Oligonucleotide-Lipid Conjugates Forming G-Quadruplex Structures Are Potent and Pangenotypic Hepatitis C Virus Entry Inhibitors In Vitro and Ex Vivo

ABSTRACT

A hepatitis C virus (HCV) epidemic affecting HIV-infected men who have sex with men (MSM) is expanding worldwide. In spite of the improved cure rates obtained with the new direct-acting antiviral drug (DAA) combinations, the high rate of reinfection within this population calls urgently for novel preventive interventions. In this study, we determined in cell culture and ex vivo experiments with human colorectal tissue that lipoquads, G-quadruplex DNA structures fused to cholesterol, are efficient HCV pangenotypic entry and cell-to-cell transmission inhibitors. Thus, lipoquads may be promising candidates for the development of rectally applied gels to prevent HCV transmission.

INTRODUCTION

Hepatitis C virus (HCV) infection of human immunodeficiency virus (HIV)-infected men who have sex with men (MSM) has emerged since the early 2000s as a growing epidemic worldwide (1). Although the introduction of interferon-free direct-acting antiviral (DAA) therapies improved significantly the sustained treatment responses, the rates of reinfection after treatment termination among HIV- and HCV-coinfected MSM are high (2, 3). Consequently, preventive interventions tailored to the MSM community are urgently needed. Given that this HCV epidemic is linked to high-risk sexual behaviors that include unprotected anal sex, formulations of water-soluble molecules as rectally applied gels that prevent HCV transmission would represent an ideal option. Currently, there are no prophylactic therapies for HCV in this setting.

G-quadruplexes comprise a distinct category of nucleic acid secondary structures that are formed from G-rich DNA and RNA sequences (4). Polymorphisms in these structures can be observed in the number (from one to four) and orientation of the strands, the number of stacked G-tetrads, differences in the loop (length, type, and/or location), and, finally, the dimension of the four grooves (4). The guanosine quartet AR177 (Zintevir, Aronex Pharmaceuticals, Inc.) is a 17-base oligonucleotide composed of deoxyguanosines and thymidines on a phosphodiester backbone supplemented by phosphorothioate internucleoside linkages at the 5′ and 3′ ends. AR177 is a potent inhibitor of laboratory strains and clinical isolates of HIV-1, with 50% effective concentrations (EC₅₀) ranging between 0.025 and 3 μM in cell culture tests (5, 6). The effect is an inhibition of viral entry by blocking a step before membrane fusion, and virus-resistant strains have shown mutations in the HIV gp120 gene (7). Similar G/T-rich phosphorothioate oligonucleotides have been reported to have antiviral activity against herpes simplex virus 2 (8). G-quadruplexes are polyanionic structures like sulfated polysaccharides, and hence, their inhibition mechanism may include mimetics of the glycosaminoglycans (GAGs) and other cell surface attachment receptors involved in virus-cell attachment. Thus, these molecules compete with viral envelope glycoproteins during binding with their main receptors.

HCV cell-free virions enter into hepatocytes through a highly coordinated process which involves the two viral envelope glycoproteins E1 and E2 and multiple host cell factors. HCV first associates with its target cells through interactions of basic residues in its glycoproteins with GAGs, including heparan sulfate proteoglycans (HSPG) (9–11). Apolipoprotein E (ApoE), which associates with HCV virions, also plays a role in the initial attachment through interaction with the HSPG associated with syndecan 1 (SDC1) (12) and syndecan 4 (SDC4) (13). The liver/lymph node-specific intercellular adhesion molecule 3-grabbing integrin (L-SIGN) (14, 15) and the low-density lipoprotein receptor (LDLR) (16, 17) have been also implicated in the preliminary attachment of cell-free viruses. Although the exact sequential order of receptor engagement is still unclear, some evidence suggest that HCV virions interact with scavenger receptor class B type 1 (SR-B1), CD81, tight junction proteins claudin-1 (CLDN1), occludin (OCLN), and possibly other factors (18). Virions are later internalized through clathrin-mediated endocytosis and fuse with the host membrane following endosomal acidification (19). Finally, the transferrin receptor 1 (TfR1) (20), epidermal growth factor receptor (EGFR) and ephrin receptor A2 (EphA2) (21), and Niemann-Pick C1-like 1 cholesterol absorption receptor (NPC1L1) (22) have also been implicated in HCV entry. Cell-to-cell spread has been also presented as an important route for HCV transmission within the infected liver. The requirements for this alternative route suggest that SR-B1, CLDN1, OCLN (23), EGFR, EphA2 (21), and NPC1L1 (24) are implicated for both cell-free and cell-to-cell spread.

In the present study, we characterized the anti-HCV inhibitory capacity of novel lipid-G-quadruplex conjugate structures, designated lipoquads (Fig. 1A). First, we showed that the anti-HCV potency of lipoquads is correlated with the ability of the G-rich sequences to form stable structures. Then, by using the HCV pseudoparticles (HCVpp), which are a well-established system to investigate HCV entry and neutralization (25, 26) and cell culture-produced viruses (HCVcc), we characterized the inhibition mechanism of lipoquads acting at the early steps of HCV entry, including the attachment phase, and demonstrated that this inhibition is linked to basic amino acids in hypervariable region 1 (HVR1) of the E2 glycoprotein and/or to HVR1 itself. Moreover, lipoquads also inhibit cell-to-cell HCV transmission and potently inhibit all major HCV genotypes. Finally, we provide evidence that lipoquads inhibit HCV infection ex vivo in a mucosal model based on colorectal tissue explants. Our results pave the way for the use of lipoquads in the development of prevention strategies against HCV.

FIG 1.

