Characterization of the Anti-Hepatitis C Virus Activity of New Nonpeptidic Small-Molecule Cyclophilin Inhibitors with the Potential for Broad Anti-Flaviviridae Activity

Article abstract of DOI:10.1128/AAC.00126-18

Although members of the Flaviviridae display high incidence, morbidity, and mortality rates, the development of specific antiviral drugs for each virus is unlikely. Cyclophilins, a family of host peptidyl-prolyl cis-trans isomerases (PPIases), play a pivotal role in the life cycles of many viruses and therefore represent an attractive target for broad-spectrum antiviral development. We report here the pangenotypic anti-hepatitis C virus (HCV) activity of a small-molecule cyclophilin inhibitor (SMCypI). Mechanistic and modeling studies revealed that the SMCypI bound to cyclophilin A in competition with cyclosporine (CsA), inhibited its PPIase activity, and disrupted the CypA-nonstructural protein 5A (NS5A) interaction. Resistance selection showed that the lead SMCypI hardly selected amino acid substitutions conferring low-level or no resistance in vitro. Interestingly, the SMCypI selected D320E and Y321H substitutions, located in domain II of the NS5A protein. These substitutions were previously associated with low-level resistance to cyclophilin inhibitors such as alisporivir. Finally, the SMCypI inhibited the replication of other members of the Flaviviridae family with higher 50% effective concentrations (EC₅₀s) than for HCV. Thus, because of its chemical plasticity and simplicity of synthesis, our new family of SMCypIs represents a promising new class of drugs with the potential for broad-spectrum anti-Flaviviridae activity as well as an invaluable tool to explore the role of cyclophilins in viral life cycles.

Full text of DOI:10.1128/AAC.00126-18

AAC Accepted Manuscript Posted Online 14 May 2018
Antimicrob. Agents Chemother. doi:10.1128/AAC.00126-18
Copyright © 2018 Nevers et al.
This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license.
1
2
Characterization of the anti-hepatitis C virus activity of new non-peptidic small-molecule
cyclophilin inhibitors with a potential for broad anti-Flaviviridae activity
3
4
Quentin Nevers,a* Isaac Ruiz,a* Nazim Ahnou,a,b Flora Donati,a,b Rozenn Brillet,a Laurent
Softic,a Maxime Chazal,c Nolwenn Jouvenet,c Slim Fourati,a,b Camille Baudesson,a Patrice
Bruscella,a Muriel Gelin,d Jean-François Guichou,d Jean-Michel Pawlotskya,c# and Abdelhakim
Ahmed-Belkacema#
5
6
7
8
9
aInstitut Mondor de Recherche Biomédicale (IMRB), INSERM U955 Team 18, Hôpital Henri
Mondor, Université Paris-Est, Créteil, France
10
11
12
13
14
15
16
17
18
19
20
21
22
23
bNational Reference Center for Viral Hepatitis B, C and D, Department of Virology, Hôpital
Henri Mondor, Université Paris-Est, Créteil, France
cUnité Génomique Virale et Vaccination, CNRS UMR 3569, Institut Pasteur, Paris, France
dCentre de Biochimie Structurale (CBS), INSERM, CNRS, Université de Montpellier, France.
Running Title: SMCypI, a new family of broad-spectrum antiviral
Address for correspondence
Abdelhakim Ahmed-Belkacem, hakim.ahmed-belkacem@inserm.fr.
Present address
INSERM U955 Team 18, Hôpital Henri Mondor, 51 avenue du Maréchal de Lattre de
Tassigny, 94010 Créteil, France.
*Q.N. and I.R. contributed equally to this work; AA.B. and JM.P. contributed equally to this
work#
1

24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
SUMMARY
Despite Flaviviridae viruses display high incidence, morbidity and mortality rates,
the development of specific antiviral drugs for each virus is unlikely. Cyclophilins, a family
of host peptidyl-prolyl cis/trans isomerase (PPIase), play a pivotal role in the lifecycle of
many viruses and therefore represent an attractive target for broad-spectrum antiviral
development. We report here the pan-genotypic anti-hepatitis C virus (HCV) activity of a
small-molecule cyclophilin inhibitor (SMCypI). Mechanistic and modeling studies revealed
that SMCypI bound to cyclophilin A in competition with Cyclosporin A (CsA), inhibited its
PPIase activity and disrupted the CypA/NS5A interaction. Resistance selection showed that
the lead SMCypI hardly selected amino acid substitutions conferring low-level or no
resistance in vitro. Interestingly, the SMCypI selected substitutions D320E and Y321H
substitutions, located in the domain II of NS5A protein. These substitutions have been
previously associated with low-level resistance to cyclophilin inhibitors such as Alisporivir.
Finally, the SMCypI inhibited the replication of other members of the Flaviviridae family
with higher EC50 values than HCV. Thus, because of its chemical plasticity and simplicity of
synthesis, our new family of SMCypIs represents a promising new class of drugs with a
potential for broad-spectrum anti-Flaviviridae activity, as well as an invaluable tool to
explore the role of cyclophilins in viral lifecycles.
43
44
45
46
2

