s 2 and 5 being S,S; altering the stereochemistry R,R, attenuated the
activity of 16 against genotype 1a. 2S,5SꢀBenzimidazole analog 17 was weakened in genotype 1a in the
presence of 40% human plasma, despite being an isosteric replacement for the linker found in
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 47
ombitasvir. Surprisingly, the 2R,5R isomer 18, provided potent activity against all the genotypes tested.
The most compelling data that indicated a promising path for the benzimidazole lead 18, came from
comparing the lower fold resistance of single point resistant NS5A variants in the 1a and 1b transient
replicon experiment against 18 when compared with ombitasvir (Table 2).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 2
Fold Resistance of HCV Genotype 1a/1b NS5A Transient Replicon Variants vs. WildꢀType
Against NS5A inhibitors 3 and 21
HCV Genotype 1a/1b NS5A Variants vs. Wild Type (Fold
Resistance)
cmpd
1a*
1b*
Y93H
77
M28T Q30E Q30R Y93C Y93H Y93N
8965
4
NT
61
800
14
1675 41383 66740
216 510
OMB
18
7
2
*Observed as resistant variants in ombitasvir single agent clinical studies
In mouse PK experiments, 18 demonstrated inferior PK to ombitasvir (Table 3) postulated to
result from elevated metabolism, as judged by the lower IV halfꢀlife, IV AUC (despite the higher dose),
and higher IV clearance measurements (volumes of distribution were similar) vs. ombitasvir. These
Table 3
Comparison of PK Parameters of 3 and 21 in Mouse
IV
PO
cmpd
Dose
T1/2
(h)
Vss AUC0ꢀinf
CLp
(L/h/kg)
T1/2 Tmax Cmax C24h AUC0ꢀinf
Vβ
F
(%)
(mg/kg)
(L/kg)
(ꢁg*h/mL)
(h)
(h)
(nM)
(nM)
(ꢁg*h/mL)
(L/kg)
3
5
11
2
1.8
1.67
0.65
26.44
0.11
0.38
11.7
7
636
205
11.03
41.7
10.6
OMB
18
1.11
13.13
2.61 1.67 301
0
1.39
Formulation: 3 only PO 50:20:10:20 (w/w) PEGꢀ400:Tween 80:Ploxamer 124:Vit.E TPGS; see Experimental Section for additional details
12
ACS Paragon Plus Environment

Page 13 of 47
Journal of Medicinal Chemistry
data then set the stage for subsequent program development, as 2,5ꢀR,Rꢀpyrrolidine core, benzimidazole
linkers were viewed as a structural platform in which broad HCV genotype coverage might be attainable
along with an improved resistance profile. The ultimate objective was to further improve on the initial
virological observations seen with 18, while significantly improving the PK profile of emerging analogs.
Our initial mission sought to achieve greater potencies with comparable activities across the
various HCV genotypes, as well as retaining activity against amino acid substitutions commonly
selected as resistant variants by first generation NS5A inhibitors. Table 4 summarizes the antiꢀviral
activity of analogs resulting from modifying substituents X, Y and Z in wildꢀtype HCV genotypes 1ꢀ4,
as well as genotype 6 for selected analogs. A comparison of smaller or comparably sized groups to the tꢀ
Bu borne by OMB or 18 in the “Z” position revealed that F, Cl, cyclopropyl, cyclohexyl and phenyl all
provided picomolar inhibition of wildꢀtype genotype 1ꢀ4 HCV replicons, and demonstrated a significant
improvement relative to OMB in genotype 6a. This was surprising, as groups such like F, Cl, and
cyclopropyl lost significant genotype 1a activity in the prior bisꢀanilide series relative to genotype 1a and
larger groups such as cyclohexyl and phenyl lost significant potency in the presence of 40% human
plasma.25 In the presence of 40% human plasma, the genotype 1a and 1b EC50s remained below 500 pM.