Quadruplex DNA structures (lipoquads) are potent HCV inhibitors. (A) Sequence and schematic structure of the parallel tetramolecular lipoquads of the present study, formed by the association of four 17-mer oligonucleotides containing the GGGGGG (orange) repeat. The structure is held by six stacked G-quartets. The 3′ extremities of the individual strands are conjugated with a cholesteryl function (magenta), forming a lipophilic tail. (B) Anti-HCV activity of lipoquads and viability of Huh7/Scr cells infected with Luc-Jc1 HCVcc viruses (at an MOI of 0.01 to 0.03 TCID₅₀/cell) and treated with increasing concentrations of lipoquads or dasatinib (positive control). (C) Anti-HCV activity and viability on Huh7/Scr cells infected with Luc-Jc1 HCVcc viruses and treated with increasing concentrations of lipoquads or A-rich control sequence. Infectivity and viability were determined 72 h postinfection by luciferase assays. Results are the means (±SEM) from two replicate infections measured in duplicates and expressed as RLU compared to the infection of control (mock) cells. The data presented are from a single experiment and are representative of those from three independent experiments.

RESULTS

Lipoquads inhibit HCVcc infection.

To evaluate the inhibitory effects of lipoquads (Fig. 1A) on HCV infection, we used the HCVcc system (27–29). Unless otherwise stated, we used the highly permissive Huh7/Scr cells (30) for HCV propagation in vitro. Furthermore, we used the highly infectious genotype 2a HCVcc virus (Jc1 chimera [31]). To simplify the quantification of infection, we used the bicistronic Jc1 luciferase reporter construct designated Luc-Jc1 HCVcc (10). Cell viability was monitored in parallel by a commercial ATP assay (32). Briefly, cells were preincubated for 1 h at 37°C with lipoquads and then inoculated with the virus in the presence of lipoquads for 4 h at 37°C. After this, virus-containing medium was replaced with a fresh medium-lipoquad mixture. Luciferase activity was assayed 72 h postinfection. As shown in Fig. 1B, Luc-Jc1 HCVcc infection of Huh7/Scr cells was completely inhibited by lipoquads at concentrations in the 5 to 10 μM range (maximum percent inhibition [MPI]) without affecting cell viability. The half-maximal inhibitory concentration (IC₅₀) for lipoquads was estimated to be ∼0.8 μM, with a half-maximal cytotoxic concentration (CC₅₀) was above 10 μM. Under the same infection conditions, dasatinib (21), an FDA-approved anticancer kinase inhibitor which has been shown to inhibit HCV entry, inhibited HCV with an IC₅₀ of ∼3.2 μM and a CC₅₀ of 51.8 μM. Disruption of the G-quadruplex structure by exchanging three of the guanines with adenines resulted in an ∼7.8-fold increase of the IC₅₀ (Fig. 1C), indicating a certain specificity mediated by the G-quadruplex element.

Lipoquads inhibit the infection of HCVpp.

To examine if lipoquads inhibit HCV entry, we used HCVpp that carried the same E1E2 glycoproteins as Luc-Jc1 HCVcc. HCVpp are a well-established system to investigate HCV entry and neutralization (25, 26). HCVpp infection conditions with lipoquads were similar to those used with Luc-Jc1 HCVcc, except that Huh7/Scr cells were inoculated with HCVpp for 6 h. Dasatinib was used again as a positive control. At a 5 μM concentration, which has been shown previously to inhibit Luc-Jc1 HCVcc infection potently, lipoquads inhibited HCVpp entry ∼8-fold, while pseudoparticles carrying the vesicular stomatitis virus glycoprotein (VSV-Gpp) were inhibited only ∼2-fold (Fig. 2A). Thus, lipoquads inhibit E1E2-mediated HCV entry into target cells with a preference over the VSV-G envelope.

FIG 2.

Lipoquads preferentially inhibit HCV pseudoparticles (HCVpp). Huh7/Scr cells were pretreated for 1 h with 5 μM lipoquads or dasatinib. Then cells were inoculated with HCVpp or VSV-Gpp in the presence of compounds. Six hours later, the medium was replaced with a fresh medium-compound mixture. Infectivity and viability were determined 72 h postinfection by Renilla luciferase for infectivity (A) or Firefly luciferase assays for cell viability (B). Results are the means (±SEM) from two replicate infections measured in duplicate and expressed as RLU compared to the infection of control (lipoquads solvent for lipoquads or DMSO for dasatinib) cells for both infectivity and viability. The data presented are from a single experiment and are representative of those from three independent experiments.

Lipoquads interact with HCV and inhibit early entry steps of HCV life cycle.

To identify which HCV entry step was inhibited, we carried out time-of-addition experiments using Luc-Jc1 HCVcc viruses. To this end, we carried out 4 different incubation protocols (Fig. 3A): (i) Huh7/Scr cells were preincubated with lipoquads for 1 h prior to inoculation, (ii) Luc-Jc1 HCVcc viruses were preincubated for 1 h with lipoquads prior to inoculation and then viruses containing lipoquads were added to the cells, (iii) lipoquads were present only during infection for 4 h, or (iv) lipoquads were added to cells postinoculation (from 4 h until the time of the luciferase assays). As shown in Fig. 3B, lipoquads inhibited Luc-Jc1 HCVcc infection only when they were preincubated with Luc-Jc1 HCVcc viruses or added simultaneously to the cell-virus mixture, indicating that lipoquads act on the viral particles and/or initial steps of HCV entry. Further, in order to test if lipoquads inhibits HCV attachment on the surface of target cells, Huh7/Scr cells were incubated with Jc1 HCVcc without reporters in the presence or absence of inhibitors for 2 h at 4°C. Under these conditions, virus attaches to the cells but does not efficiently enter. Heparin sodium salt, which is known to inhibit HCV entry at the attachment step, was used as a positive control (10). After 2 h, viruses were removed, cells were washed extensively to remove the unbound virus, and the bound HCV was quantified with quantitative reverse transcription-PCR (qRT-PCR). As shown in Fig. 4, lipoquads efficiently inhibited HCV attachment, indicating a role at least in early entry steps.

FIG 3.