47
48
INTRODUCTION
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
Urbanization, human migrations and climate changes facilitate emergence or re-
emergence of a large number of viruses causing unexpected illnesses and epidemics, while
the capacity to identify and control emerging diseases is limited in poorer regions where
many of them have their origin (1). Among emerging and re-emerging viruses, members of
the Flaviviridae family represent a global public health issue. The Flaviviridae family
consists of four genera (Flavivirus, Hepacivirus, Pegivirus and Pestivirus) which include
viruses responsible for important animal and human diseases, such as yellow fever virus
(YFV), zika virus (ZIKV), Japanese encephalitis virus (JEV), West-Nile virus (WNV), hepatitis
C virus (HCV) and dengue virus (DENV). YFV, ZIKV, JEV, WNV and DENV are leading causes
of arthropod-borne human diseases worldwide. According to the World Health
Organization (WHO), they globally infect 400 million individuals each year, causing
approximately 80,000 deaths each year. HCV is a blood-borne Hepacivirus responsible for
chronic liver diseases causing approximately 700,000 deaths annually (2). Approximately
71 million individuals are infected worldwide, representing 1% of the global population
(Global Hepatitis Report from WHO, 2017). The recent approval of a large number of direct-
acting antiviral agents (DAAs) active against HCV, including generic compounds, has
revolutionized the treatment of this infection with more than 95% rates of infection cure
(3). In contrast, no antiviral drugs are available so far to cure infections caused by
arthropod-borne members of the Flaviviridae family, despite their global public health
importance.
3

69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
During the past 10 years, two different types of antiviral agents, including DAAs and
host-targeting antiviral (HTA) agents, have been developed for the treatment of HCV
infection. Among the HTAs, non-immunosuppressive derivatives of cyclosporin A (CsA) that
target host cyclophilins (Cyps) yielded the most promising results. Alisporivir (ALV) was
the first HTA to enter HCV clinical development and reach Phase III clinical trials (4, 5). Its
development was halted following the report of seven cases of acute pancreatitis, including
a lethal one (6). These events were independent from Cyp inhibition, most likely due to
ALV-induced hypertriglyceridemia that potentiated the pancreatic toxicity of interferon
alpha that was part of the combination regimens. Although cyclophilin inhibitor (CypI)
failed to reach the market for the indication of hepatitis C treatment, they remain attractive
to combat other viral infections (7-10). Indeed, CypIs have been reported to be involved in
the lifecycle of viruses other than HCV (11), while having a high barrier to resistance, broad
antiviral activity and possibly additive or synergistic effects with other antiviral compounds
in various models.
Cyps are peptidyl-prolyl cis/trans isomerases (PPIases) that catalyze the
interconversion of the two energetically preferred conformers (cis and trans) of the planar
peptide bond preceding an internal proline residue. Seventeen human Cyps have been
identified but more may exist. Cyps have been convincingly shown to play a pivotal role in
the lifecycle of a large number of viruses from different families (11). However, there are
few data available regarding the anti-Flaviviridae activity of CypIs, all of which were
obtained with CsA and ALV (8, 12-14). The molecular mechanisms of CypI anti-HCV activity
are not yet fully understood. It is believed that they exert their antiviral effect by disrupting
the CypA-NS5A interaction that regulates multiple phases of HCV replication (15) (16).
4

92
93
94
95
96
97
98
99
We previously reported our rational design of a new family of small-molecule, non-peptidic
CypI (SMCypI) unrelated to CsA by means of a complex fragment-based drug discovery
approach (17). Our SMCypIs displayed antiviral effectiveness not only against HCV, but also
HIV and coronaviruses, suggesting, together with data reported in the literature, that they
could act as broad-spectrum antiviral agents, effective against a number of different viruses
from different viral families. The present study aims at characterizing the anti-HCV activity
of the new family of SMCypIs, unraveling their molecular antiviral mechanism and
evaluating their spectrum of anti-Flaviviridae activity.
5

100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
MATERIALS AND METHODS
Drugs
Alisporivir (ALV) and ledipasvir (LDV) were purchased from AGV Discovery
(Clapiers, France), while cyclosporine A (CsA) was purchased from Sigma-Aldrich (Saint-
Louis, Missouri, USA).
Compound synthesis
SMCypI compound synthesis is described in the Supplementary Materials and
Methods. Chemical reagents were obtained from Aldrich Chemical (Saint-Louis, Missouri,
USA), Acros Organics (Geel, Belgium), abcr GmBh (Karlsruhe, Germany), acbblocks
(Toronto, Canada) and Chembridge (San Diego, California) and used without further
purification.
HCV-SGR plasmids
Plasmids H77/SG-Feo, S52/SG-Feo and SA1/SG-Feo, that contain a Firefly luciferase
reporter gene and a genotype (GT) 1a, 3a or 5a HCV subgenome, respectively, were kindly
provided by Dr Charles M. Rice (Rockefeller University, New York, New York) (18, 19).
Plasmid p1071-NS5A(Ni)-S2204I contains a Firefly luciferase reporter gene and a genotype
1b HCV subgenome (Con1 strain) with an NS5A cassette from the genotype 1b HCV-N strain
6

