Discovery of the pan-genotypic hepatitis C virus NS3/4A protease inhibitor voxilaprevir (GS-9857): A component of Vosevi

Article abstract of DOI:10.1016/j.bmcl.2019.03.037

Treatment of hepatitis C virus (HCV) infection has been historically challenging due the high viral genetic complexity wherein there are eight distinct genotypes and at least 86 viral subtypes. While HCV NS3/4A protease inhibitors are an established treatment option for genotype 1 infection, limited coverage of genotypes 2 and/or 3 combined with serum alanine transaminase (ALT) elevations for some compounds has limited the broad utility of this therapeutic class. Our discovery efforts were focused on identifying an NS3/4A protease inhibitor with pan-genotypic antiviral activity, improved coverage of resistance associated substitutions, and a decreased risk of hepatotoxicity. Towards this goal, distinct interactions with the conserved catalytic triad of the NS3/4A protease were identified that improved genotype 3 antiviral activity. We further discovered that protein adduct formation strongly correlated with clinical ALT elevation for this therapeutic class. Improving metabolic stability and decreasing protein adduct formation through structural modifications ultimately resulted in voxilaprevir. Voxilaprevir, in combination with sofosbuvir and velpatasvir, has demonstrated pan-genotypic antiviral clinical activity. Furthermore, hepatotoxicity was not observed in Phase 3 clinical trials with voxilaprevir, consistent with our design strategy. Vosevi (sofosbuvir, velpatasvir, and voxilaprevir) is now an approved pan-genotypic treatment option for the most difficult-to-cure individuals who have previously failed direct acting antiviral therapy.

Full text of DOI:10.1016/j.bmcl.2019.03.037

Accepted Manuscript
Discovery of the Pan-genotypic Hepatitis C Virus NS3/4A Protease Inhibitor
Voxilaprevir (GS-9857): A Component of Vosevi ®
James G. Taylor, Sheila Zipfel, Kyla Ramey, Randy Vivian, Adam Schrier,
Kapil K. Karki, Ashley Katana, Darryl Kato, Tetsuya Kobayashi, Ruben
Martinez, Michael Sangi, Dustin Siegel, Chinh V. Tran, Zheng-Yu Yang, Jeff
Zablocki, Cheng Y. Yang, Yujin Wang, Kelly Wang, Katie Chan, Ona
Barauskas, Guofeng Cheng, Debi Jin, Brian E. Schultz, Todd Appleby,
Armando G. Villaseñor, John O. Link
PII:
S0960-894X(19)30178-7
https://doi.org/10.1016/j.bmcl.2019.03.037
BMCL 26351
DOI:
Reference:
To appear in:
Bioorganic & Medicinal Chemistry Letters
Received Date:
Revised Date:
Accepted Date:
8 October 2018
22 March 2019
25 March 2019
Please cite this article as: Taylor, J.G., Zipfel, S., Ramey, K., Vivian, R., Schrier, A., Karki, K.K., Katana, A., Kato,
D., Kobayashi, T., Martinez, R., Sangi, M., Siegel, D., Tran, C.V., Yang, Z-Y., Zablocki, J., Yang, C.Y., Wang, Y.,
Wang, K., Chan, K., Barauskas, O., Cheng, G., Jin, D., Schultz, B.E., Appleby, T., Villaseñor, A.G., Link, J.O.,
Discovery of the Pan-genotypic Hepatitis C Virus NS3/4A Protease Inhibitor Voxilaprevir(GS-9857): A Component
of Vosevi ®, Bioorganic & Medicinal Chemistry Letters (2019), doi: https://doi.org/10.1016/j.bmcl.2019.03.037
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.

Discovery of the Pan-genotypic Hepatitis C Virus NS3/4A Protease Inhibitor Voxilaprevir (GS-
857): A Component of Vosevi ®
9
*
,a
a
a
a
a
a
James G. Taylor , Sheila Zipfel , Kyla Ramey , Randy Vivian , Adam Schrier , Kapil K. Karki ,
a,e
a
a
a,e
a
a
Ashley Katana , Darryl Kato , Tetsuya Kobayashi , Ruben Martinez , Michael Sangi , Dustin Siegel ,
a,e
a
a,e
b,e
b,e
b,
Chinh V. Tran , Zheng-Yu Yang , Jeff Zablocki , Cheng Y. Yang , Yujin Wang , Kelly Wang ,
Katie Chan , Ona Barauskas , Guofeng Cheng , Debi Jin , Brian E. Schultz , Todd Appleby , Armando
G. Villaseñor , and John O. Link
c,e
c,e
c
c
c
d
d
a
a
b
c
d
Medicinal Chemistry, Drug Metabolism, Biology, Structural Chemistry, Gilead Sciences, 333
Lakeside Drive, Foster City, California 94404, United States
*
Corresponding author. E-mail address: james.taylor@gilead.com
e
Former Gilead Employees.
Keywords: Vosevi®, voxilaprevir, VOX, GS-9857, HCV, Hepatitis C virus, protein adducts
Abstract: Treatment of Hepatitis C virus (HCV) infection has been historically challenging due the high
viral genetic complexity wherein there are eight distinct genotypes and at least 86 viral subtypes. While
HCV NS3/4A protease inhibitors are an established treatment option for genotype 1 infection, limited
coverage of genotypes 2 and/or 3 combined with serum alanine transaminase (ALT) elevations for some
compounds has limited the broad utility of this therapeutic class. Our discovery efforts were focused on
identifying an NS3/4A protease inhibitor with pan-genotypic antiviral activity, improved coverage of
resistance associated substitutions, and a decreased risk of hepatotoxicity. Towards this goal, distinct
interactions with the conserved catalytic triad of the NS3/4A protease were identified that improved
genotype 3 antiviral activity. We further discovered that protein adduct formation strongly correlated
with clinical ALT elevation for this therapeutic class. Improving metabolic stability and decreasing
protein adduct formation through structural modifications ultimately resulted in voxilaprevir.
Voxilaprevir, in combination with sofosbuvir and velpatasvir, has demonstrated pan-genotypic antiviral
clinical activity. Furthermore hepatotoxicity was not observed in Phase 3 clinical trials with voxilaprevir,
consistent with our design strategy. Vosevi ® (sofosbuvir, velpatasvir, and voxilaprevir) is now an
approved pan-genotypic treatment option for the most difficult-to-cure individuals who have previously
failed direct acting antiviral therapy.
Graphical Abstract:
Voxilaprevir
HCV GT1–6a EC = 1.5–6.6 nM
50
Human hepatocyte Pred CL = <0.07 L/h/kg
Protein adducts = <1 pmol equivalent/mg protein