Lipoquads interact with HCV particles and selectively inhibit early HCV entry events. Huh7/Scr cells were inoculated with Luc-Jc1 HCVcc viruses (at an MOI of 0.01 to 0.03 TCID₅₀/cell) prepared in the absence of drugs. Lipoquads (final concentration, 1 μM) were added to the cells only before inoculation for 1 h (black), preincubated with the viruses (hatched), only during inoculation (gray), or selectively after infection (white), as schematically depicted in panel A. Infectivity and viability (B) were determined 72 h later by Firefly luciferase assays. Results are the means (±SEM) from two replicate infections measured in duplicate and expressed as RLU compared to the infection of control (mock) cells. The data are from a single experiment and are representative of those from three independent experiments. ns, not significant.

FIG 4.

Lipoquads block virus attachment to the cell. Huh7/Scr cells were placed on ice for 30 min to cool down. Cells were then incubated for 2 h with prechilled Jc1 HCVcc (MOI, ∼10 TCID₅₀s/cell) at 4°C. Lipoquads (5 μM), heparin (100 μg/ml), or buffer (mock) was added to the cells simultaneously with virus inoculation. Cell monolayers were washed 3 times with cold PBS and then lysed and subjected to RNA extraction. HCV RNA was quantified by qRT-PCR in 25 ng of total RNA. Results are the means (±SEM) from two replicate infections measured in triplicate and expressed as relative RNA compared to the viral attachment of control (mock) cells. The data are from a single experiment and are representative of those from three independent experiments.

Lipoquads inhibit neither HCV RNA translation nor HCV RNA replication.

To test whether lipoquads exert an additional effect on HCV translation and/or replication, we transfected Huh7/Scr cells with a subgenomic JFH1 luciferase replicon (SGR-JFH1). Lipoquads were present for a preincubation period of 1 h, during transfection (4 h), and posttransfection, as described previously for Luc-Jc1 HCVcc infections. RNA replication was monitored 24 h posttransfection by luciferase assays. As a control, we incubated the cells with increasing doses of the NS3-4A protease inhibitor telaprevir (VX-950). As shown in Fig. 5, lipoquads did not affect luciferase expression, while telaprevir showed a sharp reduction in luciferase expression, indicating that lipoquads do not inhibit HCV RNA translation and/or replication.

FIG 5.

Lipoquads do not inhibit HCV subgenomic replicons. Huh7/Scr cells were pretreated with lipoquads or telaprevir (VX-950) at increasing concentrations (0.004, 0.016, 0.08, 0.4, 2, and 10 μM) for 1 h, transfected with SGR RNA (compounds present), and seeded in 96-well plates. Four hours posttransfection, Lipofectamine-RNA-containing medium was replaced with a fresh medium-compound mixture, and Firefly luciferase activity was assayed 24 h later. In both panels A and B, results are the means (±SEM) from two replicate transfections measured in duplicate and expressed as RLU compared to the transfection of control (mock) cells for both infectivity and viability. The data are from a single experiment and are representative of those from three independent experiments.

Basic residues of HVR1 and HVR1 itself play a role in lipoquad-mediated HCV neutralization.

Hypervariable region 1 (HVR1) of the HCV E2 glycoprotein is known to facilitate virus-host cell interactions (33). To investigate whether this region is involved in the lipoquad-mediated HCV inhibition, we performed inhibition experiments with HCV mutants carrying altered HVR1 regions (Fig. 6). To this end, we used Luc-Jc1 wild-type (WT) HCVcc viruses, Luc-Jc1 HCVcc viruses which harbor a total deletion of the HVR1 region (ΔHVR1 [34]), or a Luc-Jc1 HCVcc virus mutant that possesses alanines instead of the basic amino acids in HVR1 (basic− [77]). In order to observe subtle differences in lipoquad inhibition, we used a 1 μM final concentration in this setting, which according to Fig. 1 is higher than the IC₅₀ but below 100% inhibition. As shown in Fig. 6, WT virus was inhibited up to ∼80%, while both mutants (ΔHVR1 and basic−) were not affected. This suggests a direct role of the basic amino acids of the HVR1 and/or of the HVR1 region itself in lipoquad antiviral activity.

FIG 6.

Basic amino acids in E2 HVR1 and HVR1 itself play a role in lipoquad-mediated HCV inhibition. Huh7/Scr cells were inoculated with Luc-Jc1 WT HCVcc viruses or with the indicated mutants (at an MOI of 0.01 to 0.03 TCID₅₀/cell). Lipoquads (final concentration, 1 μM) were added to the cells simultaneously with inoculation. The HCVcc-compound mixture was replaced 4 h postinfection with a fresh medium-compound mixture, and 72 h after infection, cells were assayed for Firefly luciferase activity. Results are the means (±SEM) from two replicate infections measured in duplicate (black bars) and expressed as RLU compared to the infection of control (mock) cells (white bars) for both infectivity (A) and viability (B). The data are from a single experiment and are representative of those from three independent experiments.

Lipoquads inhibit HCV cell-to-cell transmission.

To assess whether lipoquads can inhibit HCV cell-to-cell transmission, we used an agarose overlay-based assay which inhibits cell-free virus, using a previously described infection reporter system (35). In this infection reporter system, Jc1 HCVcc-infected Huh7/Scr cells act as virus donor cells and uninfected Huh7.5/EGFP-NLS-IPS as acceptor cells (35, 36). The latter cells stably express a chimeric enhanced green fluorescent protein (EGFP) which upon HCV infection is redistributed from mitochondria to the nucleus (a complete description of these cells is available in Materials and Methods). Moreover, HCV infection was evaluated by anti-NS5A immunofluorescence. For consistency with the previous experiments, the lipoquad incubation period was similar to that used for cell-free infections. As shown in Fig. 7, lipoquads and the control dasatinib efficiently inhibited HCV cell-to-cell transmission, as deduced by the number of NS5A-positive acceptor cells and EGFP redistribution to the nucleus.

