Development of a New Structural Class of Broadly Acting HCV Non-Nucleoside Inhibitors Leading to the Discovery of MK-8876

Article abstract of DOI:10.1002/cmdc.201700228

Studies directed at developing a broadly acting non-nucleoside inhibitor of HCV NS5B led to the discovery of a novel structural class of 5-aryl benzofurans that simultaneously interact with both the palm I and palm II binding regions. An initial candidate was potent in vitro against HCV GT1a and GT1b replicons, and induced multi-log reductions in HCV viral load when orally dosed to chronic GT1 infected chimpanzees. However, in vitro potency losses against clinically relevant GT1a variants prompted a further effort to develop compounds with sustained potency across a broader array of HCV genotypes and mutants. Ultimately, a biology and medicinal chemistry collaboration led to the discovery of the development candidate MK-8876. MK-8876 demonstrated a pan-genotypic potency profile and maintained potency against clinically relevant mutants. It demonstrated moderate bioavailability in rats and dogs, but showed low plasma clearance characteristics consistent with once-daily dosing. Herein we describe the efforts which led to the discovery of MK-8876, which advanced into Phase 1 monotherapy studies for evaluation and characterization as a component of an all-oral direct-acting drug regimen for the treatment of chronic HCV infection.

Full text of DOI:10.1002/cmdc.201700228

DOI: 10.1002/cmdc.201700228
Full Papers
Development of a New Structural Class of Broadly Acting
HCV Non-Nucleoside Inhibitors Leading to the Discovery
of MK-8876
Casey C. McComas,*[a] Anandan Palani,[a] Wei Chang,[b] M. Katharine Holloway,[a]
Charles A. Lesburg,[a] Peng Li,[b] Nigel Liverton,[a] Peter T. Meinke,[a] David B. Olsen,[a]
Xuanjia Peng,[b] Richard M. Soll,[b] Ajay Ummat,[a] Jie Wu,[b] Jin Wu,[a] Nicolas Zorn,[a] and
Steven W. Ludmerer[a]
Studies directed at developing a broadly acting non-nucleoside
inhibitor of HCV NS5B led to the discovery of a novel structural
class of 5-aryl benzofurans that simultaneously interact with
both the palm I and palm II binding regions. An initial candi-
date was potent in vitro against HCV GT1a and GT1b replicons,
and induced multi-log reductions in HCV viral load when orally
dosed to chronic GT1 infected chimpanzees. However, in vitro
potency losses against clinically relevant GT1a variants prompt-
ed a further effort to develop compounds with sustained po-
tency across a broader array of HCV genotypes and mutants.
Ultimately, a biology and medicinal chemistry collaboration led
to the discovery of the development candidate MK-8876. MK-
8876 demonstrated a pan-genotypic potency profile and main-
tained potency against clinically relevant mutants. It demon-
strated moderate bioavailability in rats and dogs, but showed
low plasma clearance characteristics consistent with once-daily
dosing. Herein we describe the efforts which led to the discov-
ery of MK-8876, which advanced into Phase 1 monotherapy
studies for evaluation and characterization as a component of
an all-oral direct-acting drug regimen for the treatment of
chronic HCV infection.
Introduction
The global prevalence of hepatitis C (HCV) infection is estimat-
ed to be nearly 3% of the world’s population.[1] The morbidity
associated with chronic HCV infection includes cirrhosis, fibro-
sis, and liver failure. In addition, chronic HCV infection is a lead-
ing cause of liver transplantation.[2] Recently, a number of ther-
apeutic regimens consisting entirely of orally bioavailable,
direct-acting antiviral agents have demonstrated high response
rates and have been approved for treatment.[3–5]
ture chain termination.[8] In addition to nucleoside analogue in-
hibitors, multiple classes of non-nucleoside inhibitors (NNIs)
have been identified which interact with one of multiple bind-
ing sites mapped to various pockets near the enzyme sur-
face.[9] It has been proposed that some of these inhibitors
modulate allosteric interactions of the enzyme, resulting in
losses of replication capacity.[10] Several of these inhibitors have
been clinically validated.[9] These NNI binding pockets typically
are less genetically constrained than the active site pocket. A
consequence is that breadth of potency across genotypes has
been difficult to achieve with these types of inhibitors, many
displaying a fragile profile with respect to the emergence of
viral resistance. To date only dasabuvir[3] (in combination the
protease inhibitor paritaprevir (boosted with ritonavir), the
NS5A inhibitor ombitasvir, and ribivarin), and beclabuvir[11] (in
combination with the protease inhibitor asunaprevir and the
NS5A inhibitor daclatasvir) are non-nucleoside inhibitors ap-
proved as a component of a multi-drug regimen, both indicat-
ed for the treatment of GT1-infected patients only.
The hepatitis C virus is a positive-strand RNA virus which has
a high degree of genetic heterogeneity as manifested in the
classification of seven major genotypes and multiple subtypes,
with considerable additional genetic variability within each
class.[6,7] The gene product of the virally encoded NS5B is an
RNA-dependent RNA polymerase. Polymerase activity can be
enzymatically disrupted by nucleoside analogues through in-
corporation into the nascent RNA genome, resulting in prema-
[a] Dr. C. C. McComas, Dr. A. Palani, Dr. M. K. Holloway, Dr. C. A. Lesburg,
Dr. N. Liverton, Dr. P. T. Meinke, Dr. D. B. Olsen, Dr. A. Ummat, Dr. J. Wu,
Dr. N. Zorn, Dr. S. W. Ludmerer
The topology of the NS5B protein is generally likened in
shape to that of a right hand with the active site located in
the region at the top of the palm. There are four well studied
allosteric binding sites conventionally referred to as the thumb
I, thumb II, palm I, and palm II sites.[9] The palm II binding site
possesses high sequence conservation across genotypes,
making it an attractive target for development of a drug that
is effective against all HCV genotypes. The initial clinical valida-
Merck & Co. Inc., Kenilworth NJ (USA)
and
Present address: International Discovery Service Unit, WuXi AppTec, Inc.,
1690 Sumneytown Pike, Suite 150, Lansdale, PA 19446 (USA)
E-mail: casey.mccomas@wuxiapptec.com
[b] Dr. W. Chang, Dr. P. Li, Dr. X. Peng, Dr. R. M. Soll, Dr. J. Wu
WuXi AppTec, Shanghai (China)
The ORCID identification number(s) for the author(s) of this article can
be found under https://doi.org/10.1002/cmdc.201700228.
ChemMedChem 2017, 12, 1 – 14
1
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
tion for NS5B as a drug target for non-nucleoside inhibitors
was achieved with HCV-796, a palm II binder, which progressed
to Phase 2 studies before being halted due to safety liabili-
ties.[12] Based on these potency data we set out to develop
compounds that would inhibit HCV NS5B by binding the
palm II site while retaining potent activity against all of the
clinically relevant major genotypes. Similar strategies conduct-
ed in parallel to this work have recently been described.[13,14]
Table 1. Anti-HCV replicon activity of sulfonamide derivatives of com-
pound 1.
