Benzimidazole-containing HCV NS5A inhibitors: Effect of 4-substituted pyrrolidines in balancing genotype 1a and 1b potency

Article abstract of DOI:10.1016/j.bmcl.2014.12.045

The treatment of HCV with highly efficacious, well-tolerated, interferon-free regimens is a compelling clinical goal. Trials employing combinations of direct-acting antivirals that include NS5A inhibitors have shown significant promise in meeting this challenge. Herein, we describe our efforts to identify inhibitors of NS5A and report on the discovery of benzimidazole-containing analogs with subnanomolar potency against genotype 1a and 1b replicons. Our SAR exploration of 4-substituted pyrrolidines revealed that the subtle inclusion of a 4-methyl group could profoundly increase genotype 1a potency in multiple scaffold classes.

Full text of DOI:10.1016/j.bmcl.2014.12.045

Accepted Manuscript
Benzimidazole-containing HCV NS5A inhibitors: Effect of 4-substituted pyr-
rolidines in balancing genotype 1a and 1b potency
James A. Henderson, Darius Bilimoria, Monica Bubenik, Caroline Cadilhac,
Kevin M. Cottrell, Evelyne Dietrich, Francois Denis, Nigel Ewing, Guy
Falardeau, Simon Giroux, Ronald Grey Jr., Lucille L’Heureux, Bingcan Liu,
Nagraj Mani, Mark Morris, Olivier Nicolas, Oswy Z. Pereira, Carl Poisson, B.
Govinda Rao, T. Jagadeeswar Reddy, Subajini Selliah, Rebecca S. Shawgo,
Louis Vaillancourt, Jian Wang, Constantin G. Yannopoulos, Nathalie Chauret,
Francoise Berlioz-Seux, Laval C. Chan, Sanjoy K. Das, Anne-Laure Grillot,
Youssef L. Bennani, John P. Maxwell
PII:
S0960-894X(14)01353-5
DOI:
http://dx.doi.org/10.1016/j.bmcl.2014.12.045
BMCL 22290
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
24 October 2014
10 December 2014
12 December 2014
Please cite this article as: Henderson, J.A., Bilimoria, D., Bubenik, M., Cadilhac, C., Cottrell, K.M., Dietrich, E.,
Denis, F., Ewing, N., Falardeau, G., Giroux, S., Grey, R. Jr., L’Heureux, L., Liu, B., Mani, N., Morris, M., Nicolas,
O., Pereira, O.Z., Poisson, C., Govinda Rao, B., Jagadeeswar Reddy, T., Selliah, S., Shawgo, R.S., Vaillancourt,
L., Wang, J., Yannopoulos, C.G., Chauret, N., Berlioz-Seux, F., Chan, L.C., Das, S.K., Grillot, A-L., Bennani, Y.L.,
Maxwell, J.P., Benzimidazole-containing HCV NS5A inhibitors: Effect of 4-substituted pyrrolidines in balancing
genotype 1a and 1b potency, Bioorganic & Medicinal Chemistry Letters (2014), doi: http://dx.doi.org/10.1016/
j.bmcl.2014.12.045
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Benzimidazole-containing HCV NS5A inhibitors: Effect of 4-
substituted pyrrolidines in balancing genotype 1a and 1b
potency
James A. Henderson*,a, Darius Bilimoriab, Monica Bubenikb, Caroline Cadilhacb, Kevin M.
Cottrella, Evelyne Dietrichb, Francois Denisb, Nigel Ewinga, Guy Falardeaub, Simon Girouxa,
Ronald Grey, Jr.a, Lucille L’Heureuxb, Bingcan Liub, Nagraj Mania, Mark Morrisa, Olivier
Nicolasb, Oswy Z. Pereirab, Carl Poissonb, B. Govinda Raoa, T. Jagadeeswar Reddyb, Subajini
Selliahb, Rebecca S. Shawgoa, Louis Vaillancourtb, Jian Wanga, Constantin G. Yannopoulosb,
Nathalie Chauretb, Francoise Berlioz-Seuxa, Laval C. Chanb, Sanjoy K. Dasb,+, Anne-Laure
Grillota, Youssef L. Bennania, and John P. Maxwella
a Vertex Pharmaceuticals Incorporated, 50 Northern Avenue, Boston, Massachusetts 02210, USA
b Vertex Pharmaceuticals Canada Incorporated, 275 Boul. Armand-Frappier, Laval, QC, H7V 4A7, Canada
+ Present address: AngioChem, 201 President-Kennedy Ave., Suite PK-2880, Montreal, QC H2X 3Y7, Canada
Abstract
The treatment of HCV with highly efficacious, well-tolerated, interferon-free
regimens is a compelling clinical goal. Trials employing combinations of direct-acting
antivirals that include NS5A inhibitors have shown significant promise in meeting
this challenge. Herein, we describe our efforts to identify inhibitors of NS5A and
report on the discovery of benzimidazole-containing analogs with subnanomolar
potency against genotype 1a and 1b replicons. Our SAR exploration of 4-substituted
pyrrolidines revealed that the subtle inclusion of a 4-methyl group could profoundly
increase genotype 1a potency in multiple scaffold classes.
Letter
Hepatitis C virus (HCV) affects an estimated 130-170 million individuals
worldwide and is the leading cause of liver transplant and hepatocellular carcinoma.1
Once only treatable with poorly tolerated and moderately efficacious regimens of
pegylated interferon alpha (PEG-IFN-α) and ribavirin (RBV), chronic HCV infection
is now being cured with high rates in clinical trials using IFN-free combinations of
direct-acting antivirals (DAAs). The search for DAAs interfering with HCV
replication has included targeting of the non-structural (NS) viral proteins NS2,
NS3ꢀ4A, NS4B, NS5B and NS5A.2-4
*
Corresponding author. Tel.: +1 617 961 7605
E-mail address: james_henderson@vrtx.com (J. A. Henderson)