Hepatitis C virus (HCV) infection remains a global health concern with an estimated 71 million people
1
chronically infected worldwide . Untreated, chronic HCV infection can result in cirrhosis, hepatocellular
2
carcinoma and liver failure, contributing to approximately 400,000 annual deaths worldwide . Thus,
curative HCV treatment options can have a profound impact on the health of those infected. Despite this
need, treatment of HCV infection has been particularly challenging due to the high viral genetic diversity
3
wherein there are eight distinct genotypes (GTs) and at least 86 viral subtypes . Consistent with this
challenge, the initial wave of direct acting antiviral (DAA) agents were frequently most potent against
HCV GT 1 while having compromised activity against GT 2 and/or 3.
Initially we advanced multiple compounds with potent GT1 antiviral activity into the clinic including
4
5
6
inhibitors of the HCV NS3/4A protease (vedroprevir and GS-9256 ), NS5B polymerase (radalbuvir and
7
8
tegobuvir ) and NS5A protein (ledipasvir ). To address the therapeutic need to treat patients regardless of
genotype, discovering compounds with potent pan-genotypic antiviral activity became a priority. Toward
this goal, we initiated a comprehensive NS5A inhibitor program that resulted in the discovery of
9
-10
velpatasvir, which has potent picomolar activity against all gentoypes tested . Upon combination of
velpatasvir with sofosbuvir into the oral once-daily single tablet regimen (STR) Epclusa®, high sustained
virological response (SVR) rates across all genotypes have been observed[4, 5]. Despite the generally
high SVR rates obtained for the multiple HCV treatment options that are now approved, there are patients
that fail these regimens with viral resistance associated substitutions (RASs, also referred to as mutants or
1
1
variants) that are less susceptible to the individual DAA agents . Thus, identifying next generation DAA
compounds with antiviral activity against these RASs, while also having potent pan-genotypic activity,
was considered essential to help treat this emerging high unmet medical need population. Voxilaprevir,
the discovery of which described herein, is a pan-genotypic HCV NS3/4A protease inhibitor (PI) with
improved antiviral activity against common RASs. Voxilaprevir has recently been approved as a
component of the pan-genotypic STR Vosevi® (sofosbuvir, velpatasvir and voxilaprevir). Vosevi®
provides a pan-genotypic treatment option for the most difficult-to-cure patients who have failed prior
12-13
DAA therapies . Additionally, for some DAA-naïve patients, Vosevi® provides an option of
13
shortened treatment duration (8 weeks) .
NS3/4A PIs are a longstanding, important therapeutic class in the treatment of HCV infection. Reversible
covalent -ketoamide containing NS3/4A PIs, telaprevir and boceprevir, were the first DAAs approved
for the treatment of HCV GT1 infection in combination with PEG and RBV. Second generation NS3/4A
PIs containing terminal carboxylic acid or acyl sulfonamide groups, along with large aromatic
hydrophobic groups extending from the central P2 amino acid, soon followed with improved GT1 cellular
potency (Figure 1). However, reduced relative enzymatic and cellular potency against genotype 2 and/or
1
4-15
3
has limited their use in these genotypes. Serum alanine transaminase (ALT) elevations have also
1
6
been identified as a potential hepatotoxicity signal with some NS3/4A PIs , contributing to clinical
discontinuations (IDX-320) , clinical dose reductions (paritaprevir , asunaprevir , and grazoprevir ),
and/or transaminase elevation warnings in the prescribing label (paritaprevir , asunaprevir , and
grazoprevir ). Our internal discovery efforts focused on identifying a potent pan-genotypic HCV
NS3/4A PI with improved coverage of RASs and a decreased risk of hepatotoxicity.
17
18
19
20
2
1
22
2
3

Figure 1. Representative HCV NS3/4A PIs
The HCV NS3/4A protein consists of an N-terminal serine protease domain and a C-terminal helicase
domain. The NS3/4A protease cleaves the viral polyprotein at four distinct sites utilizing a classic
catalytic triad (S139, H57 and D81) to generate the viral machinery required for viral replication. While
the catalytic triad is conserved in all reported NS3/4A sequences, RASs elsewhere in the NS3/4A protease
2
4
domain have rapidly emerged (including amino acid positions 155 and 168) . In GT1, an electrostatic
interaction between R155 and D168 is formed that creates an important surface for interaction with
aromatic hydrophobic P2 groups present on many NS3/4A PIs. Amino acid substitutions at either
position therefore have a significant impact on NS3/4A PI binding affinity. Amino acid 168 also plays an
important role in GT3 wherein the wild-type Q168 contributes to a significant loss in GT3 potency for
15
most NS3/4A PIs . Removing steric bulk on the P2 region proximal to amino acid positions 155 and
68, along with modifications to the macrocycle, were strategies utilized in the discovery of grazoprevir .
Even with this design, grazoprevir loses 2-3 orders of magnitude in potency against GT3 compared to
25
1
1
5, 26
GT1 in both biochemical and cellular assays
. To improve both GT3 potency and activity against
common GT1 RASs, our internal discovery efforts were directed at modifications to the central
pyrrolidine core with the goal of minimizing the dependence on interactions with the D168 and R155
residues and enhancing interactions with conserved catalytic triad region of the HCV NS3/4A protease.
We targeted direct interactions with the catalytic triad as it is conserved across genotypes and, due to its
essential role in enzymatic activity, the catalytic triad is not a site of viral polymorphism or resistance.
Realizing that the conserved H57 of the catalytic triad is positioned adjacent to the 3-position of the
central pyrrolidine we investigated the potential of introducing alkyl, substituted alkyl and cycloalkyl

groups to establish a hydrophobic interaction with H57 (Figure 2). Consistent with this design principle,
methyl substituted pyrrolidine 11 (Table 1) improved cellular potency nearly 4-fold vs the unsubstituted
comparator 10 (20 nM vs 74 nM, respectively). Ethyl substituted compound 12 possessed similarly
improved cellular potency (EC = 23 nM), while also improving biochemical potency 2.5 fold vs
50
compound 10 (67 pM vs 165 pM, respectively). A crystal structure of inhibitor 12 bound to the NS3/4A
protease highlights that the ethyl substituent is positioned in van der Waals contact with the catalytic H57
residue as we intended (Figure 2). Extending beyond ethyl to n-propyl (13) was also beneficial
compared to the unsubstituted pyrrolidine 10 (2-fold improvement in cellular potency), but was inferior in
potency to the methyl or ethyl substituted pyrrolidine containing compounds 11 and 12. Consistent with
space constraints observed in the crystal structure, cyclopropyl (14), iso-butyl (15) or benzyl (16)
substitution led to significantly compromised GT3 potency (all >1000 nM in GT3 EC assay). Since
5
0
both the 3-methyl and 3-ethyl pyrrolidine substitutions contributed to significantly improved GT3
potency, these series were further pursued in parallel.
Figure 2. Crystal structure of 3-ethyl substituted pyrrolidine compound 12 bound to GT3 surrogate (GT1
D168Q) NS3/4A protease (PDB code: 6NZV)
Table 1
Genotype 3 biochemical and cellular potency of 3-substituted pyrrolidine core containing macrocycles
2
a
b
Cpd
R
GT3 Ki
nM)
GT3a EC50
(nM)
(
1
0
1
H
0.165
0.100
74
20
1
Me

1
2
3
Et
0.067
0.120
23
36
1
n-Pr
1
4
5
4.9
1164
1125
1
0.630
16
0.870
1107
a
Ki determined by enzymatic assay using an HCV genotype 3a NS3/4A protein
b
EC determined by cell based assay using RLuc cells harboring subgenomic genotype 3a replicon
5
0
Three synthetic routes were employed to prepare pyrrolidine cores derivatized at the 3-position with
2
7
alkyls, substituted alkyls, and cycloalkyls (Scheme 1). Enaminone 17 served as a common precursor
utilized for the synthesis of both alkyl and substituted alkyl derivatives, with the synthesis of both 3-
methyl and 3-ethyl substituted pyrrolidines highlighted in Scheme 1. Hydrogenation of enaminone 17
with Pd/C yielded the desired 3-methyl pyrrolidine 18 along with the undesired diastereomer in a ratio of
2:1, which was enriched to 93:7 following recrystallization. Ketone 18 was then stereoselectively
reduced with the CBS catalyst, affording 4-hydroxy-3-methyl pyrrolidine 19 as a single diastereomer.
Synthesis of 3-ethyl-4-hydroxy-pyrrolidine 21 proceeded through initial treatment of enaminone 17 with
MeMgBr to generate enone 20. Stereoselective CBS reduction of the ketone followed by hydrogenation
of the olefin generated 3-ethyl-4-hydroxy-pyrrolidine 21. Synthesis of 3-cyclopropyl-4-hydroxy-
pyrrolidine 27 initiated from 3-hydroxy-pyrrolidine 22, which was first esterified and protected with a
trityl group to generate 23. Oxidation of 23 with NMO and TPAP, followed by treatment with LiHMDS
and Comins’ reagent, selectively generated enol triflate 24. Negishi coupling of 24 using cyclopropyl
zinc bromide furnished intermediate 25. Amine protecting group exchange yielded 26, which underwent
stereoselective hydroboration-oxidation to provide 3-cyclopropyl-4-hydroxy-pyrrolidine 27.

