Treating hepatitis C: Can you teach old dogs new tricks?


  • Potential conflict of interest: Nothing to report.


Viruses depend on host-derived factors for their efficient genome replication. Here, we demonstrate that a cellular peptidyl-prolyl cis-trans isomerase (PPIase), cyclophilin B (CyPB), is critical for the efficient replication of the hepatitis C virus genome. CyPB interacted with the HCV RNA polymerase NS5B to directly stimulate its RNA binding activity. Both the RNA interference (RNAi)-mediated reduction of endogenous CyPB expression and the induced loss of NS5B binding to CyPB decreased the levels of HCV replication. Thus, CyPB functions as a stimulatory regulator of NS5B in HCV replication machinery. This regulation mechanism for viral replication identifies CyPB as a target for antiviral therapeutic strategies.

Watashi K, Ishii N, Hijikata M, Inoue D, Murata, T, Miyanari Y, Shimotohno K. Cyclophilin B is a functional regulator of hepatitis C virus RNA polymerase. Mol Cell 2005;19:111-122. (Reprinted with permission from Elsevier.)


Therapy for hepatitis C has improved steadily since administration of type I interferons (IFNs) was first shown to have modest efficacy.1, 2 Increasing the frequency and dose of IFN improved response rates, but addition of the nucleoside analog ribavirin yielded an amazing threefold improvement in sustained virologic response3—“amazing” because ribavirin by itself showed at best transient antiviral effects.4 Despite today's 20% to 80% sustained virologic response rates depending on hepatitis C virus (HCV) genotype and ethnic background, the mechanisms of action for IFN and ribavirin are still unknown.5 IFN induces gene products that presumably block one or more steps in HCV replication. Ribavirin's mode of action is hotly debated, with possible mechanisms that range from limiting nucleoside pools by inhibiting inosine monophosphate dehydrogenase to immunomodulation or acting as a mutagen to reduce viral fitness and promote error catastrophe.4, 6 Both IFN and ribavirin were tried for hepatitis C because of their previous track records. IFN-α had been used successfully to treat HBV,7 while ribavirin had been administered for influenza viruses, respiratory syncytial virus, parainfluenza, and various hemorrhagic fever–causing viruses such as Lassa fever virus.8, 9 In a sense, these therapeutic agents are like old dogs that have learned new tricks.

Enter cyclosporin A (CsA). As every hepatologist knows, CsA has been used as an immunosuppressant after liver transplantation to prevent rejection and improve allograft survival. For many applications, CsA has been replaced by FK506 (tacrolimus).10 Both CsA and FK506 bind to cellular proteins called immunophilins; CsA binds to cyclophilins (CyPs), and FK506 binds to FK506-binding proteins. These drug–protein complexes inhibit calcineurin, a calcium-activated protein phosphatase that is required for T-cell activation through the NF-AT pathway. CyPs and FK506-binding proteins are peptidyl-prolyl cis-trans isomerases (PPIases) able to catalyze the cis-trans interconversion of peptide bonds N-terminal to proline residues11 to influence protein folding and function (see Heitman and Cullen12 and Watashi et al.13 and citations therein). CsA binding to CyPs inhibits PPIase activity.11

In a paper previously published in HEPATOLOGY, Watashi et al.14 reported a surprising inhibitory effect of CsA on HCV replication using a replicon system they had developed.15 Treatment of the replicon cell cultures with CsA but not FK506 resulted in a dose-dependent decrease in HCV RNA and protein levels that paralleled the antiviral effects seen with IFN-α. This was also seen in an HCV infection model using a cultured nonneoplastic human hepatocyte line (PH5CH8).16 Inhibition was achieved at levels of CsA within its therapeutic range and did not result from nonspecific cytotoxic effects. The absence of IFN-induced gene expression in CsA-treated cells ruled out this pathway as the explanation for CsA's anti-HCV activity. Finally, Watashi et al. showed that the anti-HCV activity of CsA was independent of its immunosuppressive activity. This very exciting finding from a potential therapeutic standpoint was demonstrated in two ways. First, CsA treatment of the replicon-containing cells did not inhibit the transcriptional activation of pathways that are blocked in T cells and responsible for immunosuppression (NF-AT, AP-1, and NF-κB). Second, potent anti-HCV activity was still seen using CsA analogs that bind CyP and inhibit PPIase activity but fail to associate with calcineurin and block the NF-AT pathway. Similar results were published independently by Nakagawa et al.17 This implicated formation of a CsA–CyP complex as being responsible for inhibiting HCV replication—but three questions remained. Could other effects of CsA on cellular pathways be ruled out? If CyP was the target of CsA's anti-HCV action, which of the 16 human CyPs12, 18 was involved? Finally, if a CyP was participating in HCV replication, how was it functioning?