Introduction of an oxygenꢀlinking atom (eg. Z = phenoxy in 24; Z = trifluoromethyloxy in 25) resulted
in analogs possessing similar “antiviral” potency to what we observed for simple halogen, alkyl and aryl
substituents. By contrast, introduction of a basic morpholino group (26) resulted in much weaker
analogs, even upon introduction of a fluorine atom at position “Y” to attenuate basicity.34 Surprisingly,
piperidinyl analogs 28, 29 and 30 gained potency, especially with the introduction of halogens at both
positions X and Y. The unanticipated potency enhancing effect of fluorine at positions X and Y is
especially dramatic when comparing analog 31 with 32 and 33. For example, 32 provides single digit
picomolar potency across all tabulated wildꢀtype HCV genotypes as well as 3 pM inhibitory potency for
genotype 5a (data not in table).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 47
Table 4
1
2
3
4
5
SAR of (R,R)ꢀBenzimidazole Linker NS5A Inhibitors
Inhibition of HCV Stable Replicons
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Inhibition of HCV Stable Replicons Containing NS5A from
Genotypes 1ꢀ4, and 6, EC50 (pM)
40% H. Plasma
X Y
Z
cmpd
1a
1b
2a
2b
3a
4a
6a
1a
1b
H
H
H
H
H
H
F
19
20
21
38
6
25
7
490
240
161
427
330
354
28
12
41
21
7
31
12
26
14
5
14
8
Cl
12
23
20
20 NT
H
H
H
H
22
23
4
4
9
3
107
75
178
108
10
11
7
9
8
9
6
NT
11
16
H
H
24
12
20
155
279
230
12
43
8
11
18
5
NT
H
H
H
H
F
H
H
F
25
26
27
28
29
30
31
32
19
379
189
18
5
12
382
196
29
6
276
902
474
199
151
213
1510
76
12
13 NT
2070 598
960 1200 316 NT
915
383
241
132
307
48
229
30
7
355 170 NT
F
44
8
17 NT
F
13
5
6
7
N
Cl
H
F
Cl
H
F
5
3
44
5
11
10
324
2
325
6
1510 468
452
6
481 243 402
183
126
8
6
4
9
8
F
F
33
3
8
78
20
13
13
21
14
ACS Paragon Plus Environment

Page 15 of 47
Journal of Medicinal Chemistry
Table 5
1
2
3
SAR of (R,R)ꢀBenzimidazole Linker NS5A Inhibitor ꢀ Physical Properties, Metabolism and Mouse PK
4
5
6
7
8
9
Human
ꢀꢀsomal
Intrinsic
Clearance
(L/h/kg)
Rat
Mouse PK
(3 mg/kg)
PO AUC
CLND
Solubility
(ꢀM)
ꢀꢀsomal
Intrinsic
Clearance
(L/h/kg)
LogD
(HPLC)
PAMPA
cmpd
(x 10ꢀ6 cm/s)
(ꢀg*h/mL)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
1.39
(5 mg/kg)
18
5.3
0.14
1.6
7.3
10
26
28
29
30
31
32
33
3.2
4.7
NT
5.9
5.4
NV
5.8
0.73
0.01
0
14.2
2.5
6.8
6.5
9.7
11.2
3
0.002
2
8.4
0.7
10.5
<1.5
<1.5
<1.5
<1.5
1.79
0.96
NT
0.03
NV
0
<0.6
0.82
1.1
2.8
3.2
3.2
6.83
4.59
0
0.9
The physical properties of multiple analogs were measured, trying to establish a correlation
between in vitro ADME characteristics and AUCs determined after oral administration (Table 5). As all
of the molecules fall outside of the boundary of Lipinski’s Rule of 5, an effort was extended to improve
the water solubility of these molecules by replacing the lipophilic tꢀbutyl “R” substituent with moieties
expected to lower the logD, increase solubility (CLND) and decrease rat microsomal clearance.
Morpholinyl analog 26 showed that as the LogD decreased, the solubility improved, although
permeability (desired PAMPA values >2 x 10ꢀ6 cm/s) did not improve to the required extent, nor was rat
microsomal clearance effected substantially. Furthermore, this less lipophilic analog unexpectedly
demonstrated even poorer plasma exposure than 18 upon oral administration. Ortho substituted aryl
halides 28, 29 and 30 were synthesized to measure the effect of attenuating piperidine basicity. While
higher LogD values, lower aqueous solubility, and poor permeability remained an issue in vitro, these
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 47
compounds demonstrated improved oral AUCs, but did not achieve the exposures required for
additional development. Despite the deleterious effects of lowering the logD upon the oral AUCs, we
nevertheless decided to test larger and more lipophilic analogs in vivo. When measurable, phenyl or 4ꢀ
fluoroꢀphenyl substituted piperidines, 31, 32 and 33, demonstrated LogDs greater than 5, with no
significant ability to provide intrinsic permeability in the PAMPA assay or solubility in the CLND
measurements, yet, for reasons that were not elucidated at that time, oral exposures were unexpectedly
improved. It should, however, also be noted that intrinsic permeability measurements in the PAMPA
assay are likely misleading due to the very low solubility of these compounds.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Of particular note, in 2016 Randolph et al. reported that 5ꢀfluoro substitution of the
benzimidazole linker of analogs 18 and 29, provided compounds of nearly identical potencies in
genotypes 1ꢀ4, and 6, while also increasing drug plasma levels after oral dosing in mice.35 The
hypothesis underlying the improvement of oral PK upon benzimidazole fluorination predicted that the
halogenation would decrease the basicity of the benzimidazole nitrogens.36 In order to verify that
sufficient antiviral activity of this potentially promising modification extended to other analogs with
respect to virology and PK, additional compounds were synthesized. We were gratified to find similar
activity across all genotypes for 34 when compared with 32, but were surprised by the improvement
observed for 35 when compared with 33 (see Table 4 and Table 6). While broad genotype in vitro
activity was achieved in the series, a second requirement of a next generation NS5A inhibitor was
effectiveness against replicon variants which confer resistance to first generation NS5A inhibitors.
Table 7 illustrates the SAR of a variety of substituent patterns that, while found to show promise
across wild type genotypes 1ꢀ6, resulted in significant variation in their ability to suppress the replication
of clinically relevant ombitasvirꢀresistant variants.26 Examination of the data reveals that the greatest
resistance emerged with the Q30E, Y93H and Y93N variants in genotype 1a. Y93 variants of genotype
1b, where tested, were susceptible to all analogs in Table 7. There appears to be a general trend that
increased size at position “Z” correlates with an improved ability to suppress replication with lower
16
ACS Paragon Plus Environment

Page 17 of 47
Journal of Medicinal Chemistry
Table 6
1
2
3
4
SAR of 6ꢀFluoroꢀ(R,R)ꢀBenzimidazole Linker NS5A
5
6
Inhibitorsꢀ Antiviral Activity in HCV Stable Replicons
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Inhibition of HCV Stable Replicons Containing NS5A from Genotypes 1ꢀ6, EC50
(pM)
40% H.