Scheme 1. (a) H , Pd/C, acetone, 62% following recrystallization at 93:7 d.r; (b) BH •Me S, (R)-Me-
2
3
2
CBS, THF, -78 °C to rt, 68% as single diastereomer; (c) MeMgBr, MeTHF/toluene, -12 °C; (d) (R)-Me-
CBS, BH -PhNEt , 80-90% over 2 steps; (e) H , Pd/C, Cs CO , IPAc, 80%; (f) i. SOCl , MeOH; ii. TrCl,
3
2
2
2
3
2
Et N, DCM, 0 °C to rt, 72% over 2 steps; (g) NMO, TPAP, 4 Å MS, DCM, 81%; (h) N-(5-chloro-2-
3
pyridyl)bis(trifluoromethanesulfonimide), LiHMDS, THF, -78 °C to rt, 63%; (i) cyclopropyl zinc
bromide, Pd(Ph ) , THF, 50 °C, 60%; (j) i. HCl, MeOH, DCM; ii. Boc O, DMAP, Et N, DCM, 68% over
3
4
2
3
2
steps; (k) BH •Me S, H O , NaOH, THF, H O, 45%.
3 2 2 2 2
The general route to prepare 3-substituted pyrrolidine core containing macrocyclic NS3/4A PIs is shown
in Scheme 2. Base promoted S Ar reaction between 3-ethyl-4-hydroxy-pyrrolidine 21 and quinoxaline
n
28 generated intermediate 29, which was then subjected to acidic Boc deprotection to afford amine 30.
Amide bond formation with acid 31 employing standard peptide coupling conditions yielded 32, which
underwent ring closing cross-metathesis using Zhan 1b catalyst to generate macrocycle 33.
Hydrogenation of the resulting macrocyclic olefin over Rh/alumina followed by t-butyl ester deprotection
afforded acid 34. Coupling of the cyclopropyl sulfonamide 35 using HATU completed the synthesis of
the 3-ethyl-pyrrolidine containing macrocyclic NS3/4A PI 36.
Scheme 2. Reagents and conditions. (a) Cs CO , MeCN, 85 °C, 68%; (b) HCl, dioxane, rt, 75-100% (c)
2
3
HATU, DIPEA, DMF, rt, 74%; (d) Zhan 1b catalyst, DCE, 80 °C, 86%; (e) Rh/Alumina, H , DCM, 86-
2
88%; (f) TFA, DCM, 95–97%; (g) HATU, DIPEA, DMF, 70–87%.
After our initial success in improving GT3 activity, we turned our attention to addressing a toxicity
concern that plagues some NS3/4A PIs. Antiviral therapy has been associated with ALT elevations, as a

potential biomarker for hepatotoxicity, contributing to label warnings and recommendations for ALT
monitoring for a number of NS3/4A PIs. While drug-induced hepatotoxicity is always a concern, the
concern is heightened when treating a liver disease. A number of factors have been associated with drug-
induced liver injury (DILI) including reactive metabolite formation, mitochondrial toxicity, host immune-
2
8-32
response pathways and inhibition of biliary transporters . Consistent with reactive metabolites
contributing to ALT elevation with NS3/4A PIs, co-administration of the cytochrome P450 3A (CYP3A)
inhibitor ritonavir (to suppress reactive metabolite formation) with danoprevir (2) resulted in reduced
33-34
clinical ALT elevations . To further investigate the association of reactive metabolites with ALT
3
14
elevation across multiple NS3/4A PIs, we synthesized and then evaluated radiolabeled ( H or C)
paritaprevir, danoprevir, asunaprevir, IDX-320, vaniprevir, vedroprevir, grazoprevir, simeprevir and
faldaprevir in protein adduct assays and correlated these results with reports of clinical ALT elevations.
Overall, protein adduct levels had a strong correlation with clinical ALT elevations (Table 3). All
compounds generating >50 pmol equivalent/mg protein adducts experienced clinical ALT elevations,
while four of the five compounds generating <50 pmol equivalent/mg protein adducts did not show
significant ALT elevations. These studies are consistent with the formation of reactive metabolites as an
important correlator with clinical ALT elevations and potential hepatotoxicity. We were additionally
concerned the reactive metabolites generated in an HCV infected patient’s liver might be an additional
driver of idiosyncratic drug reactions based on the “danger hypothesis” (that cell stress is a factor in
idiosynchratic drug reactions).[36] Therefore, minimizing protein adduct formation became an essential
element in our optimization efforts.
Table 3
Protein adduct formation and reported ALT elevation for representative HCV NS3/4A PIs.
35
Cpd
Protein adducts
pmol equivalent/mg protein)
ALT elevation
(
18, 21, 36
33-34
19, 22
17
parataprevir (1)
danoprevir (2)
asunaprevir (3)
IDX-320 (4)
295
202
114
96
Yes
Yes
Yes
Yes
3
7
vaniprevir (5)
vedroprevir (6)
grazoprevir (7)
simeprevir (8)
faldaprevir (9)
29
No
No
3
8-39
0, 23
26
2
24
Yes
4
0
22
No
No
41
15
The cyclopropyl olefin P1 moiety is a common structural feature on HCV NS3/4A PIs (Figure 1), with
all five HCV PIs that demonstrate clinical ALT elevation containing this olefin functional group either as
a terminal vinyl or a macrocyclic olefin (olefin highlighted in red in Figure 1). Mechanistically, we
hypothesized this cyclopropyl olefin could undergo metabolic activation to form a reactive epoxide that
could then react with nucleophiles (e.g. proteins and glutathione), contributing to transaminase
4
2
elevation . To explore this mechanistic hypothesis, we synthesized the vinyl containing compound 37 in
our methyl pyrrolidine series (Table 4) and then incubated this compound in human liver microsomes in
the presence of glutathione followed by metabolite identification (Met ID). Two of the metabolites
generated supported our mechanistic hypothesis; one was direct oxidation at P1 (consistent with an
epoxide being formed from the olefin) and the second was olefin oxidation followed by epoxide opening
with glutathione. Additional olefin P1 containing molecules were also evaluated in human microsomal
Met ID assays with similar resultant metabolite profiles involving olefin activation (data not shown).
Compound 37 was also evaluated in a human liver microsomal (HLM) metabolic stability assay to