FIG 7.

Lipoquads inhibit HCV cell-to-cell transmission. Huh7/Scr cells were infected with Jc1 HCVcc at an MOI of ∼10 TCID₅₀/cell and 20 h later seeded on 24-well plates. Simultaneously, Huh7.5/EGFP-NLS-IPS cells were seeded on the same 24-well plates (coculture, at a ratio 1:1 with Huh7/Scr cells), and lipoquads (10 μM) or dasatinib (10 μM) was added to the wells. The addition of the compound solvent (lipoquad solvent or DMSO) served as a negative control. Four hours later, medium was removed and cells were overlaid with 1% agarose with fresh compounds. Twenty-four hours later, the HCV infection was analyzed by immunofluorescent NS5A staining. (A) Results are the means of quantification of Huh7.5/EGFP-NLS-IPS cells in 3 independent wells by taking 3 independent pictures of different fields of each well that contained at least 200 cells in total. Cell-to-cell inhibition is expressed as the percentage of acceptor cells with the EGFP-NLS plus NS5A signal with respect to the total number of EGFP plus NS5A cells in the DMSO-treated (for dasatinib) or lipoquad buffer-treated (for lipoquads) wells. (B) Representative images of coculture cells stained with anti-NS5A antibodies. Huh7.5/EGFP-NLS-IPS cells with EGFP signal redestributed to the nucleus and cytoplasmatic NS5A signal are indicated with arrows. Magnification, ×63.

Lipoquads inhibit entry of all major HCV genotypes.

To determine the antiviral efficiency of lipoquads against all major HCV genotypes, we utilized JFH1-based reporter virus constructs (JFH1 NS3-NS5B proteins) carrying Renilla luciferase inserted at the NS5A gene and core to NS2 proteins from all major HCV genotypes: 1a (isolate TN), 1b (isolate J4), 2b (isolate J8), 3a (isolate S52), 4a (isolate ED43), 5a (isolate SA13), 6a (isolate HK6a), and 7a (isolate QC69) (37). Strikingly, lipoquads showed antiviral activity against all major HCV genotypes (Fig. 8). The estimated IC₅₀ for each HCV genotype (Table 1) was comparable to the IC₅₀ estimated in the first set of experiments for the genotype 2a Luc-Jc1 HCVcc.

FIG 8.

Lipoquads inhibit HCV isolates of different genotypes. Huh7/Scr cells were inoculated with viruses of the indicated genotypes and isolates (at an MOI of ∼0.01 to 0.03 TCID₅₀/cell). Lipoquads at the indicated concentrations were added to the cells simultaneously with inoculation. The HCVcc-compound mixture was replaced 4 h postinfection with a fresh medium-compound mixture. Seventy-two hours postinfection, cells were assayed for Renilla luciferase activity for infectivity (A) or Firefly luciferase for cell viability (B). Results are the means (±SEM) from two replicate infections measured in duplicate and expressed as RLU compared to the infection of control (mock) cells. The data are from a single experiment and are representative of those from three independent experiments.

TABLE 1.

Antiviral activity (IC₅₀) of lipoquads across all major HCV genotypes

HCV genotype	IC50 (μM)
1a (TN)	0.9
1b (J4)	0.2
2b (J8)	0.3
3a (S52)	0.3
4a (ED43)	0.9
5a (SA13)	1.2
6a (HK6a)	1.5
7a (QC69)	0.5

Lipoquads inhibit HCV ex vivo in a mucosal model.

To assess lipoquad activity ex vivo, we tested lipoquads in a mucosal model based on ex vivo viral challenge of colorectal tissue explants (38, 39). This model allowed us to evaluate the potency of lipoquads as an HCV entry inhibitor against a reporter HCV, Jc1FLAG(p7-nsGluc2A), by measurement of Gaussia luciferase expressed and secreted upon viral entry, translation, and/or replication of the HCV genome (40). Pulse exposure of explants to drug for 3 h resulted in a dose-dependent reduction of viral levels (Fig. 9A) (IC₅₀ of 11.88 ± 7.35 μM). Interestingly, the level of inhibition reached with the highest drug concentration tested was similar to that obtained with a combination of sofosbuvir (SOF) and ledipasvir (LDV) (both at a 1 μM final concentration). With sustained exposure to lipoquads (compound maintained throughout explant exposure to virus and culture), a decrease in the IC₅₀ to 1.08 ± 0.13 μM and an increase in Jc1FLAG(p7-nsGluc2A) infection inhibition up to 93% were reached within the concentration range tested (Fig. 9B). Together, these data demonstrate the potential of lipoquads as candidate drugs for topical HCV prevention strategies.

FIG 9.

Inhibitory activity of lipoquads in colorectal tissue explants against HCVcc Jc1FLAG(p7-nsGluc2A). Mucosal explants were treated for 1 h in the presence or absence of lipoquads (at the indicated concentrations) or SOF-LDV (both at a 1 μM final concentration) prior to viral exposure for 2 h (total compound presence, 3 h). Explants were then washed four times with PBS and transferred to gel foam rafts. Tissue explants were kept in culture for 48 h in the absence (pulse) (A) or in the sustained presence of compounds (B). Culture supernatants were harvested for detection of Gaussia luciferase activity, and the extent of infection was plotted as a percentage relative to the RLU obtained for explants infected with virus in the absence of compound (100% infectivity). Data represent the means (±SEM) of three independent experiments performed in triplicate.