Plasma
cmpd
R
1a
1b
2a
2b
3a
4a
5a
6a
1a
1b
Phenyl
2
1
6
3
58
70
135
172
8
3
3
2
4
1
4
1
2
1
9
3
34
35
4ꢀFꢀPhenyl
multiples of the wildꢀtype EC50. One particular exception to this trend is found for 22, whose wildꢀtype
genotype 1a EC50 was lower than the lowest concentration tested. Compounds emerging from these
studies, which came close to fulfilling the virology requirements for a next generation analog, are 32, 33
and 34. All three compounds deliver a profound improvement in performance against resistant variants
when compared to the first generation NS5A inhibitor OMB (ref. 25) or 18, and maintain reasonable
potency against genotypes 1ꢀ6. Analog 35 was not characterized against these variants.
Compounds 34 and 35 were evaluated in the mouse at 3 mg/kg to determine whether or not oral
exposures could be improved vs. the unfluorinated benzimidazole analog 32 and 33 (Table 8). It was
gratifying to discover that in both cases, the AUC increased by several multiples, thus underscoring the
importance of the fluorinatedꢀbenzimidazoles towards improving the oral PK properties of resulting
analogs, while maintaining virological potency.
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 47
1
2
3
4
5
Table 7
Z
X
A
X
A
6
7
8
9
N
SAR of NS5A Inhibitors ꢀ Fold Resistance of HCV
N
O
HN
NH
NH
O
O
N
N
HN
O
N
O
O
Genotype1a/1b NS5A Transient Replicon Variants vs.WildꢀType
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
HCV Genotype 1a/1b NS5A Variants vs. Wild Type
(FoldꢀResistance)
A X
Z
cmpd
1a*
1b*
M28T Q30E Q30R Y93C Y93H Y93N Y93H Y93N
H
H
H
H
Cl
28
83
44
22
670
1155
NT
1
1
NT
NT
20
22
>17
>90
>30
>30
>577
H
H
H
F
NT
5
47
10
13
24
9
329
241
827
510
1
1
1
23
29
134
NT
H
H
F
F
3
2
69
20
7
3
6
3
72
20
145
25
1
2
1
32
33
NT
F
F
1
12
2
6
49
50
1
1
34
* Observed as resistant variants in ombitasvr single agent clinical studies
18
ACS Paragon Plus Environment

Page 19 of 47
Journal of Medicinal Chemistry
Table 8
1
2
3
4
Oral PK Parameters of 37 and 38 in Mouse
5
6
7
PO dosing (3 mg/kg)
8
9
cmpd
AUC0ꢀinf
(ꢁg*h/mL)
T1/2 (h)
Tmax (h)
24.0
Cmax (nM)
754
C24h (nM)
754
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Unable to
determine
14.0
26.1
34
35
Unable to
determine
24.0
1708
1708
Table 9
SAR of Endcap Replacements of 6ꢀFluoroꢀ(R,R)ꢀ
Benzimidazole Linker NS5A Inhibitors Antiviral Activity in
HCV Stable Replicons
Inhibition of HCV Stable Replicons Containing NS5A from Genotypes 1ꢀ6, EC50 (pM)
40% H.
Plasma
cmpd
R1
R2
1a
1b
2a
2b
3a
4a
5a
6a
1a
1b
4
1
3
8
3
7
136
337
9
3
8
8
3
8
11
3
6
3
6
6
2
5
12
4
36
Phenyl
Phenyl
43
138
522
37
38
4ꢀFꢀ
Phenyl
220
7
8
1.8
±0.86
4.3
64
200
±54
2.3
1.9
2.1
1.9
1.4
2.8
4ꢀFꢀ
Phenyl
PIB
±1.7
±14
±0.65
±0.59
±0.66
±0.61
±0.36
±0.67
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 47
As the fluorobenzimidazoleꢀlinked compounds were previously shown to improve oral PK
characteristics,35 final synthetic efforts sought to examine the effects of varying the MocꢀValꢀ “end
capping” groups, with the aim of further improving the virological profile, while conserving the central
core structure of 33 and 34. Table 9 summarizes the antiviral activity of the best of these compounds
against replicons of wildꢀtype genotypes 1ꢀ6. All of these compounds show very potent picomolar EC50s
across all genotypes in the absence of 40% human plasma. Indeed, the EC50s of 1 maintain levels below
10 pM in genotypes 1ꢀ6.37 Addition of 40% human plasma attenuated potencies by ca. 30ꢀ80 fold, and
the best overall profile with respect to EC50 potencies across genotypes and the least protein binding
attenuation effects were assigned to 37 and pibrentasvir (PIB), both of which contained the Mocꢀ
Thr(OMe)ꢀ capping group.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 10
SAR of End Cap Replacement Analogs
Fold Resistance of Genotype1a/1b NS5A Transient Replicon Variants vs.WildꢀType
HCV Genotype 1a/1b NS5A Variants vs. Wild Type
(FoldꢀResistance)
cmpd
1a*
1b*
M28T Q30E Q30R Y93C Y93H Y93N Y93H Y93N
NT
1
1
2
1
NT
1
NT
1
3
6
2
2
5
2
NT
1
NT
0.3
NT
36
37
38
NT
NT
NT
NT
2.1
2.4
1.7
1.7
6.7
6.7
0.6
0.6
PIB
* Observed as resistant variants in ombitasvr single agent clinical studies
Table 10 provides an overview of the foldꢀresistance of HCV genotype 1a/1b NS5A variants
relative to wildꢀtype in the transient replicon of the analogs resulting from these end capping variations.
