Hepatitis C Virus-Induced Autophagy Is Independent of the Unfolded Protein Response

Abstract

Hepatitis C virus (HCV) has been shown to induce autophagy and the unfolded protein response (UPR), but the mechanistic link between the induction of these two cellular processes remains unclear. We demonstrate here that HCV infection induces autophagy, as judged by accumulation of lipidated LC3-II, and that this induction occurs rapidly after infection, preceding the stimulation of the UPR, which occurs only at later stages, after the viral envelope glycoproteins have been expressed to high levels. Furthermore, both genotype 1b and 2a subgenomic replicons expressing nonstructural (NS3-5B) proteins and JFH-1 virus lacking the envelope glycoproteins potently induced autophagy in the absence of detectable UPR. This ability was also shared by a subgenomic replicon derived from the related GB virus B (GBV-B). We also show that small interfering RNA (siRNA)-mediated silencing of the key UPR inducer, Ire1, has no effect on HCV genome replication or the induction of autophagy, further demonstrating that the UPR is not required for these processes. Lastly, we demonstrate that the HCV replicase does not colocalize with autophagosomes, suggesting that the induction of autophagy is not required to generate the membrane platform for HCV RNA replication.

INTRODUCTION

Hepatitis C virus (HCV) is a positive-strand RNA virus that establishes a chronic infection in 85% of infected individuals, leading to long-term liver diseases such as cirrhosis and hepatocellular carcinoma. The 9.6-kb genome is translated into a single polyprotein that is subsequently cleaved into 10 structural and nonstructural proteins. The recent development of an infectious cell culture system for HCV, based on the genotype 2a isolate JFH-1 (41), has allowed detailed analyses of the molecular mechanisms of virus replication. One of the outcomes of this advance has been the observation that HCV infection results in the induction of autophagy (1, 7, 31).

Autophagy is a cellular process for the bulk degradation of cytoplasmic contents, either to allow them to be recycled or to provide an energy source during times of nutrient starvation or stress (35). It is characterized by the formation of double-membraned vesicles, autophagosomes, which fuse with lysosomes to form autolysosomes, allowing the degradation of the vesicular contents. Autophagy is also triggered in response to endoplasmic reticulum (ER) stress, which results in the unfolded protein response (UPR) (5, 29). In this case, double-membraned vesicles are formed but do not fuse with lysosomes; this response serves to sequester misfolded proteins from the ER and restores homeostasis by reducing protein synthesis and upregulating membrane synthesis. Recently, the significance of autophagy for virus infection has become clear; in particular, some positive-strand RNA viruses utilize autophagy to generate the cytoplasmic membrane structures required for genome replication, although autophagy has also been implicated in the immune response to pathogens (for a review, see reference 6).

HCV has been shown to also induce the UPR (3, 16, 22), and furthermore, UPR activation was proposed to be responsible for the subsequent induction of autophagy (31). However, the mechanistic link between induction of the UPR and induction of autophagy has not yet been defined. In order to better understand these processes, we undertook a detailed analysis of the induction of the UPR and of autophagy, using both infectious virus and subgenomic replicon (SGR) systems. Our results reveal that although HCV is indeed able to induce both of these cellular processes, the induction of autophagy by HCV is independent of the induction of the UPR, suggesting that these processes are mechanistically distinct.

MATERIALS AND METHODS

Cell culture.

Huh7 and Huh7.5 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (vol/vol) fetal bovine serum (FBS), 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine, and nonessential amino acids (Gibco) at 37°C and 5% CO₂ in a humidified incubator. Huh7 cells stably harboring subgenomic replicons were maintained in the presence of G418 at 500 μg/ml (Melford).

In vitro transcription and RNA transfection.