Compd
R
EC50 [nm][a]
GT1b
GT2a
31
Results and Discussion
HCV-796
8
While numerous researchers have investigated derivatives of
HCV-796 as potential inhibitors of NS5B, we recognized that
the 5-position of the benzofuran core represented an area that
had remained largely unexplored.[15] Despite the introduction
of alkyl and alkoxy derivatives at this position, no examples of
analogues bearing aryl groups at this site were reported. To
determine if 5-aryl derivatives would impart a viable path
toward developing superior inhibitors, the parent phenyl ana-
logue was synthesized (Scheme 1). 5-Phenyl analogue
1 showed very promising activity with an EC50 of 185 nm
against the HCV GT1b replicon. This result led us to undertake
a program to fully explore the new 5-aryl benzofurans as
a chemical scaffold.
1
62
120
29
185
1a
1b
ND
98
1c
12
9
62
37
1d
1e
1 f
125
145
338
602
Commencing from 1 as a lead, we first probed the effects of
varying the pendant alkyl group attached to the sulfonamide.
Previous work in our group had established that the ethanolic
group posed a significant development liability because one of
the primary metabolites of HCV-796 was its oxidation to the al-
dehyde. More specifically, the prevalence of this reactive alde-
hyde metabolite posed a potential toxicity risk so an alterna-
tive group with a superior profile was attractive. An array of al-
ternative R-groups was examined (Table 1) and it was evident
from early SAR studies that diverse motifs were tolerated at
this position. While a number of analogues with extended side
chains, aryl groups, or heterocycles showed good potency
such as 1d and 1c, these analogues generally suffered from
poor metabolic stability. However, a simple methyl analogue
(1a) displayed promising potency and attractive overall stabili-
ty, and was selected as the standard for further optimization.
Holding the N-methyl methanesulfonamide constant, the 5-
aryl position of the benzofuran was extensively explored. How-
ever, to fully optimize the 5-position, it was necessary to devel-
op a versatile synthesis of a fully functionalized intermediate
that would allow ready access to a diverse array of aryl ana-
logues. The optimized synthesis is shown in Scheme 2.[16] Start-
ing with 2-hydroxyphenyl acetic acid, selective bromination at
the 4-position provided intermediate 3 which was then treated
1g
1h
50
>250
>1000
>1000
1i
118
250
[a] Data are the mean of nꢀ2 determinations unless otherwise noted;
ND=not determined.
with TBSCl to give silyl ether 4. Acylation with 4-fluorophenyl
acetyl chloride gave intermediate 5 followed by deprotection
of the phenol and cyclization under acidic conditions which
generated the benzofuran 7. Nitration of the benzofuran gen-
erated intermediate 8. This was followed by reduction to pro-
vide aniline 9, which was then converted into the
methyl sulfonamide 10. Subsequent hydrolysis of the
ester and condensation with methyl amine provided
amide 12 with all key functionalities in place. Alkyla-
tion of 12 allowed ready access to key intermediate
benzofuran core 13. The key core 13 was an excel-
lent substrate for Suzuki coupling and was used for
the facile exploration of a range of 5-aryl substitu-
ents. Alternatively, 13 could be readily converted into
Scheme 1. Novel 5-aryl-substituted benzofurans.
&
ChemMedChem 2017, 12, 1 – 14
www.chemmedchem.org
2
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Scheme 2. Synthesis of key intermediates 13 and 14.
Table 2. Anti-HCV replicon activity of 5-aryl-substituted analogues of 1a.
Compd
Ar
EC50 [nm][a]
Compd
Ar
EC50 [nm][a]
GT1b
GT2a
ND
GT1b
GT2a
129
1a
120
14h
14i
22
14a
25
32
69
923
1790
60
156
6
139
96
14b
14c
14d
14e
14 f
14g
14j
109
16
14k
6
72
14l[b]
14m[b]
14n[b]
14o[b]
25
17
83
76
111
127
169
270
24
48
133
22
348
48
[a] Data are the mean of nꢀ2 determinations unless otherwise noted; ND=not determined. [b] n=1.
ChemMedChem 2017, 12, 1 – 14
www.chemmedchem.org
3
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
boronate 14 to generate more complex 5-aryl derivatives.
With the key bromobenzofuran intermediate 13 now readily
available, a range of 5-aryl derivatives was generated to further
explore the SAR at this position (Table 2). A number of these
analogues showed promising activity and identified clear SAR
trends. In general, ortho substitutions such as 14 f and 14i de-
creased potency relative to corresponding meta and para ana-
logues. Meta-substituted analogues 14b and 14h demonstrat-
ed good potency against GT1b but were generally less inhibi-
tory against GT2a. Similarly 3,4-disubstituted analogues, 14j
and 14k in particular, were potent against GT1b but were sig-
nificantly less potent against GT2a. Para-substituted analogues
incorporating smaller groups (e.g., 14a, 14d, 14g) showed
promising potency against both GT1b and GT2a, whereas ana-
logues with larger groups at this position could show a loss of
activity (e.g., 14n and 14o). In addition to good potency, 14D
showed promising pharmacokinetics and stability and was se-
lected to undergo further studies as a lead compound. In Spra-
gue–Dawley (S–D) rats at 10 mgkgꢁ1, a single dose had 45%
bioavailability and an AUC0–1 of 25.9 mmꢁh. In monitoring
early clinical trials of other HCV non-nucleoside inhibitors NS5B
in development at the time, however, fragility to viral resist-
ance in vivo appeared to be a significant liability.[17–19] Thus it
became clear that to develop a viable clinical candidate, dra-
matic enhancements in potency were required to ensure po-
tency across the diverse genetic variability observed in chronic
HCV patients including potential resistant variants.
Figure 1. Overlay of HCV-796 and palm I thiadiazine inhibitor. HCV-796 locat-
ed in the palm II binding site interacts with key serine 365 via the methyl
amide at position 3 of the phenyl benzofuran core (lavender). Its partial
overlap with the palm I thiadiazine compound (purple) is evident above ar-
ginine 200.
palm I thiadiazine 16,[21] these binding pockets appeared con-
tiguous (Figure 1) such that a suitably constructed molecule
might interact with both pockets. This structure provided evi-
dence that hybrid molecules may simultaneously interact with
both pockets by addition of pendant groups attached to the
C5 aryl ring.
All previous efforts to improve potency via modification of
either the benzofuran C2 aryl group or the C3 methyl amide
group were met with failure. Even simple alterations of the C3
amide, such as methylamide to ethyl amide, led to total loss of
potency. The C2 aryl group tolerated modest changes but the
4-fluorophenyl analogues were consistently superior.
The initial approach toward accessing the palm I binding
pocket used mono-substituted aryl groups to generate simple
SAR and identify the best site and linker for introducing pend-
ant groups. As shown earlier with compound 14o, larger
pendant groups could be introduced at the 4-postion while
still maintaining some activity. Initially the potency was inferior
to compounds with smaller substituents, but as larger groups
as in 17a were added the potency improved (Scheme 3).
This result provided the first evidence that an appropriately
sized group could adopt a suitable binding orientation with
the palm I region. Analogous to prior trends, introduction of
larger pendant groups at the 2-position resulted in dramatic
potency losses, and these analogues were abandoned. Finally,
placing the pendant group at the 3-position of the aryl (17b)
resulted in a significant boost in inherent activity and provided
the most potent analogues generated to date. As additional
analogues were synthesized and tested, it became clear that
this was the favored position in all cases, with the results for
an array of representative analogues shown in Table 3.
To determine the optimal direction required to improve the
profile of new compounds, we re-examined the crystal struc-
ture of our original lead HCV-796[20], noting that the palm II
binding site is located adjacent to the palm I binding site, and
postulated that these regions could be contiguous. When
modeling overlapping crystal structures of HCV-796 and the
Diverse linkers to the pendant aryl groups were tolerated
such as ethers (17d, 17h), sulfones (17 f, 17g), amides (17k,
Scheme 3. Anti-HCV replicon activity of 17a and 17b with pendant linker.