20
ACS Paragon Plus Environment

Page 21 of 47
Journal of Medicinal Chemistry
When compared to the best analogs present in Table 7 (i.e. 33 and 34), 37 and PIB provide a
significantly improved resistance profile. Indeed, all clinically relevant single amino acid NS5A
substitutions confer less than 10ꢀfold resistance relative to wildꢀtype. More details on the virological
profile of PIB, including resistant variant “fitness”, activity against double/triple amino acid
substitutions, the compound’s much lower potential than other approved NS5A inhibitors to elicit
resistance across all genotypes, and in vitro comparisons to daclatasvir, ledipasvir, ombitasvir, elbasvir
and velpatasvir, are described by Ng, et al.37
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 11
Pharmacokinetics of PIB in Mouse, Rat, Dog and Monkey at Low Doses
IV
PO
Species
Dose
(mg/kg)
T1/2
(h)
Vc
AUClast
CLp
(L/h/kg)
Dose
(mg/kg)
T1/2
(h)
Cmax
(nM)
C24h
(nM)
AUClast
(ꢁg*h/mL)
Tmax (h)
F (%)
(L/kg) (ꢁg*h/mL)
mouse
rat
3
3
1
1
2
0.09
ꢀ
252
ꢀ
<0.003
3
ND
11
1140
467
20
7.9
6.2
0.07
3
7.0
5.3
3.67
4
250
562
260
41
3.6
9.9
dog
7.1 0.1
8.3 0.07
10.55
6.47
0.097
0.15
2.5
2.5
8.3
82
5.99
2.25
29.8
14.1
monkey
5.69
17
Dog and Monkey Formulations: IV 10:90 (v/v) DMSO:PEG-400; PO 70:20:10 (w/w) PEG-400:Tween 20:Poloxamer 124 acidified with 2 eq 1N HCl
Pharmacokinetic characterization of PIB in mouse, rat, dog and monkey is summarized in Table
11. The oral bioavailability observed in all four species is low to moderate. On a per milligram basis,
plasma concentrations at the 24 hour time point in rat, dog and monkey are lower than mouse, but are
still in the nanomolar range. Adjusted EC50s of PIB due to plasma protein binding against genotype 1a
and 1b replicons were shown to be subnanomolar and resistanceꢀfold changes to common NS5A amino
acid substitutions were less than 10ꢀfold. Consequently, suppression of HCV replication in patients by
PIB was considered achievable, especially if used in combination with another mechanistically
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 47
orthogonal agent. Towards that end, in vitro synergy with GLE was demonstrated38 with later clinical
1
2
3
studies showing this combination to be effective in patients.24
4
5
6
CONCLUSIONS
7
8
9
These studies have examined the effectiveness of altering the linker region of NS5A inhibitors
from bisꢀanilide structures to bisꢀbenzimidazoles. This alteration, in combination with the unexpected
finding that reversing the stereochemical linkage at the pyrrolidine central core, improved suppression of
the replication of HCV replicons in genotypes 1ꢀ4, prompted further structural investigations for more
clinically promising inhibitors. These improvements stressed potency across multiple genotypes and an
ability to retain activity against resistanceꢀassociated amino acid substitutions commonly selected by the
first generation NS5A inhibitors. Structural modifications that allowed us to leverage the initial findings
into a potent next generation panꢀgenotypic NS5A inhibitor drove toward larger and more lipophilic
anchor substituents, as well as selective fluorination of the anchor and benzimidazole linker groups.
Significant pharmacokinetic challenges in mice were encountered in the early use of the
benzimidazole linker in that plasma exposure of the drugs could not be detected upon oral
administration. Attempts to improve the physical properties of the agents by the introduction of
hydrophilic substituents on the anchor portion of the molecule were not successful. Surprisingly, plasma
exposures were achieved upon fluorination of the benzimidazole, as well as increasing the lipophilicity
and size of the anchor group.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Pibrentasvir (PIB) was the molecule that emerged from this study due to its possession of the
attributes necessary for a successful next generation NS5A inhibitor, namely 1) potent activity across all
common genotypes, 2) activity against HCV viral variants demonstrating resistance to OMB and other
first generation NS5A inhibitors, 3) high genetic barrier to development of drug resistance in all
common genotypes,37 4) sufficient plasma exposures upon oral administration to suppress HCV
replication, and 5) the ability to be combined with another mechanistically orthogonal antiviral agent.37
22
ACS Paragon Plus Environment

Page 23 of 47
Journal of Medicinal Chemistry
EXPERIMENTAL SECTION
1
2
3
4
Replicon assays with laboratory strains.