In clinical monotherapy studies, NS5A inhibitors have proven to produce the
most rapid viral load declines of any HCV antiviral class.5 NS5A is a large
phosphoprotein (49 kDa) with no apparent enzymatic activity or homologues in
prokaryotes and eukaryotes. Although the mechanistic role of NS5A in the HCV
replication cycle currently is unknown, it is thought to play important roles in viral
replication, virion assembly and secretion from infected cells.6 The elucidation of
NS5A’s mechanistic role remains an active area of investigation.7 Several NS5A
inhibitors are currently in various stages of development,8 including daclatasvir
(BMS-790052, Figure 1), which achieved the first clinical proof of concept in this
inhibitor class by displaying an impressive 3log10 reduction of HCV RNA in 24 h
with a single 10 mg oral dose (genotypes 1a and 1b).9
Figure 1. Structure of daclatasvir (BMS-790052)
Inspired by literature leads and internal screening, our early medicinal
chemistry efforts to identify NS5A inhibitors with balanced genotype 1a (G1a) and 1b
(G1b) potency focused on benzimidazole-containing scaffolds (Table 1).10,11 Directly
linked bis-benzimidazole 1 showed low nanomolar G1b potency and no activity
against G1a, whereas mono-alkyne inhibitor 2 exhibited a potency increase in both
genotypes, suggesting a minimum distance requirement between the two
benzimidazoles. Reduction of the alkyne linker to flexible ethyl analog 3 resulted in a
3-fold potency reduction in G1b and a 25-fold loss for G1a, suggesting the need for a
rigid linker. Increasing the linker length by introduction of a bis-alkyne (4) had no
affect on G1a and G1b replicon activity. Replacement of the alkynes with a more
lipophilic phenyl ring 5 improved G1a potency 30-fold (EC50 = 550 pM). The 1,3-
linked analog 6 was inactive but the 2,5-substituted thiophene analog 7 was
equipotent to 5, indicating that up to ~30o deviations from linearity are tolerated in
these molecules. Mono substitution appeared to be tolerated as observed with analogs
8-10 and electronic factors did not play a significant role as methyl, fluoro and

methoxy exhibited similar potencies to unsubstituted phenyl 5, and dimethylphenyl
linkers 11 and 12 also had similar potency to phenyl analog 5.
Table 1
Activity of bis-benzimidazole analogs
G1a
EC50
(nM)
G1b
EC50
(nM)
CC50
Compd
X
(µM)
None
>2000
26
>20
1
28
0.022
0.076
0.024
>20
>20
13
2
3
4
688
18
5
6
7
8
0.55
378
0.012
20
6
6
6
3
0.94
0.84
0.01
0.006
S
Me
OMe
F
0.36
0.84
0.013
0.012
6
6
9
10
Me
0.9
2.9
0.006
0.01
5
7
11
12
Me
Me
Me
Our goal of identifying a NS5A inhibitor with balanced G1a and 1b potency
necessitated further optimization of the benzimidazole-phenyl-benzimidazole scaffold