123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
(20). Plasmids APP76-Con1-SG-Neo-(I)-hRluc2aUb and APP40-J6/JFH1EMCVIRES-aRlucNeo
that contain a Renilla luciferase reporter gene and genotype 1b and 2a HCV subgenomes,
respectively, were purchased from Apath LLC (New York, New York, USA)
(http://www.apath.com/Hepatitis_C_Virus_Technology.htm#3). Plasmid I389-Neo/NS3-
3’/5.1 that contains the Neomycin resistance gene and an HCV genotype 1b subgenome was
used for resistance selection experiments and was kindly provided by Dr Ralf
Bartenschlager (University of Heidelberg, Heidelberg, Germany) (21). Plasmid DBN3acc
that contains a full-length HCV genotype 3a genome was kindly provided by Dr Jens Bukh
(University of Copenhagen, Copenhagen, Denmark) (22). Finally, chimeric plasmid fdRocco-
chimeric2a/4aNS5A consisting of a genotype 4a NS5A sequence inserted into a genotype 2a
HCV subgenome from APP40-J6/JFH1EMCVIRES-aRlucNeo has been developed in our
laboratory.
Cell cultures
Human hepatoma Huh7 cells (kindly provided by Dr Eliane Meurs) and Huh7.5 cells
(Apath LLC) were cultured in complete Dulbecco’s Modified Eagle Medium (DMEM, Thermo
Fisher Scientific, Waltham, Massachusetts, USA) supplemented with 10% fetal bovine
serum, 50 IU/ml penicillin, 100 μg/ml streptomycin and 0.1 μg/ml amphotericine-β
(ThermoFisher Scientific).
Assessment of antiviral activity in HCV-SGR models
7

146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
For transient HCV-SGR models, plasmids p1071-NS5A(Ni)-S2204I, APP40-
J6/JFH1EMCVIRES-aRlucNeo and fdRocco-chimeric2a/4aNS5A were linearized with XhoI,
XbaI and XbaI, respectively (FastDigest, ThermoFisher Scientific) and in vitro transcribed
using MEGAscript T7 Transcription Kit (ThermoFisher Scientific). Then, approximately
1.5 x 104 Huh7.5 cells were transfected with 250 ng of HCV-SGR RNA using trans-IT mRNA
transfection kit (Mirus Bio LLC, Madison, Wisconsin, USA). Four hours after transfection,
compounds were added to the culture medium. Luciferase activity was monitored 96, 48
and 72 hours post-transfection for genotype 1b, 2a and chimeric 2a/4a HCV-SGR,
respectively.
Huh7.5 cells stably harboring genotype 1a, 3a and 5a HCV-SGR were cultured in the
presence of the compounds for 48 hours before luciferase activity measurement. Plots were
fitted with a four-parameter logistic curve with SigmaPlot v11 software (Systat Software,
Inc) and the EC50s were determined from the curves.
Assessment of antiviral activity in the infectious HCV model
Huh7.5 cells were seeded at the density of 1.5 x 104 cells and incubated 24 hours
before infection with 250 µl of HCVcc (J6/JFH1 strain, GT2a/2a) in the presence of
increasing concentrations of the compounds. Eight hours after infection, the cells were
washed with PBS and incubated with fresh medium containing the inhibitors for 72 hours.
Then, luciferase activity was measured and plotted against compound concentrations. The
EC50s were determined from curves fitted with a four-parameter logistic equation.
8

169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
Assessment of the combination of compound 31 and ledipasvir
Huh7.5 cells stably harboring a genotype 1a HCV-SGR were cultured in the presence
of 10 pM LDV, 2.5 µM C31 or both drugs at the same concentrations, in the absence or
presence of 500 µg/ml of G418. After 5 passages, the remaining living cells were stained
with crystal violet.
Selection of clones resistant to compound 31
Huh7.5 cells stably harboring a genotype 1b HCV-SGR replicon that confers cell
resistance to G418 were used for selection experiments. The cells were cultured in the
presence of escalating doses of C31 (1 to 50 µM) and of 1.5 mg/ml of G418 until colonies
growing in the presence of C31 appeared. Two resistant colonies were isolated after several
passages and total RNA was extracted with the RNeasy Mini Kit (Qiagen, Hilden, Germany)
and reverse transcribed with the High Capacity cDNA reverse transcription kit
(ThermoFisher Scientific). The NS5A-coding region was amplified at baseline and in the
resistant colonies by PCR using forward oligonucleotides 5’-GTG CAG TGG ATG AAY CGG
CTG ATA GC-3’, 5’-TTC CAR GAC TCT ARC ART G-3’, 5’-ACT ATG TGC CTG AGA GCG ACG-3’,
5’-GGR TTG TAR TCC GGS CGY GCC CAT A-3’ and reverse oligonucleotides 5’-TCC CRT GYG
AGC CYG AAC CG-3’, 5’-GTG GTG ACG CAG CAA AGA GT-3’, 5’-CCC ACA TTA CAG CAG AGA
CGG C-3’, and 5‘-TTG ATG GGC AGC TTG GT-3’.
Phenotypic characterization of amino acid substitutions selected by compound 31
9