5
6
7
8
Compounds were tested against eight stable subgenomic replicon cell lines using the luciferase
reporter activity assay as described previously by Ng and Krishnan.26,37 These include genotype 1aꢀH77
(GenBank accession number NC004102), genotype 1bꢀCon1 (GenBank accession number AJ238799),
and six chimeras of genotype 1bꢀCon1 containing the Nꢀterminal region of NS5A from genotype 2a, 2b,
3a, 4a, 5a, or 6a HCV. All replicon constructs were bicistronic subgenomic replicons similar to those
described by Bartenschlager and coworkers, and the stable replicon cell lines were generated by
introducing these constructs into the Huhꢀ7 human hepatoma cell line. The inhibitory effect of
compounds on HCV replication in replicon cells was determined in DMEM containing 5% fetal bovine
serum (FBS) with or without 40% human plasma (Bioreclamation, Westbury, NY). The cells were
incubated for 3 days and were subsequently lysed and processed according to the manufacturer’s
instructions (Promega, Madison, WI) to measure the luciferase reporter activity using a Victor II
luminometer (PerkinꢀElmer, Waltham, MA). Resistanceꢀassociated substitutions in NS5A were each
introduced into the genotype 1aꢀH77 or 1bꢀCon1replicons using the ChangeꢀIT Multiple Mutation Site
Directed Mutagenesis Kit (Affymetrix, Santa Clara, CA). After the presence of the substitution was
confirmed by sequence analysis, the plasmid was linearized and the TranscriptAid T7 High Yield
Transcription Kit (Fermentas, Glen Burnie, MD) was used to transcribe the HCV subgenomic RNA from
the plasmid. In a transient assay, the replicon RNA containing the substitution was transfected via
electroporation into a Huhꢀ7 cell line. The cells were incubated for 4 days in the presence of the
compounds. The EC50 was calculated using nonlinear regression curve fitting to the 4ꢀparameter logistic
equation in the GraphꢀPad Prism 4/5 software. The protein binding effect on inhibitor potency was
assessed by including 40% human plasma in culture media containing 5% FBS during the three day
incubation period.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 47
Pharmacokinetic profiles. Unless otherwise noted, compounds were formulated in 2:5:20:73 (v/v)
DMSO:Tween 80:PEGꢀ400:D5W (5% dextrose in water) for single 3 mg/kg IV dosing and 40:20:20:20
(w/w) Phosol 53 MCT:PEGꢀ400:Ploxamer 124:Cremophor RH40 for single 3 mg/kg PO dosing in
mouse or rat.39 Heparinized plasma samples were withdrawn at 0.1 (IV only) 0.25, 0.5, 1, 2, 4, 6, 9, 12,
and 24 hours postꢀdosing. Plasma drug concentrations were determined by a liquid chromatographyꢀ
mass spectrometry (LCꢀMS) assay as follows. Compounds were separated from plasma samples using
protein precipitation with acetonitrile. A 250 µL aliquot of each sample or spiked standard was
combined with 25 µL of internal standard and 500 µL of acetonitrile. The samples were vortexed
vigorously for one minute followed by centrifugation for 10 minutes at 4 °C. Each supernatant was
transferred to a well of a 96ꢀwell plate and evaporated to dryness with a gentle stream of nitrogen. The
samples were reconstituted with sequential aliquots of acetonitrile and 20 mM ammonium acetate.
Compounds and the internal standard were separated from each other and coꢀprecipitated contaminants
on a 50 x 3 mm Clipeus 5µm column (Higgins Analytical, Inc.) with a 1:1 acetonitrile:20 mM
ammonium acetate mobile phase at a flow rate of 0.3 mL/min. Analysis was performed on a Sciex
API2000™ Biomolecular Mass Analyzer with a turboꢀionspray interface, in the negative ion mode.
Detection was in the multiple reaction monitoring (MRM) mode at m/z 516.2 ꢀ >437.0. Compound and
internal standard peak areas were determined using Sciex TurboQuan™ software. The drug
concentration of each sample was calculated by least squares linear regression analysis (nonꢀweighted)
of the peak area ratio (parent/internal standard) of the spiked standards versus concentration. The
method, generally evaluated over the concentration range 0ꢀ6.4 µg/mL, was linear (correlation
coefficient >0.999), with mean accuracy values from 96ꢀ107% of theoretical for the analysis of triplicate
standards at seven separate concentrations. The limit of quantitation was estimated to be ~20 ng/mL
from a 0.25 mL plasma sample.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
24
ACS Paragon Plus Environment

Page 25 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Chemistry. General Procedures. Reagents and solvents, including anhydrous solvents, were obtained
from commercial sources and used as supplied. Column chromatography was carried out on silica gel.
1H NMR spectra were measured using either a Brucker AMX 300 MHz or a Varian Inova 400 MHz
spectrometer. Chemical shifts are reported in ppm (δ) and referenced to an internal standard of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
tetramethylsilane (δ 0.00 ppm). H – 1H couplings are assumed to be firstꢀorder, and peak multiplicities
are reported in the usual manner. MS analysis was conducted using a Finnigan SSQ7000 (ESI) mass
spectrometer. All final compounds were purified to >95% purity as determined by reverse phase HPLC
performed on a Waters 2695 Separation Module / Waters 2489 UV/Visible Detector equipped with a
YMC ODSꢀA, 5.5 ꢁm, 120Å, 4.6 X 150 mm column using a solvent gradient of 10ꢀ100% acetonitrile in
water (0.1% TFA).