determine the predicted clearance (Pred CL). Moderate metabolic instability was observed (Pred CL =
.64 L/h/kg). The P1 vinyl was also synthesized in the ethyl pyrrolidine (41) series and similarly
0
demonstrated poor metabolic stability (Pred CL = 0.84 L/h/kg). Of particular concern, compound 41
generated 117 pmol equivalent/mg protein adducts further highlighting the potential risk with this
functional group. Thus, identifying a replacement for the P1 vinyl with acceptable potency and an
improved drug metabolism and pharmacokinetic profile became our next challenge.
The natural peptide substrates of the HCV NS3/4A protease contain a cysteine or threonine at P1,
4
3
highlighting the small nature of the S1 pocket . While replacing the P1 vinyl (37) with an ethyl (38) on
the 3-methyl pyrrolidine series improved the predicted clearance from 0.64 to 0.40 L/h/kg, the larger ethyl
group resulted in an approximate 5-fold loss in both biochemical and cellular GT3 potency. We
considered that improved oxidative metabolic stability with replacement of the vinyl group at P1 was an
initial indicator of a positive direction for removing reactive metabolites. Significantly improved
metabolic stability could be realized by adding a fluoro (39) or difluoro (40) to the P1 ethyl group (Pred
CL of 0.20 and <0.17 L/h/kg, respectively), but at the expense of further compromised GT3 potency. A
similar trend in metabolic stability improvement and GT3 potency loss was obtained in comparing the P1
vinyl (41) to P1 ethyl (42) on the 3-ethyl pyrrolidine series. Compound 43, wherein P1 is truncated to a
methyl, had improved metabolic stability vs the vinyl 41, but lost significant GT3 biochemical potency
(
~10-fold). Adding two fluorines to the P1 methyl in an attempt to further improve metabolic stability
afforded compound 44, which indeed had substantially improving metabolic stability (Pred CL = 0.20
L/h/kg) while also maintaining the potency of the P1 vinyl containing compound 41. Thus, the P1
difluoromethyl substituent appeared to be a viable replacement for the reactive vinyl group wherein GT3
potency was maintained and metabolic stability was improved with compound 44.
Table 4
1
2
Impact of P1 (R ) and P2 (R ) modifications on GT3 potency and human microsomal stability.
P2
P1’
P1
P3
1
2
3
a
b
c
Cpd
R
R
R
GT3 Ki (nM) GT3a EC50
Pred CL HLM
(L/h/kg)
(nM)
3
7
8
Vinyl
Et
Me
Me
H
H
0.094
0.530
41
0.64
0.40
3
188
3
9
0
Me
Me
H
H
1.3
419
375
0.20
4
0.92
<0.17
4
1
2
Vinyl
Et
Et
Et
H
H
0.081
0.563
20
0.84
0.44
4
272

4
3
4
Me
CF
Et
Et
Me
H
0.838
0.043
109
27
0.67
0.20
4
2
H
a
Ki determined by enzymatic assay using an HCV genotype 3a NS3/4A protein
b
EC determined by cell based assay using RLuc cells harboring subgenomic genotype 3a replicon
5
0
c
Pred CL HLM = human metabolic clearance predicted from microsomal stability
After finding the P1 difluoromethyl provided good potency and stability as a vinyl replacement, we
sought to understand the effect of P1 groups on preclinical pharmacokinetics. We envisioned combining
our NS3/4A PI with sofosbuvir and velpatasvir in an STR, which would require pharmacokinetics
consistent with once-daily dosing and a projected daily dose amenable to combination with sofosbuvir
(
400 mg) and velpatasvir (100 mg) into a single tablet. As the liver is the primary site of HCV
replication, we also considered preferential distribution to the liver to be a favorable attribute. The
combination of 3-substituted pyrrolidines with P1 modifications had a profound impact on
pharmacokinetic properties. While P1 vinyl containing compound 41 had a favorable volume of
distribution (Table 5, Vss 1.9 L/h/kg) and high liver concentration (20.9 M at 6 h post dose) in rats, the
poor metabolic stability in human microsomes (Pred CL = 0.87 L/h/kg) and high level of protein adduct
formation (117 pmol/mg protein) precluded advancement of this compound. Replacement of the P1 vinyl
(
41) with ethyl (42) improved the Vss, half-life and liver distribution in rats, but as already noted, P1
ethyl substitution was deleterious to GT3 potency. Much to our dismay, the P1 difluoromethyl group (44)
that we found to both improve metabolic stability and maintain GT3 potency also came with a detrimental
impact to V (1.0 L/kg), half-life (0.8 h) and liver concentration (1.2 M at 6 h) compared to the vinyl
ss
comparator 41. We hypothesized that decreased hydrophobicity surrounding the polar acylsulfonamide
3
motif may contribute to the reduced Vss. While modifications to the R position on P1’ would appear a
4
4
45
logical position to derivatize, our previous efforts on a related scaffold as well as prior literature
highlighted limited tolerance for P1’ derivatization and that small hydrophobic groups were generally
3
preferred. Consistent with expectations, P1’ analogs wherein R was ethyl, ally, -CH CH OH, -
2
2
CH CH OBn, or -CH CN all lost >3-fold cellular activity when investigating the unsubstituted
2
2
2
3
pyrrolidine core (data not shown). Gratifyingly, R methyl substitution was tolerated for potency on the
unsubstituted pyrrolidine core. The P1’ methylcyclopropyl substitution was therefore synthesized on our
highly optimized scaffold yielding 3-ethyl pyrrolidine (12) and 3-methyl pyrrolidine (11), both having
potent GT3 cellular potency (23 and 20 nM) and low predicted metabolic clearance (Pred CL of 0.25 and
0.18 L/h/kg). In agreement with our design, addition of a methyl to P1’ in compound 12 increased the
LogD and improved V (1.7 L/kg), MRT (1.4 h) and liver concentration (14.0 M at 6 h) vs the des-
ss
methyl P1’ compound 44. The 3-ethyl pyrrolidine (12) proved to have enhanced liver pharmacokinetic
properties vs 3-methyl pyrroline (11) and was therefore selected for further investigation (liver
concentration at 6 h of 3.1 vs 14.0 M, respectively). Compound 12 was further profiled in protein adduct
studies to probe our hypothesis that replacement of the P1 vinyl with a metabolically blocked
difluoromethyl would reduce protein adducts. Indeed, compound 12 had a >10-fold reduction in
radiolabeled protein adduct formation compared to the vinyl containing comparator 41 (10 vs 117
pmol/mg protein, respectively).
Table 5
1
2
3
R , R and R effect on drug metabolism and pharmacokinetic properties

P2
P1’
P1
1
2
3
Cpd
R
R
R
GT3a
EC50
LogD
Pred CL
HLM
(L/h/kg)
Protein adducts Rat CL Rat Vss Rat t1/2
Rat F Rat [liver]
a
(pmol equivalent (L/h/kg) (L/kg)
/mg protein)
(h)
(%)
@6h (µM)
(nM)
41
42
44
12
11
Vinyl Et
H
H
H
20
4.1
4.6
3.5
3.8
3.5
0.84
117
0.88
0.67
2.0
1.9
3.1
1.0
1.7
2.2
2.2
5.0
0.8
1.4
1.3
20.9
30.2
1.2
Et
Et
Et
Et
272
27
0.44
100
36
CF
CF
CF
2
H
2
H
2
H
0.20
Me 23
0.25
10
1.3
31
14.0
3.1
Me Me 20
0.18
2.1
100
a
Pred CL HLM = predicted clearance from human liver microsomes
With compound 12 we had improved GT3 potency by introduction of the 3-ethyl pyrrolidine core,
increased metabolic stability and decreased radiolabeled protein adduct formation by replacing the P1
vinyl with a difluoromethyl group, and improved the Vss and liver concentration at 6 h by introduction of
the P1’ methyl. To realize our goal of a low dose once-daily administered compound amenable to an
STR with sofosbuvir and velpatasvir, we next focused on further improving the predicted human
metabolic stability and GT3 potency. A human microsomal Met ID study was performed on compound
12 wherein three distinct hydroxylated metabolites were observed on the P2 to P4 macrocycle fragment.
We hypothesized the benzylic methylene connected to the 3-position of the quinoxaline would be the
most readily oxidized position. We were further concerned that if this position were to be hydroxylated a
quinone-methide could be formed, based on the position of the methoxy group on the P2 quinoxaline, and
might be another source of protein adduct formation. Therefore, a series of metabolic blocks at this
position were synthesized to determine if additional gains in metabolic stability could be realized. Upon
binding to the protease, this benzylic methylene is also positioned directly above the hydrophobic alkyl
portion of R155 so there was also the potential for van der Waals interactions with this amino acid. The
two cyclopropyl diastereomer compounds 45 and 46 had slightly improved microsomal metabolic
stability consistent with blocking of benzylic methylene oxidation, but both compounds lost significant
potency in both biochemical (>25 fold) and cellular assays (>5 fold). The benzylic monofluoro
substituted diastereomeric mixture (47) maintained GT3 potency and also had a minor improvement in
human microsomal stability. Benzylic difluorination of the P2 quinoxaline with an adjacent olefin (a
consequence of the ring closing metathesis reaction) afforded compound 48. While compound 48 had an
approximate 4-fold improvement in GT3 potency vs the unsubstituted comparator 12 (5.3 vs 23 nM,
respectively), the human microsomal metabolic stability was unchanged (Pred CL = 0.25 L/h/kg). To
mitigate concerns the macrocyclic olefin may contribute to metabolic instability and potentially generate
reactive metabolites, the olefin was reduced to the corresponding alkylene, affording compound 36.
Indeed, compound 36 had improved metabolic stability in human microsomes (Pred CL <0.17 L/h/kg),
which was further supported by a very low Pred Cl of <0.07 L/h/kg in cryopreserved human hepatocytes.
Radiolabeled protein adducts were also reduced to <1 pmol equivalent/mg protein for compound 36,
supporting a low risk of clinical ALT elevation and hepatotoxicity. This decrease in adduct formation is