192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
Candidate resistance-associated substitutions (RASs) were introduced into a wild-
type genotype 1b HCV-SGR containing the luciferase reporter gene by means of site-
directed mutagenesis (Quick Change II XL Site-Directed Mutagenesis Kit, Agilent
Technologies, Santa Clara, California, USA). Ninety-six hours post-RNA transfection, the
susceptibility to compound 31 of RAS-containing HCV-SGR was determined by measuring
luciferase activity and compare to wild-type HCV-SGR susceptibility. The replication
capacity of RAS-containing HCV-SGR was assessed by comparing luciferase activity 4 hours
and 96 hours post-transfection and expressed as a percentage of wild-type replication
capacity.
Phenotypic characterization of amino acid substitutions selected by ledipasvir
The L31V and Y93H RASs, which confer high-level resistance to LDV (23), were
introduced in the NS5A-coding region of the APP76-Con1-SG-Neo-(I)-hRluc2aUb plasmid by
means of site-directed mutagenesis. After linearization with ScaI and in vitro transcription,
the genotype 1b HCV-SGR was transfected into Huh7.5 cells. The cells were cultured in the
presence of the compounds for 84 hours and luciferase activity was measured. Then,
susceptibility of HCV-SGR(L31V/Y93H) to C31, ALV and LDV was compared to that of wild-
type HCV-SGR.
Competitive binding assay
10

215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
A 96-well plate-based TR-FRET (time-resolved fluorescence energy transfer) assay
(Selcia, Ongar, UK) was used to determine competitive CypA binding of C31 and CsA.
Briefly, CypA was tagged with a polyhistidine sequence (6His) and formed a complex with
an anti-6His antibody labelled with a fluorescent donor F(d), while CsA was tagged with a
fluorescent acceptor F(a). C31 was added to the master mix containing the
CypA/antibody/CsA complex, with a final concentration of detergent of 0.01%. After 30 min
of incubation at room temperature, the plate was read on a SpectraMax M5 (Molecular
Devices, Sunnyvale, California, USA) at 2 wavelengths to detect F(d) and F(a) emissions. The
F(a)/F(d) ratio was calculated and the values were plotted against the inhibitor
concentration in Log10 molar and fitted using one-site Ki nonlinear regression to determine
the Kd. Non-labelled CsA and ALV were used as controls.
Protein-protein interaction assay
An NS5A-Rluc/CypA-6his in vitro interaction assay was developed to measure the
efficiency of CypI disruption of the NS5A/CypA interaction by luminescence measurement.
After PCR amplification with primers 5’-AAA AAC TGC AGA TGT CCG GCT CGT GGC
T-3’ and 5’-AAA AAC CGC GGG CAG CAG ACG ACG TCC-3’ and digestion with PstI and SacII
(FastDigest, ThermoFischer Scientific), the genotype 1b NS5A-coding sequence was cloned
in fusion with the Renilla luciferase into a pRluc-N3(h) plasmid (6310009, Perkin Elmer,
Waltham, Massachussets, USA). NS5A-D320E was generated by site-directed mutagenesis
with oligonucleotides 5’-ATG GGC ACG CCC GGA ATA CAA CCC TCC ACT G-3’ and 5’-CAG
TGG AGG GTT GTA TTC CGG GCG TGC CCA T-3’ and cloned in a similar way to WT-NS5A.
11

238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
Huh7.5 cells were then transfected with the pRluc-N-NS5A plasmid. Forty-eight hours later,
cells were lysed with 4 thaw-freeze cycles and incubated 30 min at room temperature with
500 µg of Ni-NTA magnetic beads preloaded with 30 µg of purified CypA in an equilibration
buffer (imidazole 10 mM, Tween-20 0.05% in PBS (-/-)). Increasing concentrations of the
CypI were added during the incubation. The beads were then washed 3 times with a
washing buffer (imidazole 20 mM, Tween-20 0.05% in PBS (-/-)) and elution was
performed with imidazole 400 mM for 15 min at room temperature. Renilla luciferase
activity in eluates was measured with Renilla Luciferase Assay System (Promega). Results
are presented as mean ± standard deviation (SD) of at least three independent experiments.
All analyses were two-sided and considered significant when p < 0.05.
Molecular modelling and docking of C31 into CypA
Molecular modelling and docking experiment were performed using the @TOME-2
server (24), which integrates an original interface for comparative docking of small
molecules detected in the Protein Data Bank file of each template. The search for
homologous sequences and alignments was performed using the @TOME-2 server and the
CypD sequence (P30405), with a 75% identity. In each structural model, the active site
boundaries were deduced from the vicinity of the co-crystallized ligands (compounds C32,
C34, C35 and C36 were selected as templates with the corresponding Protein Data Bank
(PDB) accession numbers 4J59, 4J5C, 4J5B and 4J5E, respectively) using the @TOME-2
comparative option. In addition, the same chemical entities were used to define a shape
restraint to guide docking in the automatically computed models. The files for the ligands
12

