Modeling the antiviral activity of ribavirin against hepatitis C virus in cell culture


  • Daniel J. Felmlee,

    1. Inserm, U1110, Institut de Virologie, Strasbourg, France
    2. Université de Strasbourg, Strasbourg, France
    Search for more papers by this author
  • Fei Xiao,

    1. Inserm, U1110, Institut de Virologie, Strasbourg, France
    2. Université de Strasbourg, Strasbourg, France
    Search for more papers by this author
  • Thomas F. Baumert

    Corresponding author
    1. Université de Strasbourg, Strasbourg, France
    2. Pôle Hépato-digestif, Hôpitaux Universitaires de Strasbourg, Strasbourg, France
    • Inserm, U1110, Institut de Virologie, Strasbourg, France
    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

  • See Article on Page 1236

  • Supported by Inserm, ANRS (13081), University of Strasbourg, the European Union (INTERREG-IV-Rhin Supérieur-FEDER-Hepato-Regio-Net 2009 and 2012, FP7 HEPAMAB GAN 305600), ERC AdG HEPCENT 233130, Direction Générale de l'Offre de Soins, and the Laboratoire d'excellence HEPSYS (ANR-10-LAB-28).

  • D.J.F. supported by an EASL fellowship.

Address reprint requests to: Thomas F. Baumert, M.D., Inserm Unité 1110, Hôpitaux Universitaires de Strasbourg, Université de Strasbourg, 3 Rue Koeberle, F-67000 Strasbourg, France. E-mail:; fax: +33 3 68 85 37 24.


adenosine kinase


direct acting antiviral


guanosine triphosphate


hepatitis C virus








internal ribosome entry site


IFN stimulated genes


primary human hepatocytes




RBV monophosphate


ribavirin triphosphate


standard of care


sustained virologic response


untranslated region.

While the recent inclusion of direct-acting antiviral (DAA) therapies has recently improved the standard of care (SOC) for patients with hepatitis C virus (HCV) genotype 1 infection; the remaining limitations of efficacy, side effects, and high costs remain challenges for improving therapy. A foreseeable goal is an exclusively orally administered treatment regimen, free of interferon (IFN) and IFN-associated side effects.[1] While the current SOC for patients with genotype 1 infection is composed of pegylated IFN alpha with ribavirin (RBV) and either telaprevir or boceprevir, treatment is anticipated to improve by the inclusion of a second-generation protease inhibitor and/or DAAs targeting the viral polymerase or NS5A protein, and eventually removal of IFN.[2] A remaining arm of anticipated future treatment is the guanosine nucleotide analog, RBV. Recent results with next-generation DAAs including sofosbuvir have highlighted RBV's role in the upcoming anti-HCV regimens.[3, 4] Even though RBV has been employed in treating hepatitis C for more than 20 years, the primary mechanism of its action is still unclear. This lack of clarity is hindered by the current state-of-the-art Huh7 cell-based models of HCV infection poorly reflecting the in vivo activity of RBV at clinical concentrations.

There is evidence supporting multiple mechanisms of RBV's anti-HCV activity (Fig. 1).[5, 6] While there is an extrahepatic immunomodulatory role of RBV that alters helper T-cell responses to favor Th1 responses while suppressing Th2 responses,[7] the majority of the evidence indicates that the more important mechanisms of RBV-mediated HCV clearance occur most likely within the hepatocyte.[5] The proposed intrahepatic mechanisms of RBV action involve conversion of RBV to its monophosphate form (RMP) by host enzyme adenosine kinase (ADK),[8] followed by subsequent conversion into triphosphate form (RTP). Putative mechanisms of RBV anti-HCV activity include direct inhibition of viral polymerase activity, introduction of terminating numbers of mutations resulting in error catastrophe, and inhibition of inosine-monophosphate-dehydrogenase (IMPDH) that would result in depletion of guanosine triphosphate (GTP) pools.[5, 6] Furthermore, recent evidence suggests that an important mechanism of RBV is in potentiating inflammatory defenses through activation of IFN-stimulated genes (ISGs) beyond the stimulation of IFN alone.[9, 10] The role of RBV conversion to RTP as a preceding step for ISG activation is not well defined, but since the addition of guanosine reverses ISG stimulation, it is likely important.[10] The contribution and role of all of these possible antiviral mechanisms has been difficult to ascertain since the hepatoma cell line Huh-7, which until recently has been exclusively capable of sustaining HCV replication, is resistant to RBV treatment at clinically relevant levels.[11]

Figure 1.