&
ChemMedChem 2017, 12, 1 – 14
www.chemmedchem.org
4
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Table 3. Anti-HCV replicon activity of compound 1a analogues with pendant aryl groups.
Compd
R
EC50 [nm][a]
GT1b
Compd
R
EC50 [nm][a]
GT1b
GT2a
GT2a
17c[b]
17
13
21
17m
11
13
17d[b]
228
17n
16
95
17e[b]
11
7
13
49
7
17o
17p
17q
1
1
2
17
17 f
6
17g
2
3
17h[b]
8
19
17r
7
1
13
17i[b]
77
226
17s[b]
5
17j
5
10
36
17k
40
17i[b]
44
76
[a] Data are the mean of nꢀ2 determinations unless otherwise noted. [b] n=1.
ChemMedChem 2017, 12, 1 – 14
www.chemmedchem.org
5
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
17l), as well as heterocycles such as the tetrazole 17j. Ulti-
mately, sulfonamides linked through a benzylic amine or ethyl-
amine (17m) emerged as highly potent analogues against
both GT1b and GT2a. Extensive optimization led to inhibitors
including 17p, 17q, and 17s that possessed low single digit
nanomolar potencies (EC50) against both genotypes. Although
these compounds were selected for further investigation, they
consistently showed very poor PK profiles. Metabolite ID ex-
periments showed the compounds underwent rapid degrada-
tion with the primary metabolic pathway being oxidation at
the benzylic position. Attempts to make truncated sulfona-
mides lacking the benzylic methylene led to 17e and 17i but
manifested in a loss of potency (10- to 100-fold).
An alternative strategy to protect the benzylic position from
oxidation was to link it to the pendant aromatic ring to gener-
ate a fused aromatic heterocycle (Scheme 4). Initial compounds
showed good potency, and more importantly attractive im-
provements in metabolic stability leading to promising phar-
macokinetic profiles. Extensive optimization showed that
a range of fused heterocycles was tolerated. However, the 5,6-
fused heterocycles linked at the 2-position were consistently
the most potent analogues (Table 4). Surveying unsubstituted
Table 4. Anti-HCV replicon activity of heterocyclic analogues.
Compd
19a
HetAr
EC50 [nm][a]
Compd
19k
HetAr
EC50 [nm][a]
GT1b
GT1b
GT2a
GT2a
7
17
2
7
19b[b]
11
27
2
23
186
5
19l
4[b]
9
70
19c[b]
19d
19m[b]
19n[b]
19o[b]
14
27
9
186
31
19 f[b]
8
15
19g
19h
2
4
19p
19q
2
6
11
36
2
3
14
19i
19j
1
2
16[b]
19r
20
12
3
7
0.9
[a] Data are the mean of nꢀ2 determinations unless otherwise noted. [b] n=1.
&
ChemMedChem 2017, 12, 1 – 14
www.chemmedchem.org
6
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Scheme 4. Cyclization of aryl sulfonamide shields benzylic position from oxidation.
rings led to the finding that benzoxazoles and pyridyloxazoles
such as 19l, 19j, and 19k generally had the best potency. In
addition, it was discovered that incorporation of a 4-methoxy
substituent into the central phenyl ring removed activity in the
hERG screening assay. As the compounds were further evaluat-
ed for properties such as permeability and pharmacokinetics,
we found that incorporation of fluorine to the ring not only
improved permeability and bioavailability, but also increased
potency. Ultimately, 7-fluorobenzoxazole 20 was selected to be
further evaluated as a lead candidate due to its overall profile.
ground, although there was some attenuation of potency
against similar mutations in the GT1a background.
The pharmacokinetic properties of 20 were characterized
using Sprague–Dawley rats and beagle dogs (Table 6). Fluoro-
benzoxazole 20 showed modest bioavailability, exposure, mod-
erate clearance, and C24h drug concentrations well above the
serum adjusted GT1a or 1b EC50 values in both species. In gen-
eral, the modest bioavailability was due to the poor solubility
of 20. Evaluated in the context of its in vitro potencies, howev-
er, these parameters suggested that 20 would demonstrate ef-
ficacy in an in vivo model of chronic HCV infection. To prepare
for an efficacy proof-of-concept study in chronic HCV-infected
chimpanzees, 20 was orally dosed as a suspension in Tang to
two uninfected chimpanzees at a dose of 5 mgkgꢁ1 (Table 6).
Fluorobenzoxazole 20 achieved good exposure and a low-mi-
cromolar Cmax. Although sufficient sampling to determine clear-
ance was not possible, the mean plasma concentration at 12 h
was 0.71 mm, well above the serum-adjusted EC50 for geno-
type 1. Based on this acceptable profile, 20 was advanced to
an in vivo efficacy study in chronic HCV-infected chimpanzees.
Fluorobenzoxazole 20 was dosed to two high viral load (VL)
Characterization of fluorobenzoxazole 20
Fluorobenzoxazole 20 underwent a detailed in vitro potency
evaluation against both a genotype panel and a panel of
common resistance mutations to palm-binding non-nucleoside
inhibitors in the genotype 1 background (Table 5). Across this
genotype panel 20 showed low nanomolar potency. There was
approximately a 10-fold shift in 50% serum (data not shown).
Compound 20 also retained its potency against well-character-
ized non-nucleoside inhibitor mutations in the GT1b back-
chronic HCV-infected chimpanzees (GT1a, VL 5.73 logIUmLꢁ1
;
GT1b, VL 6.13 logIUmLꢁ1) at 2 mg per kg body weight
(ꢂ120 mg) q.d. for seven days (Figure 2). Both animals experi-
enced ꢂ2 log decrease in viral load by the end of the dosing
interval, with a modest rise in viremia by the end of an addi-
tional week post-dosing. Viral load was at its nadir at the end
of dosing, with plasma drug concentrations 24 h after the final
dose of 0.5 mm (GT1a) and 1.2 mm (GT1b), approximately 570-
and 175-fold above the serum-adjusted EC50 values. Baseline
and end-of-dosing samples were sequenced to evaluate emer-
gence of resistance; there was no evidence for either emerging
resistance or enhanced populations of pre-existing viral var-
Table 5. In vitro potency profile of fluorobenzoxazole 20.
Genotype
EC50 [nm][a]
Mutant
EC50 [nm][a]
1a
1b
2a
2b
3a
4a
2.7ꢃ2.0
3.1ꢃ1.0
25.4ꢃ16.9
9.1ꢃ2.3
3.3ꢃ2.2
3.0ꢃ1.7
1b C316Y
1b M414V
1a C316Y
1a S365A
9.2ꢃ5.5
5.1ꢃ3.4
123.9ꢃ86.1
197ꢃ107
[a] Values are the average of nꢀ3 determinations.
Table 6. Pharmacokinetic parameters of fluorobenzoxazole 20.