1-(4-Chloro-2-fluoro-5-nitrophenyl)ethanone (1b). To a solution of 4ꢀchloroꢀ2ꢀfluoroꢀ5ꢀ
nitrobenzoic acid (16.0 g, 72.9 mmol) in anhydrous dichloromethane (400 mL) was added oxalyl
chloride (9.57 mL, 109 mmol) and DMF (2 drops), and the resulting mixture was stirred at room
temperature until gas evolution ceased. The mixture was concentrated and dried under vacuum. In a
separate, heatꢀdried reaction flask, a mixture of ZnBr2 (24.6 g , 109 mmol) in anhydrous THF (300 mL)
o
was stirred at ꢀ78 C while a solution of CH3MgBr (29.1 mL, 3.0 M in Et2O, 87 mmol) was added
o
dropwise. The resulting mixture was stirred at ꢀ78 C for 15 min, and then allowed to warm to rt and
o
stirred for 30 min. The mixture was cooled to ꢀ78 C, and a solution of the previously prepared acid
chloride in anhydrous THF (100 mL) was added dropwise, followed by Pd(PPh3)4 (1.68 g, 1.46 mmol).
The resulting mixture was stirred at ꢀ78 oC for 10 min, and was allowed to warm to rt and stirred for 16
h. The reaction mixture was quenched by adding 1 N aq HCl, and was partitioned between water (100
mL) and dichloromethane (3×300 mL). The combined organic extract was dried over MgSO4, filtered
and concentrated under vacuum. The crude product was purified by column chromatography on silica
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 47
1
gel, eluting with 5% ethyl acetate in hexanes, to afford 4b (11.79 g, 74%): H NMR (400 MHz, DMSOꢀ
1
2
3
d6) δ 8.48 (d, J = 6.9 Hz, 1H), 8.04 (d, J = 10.4 Hz, 1H), 2.61 (d, J = 4.0 Hz, 3H).
4
5
6
7
8
2-Bromo-1-(4-chloro-2-fluoro-5-nitrophenyl)ethanone (2b). To a solution of 1b (11.0 g, 50.6
mmol) in THF (400 mL) was added pyridinium bromide perbromide (17.0 g, 53.1 mmol) portionꢀwise
over 10 min. The resulting orange solution was stirred at rt for 20 min, during which time the color
changed to light yellow and a precipitate formed. The precipitate was filtered off, and the solution was
concentrated under vacuum. The crude product was purified by column chromatography on silica gel
using a solvent gradient of 0ꢀ25% ethyl acetate in hexanes to give 2b as a light yellow solid (14.39 g,
96%): 1H NMR (400 MHz, DMSOꢀd6) δ 8.61 (d, J = 6.7 Hz, 1H), 8.10 (d, J = 10.4 Hz, 1H), 4.89 (d, J
= 2.1 Hz, 2H).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1,4-Bis(4-chloro-2-fluoro-5-nitrophenyl)butane-1,4-dione (3b). A mixture of ZnCl2 (11.0 g,
81.0 mmol), anhydrous benzene (80 mL), diethylamine (6.34 mL, 60.7 mmol), and tertꢀbutanol (5.81
mL, 60.7 mmol) was stirred at rt for 2 h to give a homogeneous mixture. Compounds 3b (12.0 g, 40.5
mmol) and 1b (13.2 g, 60.7 mmol) were added, and the mixture was stirred at rt under N2 for 5 days. A
solution of 5% aq H2SO4 (100 mL) was added, and the mixture was vigorously stirred for 40 min. The
resulting solid product was collected by filtration and washed with benzene, and then water. The
collected solid was suspended in 2:1 methanol:water, stirred for several minutes, and collected by
1
filtration and dried to give 3b as an offꢀwhite solid (13.47 g, 77%): H NMR (400 MHz, DMSOꢀd6) δ
8.53 (d, J = 6.8 Hz, 2H), 8.07 (d, J = 10.4 Hz, 2H), 3.40 (s, 4H).
(1S,4S)-1,4-Bis(4-chloro-2-fluoro-5-nitrophenyl)butane-1,4-diyl dimethanesulfonate (4b). A
solution of (R)ꢀ(+)ꢀα,αꢀdiphenylꢀ2ꢀpyrrolidinemethanol (0.876 g, 3.46 mmol) and trimethyl borate
(0.502 mL, 4.49 mmol) in anhydrous THF (80 mL) was stirred at rt for 1 h and then cooled to 0 oC. N,Nꢀ
o
Diethylaniline borane (6.15 mL, 34.6 mmol) was added, and the mixture was stirred at 0 C for 20 min
before a solution of 3b (3.74 g, 8.64 mmol) in anhydrous THF (10 mL) was added dropwise. The
resulting solution was allowed to slowly warm to rt and stirred overnight. The reaction was quenched
with 1 N aq HCl and extracted with ethyl acetate. The organic layer was dried over MgSO4, filtered and
26
ACS Paragon Plus Environment

Page 27 of 47
Journal of Medicinal Chemistry
concentrated under vacuum, and the crude product was purified by column chromatography on silica gel
using a solvent gradient of 0ꢀ50% ethyl acetate in hexanes to give the S,Sꢀdiol as a colorless solid (3.66
g, 97%). A portion of this solid (1.00 g, 2.29 mmol) in anhydrous dichloromethane (20 mL) was cooled
1
2
3
4
5
6
7
8
o
to 0 C, and triethylamine (0.956 mL, 6.86 mmol) was added, followed by methanesulfonyl chloride
(0.446 mL, 5.72 mmol). The resulting mixture was stirred at 0 oC for 90 min, and was then concentrated
under vacuum (avoid heating) to ~1/4 volume. Hexanes was added to give a solid that was collected by
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
filtration, washed with water and air dried to provide 4b as a colorless solid (1.34 g, 98%): H NMR
(400 MHz, DMSOꢀd6) δ 8.30 (d, J = 6.7 Hz, 2H), 7.93 (d, J = 9.8 Hz, 2H), 5.88 ꢀ 5.81 (m, 2H), 3.22 (s,
6H), 2.21 ꢀ 1.88 (m, 4H).
