Introduction

Hepatitis C virus (HCV) is a causative agent of chronic hepatitis, liver cirrhosis and hepatocellular carcinoma (HCC); therefore, the eradication of HCV from an infected liver could reduce death from HCV-related liver disease. Combination therapy of PEGylated-interferon (PEG-IFN) and ribavirin has long been the standard of care for patients with chronic hepatitis C (CH-C); however, a sustained viral response (SVR) is obtained in only ~50% of treated patients infected with genotype 1 HCV1. Recently, several classes of direct-acting antiviral agents (DAAs) have entered into clinical use. In the United States, two NS3/4A protease inhibitors, telaprevir and boceprevir, were approved for use in combination with PEG-IFN and ribavirin in 2011. Although the addition of these DAAs dramatically improves the SVR rate, 20–30% of patients still fail to eradicate HCV due to breakthrough by drug-resistant mutants or null response to therapy2. More potent DAAs are currently in late clinical development and promise much higher SVR rates even in the absence of PEG-IFN therapy; however, HCV-related HCC is likely to continue to be a significant clinical issue for many years because it will take time for potent DAAs to be distributed worldwide.

Peretinoin (generic name code: NIK-333) is an oral acyclic retinoid with a vitamin A-like structure that targets retinoid nuclear receptors, such as retinoid X receptor and retinoic acid receptor. The oral administration of Peretinoin significantly reduces the incidence of post-therapeutic HCC recurrence and improves the survival rate of patients in clinical trials3,4. In addition, Peretinoin prevents the development of hepatoma in several different hepatoma models5,6. Larger-scale clinical studies are currently ongoing in various countries to confirm its clinical efficiency. Depending on the results of these studies, Peretinoin may be used in CH-C patients to prevent HCC. Therefore, we sought to understand the effect of Peretinoin on HCV replication.

Peretinoin is categorised as a vitamin A or retinoid compound and conflicting reports have described the effects of vitamin A compounds on HCV replication. One report showed that 3 retinoids, 9-cis retinoic acid (RA), 13-cis RA and all-trans RA (ATRA), suppressed the replication of a sub-genomic HCV replicon7. However, vitamin A also reportedly enhances the replication of genome-length HCV in Huh-7 cells8. Here, we describe the impact of Peretinoin on different steps of the HCV life cycle, including translation, RNA amplification, virus assembly and secretion and its impact on host lipid metabolism in vitro. Our results clearly demonstrate that Peretinoin inhibits HCV RNA amplification and virus release by altering lipid metabolism.

Results

Inhibition of HCV RNA replication by retinoids

Several studies have tested the effects of vitamin A on HCV replication; these studies used a sub-genomic or full-genomic replicon, which contains 2 cistrons, one driven by HCV internal ribosome entry sites (IRES) and the other by encephalomyocarditis virus IRES7,8. We reported the usefulness of HCV genomes containing Gaussia princeps luciferase (GLuc) between p7 and NS2, followed by foot-and-mouth disease virus 2A, to monitor HCV RNA replication9,10 and this system is closer to physiological HCV replication than the bicistronic replicon systems (Fig. 1A). In addition to GLuc-containing HCV genomes in the backbone of genotype 1a H77S.3, a chimeric clone of H77S and genotype 2a JFH1, HJ3-511, with structural proteins from H77S and non-structural proteins from JFH1, we also constructed GLuc-containing genomes in the backbone of genotype 1b N12 and 2a JFH113 and confirmed their efficient replication in Huh-7.5 cells. Importantly, all of the strains used here are derived from cDNA clones that are infectious to chimpanzees.

Figure 1
figure 1

Antiviral effects of several retinoids and their effects on cell growth.

(A) Schematic representation of the GLuc-containing HCV genome. (B) Huh-7.5 cells were transfected with H77S.3/GLuc2A RNA and 48 h later, 0.5% DMSO or Peretinoin was added at concentrations ranging from 1 to 100 μM. Fresh medium containing Peretinoin was added every 24 h and 72 h after adding Peretinoin, secreted GLuc activity was measured. The GLuc activity from Peretinoin-treated cells was normalised to that with DMSO treatment. Data show the mean inhibition to DMSO treatment in each concentration of Peretinoin ± SD from 3 independent experiments. (C) Huh-7.5 cells were transfected with H77S.3/GLuc2A, N.2/GLuc2A, HJ3-5/GLuc2A and JFH1/GLuc2A RNAs and 48 h later, 0.5% DMSO or Peretinoin was added at the indicated concentrations. The medium was collected and replaced with fresh medium every 24 h until 72 h. GLuc activity was determined at each time point. The results shown represent the mean GLuc activity ± SD from 3 different plates. (D) Schematic representation of the bicistronic sub-genomic HCV RNA (E) Huh-7.5 cells were transfected with bicistronic sub-genomic RNA. At 48 h later, the transfected cells were treated with the indicated concentrations of Peretinoin for 72 h. Quantification of HCV RNA and 18S rRNA levels was performed and relative HCV RNA abundance normalised to the amount of 18S rRNA is presented as fold change ± SD compared to DMSO-treated cells from 3 independent experiments. (F) Huh-7.5 cells were transfected with H77S.3/GLuc2A RNA and 7 days later, HCV (H77S.3/GLuc2A)-replicating Huh-7.5 cells, depicted as ‘HCV+’, were treated with the indicated concentrations of Peretinoin and HCV-non-replicating Huh-7.5 cells, depicted as ‘HCV-’, were also treated in a same way. At 72 h after Peretinoin treatment, cell numbers were determined by using a Cell Counting Kit-8. Data represent relative cell numbers ± SD from 3 independent experiments to DMSO-treated cells. EMCV, Encephalomyocarditis virus; NeoR, Neomycin resistance gene; NT, no treatment.

