HEPATOLOGY ELSEWHERE EDITORSHartmut Jaeschke, Tucson, AZKevin Mullen, Cleveland, OHDarius Moradpour, Lausanne, Switzerland

Ribavirin Mutagenesis: Hidden Clues inMathematical Models

Dixit NM, Layden-Almer JE, Layden TJ, Perelson AS.Modelling how ribavirin improves interferon responserates in hepatitis C virus infection. Nature 2004;432:922-924. (Reprinted with permission from Nature PublishingGroup, http://www.nature.com.)

AbstractNearly 200 million individuals worldwide are currently in-

fected with hepatitis C virus (HCV). Combination therapy withpegylated interferon and ribavirin, the latest treatment for HCVinfection, elicits long-term responses in only about 50% of pa-tients treated. No effective alternative treatments exist for nonre-sponders. Consequently, significant efforts are continuing tomaximize response to combination therapy. However, rationaltherapy optimization is precluded by the poor understanding ofthe mechanism(s) of ribavirin action against HCV. Ribavirin aloneinduces either a transient early decline or no decrease in HCV viralload, but in combination with interferon it significantly improveslong-term response rates. Here we present a model of HCV dy-namics in which, on the basis of growing evidence, we assume thatribavirin decreases HCV infectivity in an infected individual in adose-dependent manner. The model quantitatively predicts long-term response rates to interferon monotherapy and combinationtherapy, fits observed patterns of HCV RNA decline in patientsundergoing therapy, reconciles conflicting observations of the in-fluence of ribavirin on HCV RNA decline, provides key insightsinto the mechanism of ribavirin action against HCV, and estab-lishes a framework for rational therapy optimization.

CommentsThe current standard of care for chronic hepatitis C is

a combination of pegylated interferon-� (Peg-IFN-�)and ribavirin (RBV). More than 50% of treated patientsachieve sustained viral response (SVR), reflecting the re-markable progress made clinically throughout the past 15years.1,2 However, the mechanisms of action for both in-terferon-� (IFN-�) and RBV remain poorly understood.

IFN-� was first approved as monotherapy for treat-ment of chronic hepatitis C in 1990. The efficacy ofIFN-� correlates with the duration of treatment, and thedose and continuity of drug exposure. Viral response toIFN-� is rapid but incomplete, resulting in a biphasicdecline in serum hepatitis C virus (HCV) RNA levels.3

Using mathematical modeling and various data-fittingtechniques, Neumann et al. concluded that the majorityof IFN-� antiviral activity resulted from inhibiting viral

production or release from infected hepatocytes.3 Thiswas further supported by in vitro studies showing thatHuh-7 cells harboring subgenomic HCV replicons couldbe “cured” by adding IFN-� to the cell cultures.4,5 Thisclearly demonstrated that IFN-� could effectively inhibitviral replication and eliminate intracellular viral RNAthrough a noncytolytic mechanism. IFN-� is known toexhibit pleiotropic antiviral activities through severalwell-characterized effector molecules, including an en-zyme known as double-strand RNA-specific adenosinedeaminase (ADAR1), which can induce A-to-G muta-tions in viral genomes.6 Therefore, IFN-� may introducedetrimental mutations into HCV genomes.7

One major drawback of IFN-� monotherapy is thehigh relapse rate. Many patients who achieve end-of-treatment response (ETR) with undetectable serum HCVRNA become viremic again during follow-up,8 suggestingan incomplete viral clearance or control. Little is knownabout the cause of this relapse. RBV, a small nucleosideinhibitor, was found to uniquely enhance ETR and pre-vent relapse, thereby increasing the rate of SVR.9 A recentstudy by Pawlotsky et al. re-examined the antiviral effectof RBV with more frequent sampling during the earlyphase of RBV monotherapy.10 They convincingly dem-onstrated that RBV was able to induce a weak, transient,yet significant antiviral response, suggesting that RBVdoes not simply serve as an adjunct therapy to IFN-�.Somehow, RBV can synergistically enhance the IFN-�action.