Parameter
S–D Rat (n=2)
Beagle Dog (n=2)
Chimpanzee (n=2)
Dose i.v. [mgkgꢁ1
]
]
2
1
CL [mLminꢁ1 kgꢁ1
2.0ꢃ0.5
3.8ꢃ1.4
2.3ꢃ1.4
27.2ꢃ14.3
t1/2 [h]
Dose p.o. [mgkgꢁ1
Cmax [mm]
tmax [h]
AUC0–1 [mmꢁh]
F [%]
]
10
5
5
3.7ꢃ1.4
5.0ꢃ1.0
29.9ꢃ11.6
21.8ꢃ8.5
0.110ꢃ0.11
0.5ꢃ0.1
4.0ꢃ0.0
6.6ꢃ0.9
14.2ꢃ1.1
0.10ꢃ0.02
1.2ꢃ0.2
6.0ꢃ2.0
25.6ꢃ12.1
ND[a]
C24h [mm]
C12h: 0.71ꢃ0.20
[a] Not determined.
ChemMedChem 2017, 12, 1 – 14
www.chemmedchem.org
7
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
loss in activity observed. The difference in potency of MK-8876
was particularly significant against the C316Y and S365A mu-
tants and even more so against the 1a C316Y:C445F double
mutant.
The pharmacokinetics of MK-8876 (23) were profiled in Spra-
gue-Dawley rats and beagle dogs (Table 9). In Sprague-Dawley
rats MK-8876 showed moderate clearance and a half-life of
6.1 h. A 10 mgkgꢁ1 oral dose in S–D rats gave a Cmax of
1.12 mm and a tmax of 7 h with 35% bioavailability. In the
beagle dog MK-8876 showed low clearance and a half-life of
4.9 h. A 5 mgkgꢁ1 oral dose in beagle dogs gave a Cmax of
1.2 mm and a tmax of 1.2 h with 26% bioavailability. The C24h
plasma drug concentrations were 0.169 mm (rat) and 0.102 mm
(dog), both well above the serum-adjusted EC50 values of the
replicon panel. The results were consistent with the profile of
a compound suitable for once daily dosing which was essential
for a lead compound that was intended to be dosed as part of
combination therapy.
Figure 2. Fluorobenzoxazole 20 in vivo efficacy study. Fluorobenzoxazole 20
was dosed at 2 mgkgꢁ1 (ꢂ120 mg) once daily for seven days as a suspen-
sion in Tang to two chronic HCV-infected chimpanzees harboring GT1a or
GT1b infections. Blood samples were collected periodically, processed to
plasma, and evaluated for HCV viral load. Plasma drug concentrations 24 h
following completion of dosing are shown.
iants in samples collected 24 h following the final dose, when
viremia was at its most suppressed. Of note, the results against
the GT1a infection were every bit as robust as those against
the GT1b infection. However, the in vitro potency losses
against some mutants in the GT1a background with 20 com-
pelled us to seek a superior analogue for clinical development.
Finally, we obtained a crystal structure of MK-8876 (23)
bound to the GT1b NS5B polymerase (Figure 3). As originally
postulated, the 2-phenylbenzofuran region of the molecule in-
teracted within the palm II binding site, as previously known
for compounds with similar core structures (e.g., HCV-796). Im-
portantly, the methyl amide at position 3 of the 2-phenylben-
zofuran engages in key interactions with serine 365 but does
not interact with residue 316. We note that the crystal struc-
ture was obtained for the GT1b BK form of NS5B, which natu-
rally encodes an asparagine at residue 316 rather than the
more common cysteine. The tetracyclic region of MK-8876
filled the hydrophobic palm I binding site, flanking the methio-
nine 414 and tyrosine 448 residues. This crystal structure clear-
ly confirmed that palm I and palm II sites were and could be si-
multaneously bound, leading to a significant improvement in
potency versus molecules without this capability. This improve-
ment in potency is particularly evident in the case of NS5B mu-
tants. The specific interactions at the residue level that are the
Discovery of MK-8876
Concurrent with the evaluation of 20, continued optimization
of the pendant heterocycle moiety was conducted to improve
potency against mutations in the GT1a background. During an
exploration of the impact of introducing ring constraints, we
learned that formation of the tetracycles shown in Table 7 gen-
erated analogues with exceptional potency. Likely due to the
ring constraint, these compounds were sensitive to the substi-
tution pattern around the ring. For example, tetracycles with
a substitution at the 6- or 7-position of the indole ring such as
22i and 22j lost potency against GT2a. In contrast, a substitu-
tion at the 4- or 5-position of the indole such as 22k and 23
maintained or improved potency against all genotypes tested.
In addition, these tetracycles showed excellent chemical and
metabolic stability as well as good pharmacokinetic profiles.
Tetracycle 23 in particular possessed excellent potency when
tested against a broad genotype and mutant panel. Based on
these early assessments, this compound was selected for ad-
vancement and underwent a full characterization, ultimately
becoming the clinical candidate MK-8876.
Characterization of MK-8876
The in vitro potency of MK-8876 (23) was determined against
a broad replicon panel of HCV genotypes and NS5B non-nu-
cleoside mutants (Table 8). Overall, MK-8876 displayed excel-
lent potency across the panel, having a nearly equipotent pro-
file in the low single digit nanomolar range across the spec-
trum. Significantly, the potency profile of MK-8876 against the
clinically relevant NS5B non-nucleoside mutants showed the
greatest differentiation from its predecessor 20. MK-8876 re-
tained its potency across all mutants tested with no significant
Figure 3. MK-8876 (23) crystal structure bound to GT1b NS5B polymerase.
The phenylbenzofuran core binds to the palm II site, projecting methyl
amide at position 3 and sulfonamide at position 6 critical for ligand potency.
The tetracyclic substituent at position 5 extends to the hydrophobic palm I
site.
&
ChemMedChem 2017, 12, 1 – 14
www.chemmedchem.org
8
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Table 7. Anti-HCV replicon activity of fused heterocyclic analogues.
Compd
Aryl
EC50 [nm][a]
GT1b
Compd
Aryl
EC50 [nm][a]
GT2a
GT1b
GT2a
22a
2
5
6
22i[b]
7
39
22b
11
18
22j[b]
4
4
24
22c[b]
2
5
8
22k
3
22d[b]
14
17
23
3
4
2
4
22e[b]
23a
22 f
22g
22h
4
3
3
4
7
4
23b
23d
23 f
5
19
23
3
13
27
[a] Data are the mean of nꢀ2 determinations unless otherwise noted. [b] n=1.
basis of this potency improvement could not been deter-
mined. However, it is clear that the tetracyclic motif of MK-
8876 results in palm I binding site induced fit associated with
large dispersion interactions resulting in the observed potency
boost. In addition, the absence of specific polar interactions
with amino acids at key positions is an advantage for pan-gen-
otypic potency and mutant resistance.
ChemMedChem 2017, 12, 1 – 14
www.chemmedchem.org
9
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
mediates were prepared as described. 1H NMR spectra were ob-
tained on a Varian VNMR System 400 (400 MHz) and are reported
as ppm downfield from TMS. The synthetic procedures of individu-
al analogues were carried out using described procedures.[16] The
optimized synthesis of MK-8876 (compound 23) was developed in
the MSD Process Laboratories and reported previously.[22]
Table 8. In vitro potency profile of MK-8876 (23).