Full size image

We initially examined the effects of 4 different retinoids, namely ATRA, 9-cis RA, 13-cis RA and Peretinoin, on HCV replication by using these 4 HCV genomes containing GLuc, according to the use of GLuc activity as an indicator of RNA replication and the structures of each retinoid were shown in Supplementary Fig. S1 online. Peretinoin inhibited the replication of H77S.3/GLuc2A in a dose-dependent manner (Fig. 1B). As the other retinoids also suppressed HCV replication, we determined the antiviral half maximal effective concentrations (EC50s) of these retinoids for each HCV genotype. Whilst Peretinoin showed the strongest antiviral effect on all genotypes tested, ATRA exerted a moderate effect and 9-cis and 13-cis RA generated a weaker effect (Table 1). Especially, Peretinoin suppressed the RNA replication of H77S.3/GLuc2A most efficiently and its EC50 was 9 μM.

Table 1 EC50 of vitamin A compounds on HCV RNA replication
Full size table

We also determined the half maximal cytotoxicity concentrations (CC50s) of these retinoids in H77S.3/GLuc2A-replicating Huh-7.5 cells by using the WST-8 assay, which reflects cell number. The CC50s of ATRA, 9-cis RA and 13-cis RA were more than 100 μM; however, the CC50 of Peretinoin was 68 μM when the cells were treated for 72 h (Table 2). Although Peretinoin had a slightly negative impact on cell growth, as it showed the strongest antiviral effect and may be used for HCC chemoprevention in HCV-infected patients in the future, we focused upon the action of Peretinoin among these retinoids.

Table 2 CC50 of vitamin A compounds on Huh-7.5 cells supporting HCV replication
Full size table

Inhibition of HCV RNA replication by Peretinoin

We examined the time dependence of the antiviral effect of Peretinoin. After HCV RNA transfection, we treated the transfected cells with Peretinoin at a range of concentrations (10–40 μM) and monitored RNA replication every 24 h until 72 h. Peretinoin started to show an antiviral effect from 24 h after treatment, which continued until 72 h. Peretinoin suppressed RNA replication in a time-dependent manner for all genotypes tested (Fig. 1C).

We also examined whether Peretinoin could also suppress RNA replication in a sub-genomic replicon system (Fig. 1D), in which infection should not occur due to the lack of structural proteins. Peretinoin was also able to suppress RNA replication in a dose-dependent manner in bicistronic sub-genomic RNA-transfected cells (Fig. 1E).

Importantly, when we treated HCV (H77S.3/GLuc2A)-replicating and HCV-non-replicating Huh-7.5 cells with Peretinoin at a range of concentrations (5–50 μM), the cell numbers were identical under the conditions tested (Fig. 1F).

As Peretinoin could suppress GLuc activity itself, we then examined directly its antiviral effect in the context of an HCV genome lacking the GLuc genome. For this purpose, Huh-7.5 cells infected with cell culture-derived HCV (HCVcc) of HJ3-5 were treated with different concentrations of Peretinoin. When we monitored HCV RNA replication by using quantitative real-time detection-polymerase chain reaction (RTD-PCR) (Fig. 2A) and protein expression by western blotting for the HCV core protein (Fig. 2B, see Supplementary Fig. S2 online), Peretinoin suppressed RNA replication and protein expression in a dose-dependent manner, which is consistent with the GLuc activity results. We also tested infectious virus production from Peretinoin-treated cells using a conventional focus forming unit (FFU) assay and found that Peretinoin also reduced this in a dose-dependent manner (Fig. 2C).

Figure 2
figure 2

Inhibition of HCV replication and infectious virus production.