consistent with the possibility of quinone-methide formation if benzylic oxidation were to occur on
compound 12. Compound 36 was also potent in both GT3 biochemical (63 pM) and cellular (6.1 nM)
assays. The improved GT3 potency for 36 compared to 12 may result from a hydrophobic interaction
between the difluoro substitution on the macrocycle and the alkyl portion of R155, as highlighted in the
co-crystal structure (Figure 3).

Table 6
Macrocycle modifications (L) to improve human microsomal stability and GT3 potency
d
d
12
45
46
47
48
36
a
GT 3 K
i
(nM)
0.067
23
1.7
9.4
0.13
35
0.053
5.3
0.063
6.1
b
GT 3a EC50 (nM)
128
0.21
829
0.23
c
Pred CL HLM (L/h/kg)
0.25
0.23
0.25
<0.17
a
Ki determined by enzymatic assay using an HCV genotype 3a NS3/4A protein
b
c
EC determined by cell based assay using RLuc cells harboring subgenomic genotype 3a replicon
5
0
Pred CL HLM = predicted clearance from human liver microsomes
Cyclopropyl diastereomers synthesized from trans olefin. Absolute stereochemistry for cyclopropyl
d
diasteromers 45 and 46 not determined.
Compound 36
GS-9857
voxilaprevir
Figure 3. Difluoromethylene/R155 interaction in crystal structure of compound 36 bound to GT3
surrogate (GT1 D168Q) NS3/4A protease (PDB code: 6NZT)
Compound 36 demonstrated potent antiviral activity across genotype 1-6a cellular replicon EC assays
5
0
(
1.5-6.6 nM), with less than a 2-fold shift between GT1a (3.9 nM) and GT3 (6.1 nM) (Table 7). Our
optimization strategy invoking interactions with the conserved catalytic triad also proved beneficial

against some common GT1 RASs including Q80K, R155K, and D168E, with potency shifts of less than
46
two-fold (Table 7) . As is common with HCV NS3/4A PIs, a high shift in potency was observed with
the A156T variant as this amino acid is positioned below the core pyrrolidine motif present in nearly all
approved NS3/4A PIs. Despite a high level of protein binding across species (>99%), 36 had a V of
ss
0.80-0.94 L/kg and the distribution of 36 into liver was high with liver to plasma ratios of >50-fold across
rat, dog and cynomolgus monkey (Table 8). Additional in vitro assays were performed to assess the
potential for preferential distribution to the liver in humans. Rapid time-dependent uptake of compound
36 into cryopreserved human hepatocytes as well as organic-anion-transporting polypeptide (OATP)
1B1/1B3 over-expressing Chinese hamster ovary (CHO) cells was also observed, supporting an
expectation for preferential liver distribution of compound 36 in humans. Compound 36, now known as
voxilaprevir (GS-9857, VOX), was selected for development due to its potent antiviral activity across
genotypes, improved activity against common GT1 RASs, decreased risk of hepatotoxicity, projected
long half-life in humans, and potential for combination in an STR.
Table 7
Compound 36 pan-genotypic potency and activity against GT1 NS3/4A RAS
Genotype (wild-type)
RAS
1b
1a
1b
2a
2b
3a
4a
5a
6a
1a
1a
1a/1b
Q80K
R155K
D168E
A156T
a
EC50 (nM)
3.9
3.3
3.7
6.6
6.1
4.9
1.9
1.5
3.1
2.7
4.0
>500
a
EC determined by cell based assays using RLuc cells harboring subgenomic genotype replicons of the
5
0
corresponding genotype
Table 8
Drug metabolism and pharmacokinetic profile of compound 36
a,b
In vitro
In vivo
%
plasma
Free
Pred
CL*
CL
(L/h/kg)
V
(L/kg)
ss
t
½
F / F
(%)
a
Liver:plasma
ratio, @2h
(hr)
(L/h/kg)
Rat
0.42
0.28
0.91
0.46
0.39
0.17
1.01
<0.07
0.82
0.19
0.89
0.84
0.94
0.80
0.93
4.2
83 / 100
27 / 30
10 / 23
>500x
55x
Dog
Monkey
Human
2.1
100x
a
Intravenous administration in rats at 1.0 mg/kg, dogs at 0.5 mg/kg, and in cyno at 0.5 mg/kg
Oral administration in rats at 3.0 mg/kg, dogs at 1.0 mg/kg, and in cynomolgus monkey at 1.0 mg/kg
b
*Pred CL = predicted clearance from liver microsomes of rat, dog and monkey, and from hepatocytes
from human
F = bioavailability
Fa = fraction absorbed
47
Voxilaprevir was advanced into a Phase 1 double-blind, placebo-controlled, single-and multiple-dose
study to evaluate the safety and pharmacokinetics at 30, 100 and 300 mg. Overall, voxilaprevir was safe
and well tolerated at all doses with no clinically significant abnormalities. Consistent with preclinical
optimization efforts focused on high metabolic stability, voxilaprevir had long median steady-state half-

lives of 28 to 41 h, supporting once-daily dosing. Dose-proportional exposure was observed over the
dose range tested after single- or multiple-dose administration (Figure 4). In a Phase 1b trial evaluating
voxilaprevir as monotherapy in GT1-4 HCV infected patients, potent antiviral activity was observed with
4
8
median maximal viral load reductions of >3 log IU/mL at the 100-mg dose across all genotypes .
1
0
Therefore, the 100-mg dose of voxilaprevir was selected for further evaluation in combination with
sofosbuvir (400 mg) and velpatasvir (100 mg) in Phase 2 and 3 clinical studies.
Figure 4. Voxilaprevir (VOX) PK in healthy volunteers after seventh daily dose.
The main goals of the clinical development program investigating voxilaprevir in combination with
sofosbuvir and velpatasvir (SOF/VEL/VOX) were to 1) demonstrate that SOF/VEL/VOX is a highly
effective and safe therapy for the most difficult-to-cure patients who had previously failed DAA-based
regimens and 2) assess the possibility of shortening treatment duration for DAA-naïve patients. Two
Phase 3 studies (POLARIS-1 and POLARIS-4) were conducted to assess the antiviral efficacy, safety,
and tolerability of SOF/VEL/VOX administered for 12 weeks in DAA-experienced patients with and
without compensated cirrhosis. POLARIS-1 was a placebo controlled study investigating
SOF/VEL/VOX in NS5A inhibitor experienced patients harboring genotype 1 to 6 chronic HCV
infection. POLARIS-4 was an open-label study assessing SOF/VEL/VOX and SOF/VEL in GT1-4
DAA-experienced patients who had not previously received a NS5A inhibitor. In the combined
POLARIS-1 and POLARIS-4 studies, the overall SVR rate was 97% in DAA-experienced patients treated
with SOF/VEL/VOX, and high SVR rates were observed irrespective of genotype, prior DAA regimen, or
49
presence of pre-existing RASs . To assess whether the addition of VOX to the SOF/VEL regimen would
allow shortening treatment duration for DAA-naïve patients, the POLARIS-2 and POLARIS-3 studies
were conducted. POLARIS-2 and POLARIS-3 were open-label studies which assessed the antiviral
efficacy, safety, and tolerability of 8 weeks of SOF/VEL/VOX compared with 12 weeks of SOF/VEL in
DAA- naïve patients with chronic HCV infection. Overall, 611 DAA- naïve patients with and without
compensated cirrhosis were treated with 8 weeks of SOF/VEL/VOX and the overall SVR12 rate was
5
0
9
5% . Treatment-emergent resistance following treatment with SOF/VEL/VOX was uncommon in both
the DAA-experienced and DAA-naïve populations consistent with the regimen having a high barrier to
resistance. In the combined Phase 3 studies, over 1000 GT 1-6 HCV-infected patients were treated with
SOF/VEL/VOX as a once-daily STR. No treatment-related serious adverse events were reported.
Gratifyingly, and consistent with the preclinical optimization efforts of voxilaprevir directed at reducing
reactive metabolites as described herein, no safety signal of hepatotoxicity was observed.