1-(2,6-Difluoro-4-nitrophenyl)piperidin-4-one
(11).
A
mixture
of
1,4ꢀdioxaꢀ8ꢀ
azaspiro[4.5]decane (8.24 mL, 64.3 mmol), 1,2,3ꢀtrifluoroꢀ5ꢀnitrobenzene (5.00 mL, 42.8 mmol), and
potassium carbonate (7.10 g, 51.4 mmol) in anhydrous DMSO (35 mL) was stirred for 4 h while heating
at 100 °C. The reaction mixture was cooled to rt and partitioned between ethyl acetate (200 mL) and
water (2 x 100 mL). The organic layer was dried over anhydrous Na2SO4, filtered, and concentrated
under vacuum to give a yellow solid (13 g). The solid was dissolved in a mixture of acetone (100 mL)
and water (20 mL), and concentrated aq HCl (18 mL) was added. The resulting solution was stirred for
16 h while heating at 50 oC. The solution was cooled to rt, and water was added to give a precipitate that
1
was collected and dried to give 11 (8.2 g, 74%): H NMR (400 MHz, CDCl3) δ 7.88 – 7.80 (m, 2H),
3.67 (t, J = 6.1 Hz, 4H), 2.63 (t, J = 6.1 Hz, 4H).
1-(2,6-Difluoro-4-nitrophenyl)-1,2,3,6-tetrahydropyridin-4-yl trifluoromethanesulfonate.
o
A solution of 11 (5.00 g, 19.5 mmol) in anhydrous THF (50 mL) was cooled to ꢀ78 C and a 1.0 M
solution of LiHMDS in THF (29.3 ml, 29.3 mmol) was added dropwise over 10 min. The resulting deep
red solution was stirred at ꢀ78 oC for
5
min and 1,1,1ꢀtrifluoroꢀNꢀphenylꢀNꢀ
(trifluoromethylsulfonyl)methanesulfonamide (7.67 g, 21.5 mmol) was added. The mixture was stirred
at ꢀ78 oC for 1 h and then allowed to warm to rt. TLC (3:1 hexanes:ethyl acetate) showed the reaction to
be complete. The mixture was diluted with ethyl acetate and washed with 1 N aq NaOH, water, and
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 47
dried over Na2SO4. The drying agent was filtered off, and the solution was concentrated under vacuum
to give a crude product that was purified by column chromatography on silica gel using a solvent
gradient of 0ꢀ40% ethyl acetate in hexanes. The title compound was obtained as a yellow oil that
crystallized under vacuum (6.12 g, 81%): 1H NMR (400 MHz, CDCl3) δ 7.88 – 7.75 (m, 2H), 7.45 –
7.28 (m, 1H), 4.08 – 4.01 (m, 2H), 3.62 – 3.53 (m, 2H), 2.67 – 2.58 (m, 2H).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1-(2,6-Difluoro-4-nitrophenyl)-4-(4-fluorophenyl)-1,2,3,6-tetrahydropyridine. A mixture of
1ꢀ(2,6ꢀdifluoroꢀ4ꢀnitrophenyl)ꢀ1,2,3,6ꢀtetrahydropyridinꢀ4ꢀyl trifluoromethanesulfonate (1.55 g, 3.99
mmol), 4ꢀfluorophenylboronic acid (0.614 g, 4.39 mmol), 2.0 M aq Na2CO3 (6.0 mL, 12.0 mmol), and
LiCl (0.507 g, 12.0 mmol) in 1,2ꢀdimethoxyethane (20 mL) was sparged with N2 for 30 min.
Tetrakis(triphenylphosphine)palladium(0) (0.23 g, 0.20 mmol) was added, and the mixture was sparged
with N2 for 10 min more. The reaction container was sealed, and the mixture was stirred for 24 h while
heating at 100 oC. The mixture was cooled to rt and partitioned between ethyl acetate and water. The aq
layer was washed with ethyl acetate, and the combined organic layers were dried over Na2SO4, filtered
and concentrated under vacuum. The crude product was purified by column chromatography on silica
gel using a solvent gradient of 0ꢀ20% ethyl acetate in hexanes. The title compound was obtained as an
1
orange solid (0.716 g, 54%): H NMR (400 MHz, CDCl3) δ 7.84 – 7.79 (m, 2H), 7.42 – 7.36 (m, 2H),
7.08 – 7.02 (m, 2H), 6.11 – 6.05 (m, 1H), 4.15 – 4.09 (m, 2H), 3.66 – 3.58 (m, 2H), 2.73 – 2.65 (m, 2H);
MS (ESI) m/z 335 (M+H)+.