261
262
were generated with MarvinSketch 6.2.2 for the SMILES and with Frog2 server for the mol2
(25). Figure 1A and 1B were generated using Pymol.
263
264
265
Assessment of compound 31 anti-Flaviviridae activity
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
Huh7 cells were infected for 48 h with the YFV strain Dakar HD 1279 (obtained from
the World Reference Center for Emerging Viruses and Arborviruses – WRCEVA, Texas,
USA), at an MOI of 7; or for 48 hours with DENV-4 strain Dominica 814669 (obtained from
the Centro de Ingeniería Genética y Biotecnologíain, Cuba) at an MOI of 10; or for 1 h with
ZIKV strain MR766 (obtained from ATCC) at an MOI of 1. Infections were performed in the
presence of different concentration of C31.
For DENV and YFV experiments, total RNA was extracted from cell cultures with the
NucleoSpin RNA II kit (Macherey-Nagel). First-strand complementary DNA (cDNA)
synthesis was performed with the RevertAid H Minus M-MuLV Reverse Transcriptase.
Quantitative real-time PCR was performed on a real-time PCR system (QuantStudio 6 Flex,
Applied Biosystems) with SYBR Green PCR Master Mix (Life Technologies). The data were
analyzed with the 2-ΔΔCT method, with all samples normalized to GAPDH. All experiments
were performed in triplicate. Genome equivalent concentrations were determined by
extrapolation from a standard curve generated from serial dilutions of the plasmid
encoding a subgenomic YFV replicon. The sequences of the 17D-NS3 primers used for the
RT-qPCR were the following: sense 5’-AGGTCCAGTTGATCGCGGC and antisense 5’-
GAGCGACAGCCCCGATTTCT. The sequences of the pan-DENV primers were the following:
13

283
284
285
286
287
288
289
290
291
sense
5′-TTGAGTAAACYRTGCTGCCTGTAGCTC
and
antisense
5′-
GAGACAGCAGGATCTCTGGTCTYTC.
For ZIKV experiment, total RNA was extracted from cell culture with the SV96 Total
RNA isolation system (Promega). Reverse transcription was performed with the High
Capacity cDNA Reverse Transcription kit (Applied Biosystems). Quantitative real-time PCR
was performed on a real-time PCR system (Applied Biosystems 7300) with TAQMAN gene
expression master mix (Applied Biosystems) and the data were normalized to GAPDH. The
sequences of the primers used were the following: sense 5’ATATCGGACATGGCTTCGGA and
antisense 5’GTTCTTTTGCAGACATATTGAGTG.
292
293
294
295
296
297
Statistical analysis
Statistical analyses were performed using SigmaPlot software. Statistics were
calculated using t-test analysis of variance. P values below 0.05 were considered
statistically significant.
14

298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
RESULTS
C31 has pan-genotype anti-HCV activity
The anti-HCV activity of the new SMCypI compound C31, our most potent inhibitor
of Cyp PPIase activity, was tested different HCV genotype models containing luciferase
reporter genes, including an infectious chimeric J6/JFH1 (2a/2a) virus, genotype 1a, 1b, 2a,
3a, and 5a HCV-SGRs, and the chimeric 2a/4a HCV-SGR containing a genotype 4a NS5A
sequence (Supplementary Figure 1). In addition, the anti-HCV activity of C31 was
evaluated in the recently developed full-length infectious HCV genotype 3a model (DBN-
3acc) (22). ALV and CsA were used as controls in all experiments.
C31 equally inhibited the replication of genotype 1a, 1b, 2a, 3a, 5a and chimeric
2a/4a HCV-SGRs, with EC50s ranging from 1.20±0.83 to 7.76±1.57 µM (Table 1). C31 also
inhibited the replication of the infectious J6/JFH1 virus, with a comparable EC50 of
2.80±0.40 µM. Finally, C31 inhibited DBN-3acc RNA replication in a dose-dependent
manner, with a maximal 244-fold HCV RNA reduction at 10 µM (Supplementary Figure 3).
C31 did not affect cell viability at its effective concentration (Supplementary Figure 2).
Altogether, these results demonstrate the pan-genotype activity of the new SMCypI.
C31 anti-HCV activity is related to its binding to CypA and inhibition of CypA PPIase
activity
15

321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
To understand the molecular mechanism of anti-HCV action of our new SMCypI, the
CypA binding modes of C31 and CsA were modelled (Figure 1A and 1B, respectively). As
shown in Figure 1B, the structure of CypA in complex with C31 revealed its expected dual
binding into the “hydrophobic pocket”, the CypA PPIase active site, on the one hand, and the
“gatekeeper pocket” on the other hand (17, 26). Interestingly, the methoxy group of
compound 31 pushes Arg55 to create a hydrogen bond with the urea moiety of the
compound.
The CypA binding sites of C31 and CsA were partially overlapping, suggesting
competitive binding to CypA. Thus, time-resolved fluorescence resonance energy transfer
(TR-FRET) was used to assess whether C31 competes with CsA for binding to the purified
CypA. Both non-labelled CsA and ALV displaced labelled CsA from its CypA binding site,
with Kd values of 8.4 and <5 nM, respectively (data not shown). As shown in Figure 1C, C31
also displaced labelled CsA with a Kd value of 105 nM, confirming competition with CsA
binding to CypA.
Finally, we assessed whether the anti-HCV effect of the SMCypIs was related to their
ability to inhibit CypA PPIase enzyme activity. For this, the anti-HCV activity of 6 different
chemically-related SMCypIs listed in Supplementary Table 1, including C31, was
determined in a genotype 1b HCV-SGR and plotted against their respective inhibitory
activities in a PPIase enzyme assay. As shown in Figure 1D, the anti-HCV activity of the
SMCypIs was strongly correlated to their ability to inhibit PPIase activity, with a Pearson’s
correlation coefficient of 0.96 and a p value <0.0001
16