Genotype
EC50 [nm][a]
Mutant
EC50 [nm][a]
1a
1b
2a
2b
3a
4a
5a
1.6ꢃ1.2
2.2ꢃ0.7
7.3ꢃ5.2
3.9ꢃ2.1
1.8ꢃ1.1
1.2ꢃ0.7
1.2ꢃ0.7
1b C316Y
1b S365A
1b M414I
1b M414T
1b M414V
1a C316Y
1a S365A
1a S365T
2.7ꢃ2.2
1.1ꢃ0.3
4.0ꢃ2.1
1.8ꢃ0.9
1.3ꢃ0.7
2.3ꢃ0.9
1.0ꢃ0.6
6.6ꢃ1.6
2.0ꢃ0.8
2.0ꢃ0.7
3.2ꢃ1.5
Procedure for the synthesis of compound 20
Step 1: 1-fluoro-3-methoxy-2-nitrobenzene: To a 08C solution of
1,3-difluoro-2-nitrobenzene (100 g, 0.63 mol) in MeOH (1.3 L) was
slowly added a solution of MeONa (0.69 mol in MeOH, freshly pre-
pared from 15.9 g of sodium metal and 200 mL of MeOH). The re-
sulting reaction was allowed to stir for about 15 h at room temper-
ature, then the reaction mixture was concentrated and diluted
with EtOAc. The organic phase was washed sequentially with
water and brine, dried over Na2SO4, then filtered and concentrated
in vacuo to provide 1-fluoro-3-methoxy-2-nitrobenzene (98 g, 91%
yield) which was used in the next step without further purification.
1H NMR (CDCl3, 400 MHz) d=7.38–7.44 (m, 1H), 6.72–6.88 (m, 2H),
3.95 ppm (s, 3H).
1a M414I
1a M414V
1a C316Y:C445F
[a] Values are the average of nꢀ3 determinations.
Table 9. Pharmacokinetic parameters of MK-8876 (23).
Parameter
S–D Rat (n=2)
Beagle Dog (n=3)
Dose i.v. [mgkgꢁ1
]
]
2
1
CL [mLminꢁ1 kgꢁ1
7.3ꢃ2.0
6.1ꢃ1.2
3.0ꢃ0.5
9.2ꢃ0.3
Step 2: 3-fluoro-2-nitrophenol: To a ꢁ408C solution of 1-fluoro-3-
methoxy-2-nitrobenzene (98 g, 0.57 mol) in dichloromethane
(500 mL) was added dropwise a solution of BBr3 (1 L, 1m in di-
chloromethane). The resulting reaction was allowed to stir for
about 15 h at room temperature then the reaction mixture was
slowly poured into ice water (500 mL). The resulting solution was
extracted with EtOAc and the combined organic layers were
washed with 5% aqueous NaHCO3, brine, and then dried over
Na2SO4. The mixture was filtered and concentrated to provide 3-
fluoro-2-nitrophenol (85 g, 95% yield) which was used in the next
step without further purification. 1H NMR (CDCl3, 400 MHz) d=
7.43–7.49 (m, 1H), 6.88 (d, J=8.0 Hz, 1H) 6.73–6.78 ppm (m, 1H).
t1/2 [h]
Dose p.o. [mgkgꢁ1
Cmax [mm]
tmax [h]
AUC0–1 [mmꢁh]
F [%]
]
10
5
1.1ꢃ0.3
7.0ꢃ1.0
13.5ꢃ2.3
35.3ꢃ5.9
0.169ꢃ0.001
1.2ꢃ0.8
1.2ꢃ0.8
12.2ꢃ4.4
24.7ꢃ6.4
0.102ꢃ0.018
C24h [mm]
Conclusions
A research program directed at developing a non-nucleoside
inhibitor of HCV NS5B based on the benzofuran core of HCV-
796 led to the discovery of a novel structural class of 5-aryl
benzofurans that interact with both the palm I and palm II
binding regions. An extensive medicinal chemistry effort result-
ed in the identification of a series of pendant heterocycles suit-
able for further development, most notably the fluorobenzoxa-
zole 20, although subsequent characterization of 20 identified
deficiencies against clinically relevant resistance variants. Fur-
ther optimization of the pendant heterocycle through incorpo-
ration of a ring constraint successfully led to the discovery of
MK-8876 (23). MK-8876 demonstrated a broadly acting geno-
typic potency profile and improved potency against clinically
relevant mutants. In addition, MK-8876 showed a good preclin-
ical pharmacokinetic profile indicating it would be suitable for
once daily dosing. MK-8876 has successfully completed
Phase 1 monotherapy studies for evaluation and characteriza-
tion as a future component of an all-oral direct-acting regimen
for the treatment of HCV.
Step 3: 2-amino-3-fluorophenol: 3-fluoro-2-nitrophenol (38 g,
0.24 mol) was dissolved in EtOH and then palladium on carbon
(5 g, 10% Pd) was added. The reaction flask was evacuated and
the reaction mixture was stirred under an H2 atmosphere (1 atm)
for 3 h at room temperature. The reaction mixture was then fil-
tered through a short pad of Celite and washed through with
EtOH. The combined filtrate was concentrated to provide 2-amino-
3-fluorophenol (26 g, 86% yield) which was used in the next step
without further purification. 1H NMR (CDCl3, 400 MHz) d=9.43 (s,
1H), 6.42–6.53 (m, 2H), 6.32–6.42 (m, 1H), 4.34 ppm (s, 2H).
Step 4: 2-(5-bromo-2-methoxyphenyl)-4-fluorobenzo[d]oxazole:
To a solution of 2-amino-3-fluorophenol (9 g, 70.8 mmol) in 10 mL
of PPA was added 5-bromo-2-methoxybenzoic acid (16.3 g,
70.8 mmol) and the resulting mixture was heated at 1408C and
stirred for 4 h. The reaction mixture was then poured into ice
water (50 mL) and extracted with EtOAc. The organic extract was
concentrated and the residue was purified via flash chromatogra-
phy on silica gel (petroleum ether/ethyl acetate 10:1) to provide 2-
(5-bromo-2-methoxyphenyl)-4-fluorobenzo[d]oxazole (16 g, 82%
yield) as a solid. 1H NMR (CDCl3, 400 MHz) d=8.29 (d, J=2.4 Hz,
1H), 7.57–7.54 (m, 1H), 7.40 (d, J=8.0 Hz, 1H), 7.27–7.33 (m, 1H),
7.07 (m, 1H), 6.96 (d, J=9.2 Hz, 1H), 3.99 ppm (s, 3H).
Step 5: 4-fluoro-2-(2-methoxy-5-(4,4,5,5-tetramethyl-1,3,2-dioxa-
borolan-2-yl)phenyl)benzo[d]oxazole: A solution of 2-(5-bromo-2-
methoxyphenyl)-4-fluorobenzo[d]oxazole (18.4 g, 57.1 mmol) and
bis(pinacolato)diboron (17.4 g, 68.5 mmol) in DMF (10 mL) was
stirred under a nitrogen atmosphere. Pd(dppf)Cl2 (500 mg) and po-
tassium acetate (10 g, 114 mmol) were added. The mixture was
Experimental Section
Chemistry
Commercially available reagents, solvents, and intermediates were
used as received unless otherwise noted. Other reagents and inter-
&
ChemMedChem 2017, 12, 1 – 14
www.chemmedchem.org
10
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
heated at 808C and stirred for 3 h. The reaction mixture was then
cooled and concentrated in vacuo. The residue was dissolved in di-
chloromethane and filtered through a pad of Celite. The solution
was washed with H2O, brine, dried over Na2SO4 then filtered and
concentrated. The residue was purified by flash chromatography
on silica gel (petroleum ether/ethyl acetate 10:1) to provide 4-
fluoro-2-(2-methoxy-5-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-
phenyl)benzo[d]oxazole (10 g, 54% yield) as a solid. 1H NMR
(CDCl3, 400 MHz) d=8.53 (d, J=1.6 Hz, 1H), 7.85–7.92 (m, 1H),
7.44 d, J=8.0 Hz, 1H), 7.20–7.28 (m, 1H), 6.96–7.05 (m, 2H), 3.97 (s,
3H), 1.29 ppm (s, 12H).