3,5-Difluoro-4-(4-(4-fluorophenyl)piperidin-1-yl)aniline (12). To a solution of 1ꢀ(2,6ꢀ
difluoroꢀ4ꢀnitrophenyl)ꢀ4ꢀ(4ꢀfluorophenyl)ꢀ1,2,3,6ꢀtetrahydropyridine (0.716 g, 2.14 mmol) in THF (20
mL) was added 10% palladium on carbon (114 mg, 0.107 mmol). The reaction container was flushed
with N2, and the mixture was stirred under 1 atm H2 overnight. The mixture was filtered through Celite,
1
and the filtrate was concentrated under vacuum to give 12 as a colorless solid (0.651 g, 99%): H NMR
(400 MHz, DMSOꢀd6) δ 7.30 (dd, J = 8.7, 5.6 Hz, 2H), 7.10 (t, J = 8.9 Hz, 2H), 6.14 (d, J = 11.8 Hz,
2H), 5.42 (s, 2H), 3.13 ꢀ 2.94 (m, 4H), 2.65 ꢀ 2.54 (m, 1H), 1.79 ꢀ 1.65 (m, 4H); MS (ESI) m/z 307
(M+H)+.
28
ACS Paragon Plus Environment

Page 29 of 47
Journal of Medicinal Chemistry
1-(4-((2R,5R)-2,5-Bis(4-chloro-2-fluoro-5-nitrophenyl)pyrrolidin-1-yl)-2,6-difluorophenyl)-
1
2
3
4-(4-fluorophenyl)piperidine (13). To a solution of 4b (0.60 g, 1.0 mmol) and 12 (0.31 g, 1.0 mmol) in
4
5
anhydrous acetonitrile (5 mL) was added Hunig’s base (0.176 mL, 1.01 mmol), and the resulting
6
7
8
o
solution was stirred at 75 C for 36 h. The mixture was cooled to rt and was partitioned between water
9
and ethyl acetate (2x). The combined organic layers were dried over Na2SO4, filtered and concentrated
under vacuum. The crude product was purified by column chromatography on silica gel using a solvent
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
gradient of 0ꢀ20% ethyl acetate in hexanes to give 13 as a colorless solid (0.364 g, 51%): H NMR (400
MHz, DMSOꢀd6) δ 7.89 (d, J = 9.7 Hz, 2H), 7.76 (d, J = 7.1 Hz, 2H), 7.29 – 7.22 (m, 2H), 7.10 – 7.03
(m, 2H), 5.98 (d, J = 11.8 Hz, 2H), 5.55 (d, J = 6.8 Hz, 2H), 3.08 – 2.90 (m, 4H), 2.61 – 2.36 (m, 5H),
1.86 – 1.57 (m, 4H).
(2S,2'S)-Di-tert-butyl
2,2'-(((((2R,5R)-1-(3,5-difluoro-4-(4-(4-fluorophenyl)piperidin-1-
yl)phenyl)pyrrolidine-2,5-diyl)bis(5-fluoro-2-nitro-4,1-phenylene))bis(azanediyl))-
bis(carbonyl))bis(pyrrolidine-1-carboxylate) (14). A mixture of 13 (0.360 g, 0.509 mmol), (S)ꢀtertꢀ
butyl 2ꢀcarbamoylpyrrolidineꢀ1ꢀcarboxylate (0.273 g, 1.27 mmol), Cs2CO3 (0.497 g, 1.53 mmol), and
Xantphos (53 mg, 0.092 mmol) in 1,4ꢀdioxane (6 mL) was degassed by sparging with N2 for 20 min.
Tris(dibenzylidineacetone)dipalladium(0) (0.014 g, 0.015 mmol) was added, and the mixture was
sparged with N2 for 5 min. The reaction container was sealed, and the mixture was stirred for 90 min
o
while heating at 100 C. The mixture was cooled to rt and partitioned between water and ethyl acetate
(3x). The combined organic layers were dried over Na2SO4, filtered and concentrated under vacuum.
The crude product was purified by column chromatography on silica gel using a solvent gradient of 0ꢀ
40% ethyl acetate in heptanes to give 14 (0.29 g, 54%): 1H NMR (400 MHz, DMSOꢀd6) δ 10.63 –
10.38 (m, 2H), 8.02 – 7.58 (m, 4H), 7.29 – 7.20 (m, 2H), 7.12 – 7.00 (m, 2H), 6.05 (d, J = 12.2 Hz, 2H),
5.56 – 5.47 (m, 2H), 4.34 – 4.18 (m, 2H), 3.45 – 3.32 (m, 4H), 3.06 – 2.89 (m, 4H), 2.59 – 2.39 (m, 3H),
2.26 – 2.10 (m, 2H), 1.97 – 1.75 (m, 8H), 1.73 – 1.57 (m, 4H), 1.41 – 1.24 (m, 18H); MS (APCI) m/z
1064.2 (M+H)+.
29
ACS Paragon Plus Environment