Huh-7.5 cells were infected with the HJ3-5 virus at a multiplicity of infection (MOI) of 1 and 72 h later, DMSO or Peretinoin was added at the indicated concentrations. The medium was replaced with fresh medium every 24 h until 72 h. (A) At 72 h after adding Peretinoin, total cellular RNA was extracted and the amount of HCV RNA and 18S rRNA was quantitated by RTD-PCR. Relative HCV RNA abundance normalised to the amount of 18S rRNA is presented as fold change ± SD compared to DMSO-treated cells from 3 independent experiments. (B) At 72 h after Peretinoin treatment, the cell lysates were collected and subjected to western blot analysis using anti-core protein and anti-β-actin antibodies. Full-length blots/gels are presented in Supplementary Fig. S2 online. (C) The medium was collected at 72 h after Peretinoin treatment and immediately, naïve Huh-7.5 cells were infected with serially diluted medium. At 72 h after infection, the infectious virus titre of HCVcc from Peretinoin-treated cells was determined by an FFU assay. Data shown here represent the mean FFUs/mL ± SD from 2 independent experiments.

Full size image

Effect of Peretinoin on translation driven by HCV IRES

We also tested the effect of Peretinoin on translation directed by HCV IRES. For this purpose, we used a mini-genome RNA which has, sequentially, the HCV 5′-untranslated region (UTR), GLuc and HCV 3′-UTR and cap-Cypridina luciferase (CLuc)-polyA RNA as a control (see Supplementary Fig. S1 online). After we treated Huh-7.5 cells with different concentrations of Peretinoin for 24 h, we co-transfected the cells with these RNAs and measured GLuc and CLuc activity every 3 h from 3 to 12 h. When we normalised GLuc activity to CLuc activity at each time point, we did not observe a significant difference among the cells treated with the different concentrations of Peretinoin (see Supplementary Fig. S3 online), suggesting that Peretinoin does not have an effect on protein expression directed by HCV IRES.

Effect of Peretinoin on cellular interferon signalling

We hypothesised that the suppression of RNA replication by Peretinoin could be due to the activation or enhancement of cellular interferon (IFN) signalling. To examine this, we treated HCV (H77S.3/GLuc2A)-non-replicating and HCV-replicating Huh-7.5 cells with either IFNα-2b (10 IU/mL) or Peretinoin (10–40 μM) and monitored the expression of total and phosphorylated signal transducer and activator of transcription 1 (STAT1). Peretinoin did not alter the expression of either total or phosphorylated STAT1 in HCV-non-replicating Huh-7.5 cells or HCV-replicating cells (see Supplementary Fig. S4 online). In addition, Peretinoin did not further enhance the amount of phosphorylated STAT1 activated by IFNα-2b in HCV-non-replicating Huh-7.5 cells or HCV-replicating cells (see Supplementary Fig. S4 online). These data suggest that Peretinoin suppresses RNA replication without either activating or enhancing cellular IFN signalling.

Impact of Peretinoin on lipid metabolism

As lipid metabolism has an important role in various aspects of HCV infection14,15,16, we examined the impact of Peretinoin on lipid metabolism. However, as it is sometimes difficult to detect small changes in lipid metabolism, we tested the effect of Peretinoin under oleic acid (OA) treatment, which amplifies changes in lipid metabolism. We treated H77S.3/GLuc2A-replicating Huh-7.5 cells with 40 μM Peretinoin and 250 μM OA, fixed and stained the cells with BODIPY 493/503 for lipid droplets (LDs) and 4′, 6-diamidino-2-phenylindole (DAPI) for nuclei and used an anti-core protein antibody to detect HCV. When we stained LDs in the presence of 250 μM OA and the absence of Peretinoin, we observed intense signals (Fig. 3A); however, when it was accompanied with 40 μM Peretinoin, the signals from LDs were dramatically reduced and at the same time, the expression of HCV core protein was also down-regulated (Fig. 3B). When we quantitated the signal strength from LDs and HCV core protein in 4 different fields, Peretinoin significantly reduced the signals from LDs and HCV core protein (two-tailed Student's t test, p<0.0001 for each) (Fig. 3C). This reduction was also confirmed by the quantitation of the 5 cells which were positive for both LDs and HCV core (see Supplementary Fig. S5 online). The reduced expression of HCV core protein was also observed by western blot analysis (Fig. 3D, see Supplementary Fig. S6 online). We next investigated the impact of Peretinoin on lipid metabolism by measuring intracellular triglyceride (TG) levels, which should mainly reflect the amount of LDs, following treatment with 0–40 μM Peretinoin with or without HCV replication and OA treatment. Peretinoin reduced intracellular TG levels in a dose-dependent manner, regardless of OA treatment and HCV replication (Fig. 4A). These effects may be primarily due to its transcriptional modulation. To address this possibility, we examined the effect of Peretinoin on the transcription of fatty acid synthase (FASN) using RTD-PCR under 0–40 μM Peretinoin with or without HCV replication and OA treatment, because FASN is a key enzyme for the synthesis of fatty acids, which are an essential component of TGs. Peretinoin reduced the mRNA levels of FASN in a dose-dependent manner, regardless of OA treatment and HCV replication (Fig. 4B). We also examined FASN protein expression as well as the levels of precursor and mature sterol regulatory element-binding protein 1c (SREBP1c), which is a critical transcription factor for FASN. Peretinoin reduced the expression of FASN protein, which is consistent with the RTD-PCR results (Fig. 4C, see Supplementary Fig. S7 online). Although Peretinoin did not have an effect on precursor SREBP1c protein expression, it dramatically reduced the levels of mature SREBP1c (Fig. 4C, see Supplementary Fig. S7 online). We also observed a reduction of FASN mRNA levels by Peretinoin in an immortalised human hepatocyte cell line (Fig. 5A) and a similar reduction was also observed for ATRA, 9-cis RA and 13-cis RA treatment of HCV-replicating Huh-7.5 cells (Fig. 5B). These results indicate that Peretinoin reduced intracellular lipid levels by reducing the amount of mature SREBP1c and, subsequently, FASN.