Table 10. Data collection and refinement statistics.
Parameter
Value
Wavelength [ꢂ]
Frame width [8]/Total sweep [8]
Space group
1.000
0.25/180
P212121
Cell dimensions
a, b, c [ꢂ]
86.1, 106.4, 126.1
90, 90, 90
126–2.80 (2.95–2.80)
0.149 (0.667)
10.1 (2.7)
a, b, g [8]
Resolution [ꢂ][a]
[a]
Rmerge
I/s(I)[a]
Step 6: 5-(3-(4-fluorobenzo[d]oxazol-2-yl)-4-methoxyphenyl)-2-
(4-fluorophenyl)-N-(methyl-6-(N-methylsulfonamido)benzofuran-
3-carboxamide (compound 20): To a solution of bromide 13 (5 g,
11.0 mmol) and 4-fluoro-2-(2-methoxy-5-(4,4,5,5-tetramethyl-1,3,2-
dioxaborolan-2-yl)phenyl)benzo[d]oxazole (5.27 g, 14.3 mmol) in
DMF (150 mL) under nitrogen was added Pd(dppf)Cl2 (200 mg) and
K3PO4 (4.66 g, 22.0 mmol). The reaction mixture was heated at
1008C and allowed to stir for 10 h then cooled and concentrated
in vacuo. The residue was dissolved in dichloromethane and fil-
tered through a pad of Celite. The filtrate was washed sequentially
with water and brine then dried over Na2SO4, filtered and concen-
trated. The crude product was purified by flash chromatography
on silica gel (petroleum ether/ethyl acetate 5:1) to provide 5-(3-(4-
fluorobenzo[d]oxazol-2-yl)-4-methoxyphenyl)-2-(4-fluorophenyl)-N-
(methyl-6-(N-methylsulfonamido)benzofuran-3-carboxamide (com-
pound 20)(3.8 g, 56% yield) as a white solid. 1H NMR (CDCl3,
400 MHz) d=8.21 (d, J=2.0 Hz, 1H), 7.91–7.95 (m, 2H), 7.83 (s,
1H), 7.68 (d, J=2.0 Hz, 1H), 7.66 (s, 1H), 7.39 (d, J=8.0 Hz, 1H),
7.14–7.27 (m, 4H), 7.06 (t, J=8.4 Hz, 1H), 5.95 (brs, 1H), 4.06 (s,
3H), 3.14 (s, 3H), 2.99 (d, J=4.8 Hz 3H), 2.77 ppm (s, 3H); MS (M+
H)+ 618.
Completeness [%][a]
Redundancy[a]
100 (99.9)
6.6 (6.8)
Refinement
Resolution [ꢂ]
Reflections
Rwork/Rfree
2.8
29173
0.18/0.24
No. non-hydrogen atoms
Protein
Waters
Ligands
8715
259
88
B-factors (average) [ꢀ2]
Protein
Waters
Ligands
46
34
54
RMSD bonds [ꢂ]/angles [8]
0.010/1.14
[a] Highest-resolution shell shown in parentheses.
Biology
All animal studies were performed according to the NIH Guide for
the Care and Use of Laboratory Animals, and experimental proto-
cols were approved by the Institutional Animal Care and Use Com-
mittees of Merck & Co., Inc., Kenilworth, NJ USA. The HCV geno-
type was determined by a line probe assay (Versant HCV genotype
assay, LiPa, Bayer Diagnostics/Innogenetics) and confirmed by RT-
PCR rescue and sequencing of HCV genetic material.[27] HCV-infect-
ed chimpanzees were dosed orally at 5 mgkgꢁ1 for a single dose
or q.d. for seven days at 2 mgkgꢁ1 by the voluntary ingestion of
compound 20 (in a Tang vehicle). Viral load determinations were
performed on plasma samples using the HCV TaqMan assay (Cene-
tron Diagnostics, Austin, TX). Viral resistance analysis of chimpan-
zee plasma samples was conducted similarly to a previously pub-
lished protocol using NS5B specific primers.[27] Initial potency
screening was conducted against GT1b or GT2a replicon cell lines
using a luciferase-based reporter assay.[28] The assay was automat-
ed which enabled rapid screening of many compounds; however,
the assay required robust expression of a co-encoded luciferase
gene which in practice restricted its use to very few genotypes.
For compounds of greater interest, additional and more rigorous
EC50 determinations were generated against broad panels of geno-
type and mutant replicon cell lines, and were conducted using
a sensitive but lower through-put TaqMan based assay.[29] Potency
values reported from the TaqMan assay may vary slightly from
those generated with the luciferase-based reporter assay due to
differences in cell lines and assay conditions. The NS5B sequences
for the genotypes are GT1a H77 (GenBank AJ238799), GT1b con1
(GenBank AB047639), GT2a JFH (GenBank AB047639), GT2b (Gen-
Bank D10988), GT3a (GenBank 17763), and GT4a,[30] and GT5a (Gen-
Crystallography
Crystals of HCV NS5B(1b/BK)D21 were prepared as previously de-
scribed.[23] Once grown, crystals were soaked for three days in re-
servoir solution supplemented with 1 mm MK-8876, then cryopro-
tected and flash-cooled as described. Diffraction data were collect-
ed on a Dectris PILATUS 6m detector installed at IMCA-CAT beam-
line 17-ID at the Advanced Photon Source. Data were processed
using autoPROC[24], and TLS refinement was conducted using auto-
BUSTER[25] and Coot[26] with HCV NS5B(1b/BK)D21 apoprotein (PDB
ID: 1C2P) as a starting model. Data collection and refinement sta-
tistics are listed in Table 10. The structure of the HCV NS5B com-
plex with MK-8876 has been deposited at the PDB with accession
code 5W2E.
Computational chemistry and modeling studies
Crystal structures were used as starting point for modeling work
executed using MOE (Molecular Operating Environment, version
2013.08; Chemical Computing Group Inc., 1010 Sherbooke St.
West, Suite #910, Montreal, QC, H3A 2R7 (Canada), 2017). Protein
preparation was applied to the initial structures and coordinates
were refined with harmonic constraints using the MMFF94x force-
field. Ligand minimization was performed using the MMFF94x
force-field with either rigid or flexible protein residues in the bind-
ing sites (up to 6 ꢂ radius around the bound ligand).
ChemMedChem 2017, 12, 1 – 14
www.chemmedchem.org
11
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
[11] G. T. Everson, K. D. Sims, P. J. Thuluvath, E. Lawitiz, T. Hassanein, M. Ro-
driguez-Torres, T. Desta, T. Hawkins, J. M. Levin, F. Hinestrosa, V. Rustgi,
H. Schwartz, Z. Younossi, L. Webster, N. Gitlin, T. Eley, S. P. Huang, F.
McPhee, D. M. Grasela, P. F. Gardiner, Liver. Int. 2016, 36, 189–197.
[12] N. M. Kneteman, A. Y. Howe, T. Gao, J. Lewis, D. Pavear, G. Lund, D.