DISCUSSION

The introduction of DAA therapies in the standard-of-care (SoC) treatment of HCV has increased significantly the sustained virological response (SVR) rates, accompanied by manageable adverse effects (41). However, these compounds target one of the nonstructural proteins of HCV (NS3-4A, NS5A, and/or NS5B) and thus cannot prevent HCV infection. Here, we report a class of novel oligonucleotide-lipid conjugates forming G-quadruplex structures, designated lipoquads, as potent and pangenotypic HCV entry inhibitors in vitro. Moreover, we were able to show anti-HCV activity ex vivo in an intestinal mucosa explant model, using genotype 2a HCVcc. Notably, lipoquads exhibited IC₅₀s comparable to those of known FDA-approved drugs like dasatinib and telaprevir (VX-950). In addition, lipoquads were able to achieve an MPI at ∼5 to 10 μM. These data for lipoquads as antiviral compounds are important if breakthrough and/or resistance needs to be tested. With their absence of toxicity at the concentrations tested and their high water solubility, lipoquads present attractive candidates for the development of HCV prevention strategies.

To characterize the lipoquads' inhibition mechanism, we used several molecular tools in diverse infection assays. First, we used the HCVpp system in order to isolate the entry process from other HCV functions, i.e., replication/translation and assembly/release. Because lipoquads inhibited HCVpp and had no effect on the subgenomic replicon, we conclude that lipoquads target viral entry. HCV entry is a complex process that requires several entry molecules (see references 42 and 43 for reviews). It comprises the steps from particle binding to the host cell up to the delivery of the viral genome to the replication site within the cell. Initial binding is mediated by interactions between HCV E1E2 envelope glycoproteins and glycosaminoglycans (GAGs) (11). Also, LDLRs on host cells may function as initial attachment factors due to the association of HCV with LDL or very-low-density lipoprotein (16, 17). Following this initial engagement, tetraspanin CD81 and SR-B1 together with the tight junction proteins CLDN1 and OCLN are the main receptors contributing to HCV uptake (44–47). Our results show that lipoquads inhibit HCV infection by negatively affecting virus-cell binding. This effect is presumably mediated via direct interaction of the compound with the virus, since preexposure of the cells to lipoquads did not result in decreased infectivity. Lipoquads are therefore responsible for blocking the interaction of the virus with the cellular receptors.

To shed more light onto the mechanism of interaction between lipoquads and HCV, we used two HCV mutants for the E2 HVR1, one in which all basic amino acids of the region were replaced by alanines (basic−) and one that lacked the region itself (ΔHVR1). Importantly, neither of them was inhibited by lipoquads. Although HVR1 is not essential for HCV productive infection, viruses lacking this domain are less infectious, both in vitro (48) and in vivo (34). HVR1 displays high genetic variability between HCV isolates, which likely contributes to immune evasion of HCV. Previous experiments performed with HVR1 deletion mutants suggest that HVR1 may act as immunological decoy since it shields conserved neutralizing epitopes (48). Indeed, ΔHVR1 mutants show increased neutralization susceptibility to HCV patient sera (48, 49). It has been proposed that a complex interplay between several regions of E2 is responsible for modulating receptor binding, possibly through intramolecular interactions (50). As deduced from our results, lipoquads may affect this interplay. Particularly, lipoquads display negative charges, so a positive-negative electrostatic interaction is likely to occur. Furthermore, alignment of the HVR1 domains of the different HCV genotypes reveals a conserved presence of positively charged residues (data not shown). Thus, HCV pangenotypic activity of lipoquads can also be explained by the interaction with basic amino acids in the E2 glycoprotein. Overall, although conclusive evidence for the target of lipoquads is lacking, we hypothesize that lipoquads use a common mechanism that involves electrostatic interactions with positively charged residues in viral entry proteins.

HCV entry into target hepatocytes, as cell-free virus, has been proven to be a well-orchestrated process with spatiotemporal requirements. Notably, in addition to infection by cell-free virus, direct cell-to-cell transmission also occurs in the liver (51) and in cultured hepatocytes (53–57). This route of viral spread may provide a way to avoid neutralization, resulting in viral persistence and hampering viral eradication (57). DAA-resistant variants have been shown to use cell-to-cell transmission as the main route of viral spread in cell culture (58). Hence, lipoquads possess a great additional value owing to their cell-to-cell virus transmission inhibition capacity. Most host entry factors are shared between the two routes. These include SR-B1, CLDN1, OCLN, and NPC1L1 as well as EGFR and its signal transducer HRas (21, 24, 59–61). The dependence on CD81 in cell-to-cell transmission remains controversial (23, 62). While the role of these factors during HCV cell-free transmission has been extensively studied, mechanistic insights about their spatiotemporal role during HCV cell-to-cell spread are still far from being satisfactory. Thus, although lipoquads appear to inhibit initial HCV cell-free attachment and therefore interfere with the interaction of HCV with attachment molecules/receptors, the exact mechanism of HCV cell-to-cell inhibition by lipoquads needs further investigation. Importantly, the majority of monoclonal antibodies targeting the viral envelope have been shown to fail to inhibit cell-to-cell transmission. In contrast, several host-targeting entry inhibitors (HTEIs) have succeeded in blocking this route of viral spread (21, 59). Since lipoquads target directly the viral particles and do not interfere with cellular processes, they present a significant advantage against HTEIs.

Antivirals formulated for topical applications are known as microbicides, and viral entry into the target cell represents an important point where microbicides can inhibit mucosal transmission. In more detail, the intestinal mucosa is composed of intestinal epithelial cells (enterocytes, secretory cells, and intestinal epithelial stem cells), stromal cells, and myofibroblasts. Colorectal tissue also contains T cells, B cells, dendritic cells, macrophages, and innate lymphoid cells in the lamina propia (63, 64). Furthermore, the expression of certain HCV entry receptors in the intestinal mucosa has been described previously (65–69). These cells could be the primary target cells for virus infection through mucosal transmission. In the field of HIV prevention, proof of concept that microbicides can block HIV transmission was obtained in a clinical efficacy trial (70), although ultimately, the efficacy of any microbicide product depends upon adherence as well as appropriate drug delivery (71). Microbicides are topical antivirals, and therefore, their inhibitory activity needs to be considered within the context of mucosal transmission. Mucosal tissue explant models are becoming an essential tool for preclinical screening of candidates for oral and topical preexposure prophylaxis against HIV (39) and are increasingly being used in HIV prevention clinical trials (72–75). In this study, we expanded the usage of this mucosal model to assess the inhibitory potency of lipoquads against HCV. Our results are important because they clearly show HCV transmission through the mucosa, although HCV primarily infects and replicates in hepatocytes. Nevertheless, we were able to inhibit HCV intestinal mucosa infection by DAA, suggesting that HCV replicates in this ex vivo model as well. Furthermore, pulse exposure of the tissue to lipoquads showed limited potency, reaching a maximum of 55% of inhibition; however, sustained exposure showed a significant increase of activity, with a decrease in the IC₅₀ and an increase in the level of inhibition reached at the highest concentration tested. Because microbicides are topically applied, higher local drug concentrations can be delivered to mucosal surfaces without significant systemic exposure, thereby reducing the risk of long-term toxicity in healthy but at-risk individuals.