The subgenomic replicons used for this study were FK5.1 (genotype 1b) Krieger (18) and the genotype 2a replicons SGR-neo-JFH-1 (15), SGR-luc-JFH-1 (38), and JFH-1Feo (43). To generate RNA, plasmids were linearized with ScaI (FK5.1) or XbaI (JFH-1), followed by mung bean nuclease digestion (JFH-1 constructs). RNA was transcribed using a T7 Ribomax Express kit (Promega). For lipofection, 10⁵ cells seeded into a 12-well plate were transfected with 1 μg RNA by use of Lipofectin (Invitrogen) following the manufacturer's instructions. For small interfering RNA (siRNA) experiments, cells were transfected with siRNA (75 pmol) by use of Lipofectamine RNAiMax (Invitrogen). Luciferase activity was measured by lysing cells in passive lysis buffer (PLB; Promega) mixed with luciferase assay reagent or Stop & Glo (Promega) and analyzing them by use of a luminometer following the manufacturer's instructions.

Western blotting.

The polyclonal sheep anti-NS5A serum was described previously (24). The polyclonal sheep anti-core serum was raised against a fragment (amino acids 1 to 122) of JFH-1 core expressed in Escherichia coli. Other antibodies were obtained from either Cell Signaling Technologies (CCAAT/enhancer-binding protein homologous protein [CHOP]) or Abcam (glyceraldehyde-3-phosphate dehydrogenase [GAPDH] and LC3). For Western blotting, cells were lysed in Glasgow lysis buffer (GLB) as previously described (26). Membranes were probed with appropriate primary antibodies followed by horseradish peroxidase-conjugated secondary antibodies (Sigma) and were visualized using an in-house enhanced chemiluminescence (ECL) reagent.

Analysis of xbp1 mRNA splicing.

One microgram of total cellular RNA was reverse transcribed at 37°C for 60 min in a total volume of 30 μl, using a TaqMan reverse transcription (RT) kit (Applied Biosystems). The products were denatured at 95°C for 5 min, and 10 μl of product was amplified by PCR under standard conditions, using primers for both xbp1 and GAPDH (details and primer sequences are available upon request).

Virus assays.

For virus infection, medium removed from cells previously transfected with the appropriate in vitro-transcribed virus RNA was clarified by centrifugation at 2,000 × g for 5 min prior to titration by a focus-forming assay as described previously (11). Cells were routinely infected at a multiplicity of infection of 0.5 focus-forming unit (FFU)/cell.

RESULTS

HCV infection induces both autophagy and the unfolded protein response.

Previously, it was described that the induction of autophagy by HCV depends on the UPR (31). However, the experiments described involved electroporation of full-length JFH-1 RNA, and we therefore investigated the induction of both autophagy and the UPR in a more physiologically appropriate scenario, namely, cells infected with JFH-1 virus. Huh7 cells were infected with JFH-1 at a multiplicity of infection of 0.5 FFU/cell, and protein expression was monitored by Western blotting. As shown in Fig. 1A, a major increase in expression of core, E2, and NS5A was observed between 24 and 48 h postinfection (hpi). Interestingly, the induction of autophagy (as judged by the accumulation of LC3-II in a Western blot) preceded this detectable accumulation of viral proteins and was observed as early as 4 hpi, suggesting that autophagy might be induced at a very early step in the virus life cycle. The induction of the UPR was measured by two assays: first, Western blotting of the transcription factor CHOP, and second, an RT-PCR assay for the splicing of xbp1 mRNA by the ER membrane-anchored endoribonuclease Ire1. Figure 1A shows that CHOP expression was first observed at 48 hpi and increased to a high level by 72 hpi. Spliced xbp1 mRNA was also robustly detectable at both 48 and 72 hpi. Mock-infected cells showed no evidence of either autophagy or the UPR at 72 hpi (lane 8). Furthermore, treatment of mock-infected cells with tunicamycin (which induces UPR by inhibiting glycosylation) resulted in the robust expression of CHOP and the appearance of spliced xbp1 mRNA (lane 9), confirming that the delay in the appearance of the UPR was not due to an inherent defect in Huh7 cells. Tunicamycin treatment also induced LC3-II accumulation. These data confirm that HCV infection induces both autophagy and the UPR, but the observation that LC3-II accumulation preceded the appearance of both CHOP and spliced xbp1 mRNA suggests that HCV-induced autophagy is not dependent on prior induction of the UPR. Similar results were obtained using a chimeric J6/JFH-1 virus that expressed Renilla luciferase (J6/JFH-1luc) (data not shown).