These questions are addressed in an elegant study recently published by Watashi et al.13 CsA has three known cellular targets; two of these, CyP and calcineurin/NF-AT, have been previously mentioned. The third target, P-glycoprotein, is a cell surface transporter involved in multidrug resistance. Using an extended set of CsA analogs and mutants—as well as unrelated CyP inhibitors Sanglifehrin A and B—a clear correlation between CyP inhibition and decreased HCV RNA levels was observed, implicating CyP in HCV replication.

The next step was to identify “the” CyP out of the more than 10 CyP subfamilies in mammalian cells.12, 18 RNA interference (the subject of a recent HEPATOLOGY editorial19) was used to knock down CyPs and assess the effect of diminished CyP levels on HCV replication. Small interfering RNAs were introduced into replicon-containing cells and HCV RNA levels were measured 5 days later. Small interfering RNAs designed to broadly target CyPs decreased CyPA and CyPB as well as HCV RNA levels to 20% of control values. CyP-specific small interfering RNAs demonstrated that decreasing CyPB but not CyPA, CyPC, CyPE, or CyPH inhibited the accumulation of HCV RNA. Although RNA interference is a near miraculous method for knocking down specific genes, induction of IFN and off-target effects have been noted.20, 21 However, overexpression of epitope-tagged CyPB restored HCV replication in the presence of the CyPB-specific small interfering RNAs ruling out these concerns and solidifying a role for CyPB in HCV replication. CyPs were also implicated in HCV inhibition by CsA in a recent study by Nakagama et al.22 However, unlike the Watashi study, the specificity of RNA interference knockdown effects was not clearly established.22

So how does CyPB participate in HCV replication? Getting to mechanism of action starting with a small molecule inhibitor and a cell-based assay is extremely difficult. Inhibition can be mediated by direct interaction of the target protein, such as CyPB, with a viral component that is blocked by CsA, or effects can be indirect—for example, by inhibiting the CyPB PPIase activity and affecting the folding of another host or viral protein required for HCV replication. This is where the field of virus–host interactions becomes challenging and where the Watashi et al. study shines.

In the next set of experiments, possible interactions between CyPB and replicon-expressed viral proteins were examined. Only one was identified: an interaction between CyPB and the NS5B RNA-dependent RNA polymerase. This interaction was seen with NS5B produced by cell-free translation or with NS5B from replicon-containing cells. Furthermore, the NS5B–CyPB interaction was inhibited by CsA in a dose-dependent manner, correlating with the HCV inhibitory effects of CsA seen in cells. They went on to show that the C-terminal region of the polymerase (residues 521-591) was important for interaction with CyPB. This portion of NS5B contains part of the polymerase catalytic core as well as a hydrophobic tail of 20 amino acids at the C terminus that anchors NS5B to cellular membranes.23

If CyPB and NS5B interact to facilitate HCV replication, they should colocalize to sites of RNA replication. CyPB is predominantly localized in the lumen of the endoplasmic reticulum.24 The HCV RNA replicase is also membrane-associated, but replication is believed to occur on the cytosolic side of modified cellular membranes.25 This creates a topological problem with CyPB on one side of the membrane separated from NS5B on the other side. By selectively permeabilizing cells with digitonin, Watashi et al. identified a fraction of CyPB that was present on the cytosolic face of intracellular membranes. Furthermore, they showed colocalization of CyPB and NS5B with sites of nascent HCV RNA synthesis. This provided further support for active involvement of CyPB in viral RNA replication.