In conclusion, permucosal HCV transmission has been confirmed as the most likely mode of HCV infection in MSM and has become a significant source of new HCV infections. In essence, HCV sexual transmission concerns the general population, independently of sexual preferences. This calls for new antiviral drugs that permit an efficient topical preexposure prophylaxis. Our work in this study demonstrates that lipoquads are efficient pangenotypic HCV entry inhibitors with interesting properties that make them candidates for further studies on drug formulations and dosing evaluations. Providing safe and acceptable microbicides to the community at risk to sexually acquire HCV will be an important step in HCV transmission control. Lipoquad presents itself as a candidate compound for development in this direction.

MATERIALS AND METHODS

Oligonucleotide synthesis and G-quadruplex preparation.

Oligonucleotide sequences 5′-TTGGGGGGTACAGTGCA-3′-cholesterol and the A-rich control sequence 5′-TTGAAAGGTACAGTGCA-3′-cholesterol were assembled using an automatic oligonucleotide synthesizer (Applied Biosystems 3400; Applied Biosystems, Foster City, CA). The solid support functionalized with cholesterol and the rest of the chemicals were from commercial sources (Link Technologies, Scotland, UK). After the assembly of the sequences, supports were treated with ammonia and the desired oligonucleotides were purified by reverse-phase high-performance liquid chromatography (HPLC). The oligonucleotides were next suspended at a concentration of 1.6 mM in an annealing buffer (lipoquad solvent, 20 mM Tris acetate [pH 7.0], and 50 mM potassium acetate), boiled for 5 min at 95°C, slowly cooled down to room temperature, and incubated at 25°C for at least 14 days. G-quadruplex formation was followed by 10% nondenaturing polyacrylamide gel electrophoresis as described previously (76) and verified by circular dichroism. The compounds were stored at −20°C until use.

Cell culture.

The human hepatocarcinoma cell lines Huh7/Scr and Huh7.5.1 Cl.2 (kindly provided by F. Chisari) and the human embryonic kidney cell line 293T (HEK293T cells; American Type Culture Collection, Manassas, VA; CRL-1573) were maintained in Dulbecco's modified Eagle medium (DMEM; Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum, 10% nonessential amino acids, 100 U/ml of penicillin, and 100 U/ml of streptomycin. Cells were grown in an incubator with 5% CO₂ at 37°C.

Plasmids. (i) Plasmids used to produce HCVcc.

Plasmid pFK-Jc1 has been described previously (31). It encodes a chimeric HCV consisting of codons 1 to 846 derived from J6/CF (genotype 2a; GenBank accession number AF177036) combined with codons 847 to 3033 of JFH1 (genotype 2a; GenBank accession number AB047639). Plasmid pFK-Luc-Jc1 (27) consists of a bicistronic construct in which the HCV polyprotein-coding region is located in the second cistron and is expressed via an internal ribosome entry site (IRES) element derived from encephalomyocarditis virus (EMCV), while the first cistron contains the Firefly luciferase reporter gene. Plasmids encoding the HVR1 mutants are also based on the Jc1 genome and have been described elsewhere (77). Briefly, these plasmids are pFK-Luc-Jc1 derivatives. The pFK-Luc-Jc1-ΔHVR1 plasmid encodes a 27-amino-acid deletion of the E2 HVR1 region, while the pFK-Luc-Jc1-basic plasmid possesses an alanine substitution in all basic amino acids of the HVR1 (basic− mutant). HCV genotype 1 to 7 plasmids are JFH1-based reporter virus constructs (NS3 to NS5B of JFH1 origin; genotype 2a), carrying Renilla luciferase inserted at the NS5A gene and expressing core nonstructural protein 2 (NS2) of genotype 1 to 7 prototype isolates (37). HCV Jc1FLAG(p7-nsGluc2A) is a cell culture-derived virus chimera of J6 and JFH1 (genotype 2a/genotype 2a chimera), which is fully infectious and carries the Gaussia luciferase gene as a reporter gene (40).

(ii) Plasmid used to produce HCV subgenomic replicon.

The subgenomic replicon plasmid carries a bicistronic construct in which a Firefly luciferase gene is expressed via HCV IRES and an EMCV IRES drives expression of JFH1 nonstructural proteins (NS3 to NS5B) (78).

(iii) Plasmids used to produce HCVpp.

Plasmid pTN7-Stopp (kindly provided by M. Dittmar) carries the HIV-1 genome with the following modifications: the Renilla luciferase reporter gene has replaced the nef gene, and the plasmid lacks a functional env gene (79). The plasmid, which encodes strain HC-J6CH E1E2 glycoproteins and is designated pcDNA3.1-ΔcE1E2-J6CH, has been described elsewhere (77). Plasmid pVPack-VSV-G, which encodes the vesicular stomatitis virus glycoprotein (VSV-G), has been purchased by Agilent Technologies (Santa Clara, CA).

In vitro transcription, electroporation, and preparation of virus stocks.