Like IFN-�, the mechanisms of action of RBV arepleiotropic and poorly characterized. Two new theorieshave been proposed to explain the effect of RBV in HCVtherapy: (1) immunomodulation toward T helper 1 re-sponses in favor of antiviral immunity and (2) lethalmutagenesis to the HCV RNA genome.11,12 Immuno-modulation may result in cytolytic “killing” as well asnoncytolytic “curing” of infected cells (Fig. 1). Lethalmutagenesis can reduce viral infectivity and fitness byincreasing mutations in the genome beyond a thresholdthat “error catastrophe” ensues.12

Proving these theories to be clinically relevant has beendifficult. The immunomodulatory activity of RBV can beovershadowed by IFN-� therapy. RBV mutagenesis hasbeen demonstrated using poliovirus as a surrogate thatallows the functional selection of a mutation against abiomarker.13 Demonstration of RBV-induced mutations

1399

in HCV has been controversial because of the lack offunctional assays. Large-scale sequencing following error-prone reverse-transcriptase polymerase chain reactionamplification is the only choice. The ability to detectRBV-increased mutations has been investigator- or sys-tem-dependent with a range of findings, from equivocalto statistically significant.7,10 The highly dynamic natureof HCV quasispecies adds complexity to the measure-ment of slightly increased mutation rate based on directsequencing of a small number of clones. There is a need todevise new methodology to study the mutagenesis theory.

Mathematical models have been used to delineatemechanisms of action of antivirals. Herrmann and col-leagues, in a study to compare viral decay kinetics betweenpegylated IFN-� alone and pegylated IFN-� or IFN-�plus RBV, were among the first to use mathematical mod-els to study the mechanisms of action of RBV.14 IFN-�effectiveness (�) measured by the log drop in viral titersduring the first day of treatment, was shown to be lower atdosages approved by the U.S. Food and Drug Adminis-tration (� � 0.36-0.67). A unique triphasic viral declinewas observed in a proportion of patients. The third phasedecline was interpreted as representing an accelerated lossrate of infected cells by RBV and correlated with im-

proved SVR. The acceleration of loss rate of infected cells(�) was modeled by an inflation factor M and attributedto an enhanced immune killing, indicating immuno-modulation as the mechanism of action of RBV (Fig. 1).14

However, this enhanced immune killing was not sup-ported by the more pronounced alanine aminotransfer-ase/aspartate aminotransferase suppression in the RBVgroups. The authors attributed this to the noncytolytic“curing” of infected cells devoid of significant liver injury(Fig. 1). Interestingly, it was shown previously that a smalland transient rise in alanine aminotransferase in earlyphases of IFN-� therapy correlated with favorable treat-ment responses.15 Nevertheless, Herrmann et al. were un-able to rule out the possibility of RBV mutagenesiscontributing to the accelerated loss of infected cells. Theeffect of RBV on de novo infection (�), more specificallyon the noninfectious virion compartment (�, fraction ofnoninfectious virions), was not analyzed. Perhaps themodel was unable to distinguish between two variablessimultaneously (i.e., effect on loss rate of infected cellsversus effect on infectivity).

In a seminal report published in Nature last December,Dixit and colleagues further investigated the mechanismof action of RBV using mathematical models.16 High-dose (10 MIU) IFN-� was administered daily to maxi-mize its effectiveness. Under the condition of high IFN-�effectivenss (� � 0.95), viral titer changes over the first 28days of therapy could be fitted to a biphasic decline curve:a very rapid log10 reduction on day 1 (first phase) followedby a slow and gradual viral decay starting from day 2(second phase).3 RBV was shown to have no observableeffect on the kinetic parameters such as � or � of eitherphase, suggesting that IFN-� dominated or saturated theimmune responses in the early phases. If � � �IFN-� ��RBV, when high-dose IFN-� was used (�IFN-� �� �RBV),the immunomodulatory activity will be saturated by thatof IFN-�, limiting RBV’s contribution to a negligiblelevel. By keeping RBV’s immunomodulatory effect to aminimum, the effect in long-term responses may be at-tributed mainly to the mutagenic effect if any, allowing amore convincing measurement of RBV mutagenesis.