342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
Altogether, these results demonstrate that our new family of SMCypIs inhibits HCV
replication by binding to both the PPIase catalytic pocket and the gatekeeper pocket of
CypA, thus blocking its PPIase catalytic activity.
C31 disrupts the CypA-NS5A interaction
Because CypA has been shown to play an important role in the HCV lifecycle through
its interaction with the non-structural NS5A protein, we measured the effect of C31 on the
NS5A/CypA interaction. Purified CypA tagged with a polyhistidine sequence (CypA-6his)
was bound to Ni-NTA (nitrilotriacetic acid) magnetic beads. A cell lysate containing WT-
NS5A or D320E-NS5A fused with the Renillia luciferase (NS5A-Rluc) was incubated with the
CypA-loaded magnetic beads in the presence or absence of C31 and of ALV, used as a
control. After washing and elution with imidazole, luciferase activity corresponding to
NS5A-Rluc/CypA-6his interacting complexes was measured. As shown in Figure 2, the
CypA/NS5A interaction was disrupted by both ALV and C31 in a dose-dependent manner.
The same result was observed with D320E-NS5A. Together these results indicate that
SMCypIs disrupt the interaction between CypA and the HCV protein NS5A through their
binding to CypA.
C31 exerts at least additive antiviral effects when combined with the NS5A inhibitor
ledipasvir and is fully active against ledipasvir-resistant viruses
17

364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
A combination study of C31 with the NS5A inhibitor LDV was performed in Huh7.5
cells stably harboring a genotype 1a HCV-SGR. As shown in Figure 3, under G418 selective
pressure, the combination of C31 and LDV at their respective EC50s of 2.5 µM and 10 pM,
was more effective in curing cells from the replicons than each compound alone at the same
concentration, suggesting at least an additive effect of C31 and LDV in combination. In
addition, a genotype 1b HCV-SGR, containing the L31V and Y93H substitutions, which
confer high-level resistance to ledipasvir (714-fold increase in LDV EC50, data not shown),
remained fully sensitive to C31 and ALV (Table 2).
Together, these results suggest that our new SMCypI family has at least additive
effects with HCV DAAs targeting domain I of the NS5A protein, without cross-resistance.
C31 hardly selects amino acid substitutions conferring low-level or no resistance in
vitro
Huh7.5 cells stably harboring a genotype 1b HCV-SGR containing the Neomycin
resistance gene were grown under G418 selective pressure in the absence or presence of
increasing doses (1 µM to 50 µM) of C31. Two cellular clones growing in the presence of
C31 were selected after 100 days. Because CsA and its non-immunosuppressive derivatives
were shown to select amino acid substitutions in the NS5A region of the HCV genome (27-
29), the NS5A-coding region of the two C31-resistant clones was sequenced. As compared
to baseline, 8 amino acid changes were identified, including I133V and L183P in domain I,
L303P, R304W, K308I, D320E and Y321H in domain II, and E442G in domain III of the NS5A
18

386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
protein (Figure 4). The NS5B-coding region was also sequenced and only the I585L change
was observed (data not shown).
Each of the 8 amino acid substitutions selected was introduced in a WT transient
genotype 1b HCV-SGR for phenotypic analysis (compound susceptibility and replication
capacity) in the presence of C31, ALV and CsA (Table 2). Among them, only D320E and
Y321H were associated with a mild increase in C31 EC50 (approximately 2 to 5-fold),
without impairment of the replicon replication capacity. L183P, located in NS5A domain I,
drastically reduced the corresponding HCV-SGR replication capacity, so its impact on
compound susceptibility could not be evaluated. Thus, only D320E and Y321H in domain II
of the NS5A protein, that were selected by serial passages at increasing concentrations of
C31, conferred low-level resistance to the SMCypI, as already reported with CsA and ALV.
C31 inhibits the replication of other members of the Flaviviridae family
We assessed whether C31 could exert antiviral activities against other members of
the Flaviviridae family. Viral replication was assessed by RT-qPCR 48 hours post-infection.
A dose-dependent decrease of viral replication was observed in the presence of C31 for
DENV (EC50: 7.3±3.5 µM), YFV (EC50: 27.2±4.7 µM) and ZIKV (EC50: 48.0±5.6 µM) with a CC50
≥100µM (Figure 5). C31 was more potent against HCV and DENV than against YFV and
ZIKV.
Although C31 is not as potent on all Flaviviridae than on HCV, this compound
represents a good candidate for further chemical optimizations.
19