Douglas, D. F. Mercer, D. L. Tyrrell, F. Immermann, I. Chaudhary, J. Speth,
S. A. Villano, J. O’Connell, M. Collett, Hepatology 2009, 49, 745–752.
[13] A. Maynard, R. M. Crosby, B. Ellis, R. Hamatake, Z. Hong, B. A. Johns,
K. M. Kahler, C. Koble, A. Leivers, M. R. Leivers, A. Mathis, A. J. Peat, J. J.
Pouliot, C. C. Roberts, V. Samano, R. M. Schmidt, G. K. Smith, A. Spalten-
stein, E. L. Stewart, P. Thommes, E. M. Turner, C. Voitenleitner, J. T.
Walker, G. Waitt, J. Weatherhead, K. Weaver, S. Williams, L. Wright, Z. Z.
Xiong, D. Haigh, J. B. Shotwell, J. Med. Chem. 2014, 57, 1902–1913.
[14] K.-S. Yeung, B. R. Beno, K. Parcella, J. A. Bender, K. A. Grant-Young, A.
Nickel, P. Gunaga, P. Anjanappa, R. O. Bora, K. Selvakumar, K. Rigat, Y.-K.
Wang, M. Liu, J. Lemm, K. Mosure, S. Sheriff, C. Wan, M. Witmer, K. Kish,
U. Hanumegowda, X. Zhuo, Y.-Z. Shu, D. Parker, R. Haskell, A. Ng, Q.
Gao, E. Colston, J. Raybon, D. M. Grasela, K. Santone, M. Gao, N. A.
Meanwell, M. Sinz, M. G. Soars, J. O. Knipe, S. B. Roberts, J. F. Kadow, J.
Med. Chem. 2017, 60, 4369–4385.
Bank 009826). Inhibitors from this study typically were modest at
best in potency when using an NS5B enzymatic assay. The data en-
abled poor discrimination among compounds, and its use was
abandoned early in the program.
Pharmacokinetic studies
Studies were performed in both rats and dogs. The study protocols
were reviewed and approved by the Institutional Animal Care and
Use Committees at Merck & Co., Inc., Kenilworth, NJ USA. For stud-
ies in which compound 20 was dosed intravenously to rats or
dogs, compound was formulated in DMSO and administered as
a bolus at 2 mgkgꢁ1 (male Sprague–Dawley rat) or formulated in
10% EtOH 70% polyethylene glycol 400 (PEG400) and adminis-
tered at 1.0 mgkgꢁ1 (male Beagle dog). For oral studies, compound
20 was dosed in PEG 400 at 10 mgkgꢁ1 as a suspension (rat) or at
5 mgkgꢁ1 as a solution (dog). For studies in which MK-8876 was
dosed intravenously to rats or dogs, compound was formulated in
DMSO and administered as a bolus at 2 mgkgꢁ1 (male Sprague–
Dawley rat) or formulated in DMSO/PEG400/water at a 2:6:2 ratio
and administered as a bolus at 1 mgkgꢁ1 (male beagle dog). For
oral studies MK-8876 was dosed as a suspension in PEG 400 at
10 mgkgꢁ1 (rat) or as a suspension in 0.5% methylcellulose/0.25%
sodium dodecyl sulfate/5 mm HCl at 5 mgkgꢁ1 (dog). For all stud-
ies, blood samples were collected in EDTA-containing tubes at ap-
propriate times, plasma separated by centrifugation and stored at
ꢁ708C until analysis. Quantitation of compound 20 or MK-8876
levels was conducted by high performance liquid chromatogra-
phy/tandem mass spectrometry (LC–MS/MS), following protein pre-
cipitation.
[15] C. J. Burns, A. M. Del Vecchio, T. R. Bailey, B. A. Kulkami, T. H. Faitg, S. R.
Sherk, C. W. Blackledge, D. J. Rys, T. A. Lessen, J Swestock, Y. Deng, T. J.
Nitz, J. A. Reinhardt, H. Feng, A. K. Saha, (Viropharma Incorporated,
Exton, USA, Wyeth, Madison, USA), Int. PCT Pub. No. WO2004041201,
2004.
[16] C. C. McComas, N. J. Liverton, J. Habermann, U. Koch, F. Narjes, P. Li, X.
Peng, R. Soll, H. Wu, (Merck Sharpe & Dohme Corp., Rahway, USA), Int.
PCT Appl. No. WO2013033901 A1, 2013; C. C. McComas, N. J. Liverton,
J. Habermann, U. Koch, F. Narjes, P. Li, X. Peng, R. Soll, H. Wu, A. Palani,
S. He, D. Xing, H. Liu, Z. Lai, C. London, D. Xiao, N. Zorn, R. P. Nargund,
(Merck Sharpe & Dohme Corp., Rahway, USA), Int. PCT Appl. No.
WO2013034048 A1, 2013; C. C. McComas, N. J. Liverton, J. Habermann,
U. Koch, F. Narjes, P. Li, X. Peng, R. Soll, H. Wu, A. Palani, S. He, D. Xing,
H. Liu, Z. Lai, C. London, D. Xiao, N. Zorn, R. P. Nargund, (Merck Sharpe
& Dohme Corp., Rahway, USA), Int. PCT Appl. No. WO2013034047 A1,
2013; C. C. McComas, N. J. Liverton, J. Habermann, U. Koch, F. Narjes, P.
Li, X. Peng, R. Soll, H. Wu, (Merck Sharpe & Dohme Corp., Rahway, USA)
Int. PCT Appl. No. WO2013033900 A1, 2013; C. C. McComas, N. J. Liver-
ton, R. Soll, P. Li, X. Peng, H. Wu, F. Narjes, J. Habermann, U. Koch, S. Liu,
(Merck Sharpe & Dohme Corp., Rahway, USA), Int. PCT Appl. No.
WO2011106992A1, 2011; C. C. McComas, N. J. Liverton, J. Habermann,
U. Koch, F. Narjes, P. Li, X. Peng, R. Soll, H. Wu, (Merck Sharpe & Dohme
Corp., Rahway, USA), Int. PCT Appl. No. WO2011106929 A1, 2011; C. C.
McComas, N. J. Liverton, R. Soll, P. Li, X. Peng, H. Wu, (Merck Sharpe &
Dohme Corp., Rahway, USA), Int. PCT Appl. No. WO2011106986 A1,
2011.
Conflict of interest
The authors declare no conflict of interest.
Keywords: hepatitis C · inhibitors · MK-8876 · non-nucleoside ·
NS5B
[1] D. Lavanchy, Clin. Microbiol. Infect. 2011, 17, 107–115.
[2] R. S. Brown, Nature 2005, 436, 973–978.
[17] P. Krishnan, R. Tripathi, G. Schnell, T. Reisch, J. Beyer, M. Irvin, W. Xie, L.
Larsen, D. Cohen, T. Podsadecki, T. Pilot-Matias, C. Collins, Antimicrob.
Agents Chemother. 2015, 59, 5445–5454.
[18] H. Mo, C. Hedskog, E. Svarovskaia, S.-C. Sun, I. M. Jacobson, D. M. Brai-
nard, J. G. McHutchison, M. D. Miller, J. Med. Virol. 2016, 23, 644–651.
[19] P. J. F. Troke, M. Lewis, P. Simpson, K. Gore, J. Hammond, C. Craig, M.
Westby, Antimicrob. Agents Chemother. 2012, 56, 1331–1341.