Figure 3
figure 3

Reduction of LD signals by Peretinoin.

Huh-7.5 cells were infected with HJ3-5 virus at an MOI of 1 and 72 h later, 250 μM OA and DMSO or 250 μM OA and 40 μM Peretinoin were added and the following assay was performed at 72 h later. (A, B) At 72 h later, the cells were fixed and stained for nuclei, LDs and HCV core protein. (A) Shows 250 μM OA and DMSO-treated cells and (B) shows 250 μM OA and 40 μM Peretinoin-treated cells. The photos in (A) and (B) were taken under exactly the same conditions. (C) The signal intensity from LDs, HCV core protein and nuclei was quantitated as described in the Methods. Data shown represent mean signal intensity ± SD from 4 different areas and the difference was analysed statistically using Student's t-test. (D) Cell lysates were collected and subjected to western blot analysis using anti-core protein and anti-β-actin antibodies. Full-length blots/gels are presented in Supplementary Fig. S6 online. N.S., not significant.

Full size image
Figure 4
figure 4

Mechanism by which Peretinoin alters lipid metabolism.

Huh-7.5 cells were transfected with H77S.3/GLuc2A RNA and 72 h later, the transfected cells, depicted as ‘HCV(+)’ and non-transfected Huh-7.5 cells, depicted as ‘HCV(-)’, were treated with or without 250 μM OA in the presence of 2% fatty acid-free BSA with 0.5% DMSO or 10–40 μM Peretinoin and the following assay was performed at 72 h later. (A) The concentration of intracellular TGs was measured. Data shown represent mean concentration ± SD from 3 independent experiments. (B) RNA was extracted and the levels of FASN mRNA and 18S rRNA were quantitated by RTD-PCR. FASN levels were normalised to those of 18S rRNA and the ratio was furthermore normalised to that from DMSO-treated cells. The results presented here represent the relative fold of FASN/18S rRNA ± SD from 3 independent experiments at the indicated conditions. (C) Lysates from the cells without OA treatment were collected and subjected to western blot analysis using anti-FASN, anti-precursor SREBP1c, anti-mature SREBP1c, anti-ApoE3 and anti-β-actin antibodies. Full-length blots/gels are presented in Supplementary Fig. S7 online.

Full size image
Figure 5
figure 5

Reduction of FASN mRNA levels by Peretinoin in a human hepatocyte cell line and the effects of ATRA, 9-cis RA and 13-cis RA on the expression of FASN mRNA.

(A) An immortalised human hepatocyte cell line, THLE-5b cells, was treated with the indicated concentrations of Peretinoin. At 72 h later, RNA was extracted and reverse transcribed and the levels of FASN mRNA and 18S rRNA were quantified by RTD-PCR. The relative amount of FASN mRNA normalised to that of 18S rRNA is presented as fold change compared to DMSO-treated cells from 3 independent experiments at the indicated conditions. (B) Huh-7.5 cells were transfected with H77S.3/GLuc2A RNA. At 72 h later, the transfected cells were treated with DMSO or 10–40 μM ATRA, 9-cis RA and 13-cis RA. At 72 h later, RNA was extracted and the levels of FASN mRNA and 18S rRNA were quantified by RTD-PCR. The relative amount of FASN mRNA was determined as described in Fig. 4 and presented as fold change compared to DMSO-treated cells from 3 independent experiments at the indicated conditions.

