Requirement for Chloride Channel Function during the Hepatitis C Virus Life Cycle

Hepatocytes express an array of plasma membrane and intracellular ion channels, yet their role during the hepatitis C virus (HCV) life cycle remains largely undefined. Here, we show that HCV increases intracellular hepatic chloride (Cl−) influx that can be inhibited by selective Cl− channel blockers. Through pharmacological and small interfering RNA (siRNA)-mediated silencing, we demonstrate that Cl− channel inhibition is detrimental to HCV replication. This represents the first observation of the involvement of Cl− channels during the HCV life cycle.

H epatitis C virus (HCV) is a major human pathogen that causes chronic liver disease, including hepatocellular carcinoma and cirrhosis (1). Between 130 and 170 million people are now chronically infected, and HCV-related disease leads to 350,000 deaths per year (2). Recent developments have led to the discovery of new direct-acting antivirals (DAA) (3), including the HCV polymerase (NS5B) inhibitor sofosbuvir. While these DAAs signify the encouraging progress of HCV treatment regimens, issues of their high cost coupled with potential resistance remain. Thus, research into new antiviral targets is still required (4).
HCV displays tropism primarily for hepatocytes, the parenchymal cells of the liver (6,8). Hepatocytes are multifunctional epithelial cells that engage in transcellular solute transport, processing of metabolites, and the synthesis and secretion of numerous important proteins. In common with all eukaryotic cells, hepatocytes possess ion channels at the plasma membrane and in multiple intracellular compartments (9,10). We have previously shown that HCV NS5A can inhibit a hepatic proapoptotic host cell K ϩ channel (Kv2.1, KCNB1) to maintain the survival of infected cells (10) by perturbing signaling leading to Kv2.1 activation (11). There are over 230 genes encoding ion channel subunits in the human genome (12), but no further functional role of these channels during HCV pathogenesis has been assigned. We therefore assessed the effects of modulating all the major cellular ion channel families on the HCV life cycle.
To determine if the activities of cellular ion channels are required during the HCV life cycle, we first assessed virus genome replication using the bicistronic JFH-1 genotype 2a subgenomic replicon (SGR), which expresses a firefly luciferase-neomycin phosphotransferase fusion protein (SGR-Feo-JFH-1) (13). A cell line stably harboring this SGR was treated with compounds previously characterized as modulating specific ion channel families. Replication was monitored through the measurement of luciferase activity and confirmed by Western blotting for NS5A. Daclatasvir (DCV) was included as a known inhibitor of virus replica-tion (14); it reduced luciferase activity by 94% Ϯ 5% (Fig. 1A) and reduced NS5A expression to undetectable levels. As shown in Fig.  1B to D, there were no effects on luciferase activity or NS5A expression when cells were treated with tetraethylammonium (TEA), a broadly acting blocker of potassium (K ϩ ) channels (15), 4-aminopyridine (4AP), a blocker of voltage-gated K ϩ channels (16), or KCl to collapse K ϩ gradients. Similarly, when cells were treated with blockers of voltage-gated Na ϩ channels (NaV), tetrodotoxin (TTX), disopyramide phosphate (DP), and lidocaine (17,18) or with inhibitors of plasma membrane Ca 2ϩ channels, namely, nifedipine (Nif) and nimodipine (Nim) (19), no reduction in genome replication 24 h after compound treatment occurred ( Fig. 1B to D). We thus concluded that the inhibition of K ϩ , Na ϩ , and Ca 2ϩ channels does not have an impact on HCV genome replication.
Since SGR-Feo-JFH-1-harboring cells express only the HCV  Figure 3A shows that NPPB and IAA-94 treatment significantly decreased J6/JFH-1 RLuc activity (67% Ϯ 20% and 63% Ϯ 5% inhibition, respectively) confirming a dependence on Cl Ϫ influx during the virus life cycle. When these assays were performed in the presence of DIDS (100 M), J6/JFH-1 RLuc activity also decreased by 77% Ϯ 4% at concentrations that did not affect SGR-Feo-JFH-1 replication (Fig. 3A). To verify these data, we directly infected Huh7 cells with full-length JFH-1 virus (25) in the presence of each Cl Ϫ inhibitor and measured the production of infectious virions by focus-forming assay. As shown in Fig. 3B, virus yields were significantly lower in IAA-94-, NPPB-, and DIDS-treated cells (87% Ϯ 14%, 81% Ϯ  Fig. 1. (E) Huh7 cells were infected with JFH-1 supernatants for 24 h. Cells were washed and replaced with medium plus compound for a further 48 h. Virus supernatants were collected, and virus production was assessed by focus-forming assays as described for panel B. (F) Compounds were mixed with virus inoculum in DMEM for 1 h and Huh7 cells infected at 37°C for 3 h. Cells were washed three times in phosphate-buffered saline (PBS) to remove unbound virus, and virus production was assessed 48 postinfection by focus-forming assays of virus supernatants. AP33 was included in these assays to inhibit HCV entry (50 mg/ml). All results were calculated relative to values for the untreated controls. **, significant difference from control value (P Ͻ 0.05); NS, no differences at the 0.05 significance level.