Fig 1.

Induction of autophagy and UPR by HCV infection. (A) Huh7 cells were mock infected or infected with JFH-1 at a multiplicity of infection of 0.5 FFU/cell. At the indicated times, lysates were prepared and analyzed by Western blotting (WB) for the indicated proteins or by RT-PCR analysis to detect GAPDH RNA (as a control) and the unspliced and spliced forms of xbp1 RNA (Xbp1u and Xbp1s, respectively). For lane 9, mock-infected cells were treated with tunicamycin (Tm) for 4 h prior to harvest. (B) Huh7.5 cells electroporated with JFH-1 RNA or the ΔE1-E2 derivative were harvested at 72 hpt and analyzed by Western blotting or RT-PCR.

Consistent with the previous report (31), we confirmed that electroporation of JFH-1 RNA induces both autophagy (LC3-II accumulation) and the UPR (CHOP expression and xbp1 mRNA splicing) (Fig. 1B). However, a mutant form of this RNA containing an in-frame deletion within the coding region for glycoproteins E1 and E2 [JFH-1(ΔE1-E2)] showed a marked reduction in the ability to induce the UPR, while fully retaining the ability to induce LC3-II accumulation. These data suggest that induction of the UPR by HCV is dependent on expression of the envelope glycoproteins and that induction of autophagy is independent of the UPR.

We reasoned that if JFH-1(ΔE1-E2) could induce autophagy, then this property might be attributed to the viral replicase. We therefore examined LC3-II accumulation in Huh7 cells stably harboring either a subgenomic replicon (SGR) derived from JFH-1 (15) or a corresponding culture-adapted genotype 1b SGR (FK5.1) (18). Figure 2A shows that compared to naïve Huh7 cells (lane 1), both JFH-1 (lane 3) and FK5.1 (lane 5) SGR-harboring cells exhibited an accumulation of LC3-II. However, neither SGR cell line exhibited any detectable UPR, as judged by a lack of both CHOP expression and xbp1 mRNA splicing. This was not due to an SGR-induced inhibition of UPR induction or a defect resulting from the selection of SGR-harboring Huh7 cells, as treatment with tunicamycin resulted in the appearance of both CHOP and spliced xbp1 mRNA (lanes 2, 4, and 6).

Fig 2.

Induction of autophagy in cells stably harboring SGR RNA. (A) Parental Huh7 cells or Huh7 cells stably harboring either a genotype 2a (JFH-1) or 1b (FK5.1) NS3-5B SGR were left untreated or treated with tunicamycin for 4 h prior to harvest for Western blotting or RT-PCR as described in the legend to Fig. 1. (B) Cells harboring the FK5.1 SGR were incubated in serum-free medium (lane 2) or medium supplemented with 10 mM leupeptin (lanes 3 and 5) for 2 h prior to harvest and analysis by Western blotting. (C) Parental Huh7 cells or Huh7 cells stably harboring either an HCV genotype 1b (FK5.1) NS3-5B SGR or a GBV-B NS3-5B SGR (25) were left untreated or treated with tunicamycin for 4 h prior to harvest for Western blotting or RT-PCR as described in the legend to Fig. 1. Due to the lack of antibodies to GBV-B nonstructural proteins, the presence of the GBV-B SGR was confirmed by Western blotting for neomycin phosphotransferase or by RT-PCR analysis of the GBV-B sequence.