408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
DISCUSSION
Anti-HCV drug research highlighted the antiviral potential of Cyp inhibition through
the clinical development of non-immunosuppressive derivatives of CsA. Unfortunately, all
CypIs in development for HCV were structurally-related, and the only compound that
reached late-phase clinical development, ALV, was halted due to severe adverse effects
unrelated to Cyp inhibition. We recently reported our use of a complex fragment-based
drug discovery approach using nucleic magnetic resonance, X-ray crystallography and
structure-based compound optimization to generate a new family of non-peptidic, small-
molecule cyclophilin inhibitors (SMCypI) unrelated to CsA, with broad cyclophilins in vitro
PPIase inhibitory activity and antiviral activity against HCV, HIV and coronaviruses (17).
Although ALV has been shown to potently inhibit HCV replication, its ability to block
the replication of other members of the Flaviviridae family is unknown. The
immunosuppressive CsA has been reported to bear antiviral activity against DENV-2, WNV
and YFV (14), with less potency against WNV. Results regarding the susceptibility of ZIKV to
CsA are limited and contradictory (30). This sparseness of data prompted us to assess the
anti-Flaviviridae activity of our new SMCypI. Here, we showed that C31, a SMCypI inhibitor
of PPIase activity with anti-HCV activity in vitro, also inhibits the replication of several
members of the Flaviviridae family in cell culture models. C31 was most effective against
HCV and DENV replication, with EC50s in the low micromolar range. C31 inhibited YFV and
ZIKV replication to a lesser extent, with EC50s in the micromolar range. Although the EC50
was above 50 µM, the highest concentration tested, a trend towards a decrease of viral
replication was also observed with WNV (data not shown). In contrast, no effect of C31
20

431
432
433
434
435
436
437
438
439
440
441
442
443
444
against JEV replication has been observed in our experiments (data not shown). To our
knowledge, C31 is the first SMCypI with a potential for broad anti-Flaviviridae activity.
Given the chemical plasticity and simplicity of synthesis of this new family of SMCypIs, a
large number of new compounds will be synthesized in the future. It is thus likely that C31
derivatives with greater effectiveness against the different members of the Flaviviridae
family will be generated.
The susceptibility of Flaviviridae family members to SMCypIs suggests that
cyclophilins play a pivotal role in their lifecycles. However, the different antiviral potency of
C31 across different viruses from the Flaviviridae family raises the question as to a similar
or different mechanism of inhibition. Our SMCypIs provide a unique, easy-to-use tool to
explore this role. In the present study, we used C31 to decipher the role of cyclophilins in
the HCV lifecycle and understand the mechanisms of its inhibition by the inhibitors. Our
results complement previous results generated with CsA or ALV in various models (27-29,
31-33).
445
446
447
448
449
450
451
452
453
We showed here that, like ALV (29, 32, 33), our family of SMCypIs has pan-
genotypic anti-HCV activity (EC50s in the low micromolar range for genotypes 1a, 1b, 2a, 3a,
4a and 5a), a high barrier to the selection of resistant viruses, and at least additive effects in
combination with HCV DAAs. It has been previously demonstrated that CypA PPIase activity
is required for HCV replication (34, 35). Our library of SMCypI (supplementary table 1),
provides us with a unique, thus far missing pharmacological tool to dissect the molecular
mechanisms of Cyp/ virus interactions and of the antiviral effects of Cyp inhibition. The
library of SMCypI proved to be particularly useful to discriminate PPIase-dependent
antiviral activity from effects related to other functional roles of the Cyps. Using several
21

454
455
456
SMCypI derivatives with different inhibitory potency of PPIase catalytic activity, we showed
that the antiviral activity of the SMCypIs strongly correlates with their PPIase inhibitory
potency, confirming that Cyp PPIase activity is required for HCV replication.
457
458
459
460
461
462
463
464
465
466
467
468
469
470
It has been shown that CypA directly interacts with the HCV NS5A protein to
regulate key processes of the HCV lifecycle (16, 34, 36). Disruption of this key interaction by
CsA or ALV impairs HCV replication, possibly explaining the anti-HCV effect of these
compounds (33, 37). Whether similar mechanisms could be involved with other families of
Cyp inhibitors remained unknown. We showed that the NS5A/CypA interaction was
disrupted by C31. We also modeled the CypA binding mode of C31 and CsA. As expected,
their superimposition suggested that both compounds shared a partially overlapping
binding site. This result was confirmed by means of a TR-FRET assay, showing that C31
displaces CsA from the CypA catalytic site. Interestingly, we also observed that the aniline
moiety of C31 was deeply buried into a pocket contiguous to the canonical catalytic site, the
gatekeeper pocket, which is out of reach for CsA, suggesting that the inhibition modes of
CsA and SMCypIs may partially differ. This could explain that CsA and ALV did not inhibit
ZIKV replication in this study, whereas C31 did, possibly through a PPIase-independent
mechanism different from that involved in HCV inhibition (data not shown).
471
472
473
474
475
Resistance experiments performed with ALV and CsA selected amino acid
substitutions essentially clustering in domain II of the HCV NS5A protein, also suggesting
that NS5A is the main viral partner of CypA (33). D320E and Y321H, both located in NS5A
domain II, have been reported to confer low-level resistance to CsA and its non-
immunosuppressive derivatives (28, 33, 38).
22