Full size image

Specific inhibition of virus secretion by Peretinoin

Recently, lipids including LDs and TG have been reported to be important for efficient infectious virus production14,15,16. Due to its huge impact on lipid metabolism, Peretinoin could affect virus assembly or secretion as well as RNA amplification. To test the effect of Peretinoin on infectious virus production, we determined intra- and extra-cellular infectivity and the virus secretion ratio by measuring the amount of intra- and extra-cellular infectious virus from HJ3-5/GLuc2A-replicating FT3-7 cells treated with various concentrations of Peretinoin. We infected naïve Huh-7.5 cells with intra- and extra-cellular virus derived from HJ3-5/GLuc2A-replicating cells after Peretinoin treatment and used GLuc activity as an indicator of infectious virus production because FFUs and GLuc activity were well correlated (see Supplementary Fig. S8 online) and a previous report also showed a good correlation between them17. Although Peretinoin did not show a significant impact on intracellular infectivity at 10–30 μM, it dramatically reduced extracellular infectivity and virus secretion from 10 μM when we normalised intra- and extra-cellular infectivity by the replication capacity of the virus producing the intra- and extra-cellular virus, as determined by GLuc activity (Fig. 6A). This result was also confirmed by using the extra-cellular virus which was prepared by centrifugation and subsequently re-suspended to fresh medium without containing Peretinoin, indicating that possible carryover of Peretinoin in the medium from extra-cellular cultures does not affect the result shown in Figure 6A (Supplementary Fig. S9 online). Interestingly, the expression of apolipoprotein E3 (ApoE3), which is essential for virus secretion, was also suppressed by Peretinoin (Fig. 4C, see Supplementary Fig. S7 online). Furthermore, we compared the buoyant density of HCVcc derived from HJ3-5/GLuc2A-replicating FT3-7 cells by equilibrium gradient ultracentrifugation. HCVcc from HJ3-5/GLuc2A-replicating cells treated with dimethyl sulfoxide (DMSO) or 30 μM Peretinoin showed exactly the same peak of infectivity at 1.107 g/cm3 (Fig. 6B, 6C). Specific infectivity, as calculated from both peaks of HCV RNA and GLuc activity, was 0.0381 ± 0.0209 (standard deviation, SD) light units (LU)/copy for DMSO-treated cells and 0.0799 ± 0.0457 LU/copy for Peretinoin-treated cells, which did not show a considerable difference. Furthermore, Peretinoin did not affect virus entry of HCVcc when we tested it by RT-PCR for HCV RNA at 5 h after infection and an FFU assay at 72 h after infection (see Supplementary Fig. S10 online). Collectively, Peretinoin seems to inhibit virus release in addition to viral RNA amplification.

Figure 6
figure 6

Impact of Peretinoin on infectious virus production.

(A) FT3-7 cells were transfected with HJ3-5/GLuc2A RNA and 7 days later, 0.5% DMSO, or 10–40 μM Peretinoin, were added. At 72 h later, extra- and intra-cellular viruses were collected and used to infect naïve Huh-7.5 cells. Replication capacity was also determined by measuring secreted GLuc activity. At 48 h after infection, we determined the amount of infectious virus from extra- and intra-cellular media by using GLuc activity as an indicator of the amount of infectious virus. Intra- and extra-cellular infectivity was normalised to replication capacity at infection and these were then normalised to those of DMSO-treated cells, which were set to 100%. The ratio of extracellular infectious virus to intracellular virus was calculated at the indicated conditions and it was then normalised to DMSO-treated cells, which were set to 100%. Data show the mean ratio to that of DMSO-treated cells ± SD from 3 independent experiments. (B, C) FT3-7 cells were transfected with HJ3-5/GLuc2A RNA and 7 days later, 0.5% DMSO or 30 μM Peretinoin were added and then 72 h later, the medium was collected and subjected to equilibrium ultracentrifugation. Fourteen fractions were taken and analysed for density (circles), HCV RNA levels (diamonds) and infectious virus titres determined by GLuc activity (grey bars). (B) shows the results from DMSO-treated cells, whilst (C) shows those for Peretinoin-treated cells.

Full size image

Discussion

In the present study, we clearly showed that Peretinoin, as well as ATRA, 9-cis RA and 13-cis RA, suppressed HCV RNA replication (Table 1). While previous reports used replicon systems to test the effects of retinoids, we used a genome-length HCV containing a GLuc-coding sequence between p7 and NS2, which is more physiological than replicons. The inhibitory effect of retinoids was universal among the HCV genotypes tested and all retinoids tested showed an inhibitory effect on HCV replication (Table 1). In addition, we also observed the antiviral effect of Peretinoin in the replicon system (Fig. 1E). Therefore, our present data strongly support the notion that retinoids exert an antiviral effect in vitro. The antiviral effect of retinoids has also been confirmed in a clinical study. Even when CH-C patients were treated with ATRA, the viral load dropped by 1–2 log units in 50% of the patients enrolled. In addition, when CH-C patients who showed no response to prior IFN/PEG-IFNα and ribavirin therapy were treated with a combination of ATRA and PEG-IFNα-2a, 30% of patients showed a significant viral reduction18. Recently, combined vitamin A and D deficiency prior to IFN-based therapy was shown to be a strong independent predictor of non-response to antiviral therapy19. Collectively, our data and the clinical findings indicate that retinoids possess inhibitory effects on HCV replication.