Clues as to how CyPB might function in HCV RNA replication came from a comprehensive series of cell biological, biochemical, and genetic experiments. CyPB might be participating in HCV replication by altering replicase formation or influencing the stability of key replicase proteins such as NS5B. However, in the presence of CsA, no differences were observed in formation of the replicase, the localization of NS5B, or the levels of replicase components such as NS5A and NS5B. This suggests that CyPB might directly regulate the function of NS5B. Further experiments demonstrated that interaction with CyPB enhanced NS5B's ability to bind RNA. This was abolished in the presence of CsA and was not seen with a mutant form of CyPB without PPIase activity or with forms of NS5B lacking the C-terminal CyPB-interacting region. Hence, CyPB may act as an NS5B cofactor to enhance RNA binding and facilitate RNA replication.

Finally, to close the loop on the importance of the CyPB–NS5B interaction, Watashi et al. conducted an alanine scanning mutational analysis of the NS5B 521-591 region and identified a single amino acid substitution (proline to alanine at NS5B position 540) that abrogated the NS5B–CyPB interaction. This mutation did not affect the stability, localization, polymerase activity, or intrinsic RNA-binding activity of NS5B, but it did compromise the replication ability of replicons harboring the mutation. Although one could argue that this mutation could be affecting other NS5B functions (like its assembly into functional replication complexes), the results are consistent with the authors' model, in which a direct NS5B–CyPB interaction is required for efficient HCV RNA replication (Fig. 1).

Figure 1.

Model of the role for CyPB in HCV RNA replication. (A) Interaction of CyPB with NS5B enhances RNA binding and promotes RNA replication. (B) CsA blocks the ability of CyPB to interact with NS5B, giving rise to weaker RNA binding and the inability to form a functional RNA replication complex. Not shown is nonstructural protein NS4B, which is involved in formation of the membraneous web—a vesicle-rich structure where HCV RNA replication occurs. Also not shown for simplicity are the additional sites of membrane attachment mediated by hydrophobic segments in nonstructural proteins NS4A, NS4B, NS5A, and NS5B. CyPB, cyclophilin B; CsA; cyclosporin A.

This study sets a high standard for the study of HCV–host interactions, but it also raises many questions. How does enhanced RNA binding facilitate HCV RNA replication? What steps are affected—initiation, processivity, elongation? Is enhanced RNA binding due solely to NS5B determinants, or does CyPB also interact with RNA when complexed with NS5B? The PPIase activity of CyPB is dispensable for interaction with NS5B but is required for RNA binding enhancement. What is the substrate for PPIase activity? Is proline isomerization the only function of CyPB, or does it make other contributions to HCV RNA replication? Is CyPB activity required for events downstream of RNA replication such as virus particle assembly and release?

For clinicians desperate for new hepatitis C treatment options, will nonimmunosuppressive CsA derivatives or other CyPB inhibitors be part of the answer? An antiviral effect of CsA was not observed in the setting of recurrent HCV infection following liver transplantation. In these immunosuppressed patients, HCV viremia is typically high and liver disease is often rapidly progressive.26 Studies reporting clinical benefits of CsA for treating chronic hepatitis C are mixed27 but are sufficiently promising28, 29 to warrant further study. Many issues have yet to be settled. Will the NS5B–CyPB interaction be important for all HCV genotypes and isolates? Will resistant HCV variants arise at low fitness cost to the virus? How would such inhibitors be used—alone or in combination with existing therapies or HCV-specific antivirals that are in the pipeline? To return to our burning question—can you teach an old dog new tricks? You can, depending on the dog and the trainer. Will an old drug (or a modification thereof) be used for a new indication? Time will tell, but the wait should not be long.


Thanks to Sadie, Wrangler, and Margaret MacDonald for their helpful input.