We also compared LC3-II abundances in FK5.1 SGR-harboring cells and in naïve Huh7.5 cells in which autophagy was induced by a physiological stimulus, i.e., amino acid starvation, in the presence or absence of the cysteine protease inhibitor leupeptin, which has been shown to block the subsequent degradative process of autophagic flux. Steady-state levels of LC3-II were significantly higher in FK5.1 SGR-harboring cells than in starved Huh7.5 cells; however, in both cases, leupeptin treatment resulted in a further increase in LC3-II abundance (Fig. 2B). This suggested that the accumulation of LC3-II in FK5.1 SGR-harboring cells was most likely due to stimulation of autophagosome generation rather than to a block to autophagic flux and protein degradation following fusion of these autophagosomes with lysosomes. If LC3-II accumulation was due to a block in autophagic flux, then leupeptin treatment would not have elicited a further increase in LC3-II levels.

Until the recent identification of the nonprimate hepacivirus (2), the closest relative to HCV was GB virus B (GBV-B) (33). As we had access to Huh7 cells stably harboring a GBV-B SGR (25), we asked whether the ability to induce autophagy but not the UPR was shared by this close relative of HCV. As shown in Fig. 2C, the presence of the GBV-B SGR (as judged by Western blot detection of neomycin phosphotransferase [NPT] and RT-PCR detection of GBV-B RNA) correlated with an elevation of LC3-II levels. In contrast, the GBV-B SGR-harboring cells did not exhibit a UPR response but remained responsive to tunicamycin-induced UPR. We concluded that the ability to induce autophagy is indeed shared among the hepaciviruses.

To confirm that the induction of autophagy in stable SGR-harboring cells was not an artifact of the long-term selection process, we transiently transfected Huh7.5 cells with a JFH-1-derived luciferase-containing SGR (SGR-luc-JFH-1) or a replication-defective GND mutant derivative (Fig. 3). Since in our hands electroporation alone resulted in an elevation of basal LC3-II accumulation as early as 4 h postelectroporation (hpt) (data not shown), we used lipofection in this experiment. In contrast to virus infection (Fig. 1A), the initial introduction of SGR RNA into cells did not induce autophagy; however, once virus proteins accumulated to detectable levels (as judged by Western blotting for NS5A), at both 48 and 72 hpt (Fig. 3, lanes 17, 18, 23, and 24), levels of LC3-II were elevated above the baseline.

Fig 3.

Induction of autophagy in cells transiently transfected with SGR RNA. Huh7.5 cells were either mock transfected or transfected with wild-type or GND SGR-luc-JFH-1 (38) RNA (2 μg) by use of Lipofectin. Cells were either left untreated (odd-numbered lanes) or treated with tunicamycin (even-numbered lanes) for 4 h prior to harvest for Western blotting or RT-PCR at the indicated times after transfection.

LC3-II did not accumulate significantly in cells transfected with the GND SGR, i.e., in the absence of RNA replication and NS5A expression (Fig. 3, lanes 15, 16, 21, and 22). Confirming the data obtained with the stable SGR-harboring cell lines, neither CHOP expression nor xbp1 mRNA splicing was observed in any of the cells at any time point. However, treatment with tunicamycin was able to effectively induce the UPR (Fig. 3, even-numbered lanes)—both CHOP expression and xbp1 mRNA splicing—indicating that the UPR remained intact and responsive during viral RNA replication.

Inhibition of the UPR does not block HCV genome replication.