Proposed intrahepatic mechanisms of RBV's anti-HCV activity include the conversion of RBV to RMP by ADK. Extrahepatic roles of RBV may include an immunomodulatory role that favors T helper (TH) cells to have a predominance of TH1 cells relative to TH2 cells,[7] favoring induction of cytotoxic T lymphocytes (CTL) that can recognize and inhibit HCV-infected cells. RBV is transported into hepatocytes and converted to RMP, then subsequently RBV diphosphate (RDP) and RBV triphosphate (RTP). Downstream proposed mechanisms of RBV action dependent on ADK (represented in gray boxes) include direct inhibition of IMPDH leading to accumulation of IMP and depletion of guanosine mono- and triphosphate (GMP, GTP) pools, direct inhibition of the viral RNA dependent RNA polymerase (RdRp), and incorporation of RTP into replicating genomes resulting in hypermutation and error catastrophe generating a preponderance of defective HCV particles. Depleted GTP pools may also contribute to increased RTP incorporation during HCV replication. It remains to be investigated if the important induction of ISGs that potentiates IFN responses,[9, 10, 20] or the immunomodulation of TH populations is dependent on ADK (white boxes). Mori et al. found that the ADK messenger RNA (mRNA) may contain an IRES in its 5′ UTR. It is possible that this could contribute to the abundance of ADK in RBV-effective cells such as Li23 and primary hepatocytes. Huh-7-derived cells could have RBV-effectiveness against HCV restored by the ectopic expression of ADK containing the putative IRES. Figure adapted and updated from Feld et al. and Thomas et al.[5, 6]

In this issue of Hepatology, Mori et al.[13] contribute a scientific advance to address this puzzle by investigating RBV's anti-HCV activity utilizing a previously identified hepatoma cell line, distinct from Huh7, which is capable of sustaining HCV replication. This cell line, Li23, displayed a different gene expression profile from Huh7 and, in contrast to Huh7-based cell lines, the anti-HCV activity of RBV was effective at more clinically relevant concentrations.[13, 14] The authors took advantage of this RBV-sensitive phenotype by conducting a comparative microarray analysis of transcript levels between an Huh-7-derived line and an Li23-derived line, both altered to report HCV replication. The RBV-sensitive line had more than 4-fold more ADK transcript expression than the RBV-resistant line, and ∼16-fold more ADK protein.[13] ADK converts RBV into RMP, an inhibitor of IMPDH. Since IMPDH converts inosine-5-monophosphate (IMP) into a precursor of GTP synthesis, Li23-derived cells treated with RBV were more dramatically depleted of GTP and accumulated more IMP levels than Huh-7 derived cells. This depletion of GTP pools is a mechanism that may contribute to RBV antiviral activity, especially in the Li23 cell line.[14] The authors defined ADK's role in mediating RBV sensitivity by ectopically expressing ADK in the Huh-7-derived cell line. Importantly, RBV treatment at clinically relevant concentrations now had an anti-HCV effect. Specific inhibitors of ADK in the Huh-7-expressing ADK again reversed the antiviral effect of RBV, demonstrating ADK as a mediator of RBVs antiviral effect in cell culture.

In searching for the reasons behind the differential expression of ADK between the two cell lines, the authors found by 5′ RACE three predominantly expressed ADK transcripts with different 5′ untranslated region (UTR) lengths: 125 nucleotides (nt), 187 nt, and 319 nt. Interestingly, the quantity of the long 319 nt 5′ UTR correlated with increased ADK quantities among different hepatoma cell lines and was highly expressed in the primary human hepatocytes (PHH) tested. The authors examined the 319 nt 5′ UTR, and seeing that it was highly structured and GC-rich, tested it for internal ribosome entry site (IRES) activity using a bicistronic IRES reporter assay. Surprisingly, the 319 nt 5′ UTR of ADK has a more robust IRES activity than the HCV IRES, which may contribute to the difference in ADK quantity between the cell lines.