(e.g., compound 5) and ultimately led to the investigation of substituted pyrrolidines.
Activity profiles from our exploration of 4-substituted pyrrolidines (13-23) are
compared to unsubstituted pyrrolidine 5 in Table 2. Modifications at the 4-position
for compounds 13-23 had modest effects on G1b potency, yet a profound effect was
observed with respect to G1a potency. While mono-fluorinated pyrrolidines 13 and
14 were less potent than 5, geminal difluoride 15 was equipotent to 5 in G1a activity.
A marked drop in activity for both genotypes was observed with 4R-hydroxyl 16,
presumably due to intolerance of polar functionality at this position given that activity
returned with the corresponding methyl ether 17. The 4S-methyl ether 18
demonstrated an unanticipated 10-fold increase in G1a activity and further exploration
revealed 4-methyl pyrrolidines 19 and 20 were 20-fold more potent (G1a EC50 = 26
pM and 28 pM, respectively) than 5, providing analogs with exceptional G1a/G1b
Table 2
Activity of analogs containing 4-substituted pyrrolidines
G1a
EC50
(nM)
G1b
EC50
(nM)
CC50
Compd
R
(µM)
H
0.546
4.9
0.012
0.082
0.018
0.004
0.1
6
5
(S)-F
11
13
14
15
16
17
18
19
20
21
22
23
(R)-F
1.1
14
gem-di-F
(R)-OH
(R)-OMe
(S)-OMe
(R)-Me
(S)-Me
(S)-CF3
(S)-Et
0.5
13
15
>20
6
1.22
0.053
0.028
0.026
0.030
0.085
0.03
0.034
0.031
0.007
0.037
0.005
0.019
0.011
>20
14
>20
13
>20
>20
gem-di-Me

potency ratios. The 4S-trifluoromethyl analog 21 displayed similar G1a activity to 20,
while 4S-ethyl 22 showed a 3-fold drop in activity. Substitution of the 4-position of
the pyrrolidines with geminal dimethyl (23) also resulted in potent G1a and 1b
activity.
Gratifyingly, the incorporation of 4-methyl substituted pyrrolidines into
additional scaffolds was also effective in boosting G1a potency (Figure 2).
A
dramatic 1000-fold increase in G1a potency is observed when diyne-linked bis-
benzimidazole 4 (G1a EC50 = 18 nM) is compared to 4S-methyl pyrrolidine-
containing analog 24 (G1a EC50 = 18 pM). In order to investigate the methyl effect12
in an analog where scaffold symmetry had been broken, we replaced one of the
benzimidazoles of 5 with a phenyl-imidazole unit to provide compound 25.
Incorporation of 4S-methyl pyrrolidines into 25 (G1a EC50 = 139 pM) resulted in a
35-fold increase in G1a activity for 26 (G1a EC50 = 4 pM), providing a single digit
picomolar inhibitor of NS5A G1a and 1b (CC50 = >20 µM).13,14
Figure 2: Activity of 4(S)-Me pyrrolidine-containing analogs
With excellent G1a and 1b potencies in hand, the ADME properties of 26
were investigated. Metabolic stability in both rat and human hepatocytes was high,
with > 90% remaining after incubation at 5 µM for 60 minutes. The cellular
permeability as measured in the Caco-2 cell line was good, and evaluation of PGP
efflux potential in the human Mdr1 overexpressing cell line MDCK-Mdr1 indicated
26 was not a PGP efflux substrate (ER = 1.8). Compound 26 was moderately bound
to plasma proteins in vitro (98.6% human, 91.7% rat), displayed no hERG channel
inhibition (IC50 > 40 µM), and did not inhibit a panel of CYP enzymes (IC50 > 30 µM
across tested isoforms). The in vivo pharmacokinetic profile of 26 in rats indicated