In summary, discovery efforts focused on forming favorable binding interactions with the conserved
catalytic triad of the HCV NS3/4A protease resulted in compounds with improved genotype 3 potency.
Additional modifications to the macrocycle resulted in further improved potency. To mitigate risk of
clinical hepatotoxicity associated with ALT elevations, distinct structural modifications were introduced
that drove protein adduct formation to the lower limit of detection (<1 pmol equivalent/mg protein).
Consistent with this strategy, hepatotoxicity was not been observed in Phase 3 clinical trials with
voxilaprevir involving over 1000 patients. The potent pan-genotypic activity, safety, and low once-daily
dose has enabled voxilaprevir to be combined with sofosbuvir and velpatasvir into a once daily STR
treatment for HCV infection known as Vosevi®, approved mid-2017 in the both the US and Europe.
Vosevi® represents a treatment option for some of the most difficult to cure patient populations who have
failed previous DAA containing regimens.
Acknowledgments
The authors thank Johnny Lee, Yang Tian, and Ruoyu Gang for cellular activity measurements, Elham
Nejati for compound formulation, and Nina Soltero and Kathy Brendza for HRMS determination. The
authors would also like to thank the patients and their families as well as study site staff who participated
in the clinical trials in support of Vosevi.

References:
1
.
Polaris Observatory, H. C. V. C., Global prevalence and genotype distribution of hepatitis C virus infection
in 2015: a modelling study. Lancet Gastroenterol Hepatol 2017, 2 (3), 161-176.
2
3
.
.
WHO, Global hepatitis report, 2017. 2017.
Hedskog, C.; Bhardwaj, N.; Chang, S.; Parhy, B.; Martin, R.; Zeuzem, S.; Moreno, C.; Shafran, S.; Borgia,
S.; Hyland, R. H.; McNally, J.; Mo, H.; Svarovskaia, E. S., Identification of Novel HCV Genotype and Subtypes in
Patients Treated with Sofosbuvir-Based Regimens. Journal of Hepatology 2017, 66.
4
.
Sheng, X. C.; Appleby, T.; Butler, T.; Cai, R.; Chen, X.; Cho, A.; Clarke, M. O.; Cottell, J.; Delaney, W. E.
t.; Doerffler, E.; Link, J.; Ji, M.; Pakdaman, R.; Pyun, H. J.; Wu, Q.; Xu, J.; Kim, C. U., Discovery of GS-9451: an
acid inhibitor of the hepatitis C virus NS3/4A protease. Bioorganic & medicinal chemistry letters 2012, 22 (7),
2
5
629-34.
.
Sheng, X. C.; Casarez, A.; Cai, R.; Clarke, M. O.; Chen, X. W.; Cho, A.; Delaney, W. E.; Doerffler, E.; Ji,
M. Z.; Mertzman, M.; Pakdaman, R.; Pyun, H. J.; Rowe, T.; Wu, Q. Y.; Xu, J.; Kim, C. U., Discovery of GS-9256:
A novel phosphinic acid derived inhibitor of the hepatitis C virus NS3/4A protease with potent clinical activity.
Bioorganic & medicinal chemistry letters 2012, 22 (3), 1394-1396.
6
.
Lawitz, E.; Hazan, L.; Gruener, D.; Hack, H.; Backonja, M.; Hill, J.; German, P.; Dvory-Sobol, H.; Jain,
A.; Arterburn, S.; Watkins, W. J.; Rossi, S.; McHutchison, J.; Rodriguez-Torres, M., Gs-9669, a Novel Ns5b Non-
Nucleotide Site Ii Inhibitor, Demonstrates Potent Antiviral Activity, Favorable Safety Profile and Potential for
Once-Daily Dosing. Journal of Hepatology 2012, 56, S471-S471.
7
.
Hebner, C. M.; Han, B.; Brendza, K. M.; Nash, M.; Sulfab, M.; Tian, Y.; Hung, M.; Fung, W.; Vivian, R.
W.; Trenkle, J.; Taylor, J.; Bjornson, K.; Bondy, S.; Liu, X.; Link, J.; Neyts, J.; Sakowicz, R.; Zhong, W.; Tang, H.;
Schmitz, U., The HCV non-nucleoside inhibitor Tegobuvir utilizes a novel mechanism of action to inhibit NS5B
polymerase function. PloS one 2012, 7 (6), e39163.
8
.
Link, J. O.; Taylor, J. G.; Xu, L. H.; Mitchell, M.; Guo, H. Y.; Liu, H. T.; Kato, D.; Kirschberg, T.; Sun, J.
Y.; Squires, N.; Parrish, J.; Keller, T.; Yang, Z. Y.; Yang, C.; Matles, M.; Wang, Y. J.; Wang, K.; Cheng, G. F.;
Tian, Y.; Mogalian, E.; Mondou, E.; Cornpropst, M.; Perry, J.; Desai, M. C., Discovery of Ledipasvir (GS-5885): A
Potent, Once-Daily Oral NS5A Inhibitor for the Treatment of Hepatitis C Virus Infection. Journal of medicinal
chemistry 2014, 57 (5), 2033-2046.
9
.
Link, J. O., Case History: Curing Genotype 1-6 Hepatitis C: Velpatasvir, the NS5A inhibitor in Epclulsa
and Vosevi. Med. Chem. Rev. 2018, 53, 541-564.
0. Link, J. O.; Taylor, J. G.; Trejo-Martin, T. A.; Kato, D.; Katana, A. A.; Krygowski, E. S.; Yang, Z.-Y.;
1
Zipfel, S.; Cottell, J. J.; Bacon, E. M.; Tran, C. V.; Yang, C. Y.; Wang, Y.; Wang, K.; Zhao, G.; Cheng, G.; Yian,
Y.; Gong, R.; Lee, J.; Yu, M.; Gorman, E.; Mogalian, E.; Perry, J., Bioorganic & medicinal chemistry letters
submitted.
1
1.
Regimens. Gastroenterology 2016, 151 (1), 70-86.
2. Vosevi (sofosbuvir, velpatasvir, and voxilaprevir) [package insert]. Gilead Sciences, Inc, Foster City, CA,
USA; July 2017.
3. Vosevi (sofosbuvir, velpatasvir, and voxilaprevir) [EMA Summary of Product Characteristics]. Gilead
Sciences Ireland UC, Carrigtohill, County Cork, Ireland; September 2017.
4. Gottwein, J. M.; Scheel, T. K.; Jensen, T. B.; Ghanem, L.; Bukh, J., Differential efficacy of protease
inhibitors against HCV genotypes 2a, 3a, 5a, and 6a NS3/4A protease recombinant viruses. Gastroenterology 2011,
Pawlotsky, J. M., Hepatitis C Virus Resistance to Direct-Acting Antiviral Drugs in Interferon-Free
1
1
1
1
1
41 (3), 1067-79.
5. Soumana, D. I.; Kurt Yilmaz, N.; Ali, A.; Prachanronarong, K. L.; Schiffer, C. A., Molecular and Dynamic
Mechanism Underlying Drug Resistance in Genotype 3 Hepatitis C NS3/4A Protease. J Am Chem Soc 2016, 138
36), 11850-9.
6. Mishra, P.; Chen, M., Direct-Acting Antivirals for Chronic Hepatitis C: Can Drug Properties Signal
Potential for Liver Injury? Gastroenterology 2017, 152 (6), 1270-1274.
7. Parsy, C. C.; Alexandre, F. R.; Bidau, V.; Bonnaterre, F.; Brandt, G.; Caillet, C.; Cappelle, S.; Chaves, D.;
(
1
1
Convard, T.; Derock, M.; Gloux, D.; Griffon, Y.; Lallos, L. B.; Leroy, F.; Liuzzi, M.; Loi, A. G.; Moulat, L.; Chiara,
M.; Rahali, H.; Roques, V.; Rosinovsky, E.; Savin, S.; Seifer, M.; Standring, D.; Surleraux, D., Discovery and
structural diversity of the hepatitis C virus NS3/4A serine protease inhibitor series leading to clinical candidate
IDX320. Bioorganic & medicinal chemistry letters 2015, 25 (22), 5427-36.