476
477
478
479
480
481
482
However, NS5A mutant proteins harboring the D320E substitution keep their
capacity to interact with CypA in vitro and this interaction remains fully sensitive to ALV
disruption (33). In our experiments, 100 days have been necessary to select two cellular
clones growing in the presence of 50 µM of C31 under selective pressure of G418. The
maximum selective pressure obtained with C31 (17-fold) was in keeping with data
previously reported with ALV and CsA (65-fold and 10-fold, respectively) (33). These
results confirmed the high barrier of resistance of SMCypIs.
483
484
485
486
487
488
489
Among the 8 amino acid changes identified in these clones, only D320E and Y321H
in domain II of the NS5A protein were associated with a modest increase in C31 EC50
(approximately 2 to 5-fold), without impairment of the replicon replication capacity. These
properties were similar to those reported with CsA and ALV. The interaction of NS5A-
D320E with CypA remained fully sensitive to C31 disruption. Altogether, these findings
confirm the that the mode of antiviral action of our SMCypIs is identical to that of CsA and
its derivatives.
490
491
492
493
494
495
496
The fact that SMCypI do not target a viral function, but instead a host protein
involved in a key step of the viral lifecycle suggested that they could bear additive or
synergistic properties in combination with DAAs. We confirmed this hypothesis by
combining C31 with the potent HCV NS5A inhibitor ledipasvir. Both drugs together were
more efficient in curing cells from HCV-SGR than each drug alone, suggesting at least an
additive effect. In addition, C31 remained efficient against HCV-SGR harboring amino acid
substitutions known to confer high-level resistance to LDV.
23

497
498
499
500
501
502
503
504
505
506
507
508
509
In conclusion, our new family of SMCypIs presumably exhibit broad-spectrum
antiviral properties against several members of the Flaviviridae family that represent
important public health problems worldwide and remain without any therapeutic option.
Their mechanism of antiviral action against one of these viruses, HCV, is related to Cyp
binding, inhibition of the PPIase catalytic activity and disruption of the CypA/NS5A
interaction, a mechanism common to other cyclophilin inhibitors derived from CsA. Non-
immunosuppressive analogues of CsA suffer from serious caveats, including their large size
resulting in poor cell permeability, their side effects unrelated to cyclophilin inhibition,
their drug-drug interactions, and manufacturing issues. Thus, because of its chemical
plasticity, low cellular toxicity and simplicity of synthesis, our new family of SMCypIs
represents a promising new class of drugs with broad-spectrum anti-Flaviviridae
properties, as well as an invaluable tool to explore the role of cyclophilins in viral lifecycles
and the mechanisms to block them.
24

510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
FIGURES
B
A
Catalytic site
Gatekeeper pocket
Catalytic
Gatekeeper pocket
R55
R55
C
D
25
20
15
10
5
0,18
0,16
0,14
0,12
0,10
0,08
0,06
0,04
0,02
0,00
R = 0.96
p < 0.00001
0
0
1
2
0
1
2
3
4
5
Concentration of compound 31 (µM)
IC50 of PPIase inhibition (µM)
Figure 1
25

533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
Figure 2
26

556
557
558
559
Figure 3
560
561
562
563
564
565
566
567
568
569
570
571
572
27

573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
Figure 4
28

596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
DENV
YFV
ZIKV
100
50
0
100
50
0
100
50
0
EC50 =7.3 ± 3.5
EC50 =27.2 ± 4.7 µM
EC50 =48 ± 5.6 µM
0
1
2
-1
0
1
2
0
1
2
Log10 Compound 31 (µM)
Log10 Compound 31 (µM)
Log10 compound 31 (µM)
Figure 5
29

615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
Keywords
Cyclophilin inhibitors, small-molecule, hepatitis C virus, Flaviviridae, resistance, broad-
spectrum antiviral activity
Authors email addresses
Quentin Nevers (quentin.nevers@orange.fr), Isaac Ruiz (isaac.ruiz@me.com), Nazim Ahnou
(nazim.ahnou@inserm.fr), Flora Donati (flora.donati@aphp.fr), Rozenn Brillet
(rozenn.brillet@inserm.fr), Laurent Softic (laurent.softic@inserm.fr), Maxime Chazal
(maxime.chazal@pasteur.fr), Nolwenn Jouvenet (nolwenn.jouvenet@pasteur.fr), Slim
Fourati (slim.fourati@aphp.fr), Muriel Gelin (gelin@cbs.cnrs.fr), Camille Baudesson
(camille-f.baudesson@inserm.fr), Patrice Bruscella (patrice.bruscella@inserm.fr), Jean-
François
Guichou
(guichou@cbs.cnrs.fr),
Abdelhakim
Jean-Michel
Pawlotsky
(jean-
michel.pawlotsky@aphp.fr),
belkacem@inserm.fr).
Ahmed-Belkacem
(hakim.ahmed-
Financial Support
National Agency for Research on AIDS and Viral Hepatitis (ANRS); Mexican National Council
of Science and Technology (CONACYT); National Agency for Research (ANR).
Prior Presentation
30