23%, and 72% Ϯ 22% inhibition, respectively). This was paralleled by a decrease of both NS5A and core protein expression in virus lysates as assessed by Western blot analysis (Fig. 3C). No effects on JFH-1 virus production were observed when TEA or KCl was assessed in these assays (Fig. 3D). We subsequently performed time-of-addition focus reduction assays using JFH-1 virus inoculum to assess the effects of DIDS over the time course of HCV infection. Cells were treated with each inhibitor 24 h p.i., and virus production was assessed 72 h p.i. Figure 3E shows that DCV, NPPB, and IAA-94 reduced JFH-1 virus production when added postinfection (92% Ϯ 9%, 81% Ϯ 23%, and 72% Ϯ 22% inhibi-tion, respectively), consistent with a block of HCV replication. DIDS however, failed to reduce virus production relative to that in the untreated wells, consistent with a lack of inhibition of HCV replication. To further determine which steps of the HCV life cycle are impaired by DIDS, we examined the effects of each Cl Ϫ channel inhibitor on virus entry by adding them to JFH-1 inoculum during the initial 3 h of virus infection (26). The HCV-neutralizing mouse monoclonal E2 antibody AP33, a characterized inhibitor of HCV entry, was included in these assays for verification (27). Figure 3F shows that, while AP33 (50 g/ml) inhibited HCV entry by 72% Ϯ 11%, IAA-94, NPPB, and DIDS did not impede viral entry. These observations suggest that a DIDS-sensitive Cl Ϫ channel can inhibit early postentry virion trafficking and/or early replication events but does not inhibit virus entry or replication following the establishment of infection. Given these data, we investigated the molecular identity of the Cl Ϫ channel(s) required during the HCV life cycle. To date, nearly 40 different genes that, when expressed, increase Cl Ϫ conductance have been cloned. These include the Cl Ϫ intracellular-channel (CLIC) proteins cyclic AMP (cAMP) (CFTR)-, calcium (CaCC)-, voltage-activated Cl Ϫ channels and Cl Ϫ /H ϩ exchangers (CLCs) as well as ligand-gated Cl Ϫ channels [GABA(A), GABA(C), and glycine]. In hepatocytes, CLIC-1, ClC-2, ClC-3, ClC-5, and ClC-7 are expressed (9). We confirmed this by reverse transcription-PCR (RT-PCR) analysis (primer sequences are available upon request) and silenced this expression through small interfering RNA (siRNA) transfection (Fig. 4A). Figure 4B shows that ClC-2, ClC-3, ClC-5, and ClC-7 silencing significantly suppressed SGR-Feo-JFH-1 replication (52% Ϯ 6%, 31% Ϯ 16%, 48% Ϯ 2%, and 50% Ϯ 10% inhibition of luciferase activity, respectively). CLIC-1 knockdown displayed no discernible effects. Since some of these CLC channels and transporters are sensitive to NPPB and IAA-94; this confirmed the importance of Cl Ϫ influx during HCV replication.
It is interesting to address what might be the molecular mechanisms underpinning the differential effects of CLIC-1, ClC-5, and ClC-7 silencing. While data on the biological role of CLIC-1 are limited, its activity has been shown to be required for the regulation of endosomal/lysosomal pH (9,28). This may explain our observations, since the acidic late endosome/lysosome pH is crucial for induction of HCV glycoprotein (E1/E2) membrane fusion during early HCV postentry events to allow HCV genome release (29). The fact that ClC-5 and -7 are dispensable for J6/ JFH-1 RLuc infectivity suggests that their function may be compensated for by HCV core-NS2 expression. ClC-5 is a known 2Cl Ϫ /H ϩ exchanger rather than a Cl Ϫ channel, the function of which is to control endosomal acidification (30,31). The HCV viroporin p7 is thought to form cationic intracellular channels that promote a global loss of organelle acidity (32,33). p7 activity may thus prevent the buildup of an excess positive charge in specific organelles, a principle typically achieved by the import of Cl Ϫ via anion transporters, including ClC-5.
Considering our findings together, we have confirmed the role of several Cl Ϫ channel proteins during the HCV life cycle. Of note, we have identified ClC-2 and ClC-3, whose activities are required during HCV replication. Endosomal acidification and [Cl Ϫ ] i accumulation are significantly impaired in hepatocytes from ClC-2/3 knockout mice (20), and fractionation studies have suggested that these channels reside in early/late endosomes (34). The organization, composition, and functions of membrane structures in-duced by positive-strand RNA viruses remain largely ill defined but are generally accepted to require endosome integrity to recruit endosomal host cell factors and concentrate virus proteins to produce viral factories. Here, for the first time, we implicate host cell Cl Ϫ influx through CLC channels/transporters during this process. The challenge will now be to define the specific virus-host interactions that require Cl Ϫ channel functionality.