Plasmids carrying Jc1 constructs were linearized with the MluI enzyme, while plasmids carrying genotype 1-7/JFH1 chimeric viruses and the Gaussia reporter were linearized with the XbaI enzyme. Plasmid DNA was purified with a QIAquick PCR purification kit (Qiagen, Düsseldorf, Germany). Purified DNA was subjected to an in vitro transcription reaction with a MEGAscript T7 kit (Applied Biosystems, Foster City, CA) according to the manufacturer's protocol. RNA from the in vitro transcription reaction was purified with a Nucleospin RNA II kit (Macherey-Nagel, Düren, Germany), RNA integrity was verified by formaldehyde agarose gel electrophoresis, and the concentration was determined by measurement of the optical density at 260 nm. For RNA electroporations, single cell suspensions of Huh7.5.1 Cl.2 cells were prepared by trypsinization of cell monolayers. Cells were washed with phosphate-buffered saline (PBS), counted, and resuspended at 1.5 × 10⁷ per ml in cytomix (120 mM KCl, 0.15 mM CaCl₂, 10 mM K₂HPO₄/KH₂PO₄ [pH 7.6], 25 mM HEPES, 2 mM EGTA, 5 mM MgCl₂; final pH of 7.6, adjusted with KOH) (80) containing 2 mM ATP and 5 mM glutathione. Ten micrograms of in vitro-transcribed RNA was mixed with 400 μl of the cell suspension. Cells were then electroporated, immediately transferred to 10 ml of culture medium, and seeded in a 10-cm dish. Electroporation conditions were 975 μF and 270 V, using a Gene Pulser Xcell system (Bio-Rad, Munich, Germany) and a cuvette with a gap width of 0.4 cm (Bio-Rad) were used. Supernatants of the electroporated cells were harvested 72 h postelectroporation, cleared by passing them through 45-μm-pore-size filters, and stored at −80°C.

For the determination of viral titers, Huh7/Scr cells were seeded at a concentration of 4 × 10⁴ per well in a 96-well plate in a total volume of 200 μl. Twenty-four hours later, serial dilutions of virus containing supernatant were added (6 wells per dilution). Three days later, cells were washed with PBS and fixed for 20 min with ice-cold methanol at −20°C. After three washes with PBS, NS5A was detected with a 1:2,000 dilution of mouse anti-NS5A antibody 9E10 (kindly provided by C. Rice, The Rockefeller University, NY) in PBS supplemented with 5% bovine serum albumin (BSA) for 1 h at room temperature. Cells were washed again three times with PBS, and the bound primary antibodies were detected by incubation in PBS plus 5% BSA with goat anti-mouse IgG peroxidase-conjugated antibody (Sigma-Aldrich, St. Louis, MO) at a 1:400 dilution. After 1 h of incubation at room temperature, cells were washed three times with PBS; the Vector NovaRED substrate kit (Linaris Biologische Produkte GmbH, Wertheim, Germany) was used for detection of peroxidase. Virus titers (50% tissue culture infective dose [TCID₅₀] per milliliter) were calculated based on the method described by Spearman (81) and Kärber (82).

Preparation of HCVpp and VSV-Gpp.

HIV-based pseudoparticles bearing HCV and VSV-G glycoproteins (HCVpp and VSV-Gpp) were generated by calcium phosphate cotransfection of 293T cells. Briefly, 3.6 × 10⁶ 293T cells were seeded in 10-cm dishes 1 day before transfection with equal amounts of pTN7-Stopp plasmid and pcDNA3.1-ΔcE1E2-J6CH or pVPack-VSV-G (Agilent Technologies, Santa Clara, CA) for HCVpp or VSV-Gpp, respectively. A total amount of 20 μg of DNA was mixed with a 2 M CaCl₂ solution and then 2× HEPES-buffered saline (HBS) was added dropwise to form a precipitate which was added to the cells. The medium was replaced the following day and supernatants containing the pseudoparticles were harvested 48 h later, cleared by passage through 0.45-μm-pore-size filters, and used for luciferase infection assays.

Luciferase infection assays.

For standard infection assays. Huh7/Scr cells were seeded at a density of 4 × 10⁴/well in 96-well plates. One day later, cells were preincubated for 1 h at 37°C with the pertinent compounds and then inoculated with the virus and the compounds for 4 h at 37°C. HCVpp were left for 6 instead of 4 h. Finally, virus-containing medium was replaced with a fresh medium-compound mixture. Luciferase activity was assayed 72 h postinfection. Cells were washed with PBS, lysed in 150 μl of passive lysis buffer, and frozen. Upon thawing, lysates were resuspended by pipetting and 50 μl was mixed with 25 μl of a luciferin solution and measured in a luminometer for 2 s. The luciferin solution was LARII for Firefly luciferase assays and Stop&Glo reagent (Promega) for Renilla luciferase assays. The BioLux Gaussia luciferase assay kit (Promega) was used to assess Gaussia luciferase activity according to the manufacturer's instructions. Cytotoxicity (viability) was measured in all infection assays using the CytoTox-Glo cytotoxicity assay (Promega) as described by the manufacturer with a FLUOstar OPTIMA plate luminometer (BMG LABTECH) according to the manufacturer's instructions. The multiplicity of infection (MOI) used in the infections was 0.01 to 0.03 TCID₅₀/cell. Unless otherwise stated, results for both infectivity and viability are the means (±standard errors of the means [SEM]; n = 4) from two replicate infections measured in duplicate and expressed as relative light units (RLU) compared to those obtained in the infection of control (mock) cells.

Attachment assay and qRT-PCR.