Peretinoin showed the strongest antiviral effect among the retinoids tested (Table 1); thus, we focused on Peretinoin to clarify its antiviral mechanism. A previous report showed that 9-cis RA enhanced the antiviral effect of IFNα by increasing the expression of the IFNα receptor20; however, another study showed that ATRA did not induce the activation of dsRNA-activated protein kinase R, which is a key player in the IFN-induced antiviral response7. In the present study, Peretinoin did not increase the amount of the activated form of STAT-1, which is pSTAT1, contrary to IFNα-2b, both in HCV-replicating and HCV-non replicating Huh-7.5 cells and dual treatment of Huh-7.5 cells with IFNα-2b and Peretinoin did not show a further increase of the pSTAT1 levels induced by only IFNα-2b (see Supplementary Fig. S4 online), indicating that Peretinoin did not activate or enhance cellular IFN signalling. Our results also indicate that the antiviral effect of Peretinoin is not due to the suppression of HCV translation directed by HCV IRES (see Supplementary Fig. S3 online). As Peretinoin suppressed the RNA replication of bicistronic sub-genomic replicons (Fig. 1E), it seems to suppress RNA amplification itself (see also the later description of FASN). A report showed that retinoids inhibited HCV RNA replication by enhancing the expression of gastrointestinal-glutathione peroxidase (GI-GPx) only in the presence of sodium selenite7; however, in the present study, we demonstrated that all retinoids tested inhibited HCV replication, even in the absence of sodium selenite. Thus, our results support the notion that the observed antiviral effects could be independent of GI-GPx, although supplementation with sodium selenite may further enhance the antiviral effects of retinoids.

To clarify the mechanism underlying the antiviral effect of Peretinoin further, we focused on the effect of Peretinoin on lipid metabolism because it has been shown to modify multiple aspects of HCV infection14,15,16 and we detected a significant reduction of FASN mRNA levels by Peretinoin in a mouse hepatoma model, implying its possible effect on lipid metabolism5. Surprisingly, Peretinoin strongly reduced the signal from LDs in the presence of OA and intracellular TGs (Fig. 3A–C, 4A). LDs are known to have an essential role in the assembly of HCV virus particles by interacting with HCV core protein and NS5A21,22. Therefore, we examined the effect of Peretinoin on several steps of infectious virus production, such as assembly and secretion. Interestingly, Peretinoin specifically impaired virus secretion without affecting assembly at 10–30 μM, whilst 40 μM Peretinoin impaired virus secretion and assembly (Fig. 6A). The role of LDs in virus secretion has not been fully characterised, but virus should be secreted through the production and release of very low-density lipoproteins. In addition to microsomal triglyceride transfer protein and several apolipoproteins, such as ApoB and ApoE23, small interfering RNA screening revealed that multiple components of the secretary pathway, including endoplasmic reticulum to Golgi trafficking and lipid and protein kinases, are involved in HCV secretion24. Thus, the mechanism underlying this specific inhibition of virus secretion by Peretinoin remains to be addressed. One possible explanation for its action is the reduction of ApoE3 expression (Fig. 4C), because ApoE3 was shown to have an important role in virus secretion with a minimal impact on assembly25.

Several reports showed that LDs play an essential role in RNA amplification and virus assembly. The hypolipidemic agent nordihydroguaiaretic acid reduced the number of LDs, resulting in the suppression of RNA amplification and virus secretion, as Peretinoin did26. Furthermore, inhibition of tail-interacting protein 47, which coats LDs and is involved in their generation and turnover, suppressed HCV RNA replication and assembly27,28. Thus, the inhibition of RNA replication by Peretinoin could be explained by its direct effect on LDs. In addition, a recent report suggested that FASN may localise within HCV replication complexes through an interaction with NS5B, thereby increasing its RNA-dependent RNA polymerase activity29. Thus, Peretinoin may inhibit RNA replication not only by reducing the signalling of LDs but also inhibiting the expression of FASN.