A number of recent studies have demonstrated that autophagy-related proteins are required for efficient HCV RNA replication, as siRNA-mediated ablation of LC3, ATG4B, ATG7, or Beclin-1 expression results in a reduction of virus RNA accumulation (7, 30, 31). Our observations broadly agree with these studies, as transfection of siRNAs targeted to ATG5, ATG7, or Beclin-1 inhibited replication of the JFH-1 SGR (data not shown). siRNA-mediated ablation of Ire1, the endoribonuclease responsible for xbp1 mRNA splicing, has also been reported to inhibit HCV RNA replication in the context of virus infection (16, 31); however, this result seemed at odds with our observation that HCV RNA replication did not induce the UPR. To confirm whether the UPR had any role in HCV RNA replication, we also used an siRNA approach to ablate Ire1 expression. We first confirmed that this silencing procedure was effective: as shown in Fig. 4, siRNA-mediated inhibition of Ire1 expression (lanes 3 and 4) was concomitant with both the absence of detectable xbp1 mRNA splicing and a significant (60%) reduction in CHOP expression following tunicamycin treatment. Because CHOP expression is regulated by factors other than Ire1 (e.g., PERK and ATF6α), we did not expect to see a complete abrogation of CHOP expression following Ire1 silencing.

Fig 4.

Silencing of Ire1 inhibits induction of the UPR. Huh7 cells were transfected with control siRNA or an siRNA against Ire1 by use of RNAiMax. Cells were either left untreated (odd-numbered lanes) or treated with tunicamycin (even-numbered lanes) for 4 h prior to harvest for Western blotting or RT-PCR at 72 h posttransfection. The ratio of CHOP to GAPDH expression was determined by densitometry for multiple experiments and is presented graphically.

To assess the effect of Ire1 ablation on HCV RNA replication, we utilized a bicistronic JFH-1 SGR expressing a neomycin phosphotransferase-luciferase fusion protein (JFH-1-Feo) (43). A cell line stably harboring this SGR was transfected with either a scrambled control or Ire1-targeted siRNA and analyzed at 72 hpt by Western blotting, RT-PCR, and a luciferase assay. As shown in Fig. 5A and B, there was no effect on either NS5A or luciferase expression (compare lanes 6 and 8), even under conditions of efficient ablation of Ire1 expression. As positive controls, the JFH-1-Feo-harboring cells were treated with two well-characterized inhibitors of HCV genome replication, namely, cyclosporine (CsA) (4) and the NS5A inhibitor BMS-790052 (10). Both inhibitors effectively blocked HCV RNA replication, as judged by reductions in both NS5A and luciferase expression (lanes 9 and 10). Interestingly, CsA induced both the UPR and autophagy, consistent with its cellular target, i.e., the cyclophilin family of peptidyl-prolyl isomerases. Inhibition of cyclophilins would likely result in accumulation of unfolded proteins within the cell. The activation of both UPR and autophagy by CsA was not dependent on the presence of the SGR, as it was also seen in naïve Huh7 cells (lane 4). BMS-790052 did not induce either UPR or autophagy, consistent with its specificity for inhibition of HCV RNA replication, which is reassuring given that this compound is currently in clinical trials (9).

Fig 5.

Silencing of Ire1 does not affect HCV RNA replication. (A and B) Naïve Huh7 cells or cells stably harboring a Neo/Luc fusion-expressing SGR (SGR-Feo-JFH-1) (43) were transfected with control siRNA or an siRNA against Ire1 by use of RNAiMax. At 72 h postlipofection, silenced cells were harvested for luciferase assay, Western blotting, and RT-PCR. As positive controls, cells were treated with well-characterized inhibitors of HCV RNA replication, namely, CsA (8) and BMS-790052 (10). Note that CsA also induces autophagy and UPR, presumably by blocking protein folding.

To further verify that induction of autophagy and induction of the UPR by HCV are mechanistically distinct processes, we determined whether HCV infection was still able to induce autophagy when the UPR was disabled by siRNA-mediated silencing of Ire1. To test this, Huh7 cells were transiently transfected with either control or Ire1-targeted siRNA and subsequently infected with J6/JFH-1luc (Fig. 6). Interestingly, Ire1 silencing resulted in a modest reduction of both luciferase expression and E2 expression, possibly due to an effect on an early stage of the virus life cycle. However, despite this reduction, J6/JFH-1luc infection induced similar levels of autophagy induction (LC3-II expression) under all conditions (lanes 4 to 6). We concluded that the induction of autophagy by HCV infection does not involve the UPR.