Huh7/Scr cells were seeded in 24-well plates at 1.5 × 10⁵/well. The following day, cells were set on ice for 30 min to cool down and then incubated with prechilled Jc1 HCVcc (at an MOI of ∼10 TCID₅₀/cell) in the presence or absence of compounds for 2 h at 4°C. Cells were washed 3 times with ice-cold PBS and lysed, and RNA was extracted using the Nucleo Spin RNA II kit (Macherey-Nagel, Düren, Germany) by following the manufacturer's protocol. RNA concentration was determined by measurement of the optical density at 260 nm. Twenty-five micrograms of the total RNA sample was used for quantitative PCR analysis using a 7500 real-time PCR sequence detector system (Applied Biosystems, Waltham, MA). HCV-specific qRT-PCRs were conducted in duplicate for each sample with the OneStep RT-PCR kit (Qiagen, Hilden, Germany) using the following HCV 5′ untranslated region (UTR)-specific probe S-29 (5′-6-carboxyfluorescein-CCTGATAGGGTGCTTGCGAGTGCC-tetrachloro-6-carboxyfluorescein-3′) and primers S-271 (5′-GCGAAAGGCCTTGTGGTACT-3′) and A-337 (5′-CACGGTCTACGAGACCTCCC-3′) (Biomers, Ulm, Germany). Reactions were performed in three stages by using the following conditions: stage 1, 60 min at 55°C (reverse transcription); stage 2, 15 min at 95°C (heat inactivation of reverse transcriptase and activation of Taq polymerase); and stage 3, 40 cycles of 15 s at 95°C and 1 min at 60°C (amplification). The total volume of the reaction mixture was 15 μl, and it contained the following components: 2.66 μM 6-carboxy-X-rhodamine (passive reference), 4 mM MgCl₂, 0.66 mM deoxynucleoside triphosphates, 0.266 μM HCV probe, 1 μM each HCV primer, and 0.6 μl of enzyme mix. The amount of HCV RNA was calculated by comparison to serially diluted in vitro transcripts.

Subgenomic replicon assay.

Huh7/Scr cells were seeded in 24-well plates at 1.5 × 10⁵/well. The following day, cells were preincubated for 1 h at 37°C with the pertinent compounds (lipoquads or telaprevir [VX-950], the latter has been purchased by Selleck Chemicals, Houston, TX) and then transfected with subgenomic replicon RNA using the Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Four hours posttransfection, Lipofectamine-RNA-containing medium was replaced with a fresh medium-compound mixture and Firefly luciferase activity was measured 24 h later as described above.

Cell-to-cell transmission assays.

The first day, Huh7/Scr cells were infected with Jc1 HCVcc at an MOI of ∼10 TCID₅₀/cell. These cells were then used as HCV donor cells, while Huh7.5-EGFP-NLS-IPS cells were used as acceptor cells. The latter cells stably express a chimeric protein that encompasses the enhanced green fluorescent protein (EGFP) associated with the simian virus 40 (SV40) nuclear localization sequence (NLS) followed by the mitochondrially tethered interferon beta promoter stimulator protein 1 (IPS-1; all together, EGFP-NLS-IPS), which upon HCV infection is redistributed from mitochondria to the nucleus (35, 36). The following day, a 1:1 ratio of donor to acceptor cells was used and a total of 2 × 10⁵ cells/well were plated on 24-well chambered coverglasses with a medium containing lipoquads or dasatinib that inhibits HCV cell-to-cell spread (21), lipoquad buffer (0.1% vol/vol), or dimethyl sulfoxide (DMSO; 0.1% [vol/vol]). Cells were covered with fresh medium containing 1% low-melting-temperature agarose and compounds 4 h after seeding and further cultured for 20 h. Finally, cells were fixed with 4% paraformaldehyde and stained with anti-NS5A antibodies. Cell-to-cell spread was analyzed in a Leica TCS-SP5 confocal microscope. Quantification of Huh7.5/EGFP-NLS-IPS cells was done in 3 independent wells by taking 3 independent pictures of different fields of each well that contained at least 200 cells in total. Data are expressed as percentages in relation to lipoquad solvent-treated (for lipoquads) or DMSO-treated (for dasatinib) cells and represent mean values of the 3 independent fields of three biological replicates (±SEM).

Patients and tissue explants.

Surgically resected specimens of colorectal tissue were collected at St. Mary's Hospital, Imperial College London, UK. All tissues were collected after receiving signed informed consent from all patients and under protocols approved by the Local Research Ethics Committee. The tissue was obtained from patients undergoing rectocele repair and colectomy for colorectal cancer. Only healthy tissue that was 10 to 15 cm away from any tumor was used. All patients were HIV and HCV negative. On arrival in the laboratory, resected tissue was cut into 2- to 3-mm³ explants comprising both epithelial and muscularis mucosae as described previously (38). Colorectal explants were maintained with DMEM containing 10% fetal calf serum, 2 mM l-glutamine, and antibiotics (100 U of penicillin/ml, 100 μg of streptomycin/ml, and 80 μg of gentamicin/ml) at 37°C in an atmosphere containing 5% CO₂.

Tissue inhibition assays.

Tissue explants were incubated with lipoquads or DAA (sofosbuvir [SOF] and ledipasvir [LDV], both at a 1 μM final concentration; SOF and LDV have been purchased by Selleck Chemicals, Houston, TX) for 1 h prior to virus addition for 2 h. Explants were then washed 4 times with PBS to remove unbound virus and drug, transferred on gel foam rafts (Welbeck Pharmaceuticals, UK), and cultured in complete medium in the presence (sustained) or absence (pulse) of drug for 48 h at 37°C. Viral levels were measured by Gaussia luciferase quantification (Promega, Madison, WI) in a Synergy HT multidetection microplate reader (BioTek Instruments, Inc., Burlington, VT) as described above.

Statistical and mathematical analysis.

Statistical comparison between two groups was made by an unpaired t test. Statistically significant differences are shown in figures as follows: *, P < 0.05; **, P < 0.01; and ***, P < 0.001. IC₅₀s were calculated from sigmoid curve fitting (GraphPad Prism; GraphPad Software, La Jolla, CA) fulfilling the criterion of R² > 0.7.