We also demonstrated that Peretinoin reduced the levels of mature SREBP1c by inhibiting the proteolysis of its precursor and subsequently the transcription and expression of FASN (Fig. 4C), which could be the main reason for the alteration of lipid metabolism by Peretinoin; however, the mechanism by which it inhibits proteolysis should be addressed in a future study. Several reports have shown that the expression of SREBP1c and/or FASN is increased in HCV-infected patients30, Huh-7 cells31 and a transgenic mouse expressing the full-length HCV polyprotein32. In addition, HCV infection was shown to enhance the proteolytic cleavage of precursor SREBP1c, resulting in an increase in its mature form33. Taken together, HCV induces lipogenesis to make infected cells more supportive for its propagation. In contrast to HCV, Peretinoin seems to suppress lipogenesis by inhibiting the SREBP1c-FASN axis, which is highly activated by HCV infection. It is also important to note that this effect did not depend on HCV infection, indicating that Peretinoin should exert a hypolipidemic effect, as we also observed a reduction of FASN mRNA levels following Peretinoin treatment of an immortalised human hepatocyte cell line (Fig. 5A). Interestingly, this effect was universal among retinoids because the other retinoids examined also reduced FASN mRNA levels (Fig. 5B). These findings suggest that Peretinoin could also be useful for the treatment of non-alcoholic fatty liver disease, whose hallmark is hepatic fat accumulation.

The antiviral EC50 of Peretinoin seems to be closer to its CC50 than that of the other retinoids in Huh-7.5 cells because several papers have shown that Peretinoin inhibits the growth of hepatoma cells in vivo and in vitro34,35, induces apoptosis in human hepatoma cell lines36 and causes an arrest of the cell cycle in G0-G1 in human hepatoma cell lines35, indicating that Peretinoin should selectively suppress the growth of hepatoma cells, although the mechanism has not been fully understood. However, pharmacokinetic data from humans showed that the mean plasma concentration of lipid-bounded Peretinoin is 7.3 μM when patients received 600 mg Peretinoin daily for 8 weeks37. This concentration is very close to the antiviral EC50 and could have an inhibitory effect on HCV replication, indicating that we could expect an antiviral effect at this dose in humans. Peretinoin showed an additive antiviral effect when combined with IFNα-2b (data not shown); furthermore, HCV did not acquire resistance to Peretinoin after 14 days treatment with 10–40 μM Peretinoin (see Supplementary Fig. S11 online). Although it could be difficult to eradicate HCV only by Peretinoin due to its low selective index (CC50/EC50), combination therapy with Peretinoin plus PEG-IFN, ribavirin, or DAAs may further improve the SVR rate, as vitamin D has been proved to do38,39.

In summary, we have demonstrated that Peretinoin, which may in the future be administered to patients infected with HCV to prevent HCC, inhibits HCV RNA replication and infectious virus release by modifying several aspects of lipid metabolism.

Methods

Cell lines

Huh-7.5 (kindly provided by Professor C. M. Rice, Rockefeller University, New York, NY) and FT3-7 cells (both clonal derivatives of Huh-7 cells) were maintained as described previously9. We used an immortalised human hepatocyte cell line, THLE-5b cells40, for the indicated experiments.

Reagents

Peretinoin and IFNα-2b were kindly provided by KOWA Company, Ltd. (Tokyo, Japan). ATRA, 9-cis RA and 13-cis RA, were purchased from Sigma-Aldrich Japan K.K. (Tokyo, Japan). Stock solutions were prepared in DMSO and all final dilutions contained 0.5% DMSO.

Plasmids

The GLuc coding sequence, followed by the FMDV2A sequence, was inserted between p7 and NS2 in pJFH1 and pHCV-N.2, which encode cDNA of genotype 2a JFH113 and genotype 1b N12, carrying several replication-enhancing mutations to be described elsewhere, respectively, by the same strategy adopted previously for H77S10. pH77S.3/GLuc2A10, pHJ3-5/GLuc2A9 and pHJ3-511 have been described previously.

Antiviral activity assay

The indicated HCV RNAs were transfected by electroporation. The medium was replaced with fresh medium containing serial dilutions of the antiviral compounds at 48 h and at 24 h intervals thereafter. Secreted GLuc activity was determined at 72 h after adding the antiviral compounds. The concentration of each compound required to reduce the amount of secreted GLuc activity by 50% (EC50) was determined using a 3-parameter Hill equation (Sigma Plot 10.0).

Cell number determination

Huh-7.5 cells were seeded in 96-well plates at a density of 5,000 cells/well and at 24 h later, the indicated compounds were added. Cell numbers were determined by a WST-8 assay using Cell Counting Kit-8. The concentration of each compound required to reduce the amount of cell number by 50% (CC50) was determined using a 3-parameter Hill equation (Sigma Plot 10.0).

RNA transcription

HCV RNAs were synthesised using a MEGAscript T7 Kit and synthesised RNA was purified using an RNeasy Mini Kit.