1
8.
Poordad, F.; Lawitz, E.; Kowdley, K. V.; Cohen, D. E.; Podsadecki, T.; Siggelkow, S.; Heckaman, M.;
Larsen, L.; Menon, R.; Koev, G.; Tripathi, R.; Pilot-Matias, T.; Bernstein, B., Exploratory study of oral combination
antiviral therapy for hepatitis C. The New England journal of medicine 2013, 368 (1), 45-53.
1
9.
Bronowicki, J. P.; Pol, S.; Thuluvath, P. J.; Larrey, D.; Martorell, C. T.; Rustgi, V. K.; Morris, D. W.;
Younes, Z.; Fried, M. W.; Bourliere, M.; Hezode, C.; Reddy, K. R.; Massoud, O.; Abrams, G. A.; Ratziu, V.; He,
B.; Eley, T.; Ahmad, A.; Cohen, D.; Hindes, R.; McPhee, F.; Reilly, B.; Mendez, P.; Hughes, E., Randomized study
of asunaprevir plus pegylated interferon-alpha and ribavirin for previously untreated genotype 1 chronic hepatitis C.
Antivir Ther 2013, 18 (7), 885-93.
2
0.
Caro, L. D., L.; Huang, S.; Wenning, L.; Su, J.; Hwang, P.; Valesky, R.; Gilbert, C.; Gress, J.; Klaassen, F.;
Gendrano, I. N.; Brunhofer, J.; Cooreman, M.; Huisman, J.; Mobashery, N. In Relationship Between Transaminase
Levels and Plasma Pharmacokinetics Following Administration of MK-5172 with Pegylated Interferon Alfa-2b and
Ribavirin to GT1 Treatment-Naive HCV Patients, 8th International Workship on Clinical Pharmacology of Hepatitis
Therapy, 2013.
2
1.
Viekira Pak (ombitasvir, paritaprevir, and ritonavir tablets; dasabuvir tablets) [package insert]. AbbVie,
Inc, Chicago, Ill, USA; December 2014.
2
2
2
2
2.
3.
016.
4.
Sunvepra (asunaprevir), Bristol-Myers Squibb Australia Pty Ltd, Mulgrave, VIC, Australia; January 2017.
Zepatier (elbasvir and grazoprevir) [package insert], Merck & Co., Inc, Whitehouse Station, NJ, USA.
Romano, K. P.; Ali, A.; Aydin, C.; Soumana, D.; Ozen, A.; Deveau, L. M.; Silver, C.; Cao, H.; Newton, A.;
Petropoulos, C. J.; Huang, W.; Schiffer, C. A., The molecular basis of drug resistance against hepatitis C virus
NS3/4A protease inhibitors. PLoS Pathog 2012, 8 (7), e1002832.
2
5.
Harper, S.; McCauley, J. A.; Rudd, M. T.; Ferrara, M.; DiFilippo, M.; Crescenzi, B.; Koch, U.; Petrocchi,
A.; Holloway, M. K.; Butcher, J. W.; Romano, J. J.; Bush, K. J.; Gilbert, K. F.; McIntyre, C. J.; Nguyen, K. T.; Nizi,
E.; Carroll, S. S.; Ludmerer, S. W.; Burlein, C.; DiMuzio, J. M.; Graham, D. J.; McHale, C. M.; Stahlhut, M. W.;
Olsen, D. B.; Monteagudo, E.; Cianetti, S.; Giuliano, C.; Pucci, V.; Trainor, N.; Fandozzi, C. M.; Rowley, M.;
Coleman, P. J.; Vacca, J. P.; Summa, V.; Liverton, N. J., Discovery of MK-5172, a Macrocyclic Hepatitis C Virus
NS3/4a Protease Inhibitor. ACS Med Chem Lett 2012, 3 (4), 332-6.
2
6.
Yu, M.; Corsa, A. C.; Xu, S.; Peng, B.; Gong, R.; Lee, Y. J.; Chan, K.; Mo, H.; Delaney, W. t.; Cheng, G.,
In vitro efficacy of approved and experimental antivirals against novel genotype 3 hepatitis C virus subgenomic
replicons. Antiviral research 2013, 100 (2), 439-45.
2
7.
Chabaud, P.; Pepe, G.; Courcambeck, J.; Camplo, M., Stereoselective synthesis of (3S,4S)-tert-butyl-N-
Boc-3-ethyl-4-hydroxy-L-prolinate and (3S,4R)-tert-butyl-N-Boc-3-ethyl-4-hydroxy-L-prolinate. Tetrahedron 2005,
6
2
1, 3725-3731.
8. Thompson, R. A.; Isin, E. M.; Ogese, M. O.; Mettetal, J. T.; Williams, D. P., Reactive Metabolites: Current
and Emerging Risk and Hazard Assessments. Chem Res Toxicol 2016, 29 (4), 505-33.
9. Tujios, S.; Fontana, R. J., Mechanisms of drug-induced liver injury: from bedside to bench. Nat Rev
Gastroenterol Hepatol 2011, 8 (4), 202-11.
0. Stephens, C.; Andrade, R. J.; Lucena, M. I., Mechanisms of drug-induced liver injury. Curr Opin Allergy
Clin Immunol 2014, 14 (4), 286-92.
1. Evans, D. C.; Watt, A. P.; Nicoll-Griffith, D. A.; Baillie, T. A., Drug-protein adducts: an industry
2
3
3
perspective on minimizing the potential for drug bioactivation in drug discovery and development. Chem Res
Toxicol 2004, 17 (1), 3-16.
3
2.
Methods of Characterization. Chem Res Toxicol 2016, 29 (12), 2040-2057.
3. Goelzer, P.; Morcos, P. N.; Tran, J. Q.; Wen, B.; Shulman, N. S.; Brennan, B. J.; Hammond, J.; Singer, T.;
Gan, J.; Zhang, H.; Humphreys, W. G., Drug-Protein Adducts: Chemistry, Mechanisms of Toxicity, and
3
Smith, P., Coadministration of ritonavir with the HCV protease inhibitor danoprevir substantially reduces reactive
metabolite formation noth in vitro and in vivo (abstract 796). Journal of Hepatology 2012, 56 (2), 580A.
3
4.
Gane, E. J.; Rouzier, R.; Wiercinska-Drapalo, A.; Larrey, D. G.; Morcos, P. N.; Brennan, B. J.; Le Pogam,
S.; Najera, I.; Petric, R.; Tran, J. Q.; Kulkarni, R.; Zhang, Y.; Smith, P.; Yetzer, E. S.; Shulman, N. S., Efficacy and
safety of danoprevir-ritonavir plus peginterferon alfa-2a-ribavirin in hepatitis C virus genotype 1 prior null
responders. Antimicrob Agents Chemother 2014, 58 (2), 1136-45.
3
5.
ALT elevation status derived from package insert for approved compounds (paritaprevir, asunaprevir,
grazoprevir and simeprevir). For compounds not approved in the US or EMA, positive ALT elevation noted when
Grade ≥3 ALT elevation occurred in ≥1% of patients or when publications document significant ALT elevation
(
danoprevir) or serious liver enzyme elevation (IDX-320). .