Fig 6.

The induction of autophagy by HCV is independent of Ire1 and the UPR. Naïve Huh7 cells were transfected with control siRNA or an siRNA against Ire1 by use of RNAiMax for 24 h prior to infection with a chimeric J6/JFH-1luc reporter virus. At 72 h postinfection, cells were harvested for luciferase assay and Western blotting.

HCV genome replication does not occur on autophagosomes.

An increasing body of evidence has recently pointed toward a role for the cellular autophagic response in generating the cytoplasmic membrane structures required for RNA replication of positive-strand RNA viruses, such as picornaviruses, coronaviruses, and flaviviruses (14, 20, 28, 39, 42). Many of these reports demonstrate colocalization of autophagosomes and viral proteins, suggesting that for these viruses RNA replication occurs either on or within autophagosomes. For HCV, there are conflicting reports in the literature—two studies utilized a green fluorescent protein (GFP)-LC3 fusion protein to either support (32) or contest (7) the hypothesis that HCV replication complexes are localized on autophagic membranes. Because ectopic expression of GFP-LC3 has been reported to lead to incorporation of the fusion protein into insoluble cytoplasmic aggregates (17, 19), potentially confounding any interpretation of the data, we analyzed the subcellular distribution of autophagosomes and HCV replication complexes by an indirect coimmunofluorescence assay using antibodies to endogenous LC3 and NS5A. We undertook this analysis with Huh7 cells stably harboring either a JFH-1 or FK5.1 SGR, as well as with cells infected with JFH-1. As expected, NS5A staining revealed an extensive punctate distribution throughout the cytoplasm; the majority of these puncta represent replication complexes, although a small proportion of NS5A staining was found in other locations. However, the NS5A staining was more widespread than that of LC3 puncta (predominantly representing autophagosomes), and there was no significant colocalization of the two (Fig. 7C, E, G, and H). This was despite the fact that, as expected, LC3 puncta were more numerous in Huh7 cells either stably harboring the SGR or infected with JFH-1 than in the parental or uninfected cells (compare Fig. 7A with Fig. 7C, E, G, And H). Treatment of SGR-harboring cells with tunicamycin did not result in the colocalization of NS5A and LC3, despite the induction of the UPR (Fig. 7B, D, And F). We concluded that HCV induces autophagy but does not utilize the resulting autophagosomes as a platform for genome replication.

Fig 7.

HCV RNA replication sites do not colocalize with autophagosomes. Naïve Huh7 cells (A and B) or Huh7 cells stably harboring SGR-neo-JFH-1 (C and D) or SGR-FK5.1 (genotype 1b) (E and F) or infected with JFH-1 (G and H) were fixed and stained with antibodies to LC3 (red) and NS5A (green). Where appropriate, cells were treated with tunicamycin (TM). Representative confocal images are shown.

DISCUSSION

Our data presented here demonstrate that both autophagy and the UPR can be induced efficiently by HCV infection, in agreement with previous studies (16, 31). However, four lines of evidence led us to conclude that induction of autophagy is independent of the UPR. First, a time course analysis of virus infection revealed that autophagy was a very early response, occurring within 4 h, whereas the UPR was apparent only at 48 h or more postinfection (Fig. 1). Second, Huh7 cells harboring a variety of SGRs (both transient and stable) showed a robust induction of autophagy, but the UPR was not induced (Fig. 2 and 3). Third, induction of autophagy by HCV infection or SGRs was also observed in cells in which the UPR was inhibited by silencing of Ire1 expression (Fig. 5 and 6). Lastly, induction of the UPR was dependent on expression of the viral envelope glycoproteins (Fig. 1B), which seems appropriate given that these accumulate within the ER. Although our data are compelling, we of course cannot absolutely rule out the possibility that very low levels of UPR (below the detection threshold) could play a role in HCV induction of autophagy. We concluded that although there is extensive rearrangement of cytoplasmic membranes in SGR-harboring cells, with the majority of the nonstructural proteins associating with these membranes, these events do not induce the UPR. However, in contrast to a previous report (37), the UPR pathway remained intact in the presence of the SGR, as treatment of cells with tunicamycin could efficiently induce both xbp1 mRNA splicing and CHOP expression.