[3] P. Ferenci, D. Bernstein, J. Lalezari, D. Cohen, Y. Luo, C. Cooper, E. Tam,
R. T. Marinho, T. Naoky, A. Nyberg, T. D. Box, Z. Younes, P. Enayati, S.
Green, Y. Baruch, B. R. Bhandari, F. A. Caruntu, T. Sepe, V. Chulanov, E.
Janczewska, G. Rizzardini, J. Gervain, R. Planas, C. Moreno, T. Hassanein,
W. Xie, M. King, T. Podsadecki, K. R. Reddy, N. Engl. J. Med. 2014, 370,
1983–1992.
[20] J. Q. Hang, Y. Yang, S. F. Harris, V. Leveque, H. J. Whittington, S. Rajya-
guru, G. Ao-Ieong, M. F. McCown, A. Wong, A. M. Giannetti, S. Le Po-
gam, F. Talamas, N. Cammack, I. Najera, K. Klumpp, J. Biol. Chem. 2009,
284, 15517–15529.
[21] F. Ruebsam, D. E. Murphy, C. V. Tran, L. S. Li, J. Zhao, P. S. Dragovich,
H. M. McGuire, A. X. Xiang, Z. Sun, B. K. Ayida, J. K. Blazel, S. H. Kim, Y.
Zhou, Q. Han, C. R. Kissinger, S. E. Webber, R. E. Showalter, A. M. Shah,
M. Tsan, R. A. Patel, P. A. Thompson, L. A. Lebrun, H. J. Hou, R. Kamran,
M. V. Sergeeva, D. M. Bartkowski, T. G. Nolan, D. A. Norris, J. Khandurina,
J. Brooks, E. Okamoto, L. Kirkovsky, Bioorg. Med. Chem. Lett. 2009, 19,
6404–6412.
[4] K. V. Kowdley, S. L. Gordon, K. R. Reddy, L. Rossaro, D. E. Bernstein, E.
Lawitz, M. L. Shiffman, E. Schiff, R. Ghalib, M. Ryan, V. Rustgi, M. Chojkier,
R. Herring, A. M. Di Bisceglie, P. J. Pockros, G. M. Subramanian, D. An, E.
Svarovskaia, R. H. Hyland, D. Phil, P. S. Pang, W. T. Symonds, J. G.
McHutchison, A. J. Muir, D. Pound, M. W. Fried, N. Engl. J. Med. 2014,
370, 1973–1982.
[5] S. Zeuzem, R. Ghalib, K. R. Reddy, P. J. Pockros, Z. B. Ari, Y. Zhao, D. D.
Brown, S. Wan, M. J. DiNubile, B. Y. Nguyen, M. N. Robertson, J. Wahl, E.
Barr, J. R. Butterton, Ann. Intern. Med. 2015, 163, 1–13.
[6] R. Bartenschlager, F. L. Cosset, V. Lohmann, J. Hepatol. 2010, 53, 583–
585.
[22] M. J. Williams, Q. Chen, L. Codan, R. K. Dermenjian, S. Dreher, A. W.
Gibson, H. Xianliang, Y. Jin, S. P. Keen, A. Y. Lee, D. R. Lieberman, W. Lin,
G. Liu, M. McLaughlin, M. Reibarkh, J. P. Scott, S. Strickfuss, L. Tan, R. J.
Varsolona, F. Wen, Org. Process Res. Dev. 2016, 20, 1227–1238.
[23] C. A. Lesburg, M. B. Cable, E. Ferrari, Z. Hong, A. F. Mannarino, P. C.
Weber, Nat. Struct. Biol. 1999, 6, 937–943.
[7] D. B. Smith, J. Bukh, C. Kuiken, A. S. Muerhoff, C. M. Rice, J. T. Stapleton,
P. Simmonds, Hepatology 2014, 59, 318–327.
[8] S. S. Carroll, D. B. Olsen, Infect. Disord. Drug Targets 2006, 6, 17–29.
[9] L. Gerber, T. M. Welzel, S. Zeuzem, Liver Int. 2013, 33, 85–92.
[10] C. Caillet-Saguy, P. C. Simister, S. Bressanelli, J. Mol. Biol. 2011, 414, 370–
384.
&
ChemMedChem 2017, 12, 1 – 14
www.chemmedchem.org
12
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
[24] C. Vonrhein, C. Fiensburg, P. Keller, A. Sharff, O. Smart, W. Paciorek, T.
Womack, G. Bricogne, Acta Crystallogr. Sect. D 2011, 67, 293–302.
[25] G. Bricogne, E. Blanc, M. Brandl, C. Flensburg, P. Keller, W. Paciorek, P.
Roversi, A. Sharff, O. S. Smart, C. Vonrhein, T. O. Womack, BUSTER (ver-
sion 2.11.2), Global Phasing Ltd. (Cambridge, UK), 2016.
[29] N. J. Liverton, S. S. Carroll, J. DiMuzio, C. Fandozzi, D. J. Hraham, D.
Hazuda, M. K. Holloway, S. W. Ludmerer, J. A. McCauley, C. J. McIntyre,
D. B. Olsen, M. T. Rudd, M. Stahlhut, J. P. Vacca, Antimicrob. Agents Che-
mother. 2010, 54, 305–311.
[30] K. J. Herlihy, J. P. Graham, R. Kumpf, A. K. Patick, R. Duggal, S. T. Shi, Anti-
microb. Agents Chemother. 2008, 52, 3523–3531.
[26] P. Emsley, B. Lohkamp, W. G. Scott, K. Cowtan, Acta Crystallogr. Sect. D
2010, 66, 486–501.
[27] S. S. Carroll, S. W. Ludmerer, L. Handt, K. Koeplinger, N. R. Zhang, D.
Graham, M. E. Davies, M. MacCoss, D. J. Hazuda, D. B. Olsen, Antimicrob.
Agents Chemother. 2009, 53, 926–934.
Manuscript received: April 10, 2017
Revised manuscript received: June 26, 2017
[28] J. M. Vrolijk, A. Kaul, B. E. Hansen, V. Lohmann, B. L. Haagmans, S. W.
Schalm, R. Bartenschlager, J. Virol. Methods 2003, 110, 201–209.
Accepted manuscript online: June 26, 2017
Version of record online: && &&, 0000
ChemMedChem 2017, 12, 1 – 14
www.chemmedchem.org
13
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

FULL PAPERS
C. C. McComas,* A. Palani, W. Chang,
M. K. Holloway, C. A. Lesburg, P. Li,
N. Liverton, P. T. Meinke, D. B. Olsen,
X. Peng, R. M. Soll, A. Ummat, J. Wu,
J. Wu, N. Zorn, S. W. Ludmerer
Both palms open: Studies directed at
developing a broadly acting non-nu-
cleoside inhibitor of HCV NS5B led to
the discovery of a novel structural class
of 5-aryl benzofurans that simultaneous-
ly interact with both the palm I and
palm II binding regions. Herein we de-
scribe the efforts that led to the discov-
ery of MK-8876, which advanced into
Phase 1 monotherapy studies for evalu-
ation and characterization as a compo-
nent of an all-oral direct-acting drug
regimen for the treatment of chronic
HCV infection.
&& – &&
Development of a New Structural
Class of Broadly Acting HCV Non-
Nucleoside Inhibitors Leading to the
Discovery of MK-8876
&
ChemMedChem 2017, 12, 1 – 14
www.chemmedchem.org
14
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!