Virus yield determination

Huh-7.5 cells were seeded in 48-well plates at a density of 4.0 × 104 cells/well at 24 h prior to inoculation with 100 μL of virus-containing medium. The cells were maintained at 37°C in a 5% CO2 environment and fed with 300 μL medium at 24 h later. Following 48 h of additional incubation, the cells were fixed in methanol-acetone (1:1) at room temperature for 9 min and stained with a C7-50 monoclonal antibody to the HCV core protein (1:300). After extensive washing, the cells were stained with Alexa Fluor 568-conjugated anti-mouse IgG antibodies. A cluster of infected cells staining for core antigen was considered to constitute a single infectious FFU; virus titres are reported as FFUs/mL.

Western blotting and immunostaining

Western blotting and immunostaining were performed as described previously41,42. Briefly, the cells were washed in phosphate-buffered saline (PBS) and lysed in a radioimmunoprecipitation assay buffer containing complete Protease Inhibitor Cocktail and PhosSTOP. The membranes were blocked in Blocking One or Blocking One-P solution and the expression of HCV core protein, FASN, precursor and mature SREBP1c, ApoE3 and β-actin was evaluated with mouse anti-core protein, rabbit anti-FASN, rabbit anti-SREBP1c, goat anti-ApoE3 and rabbit anti-β-actin antibodies, respectively.

For immunofluorescence staining, the cells were washed twice with PBS and fixed in 4% paraformaldehyde for 15 min at room temperature. After washing again with PBS, the cells were permeabilised with 0.05% Triton X-100 in PBS for 15 min at room temperature. They were incubated in a blocking solution (10% foetal bovine serum and 5% bovine serum albumin [BSA] in PBS) for 30 min and then with the anti-core protein monoclonal antibodies. The fluorescent secondary antibodies were Alexa Fluor 568-conjugated anti-mouse IgG antibodies. Nuclei were labelled with DAPI and LDs were visualised with BODIPY 493/503. Imaging was performed on a BIOREVO fluorescence microscope (Keyence Corporation, Osaka, Japan). The signal strength of LDs, core protein and nuclei was quantitated by using Measurement Module BZ-H1M (Keyence Corporation).

Quantitative RTD-PCR

The primer pairs and probes for FASN and 18S rRNA were obtained from the TaqMan assay reagents library. HCV RNA was detected as described previously43.

Secreted luciferase assay

Cell culture supernatant fluids were collected at intervals after RNA transfection and the cells were re-fed fresh medium. Secreted GLuc was measured as described previously9.

Fatty acid treatment and measurement of TGs

The cells were treated with the indicated concentrations of OA in the presence of 2% fatty acid-free BSA. Intracellular TG content was measured using a TG Test according to the manufacturer's instructions.

Intra- and extra-cellular infectivity assay

To determine the amount of intra-cellular infectious virus, cell pellets of HJ3-5/GLuc2A-replicating FT3-7 cells harvested after trypsinization were resuspended in complete medium, washed twice with PBS and lysed by 4 cycles of freezing and thawing. The lysates were clarified by centrifugation at 2,300 × g for 5 min prior to inoculation onto naïve Huh-7.5 cells. At the same time, extra-cellular medium was also collected. The medium derived from extra- and intra-cellular cultures was used to infect naïve Huh-7.5 cells, which were plated in 48-well plates at a density of 4.0 × 104 cells/well at 24 h prior to infection. After 6 h inoculation, medium containing virus and possible carryover of Peretinoin was removed by extensive wash and medium was replaced with fresh one every 24 h until 48 h. At 48 h after infection, we determined GLuc activity and used it as an indicator of the infectious virus titre.

Equilibrium ultracentrifugation of HJ3-5/GLuc2A virus particles using an isopycnic iodixanol gradient

Filtered supernatant fluids collected from HJ3-5/GLuc2A virus-replicating FT3-7 cells treated with DMSO or 30 μM Peretinoin for 72 h were concentrated 30-fold using a Centricon PBHK Centrifugal Plus-20 Filter Unit with an Ultracel-PL membrane (100-kDa exclusion; Merck Millipore, Billerica, MA), then layered on top of a preformed continuous 10–40% iodixanol gradient in Hanks' balanced salt solution. The gradients were centrifuged in an SW41 rotor at 209,678 × g for 16 h at 4°C and fractions (500 μL each) were collected from the top of the tube. The density of each fraction was determined using a digital refractometer. Virus RNA was isolated from each gradient fraction using a QIAamp Viral RNA Kit and cDNA was synthesised using a High Capacity cDNA Reverse Transcription Kit. RTD-PCR to quantitate the amount of HCV RNA was performed using a 7500 Real Time PCR System. Each fraction was used to infect naïve Huh-7.5 cells for 6 h, followed by extensive washing to ensure GLuc activity was reduced to background. The infected cells were inoculated and the medium was replaced with fresh medium every 24 h. GLuc activity, which was used an alternative to the infectious virus titre, was determined at 72 h after infection.