These findings contribute to the understanding of the action of RBV against HCV, reveal a possible regulatory mechanism of a critical step of RBV activity, and provide a new model in which the mechanisms of clinically relevant concentrations of RBV against HCV can be further defined. These are important steps forward considering that RBV is a critical component of anti-HCV triple therapy and is anticipated to remain a component of antiviral cocktails for years to come.[15] The robust antiviral activity of RBV in vitro occurred by way of ADK in a dose-dependent and reversible manner, highlighting that ADK clearly mediates RBV's anti-HCV activity. This finding is expected since all the proposed intrahepatic mechanisms involve downstream products of ADK activity on RBV,[8] but the authors definitively confirm ADK's role. The true contribution of ADK's IRES in increasing protein expression remains to be determined, and use of the bicistronic reporter assay outside of stress conditions has been criticized due to cryptic promoters and unanticipated splicing.[16] Although the authors intensively searched, the cause of the more than 4-fold increase of ADK transcript was not determined, and while the amplification of 16-fold more ADK protein may be due to the presence of the IRES, it remains unknown why this translation initiation would be favored under typical cell growth conditions over cap-mediated translation. As with other genes containing IRES activity, ADK is an enzyme that would be critical to preserve in conditions of stress or nutrient starvation when cap-mediated translation is compromised,[16] in ADK's case nucleotide metabolism. However, the authors ruled out ADK's IRES induction by stress caused by HCV infection, since cured cells had similar IRES activity as HCV replicating cells.[13]

Perhaps the most relevant contribution of this work is the establishment of a system to analyze the effects of RBV with clinically relevant concentrations in classically studied Huh-7-derived cell lines. Simple ectopic expression of ADK conferred Huh7 cells to be RBV-effective against HCV at clinically relevant concentrations, and since the mechanism of action of RBV is likely multifactoral, future studies will likely use this system to weigh the contribution of each mechanism to determine the final anti-HCV effect. Likewise, the manner in which RBV synergizes with IFN and with DAAs in vitro will be an intense area of investigation. Furthermore, there may be therapeutic potential in boosting ADK levels to stimulate RBV's effect despite the occurrence of ADK mutation or deficiency being rare,[17] and the PHH tested having high expression levels of ADK.[13]

The upcoming goal for treatment of HCV infection is a completely orally administered, IFN-free regimen.[2] RBV fits well into this goal, and is included in many future IFN-free combinations anticipated.[1, 4] The PROVE 2 trial demonstrated that including RBV along with a DAA improved the sustained virologic response (SVR) from 36% to 69%.[18] Both sofosbuvir[3] and ABT-333/ABT-450[4] have proven quite effective when used in combination with RBV. While RBV has some direct antiviral effect as a monotherapy,[19] it functions clinically to synergize with other therapies and inhibit viral relapse and breakthrough mutations.[15, 20] The precise mechanisms of RBV may be difficult to distinguish, as different mechanisms may act synergistically when coupled; for example, the diminution of GTP pools by way of IMPDH inhibition may work to increase the incorporation of mutations leading to error catastrophe.[5] Recent HCV deep-sequencing data from patients under RBV monotherapy supports this mutagenic hypothesis,[21] while in vitro and in vivo data show an IFN potentiating role.[9, 10] Mori et al. have contributed an important piece of the puzzle in studying RBV mechanisms in cell culture models and have revealed how much work is still needed to definitively identify RBV function. The anticipated results of these future studies lend hope that similar agents can be developed with improved efficacy and fewer side effects that represent an improvement over RBV.


The authors thank Dr. T. Jake Liang, Liver Diseases Branch, NIDDK, Bethesda, MD, for helpful discussions.

  • Daniel J. Felmlee, Ph.D.1,2Fei Xiao, M.D.1,2Thomas F. Baumert, M.D.1,2,3

  • 1Inserm, U1110, Institut de Virologie, Strasbourg, France

  • 2Université de Strasbourg, Strasbourg, France

  • 3Pôle Hépato-digestif, Hôpitaux Universitaires de Strasbourg, Strasbourg, France