low clearance (Cl = 9.0 mL/min/kg) and moderate bioavailability (F = 26%) after
dosing as a 3 mg/kg solution in 30% aqueous PEG-400. Key pharmacokinetic data
for compound 26 are included in Table 3.
Table 3
Selected in vitro and in vivo DMPK data for compound 26.
Selected In vitro Data
Selected In vivo Data
IV Rat PK
CYP Isoform Inhibition
IC50 (µM)
> 30
3A4, 2C9, 2D6, 2C19, 1A2
hERG Channel
Planar-, QPatch
Cl (mL/min/kg)
T1/2(h)
Vss (L/kg)
9.0
2.0
1.4
IC50 (µM)
> 30
Plasma Protein Binding
Human
Rat
% Bound
98.6
91.7
PO Rat PK
2.08
3.5
AUC0-8h(µg h/mL)
T1/2 (h)
Permeability, Efflux
Caco-2, A-B (10-6cm/sec)
Mdr1, Efflux Ratio
4.1
1.8
Cmax (ng/mL)
% F
0.03
26
The benzimidazole-containing NS5A inhibitors described in this letter were
prepared by the methods outlined in Schemes 1-3. The syntheses of iodo-
benzimidazole building blocks, represented by 29a, were accomplished in a
straightforward manner by coupling of 4-iodo-benzene-1,2-diamine with N-Boc-L-
prolines, followed by amidation of the deprotected pyrrolidine with (S)-2-
((methoxycarbonyl)-amino)-3-methylbutanoic acid under standard conditions
(Scheme 1). Symmetrical bis-benzimidazole analogs 1-5 were prepared from iodo-
benzimidazole building block 29a (or the corresponding pinacolboronate derivative)
by a series of Sonagashira and Suzuki dimerization couplings (Scheme 2). The
preparation of nonsymmetrical benzimidazole/imidazole-containing analogs 25 and
26 was ultimately accomplished by the union of boronates 31a-b with iodo-
benzimidazoles 29a-b under Suzuki conditions (Scheme 3).
In summary, we have described the identification of symmetrical and
nonsymmetrical benzimidazole-containing NS5A inhibitors with subnanomolar
potency against both genotype 1a and 1b replicons. In the bis-benzimidazole series,
several linkers designed to explore aspects of varied geometry, length and polarity
were prepared, with most exhibiting excellent potency against the genotype 1b HCV
replicon. A major breakthrough in genotype 1a potency was realized when the
incorporation of 4-substituted pyrrolidines into these scaffolds was explored.

Specifically, it was demonstrated that incorporation of 4-methyl proline into both
symmetrical and nonsymmetrical scaffolds has a profound effect on increasing G1a
potency, delivering analogs with balanced G1a and 1b potency and good rodent PK
profiles.
Scheme 1: Synthesis of building blocks 29a-b. Reagents and conditions: (a) 4-iodo-benzene-1,2-diamine,
HATU, DIPEA, THF, 0 °C; (b) TFA, DCM, rt; (c) (S)-2-((methoxycarbonyl)-amino)-3-methylbutanoic acid,
HATU, DIPEA, DMF, rt.
Scheme 2: Synthesis of analogs 1-5. Reagents and conditions: (a) bis(pinacolato)diboron, PdCl2dppf, KOAc,
DMF, 85 °C; (b) 29a (1.0 eq), Pd(PPh3)4, Na2CO3 (aq), iPrOH, 80 °C; (c) ethynyltrimethylsilane,
PdCl2(PPh3)2, CuI, DBU, H2O, CH3CN, 60 °C; (d) 10 wt % Pd/C, H2, EtOH (e) ethynyltrimethylsilane,
PdCl2(PPh3)2, CuI, Et3N, MeCN, rt; (f) K2CO3, MeOH, rt; (g) PdCl2(PPh3)2, CuI, Et3N, THF, rt; (h) 1,4-
benzenediboronic acid (0.5 eq), PdCl2dppf (10 mol%), Na2CO3, MeCN, H2O, 85 °C.