3
6.
US Food and Drug Administration. Drugs@FDAdrug approval package –Clinical Pharmacology and Bio-
Pharmaceutics Review for Viekira Pak (ombitasvir, paritaprevir, and ritonavir tablets; dasabuvir tablets),co-
packaged for oral use.
Available at: www.accessdata.fda.gov/drugsatfda_docs/nda/2014/206619Orig1s000ClinPharmR.pdf. Accessed July
1
3
, 2016.
7.
Hayashi, N.; Nakamuta, M.; Takehara, T.; Kumada, H.; Takase, A.; Howe, A. Y.; Ludmerer, S. W.;
Mobashery, N., Vaniprevir plus peginterferon alfa-2b and ribavirin in treatment-naive Japanese patients with
hepatitis C virus genotype 1 infection: a randomized phase III study. Journal of gastroenterology 2016, 51 (4), 390-
4
3
03.
8.
Abutaleb, A.; Kottilil, S., Vedroprevir in the management of hepatitis C virus infection. Expert Opin
Investig Drugs 2017, 26 (12), 1399-1402.
9. Wyles, D. L.; Rodriguez-Torres, M.; Lawitz, E.; Shiffman, M. L.; Pol, S.; Herring, R. W.; Massetto, B.;
3
Kanwar, B.; Trenkle, J. D.; Pang, P. S.; Zhu, Y.; Mo, H.; Brainard, D. M.; Subramanian, G. M.; McHutchison, J. G.;
Habersetzer, F.; Sulkowski, M. S., All-oral combination of ledipasvir, vedroprevir, tegobuvir, and ribavirin in
treatment-naive patients with genotype 1 HCV infection. Hepatology 2014, 60 (1), 56-64.
4
4
0.
1.
Olysio, Janssen Therapeutics, Division of Janssen Products, Titusville, NJ, USA, November 2013.
Across Phase 3 studies, Grade 3/4 ALT elevations occurred in <0.2% of individuals (4/2797). EMA.
Withdrawal assessment report. Faldaprevir Boehringer Ingelheim. March 2014.
4
4
2.
3.
It is probable that the cyclopropyl group would increase the reactivity of the epoxide electrophile.
Lin, C., HCV NS3-4A Serine Protease. In Hepatitis C Viruses: Genomes and Molecular Biology, Tan, S.
L., Ed. Norfolk (UK), 2006.
4. Cottell, J.; Link, J.; Schroeder, S.; Taylor, J.; Tse, W.; Vivian, R.; Yang, Z. Antiviral compounds. WO
application 2009005677; published 2009-01-08.
5. Jiang, Y.; Andrews, S. W.; Condroski, K. R.; Buckman, B.; Serebryany, V.; Wenglowsky, S.; Kennedy, A.
4
4
L.; Madduru, M. R.; Wang, B.; Lyon, M.; Doherty, G. A.; Woodard, B. T.; Lemieux, C.; Geck Do, M.; Zhang, H.;
Ballard, J.; Vigers, G.; Brandhuber, B. J.; Stengel, P.; Josey, J. A.; Beigelman, L.; Blatt, L.; Seiwert, S. D.,
Discovery of danoprevir (ITMN-191/R7227), a highly selective and potent inhibitor of hepatitis C virus (HCV)
NS3/4A protease. Journal of medicinal chemistry 2014, 57 (5), 1753-69.
4
6.
Lawitz, E.; Yang, J. C.; Stamm, L. M.; Taylor, J. G.; Cheng, G.; Brainard, D. M.; Miller, M. D.; Mo, H.;
Dvory-Sobol, H., Characterization of HCV resistance from a 3-day monotherapy study of voxilaprevir, a novel
pangenotypic NS3/4A protease inhibitor. Antivir Ther 2018, 23 (4), 325-334.
4
7.
Fourth edition. ed.; 2010; p 864 pages.
8. Rodriguez-Torres, M.; Glass, S.; Hill, J.; Freilich, B.; Hassman, D.; Di Bisceglie, A. M.; Taylor, J. G.;
Rowland, M.; Tozer, T. N., Clinical pharmacokinetics and pharmacodynamics : concepts and applications.
4
Kirby, B. J.; Dvory-Sobol, H.; Yang, J. C.; An, D.; Stamm, L. M.; Brainard, D. M.; Kim, S.; Krefetz, D.; Smith, W.;
Marbury, T.; Lawitz, E., GS-9857 in patients with chronic hepatitis C virus genotype 1-4 infection: a randomized,
double-blind, dose-ranging phase 1 study. J Viral Hepat 2016, 23 (8), 614-22.
4
9.
Bourliere, M.; Gordon, S. C.; Flamm, S. L.; Cooper, C. L.; Ramji, A.; Tong, M.; Ravendhran, N.; Vierling,
J. M.; Tran, T. T.; Pianko, S.; Bansal, M. B.; de Ledinghen, V.; Hyland, R. H.; Stamm, L. M.; Dvory-Sobol, H.;
Svarovskaia, E.; Zhang, J.; Huang, K. C.; Subramanian, G. M.; Brainard, D. M.; McHutchison, J. G.; Verna, E. C.;
Buggisch, P.; Landis, C. S.; Younes, Z. H.; Curry, M. P.; Strasser, S. I.; Schiff, E. R.; Reddy, K. R.; Manns, M. P.;
Kowdley, K. V.; Zeuzem, S.; Polaris; Investigators, P.-. Sofosbuvir, Velpatasvir, and Voxilaprevir for Previously
Treated HCV Infection. The New England journal of medicine 2017, 376 (22), 2134-2146.
5
0.
Jacobson, I. M.; Lawitz, E.; Gane, E. J.; Willems, B. E.; Ruane, P. J.; Nahass, R. G.; Borgia, S. M.;
Shafran, S. D.; Workowski, K. A.; Pearlman, B.; Hyland, R. H.; Stamm, L. M.; Svarovskaia, E.; Dvory-Sobol, H.;
Zhu, Y.; Subramanian, G. M.; Brainard, D. M.; McHutchison, J. G.; Brau, N.; Berg, T.; Agarwal, K.; Bhandari, B.
R.; Davis, M.; Feld, J. J.; Dore, G. J.; Stedman, C. A. M.; Thompson, A. J.; Asselah, T.; Roberts, S. K.; Foster, G.
R., Efficacy of 8 Weeks of Sofosbuvir, Velpatasvir, and Voxilaprevir in Patients With Chronic HCV Infection: 2
Phase 3 Randomized Trials. Gastroenterology 2017, 153 (1), 113-122.