So how does HCV induce autophagy? It is possible that one or more of the nonstructural proteins expressed from the SGR could be implicated in this induction. Indeed, it was recently shown that NS4B expression alone is able to induce autophagy via interactions with both the early endosome-associated GTPase Rab5 and a class III phosphoinositide 3-kinase, Vps34 (34); additionally, NS5B has been reported to interact with ATG5 (12). However, our observation that autophagy is induced very early after infection (4 h), at least 20 h prior to the detectable expression of the nonstructural proteins, is indicative that a component of the input virus might initially be involved in inducing autophagy. In this regard, there is a clearly described link between the innate immune response and induction of autophagy; in particular, endosomal recognition of single-stranded RNA (ssRNA) or the small-molecule agonist imiquimod can activate autophagy (reviewed in reference 27). It is therefore possible that the input viral RNA could initially stimulate autophagy; in support of this hypothesis, we have observed that transfection of a variety of RNA species (including the HCV 5′- and 3′-untranslated regions [UTRs], yeast tRNA, and Renilla luciferase transcripts) can induce LC3-II accumulation within 4 h (data not shown). Many groups have used the siRNA silencing approach to show that autophagy is required for HCV RNA replication (7, 16, 30, 31, 36); furthermore, the autophagy machinery has been reported to be required for the translation of incoming viral RNA (7). Stimulation of autophagy by the input virus would therefore facilitate the establishment of a productive replication cycle. At a later stage, once NS3 has been expressed and has acted to inhibit the innate immune response by cleaving TRIF and MAVS (21, 23), NS4B, NS5B, and/or other nonstructural proteins could act to maintain autophagy.

An additional question relates to why HCV might induce autophagy. A number of recent studies have alluded to a role for the induction of autophagy in the suppression of the innate immune response (16, 30); alternatively, it has been suggested that autophagy protects cells from defects in lipid metabolism induced by HCV infection (40). Interestingly, dengue virus also induces autophagy as a mechanism to regulate cellular lipid metabolism (13), suggesting that this might be a common rationale for induction of autophagy by related viruses. However, for a number of other viruses, e.g., picornaviruses (39) and coronaviruses (28), autophagosomes are the platform for RNA replication, and there is a strong colocalization of LC3-II puncta with the viral replicase complex. This is a controversial topic in the case of HCV, as the replicase has been shown both to colocalize with (32) and to be distinct from (7) autophagosomes. Our data agree with the latter study, as we did not observe colocalization of autophagosomes and sites of virus RNA replication in either virus-infected cells or stable SGR-harboring cells (Fig. 7). We cannot formally exclude the possibility that HCV replication in autophagosomes blocks accessibility to the LC3 antibody or that a component of the HCV replicase redistributes LC3-II away from autophagosomes; however, these explanations are less likely given that LC3-II is localized on the cytoplasmic face of autophagosomes and that the HCV replicase proteins are contained within membranous vesicles. Whatever the explanation, the controversy surrounding the association between HCV replication complexes and autophagosomes remains. However, it is important that our data were derived from analyses of endogenous LC3-II, whereas other studies have utilized either stable (32) or transient (7) expression of a GFP-LC3 fusion protein. Caution has been urged for the interpretation of data derived from analyses of GFP-LC3 due to its potential to associate with protein aggregates (17), and thus a more detailed analysis is warranted and most likely should involve electron microscopic visualization of autophagosomes and sites of HCV replication.

In conclusion, our data clearly show that HCV induces autophagy in a UPR-independent fashion. The precise mechanisms by which this induction is effected and the implications of autophagy for the virus life cycle remain to be elucidated. These studies are ongoing in our laboratory.