Scheme 3: Synthesis of analogs 25 and 26. Reagents and conditions: (a) 2-bromo-1-[4-(4-
bromophenyl)phenyl]ethanone, DIPEA, MeCN, rt; (b) ammonium acetate, PhMe, 100 °C; (c) TFA, DCM,
rt; (d) (S)-2-((methoxycarbonyl)-amino)-3-methylbutanoic acid, HATU, DIPEA, DMF, rt; (e)
bis(pinacolato)diboron, PdCl2dppf, KOAc, DMF, 85 °C; (f) 29a or 29b (1.0 eq), PdCl2dppf (10 mol%),
Na2CO3, MeCN, H2O, 85 °C.
Acknowledgments
The authors would like to thank Mark N. Namchuk, Ann D. Kwong, and Jean Bédard
for insightful discussions.
References and notes
1. For recent reviews, see: (a) Thomas, D. L. Nat. Med. 2013, 19, 850-858, (b)
Scheel, T. K. H.; Rice, C. M. Nat. Med. 2013, 19, 837-849.
2. Liang, T. J.; Ghany, M. G. N. Engl. J. Med. 2013, 368, 1907-1917.
3. Dabbouseh, N. M.; Jensen, D. M. Nat. Rev. Gastroenterol. Hepatol. 2013, 10,
268-276.
4. Aghemo, A.; De Francesco, R. Hepatology 2013, 58, 428-438.
5. Belema, M.; Lopez, O. D.; Bender, J. A.; Romine, J. L.; St. Laurent, D. R.;
Langley, D. R.; Lemm, J. A.; O’Boyle, D. R., II; Sun, J.-H.; Wang, C.; Fridell,
R. A.; Meanwell, N. A. J. Med. Chem. 2014, 57, 1643-1672.
6. Guedj, J.; Dahari, H.; Rong, L.; Sansone, N. D.; Nettles, R. E.; Cotler, S. J.;
Layden, T. J.; Uprichard, S. L.; Perelson, A. S. Proc. Natl. Acad. Sci. U.S.A.
2013, 110, 3991-3996.
7. (a) Reghellin, V.; Donnici, L.; Fenu, S.; Berno, V.; Calabrese, V.; Pagani, M.;
Abrignani, S.; Peri, F.; De Francesco, R.; Neddermann, P. Antimicrob. Agents
Chemother. 2014, doi:10.1128/AAC.03293-14; (b) Mani, N.; Yuzhakov, A.;

Yuzhakov, O.; Coll, J. T.; Black, J.; Saxena, K.; Fulghum, J. R.; Lippke, J. A.;
Govinda Rao, B.; Rijnbrand, R.; Kwong, A. D. J. Virol. 2014, doi:
10.1128/JVI.01677-14. (c) for an editorial review, see: Elazar, M.; Glenn, J. S.
Gastroenterology 2014, 147, 273-277.
8. On October 10th, 2014, the FDA approved Gilead Sciences’ Harvoni®
(sofosbuvir and ledipasvir) for the treatment of chronic HCV genotype 1
infection.
http://www.fda.gov/newsevents/newsroom/pressannouncements/ucm418365.h
tm
9. (a) Belema, M.; Meanwell, N. A. J. Med. Chem. 2014, 57, 5057-5071. (b)
Gao, M.; Nettles, R. E.; Belema, M.; Snyder, L. B.; Nguyen, V. N.; Fridell, R.
A.; Serrano-Wu, M. H.; Langley, D. R.; Sun, J.-H; O’Boyle, D.R., II; Lemm,
J. A.; Wang, C.; Knipe, J. O.; Chien, C.; Colonno, R. J.; Grasela, D. M.;
Meanwell, N. A.; Hamman, L.G. Nature 2010, 465, 96-100.
10. Henderson, J. A.; Maxwell, J.; Vaillancourt, L.; Morris, M.; Grey, R., Jr.;
Giroux, S.; Kong, L. C. C.; Das, S. K.; Liu, B.; Poisson, C.; Cadilhac, C.;
Bubenik, M.; Reddy, T. J.; Falardeau, G.; Yannopoulos, C.; Wang, J.; Pereira,
O. Z.; Bennani, Y. L.; Pierce, A. C.; Bhisetti, G. R.; Cottrell, K. M.; Morone,
V. Patent US 8,765,731 B2.
11. For details of the NS5A replicon assays used in this letter, see: Giroux, S.; Xu,
J.; Reddy, T. J.; Morris, M.; Cottrell, K. M.; Cadilhac, C.; Henderson, J. A.;
Nicolas, O.; Bilimora, D.; Denis, F.; Mani, N.; Ewing, N.; Shawgo, R.;
L’Heureux, L.; Selliah, S.; Chan, L.; Chauret, N.; Berlioz-Seux, F.; Namchuk,
M. N.; Grillot, A-L.; Bennani, Y. L.; Das, S. K.; Maxwell, J. P. ACS Med.
Chem Lett. 2014, 5, 240-243.
12. Schönherr, H.; Cernak, T. Angew. Chem. Int. Ed. 2013, 52, 2-14.
13. Compound 26 was screened against recombinant replicon cell lines harboring
these mutations (fold shift from wt activity): 1a Y93H, 8000-fold; 1a Q30R,
42-fold; 1a M28T, 55-fold; and 1b Y93H, 100-fold.
14. CC50 in Huh-7 ET cells were determined using a [3H]-thymidine incorporation
assay from 12 concentrations of the test compound in duplicate, and data were
considered accurate when the CC50 of the positive control was within 40% of
the mean value.

Graphical Abstract