In the Nature paper the authors theorized that the ac-tion of RBV was due to lethal mutagenesis within twoviral compartments: infectious versus noninfectious viri-ons. A key parameter � (related to �), representing thefraction of newly produced noninfectious virions due toRBV action, was termed the effectiveness of RBV mu-tagenesis.16 Based on this model, total viral production isquantitatively blocked by IFN-� by a factor of (1-�) but isqualitatively reduced by RBV during the next round ofinfection by a factor of (1-�) (Fig. 1). The clinical effect ofRBV seems to be more cumulative and delayed with the

Fig. 1. HCV dynamics and mechanisms of action of interferon-� (IFN)and ribavirin (RBV). Productively infected hepatocytes produce andrelease both infectious virions (F) and noninfectious virions (E) at a rateof p. IFN inhibits HCV production by a factor of (1-�) and � represents IFNeffectiveness. RBV introduces mutations into the HCV genome andincreases the production of non-infectious virions by a factor of �. Thiswill in turn reduce the infectivity measured at the next round of infectionby a factor of (1-�). The fraction of noninfectious virions induced by RBVis represented by �. Both IFN and RBV are proposed to turn over infectedcells by immune killing (k) through T lymphocytes or natural killer (NK)cells or noncytolytic “curing” (q) through cytokines. Hence, � (loss rate oninfected cells) � k (killing rate) � q (curing rate). Target cells (unin-fected hepatocytes) can be infected productively by the infectious viri-ons, but not by noninfectious virions. Target cells can also originate frominfected cells “cured” by IFN and RBV. Inhibition is represented by aperpendicular sign (�), whereas an arrow with a plus sign indicatesenhancement.

1400 HEPATOLOGY ELSEWHERE HEPATOLOGY, June 2005

characteristic “memory” effect that can prevent residualviruses from coming back after cessation of therapy. Byfactoring in a � value, the authors successfully predictedlong-term response (ETR and SVR) as a function ofIFN-� effectiveness (�) and RBV effectiveness (�). When� was set at 0.5, the model best predicted both ETR andSVR in a range of IFN-� effectiveness (0.4-0.99) withmaximal enhancement by RBV at approximately 25% to30%, consistent with clinical observations. In contrast tothe report by Herrmann et al., the loss rate of infected cells(�) in this report was not affected by RBV, suggesting thatboth immune killing (k, mediated by T cells and NKcells) and curing (q, through noncytolytic removal of in-tracellular viruses) (Fig. 1) were derived mainly fromhigh-dose IFN-�. The effect of RBV on long-term re-sponses predicted by this model derived solely from RBVmutagenesis.16

How does one reconcile this apparently conflictingconclusion by Dixit et al. with that reported by Herrmannet al.? Previously, RBV was seen to have an apparent effecton viral kinetics by some but not by others, which led to adifferent conclusion regarding the mechanism of action.A closer look at this suggests that the difference may be afunction of IFN-� effectiveness. For those studies usinghigh-dose daily IFN-�, the IFN-� effectiveness is gener-ally high (� �0.9); therefore, the immunomodulatory ef-fect of RBV becomes insignificant or undetectable. Thiswas true when Dixit et al. modeled high-dose IFN-� (� �0.95) with RBV (assuming maximum RBV effectiveness�max � 1) or without RBV (� � 0): the difference inantiviral effect (log10 reduction) was only about 0.1. Forthose studies using standard or approved IFN-� doses(IFN-� three times a week or pegylated IFN-� once aweek), the IFN-� effectiveness was lower (�0.9, withmean values ranging from 0.36 to 0.67).14 Therefore, theimmunomodulatory effect of RBV became apparent,mainly reflected by an increase in the net loss rate ofinfected cells. Using the model by Dixit et al., RBV couldhave a significant impact on viral decay, contributing toan additional 2 log10 reduction in viral titer at the end of2 months (assuming � � 0.5, median � � 0.14 day�1).16

The observed immunomodulatory activity made it moredifficult for Herrmann et al. to draw a definitive conclu-sion on the mechanism of action between immunomodu-lation and mutagenesis.14

The state of viral kinetic modeling provides a basic yetincomplete understanding of the underlying processes,but nevertheless allows one to grasp trends and deviseupper and lower bounds on virus kinetic parameters.17

Assumptions were made to simplify curve-fitting, butthey may not be correct. Furthermore, viral dynamicswere built on patients who responded to therapy but not

on those who failed to respond. There is more to learnabout the dynamic nature of HCV replication in humans.We still do not understand the mechanism of alanineaminotransferase normalization induced by RBV, whichhinders the interpretation of the observed immune re-sponse. Studies on mechanisms of IFN-� action at themolecular level may also shed light on the interpretationof antiviral kinetics. For example, is ADAR1 induced dur-ing IFN-� therapy? Can ADAR1 be mutagenic and in-troduce mutations into HCV genomes? A � value of 0.5was assumed to predict the ETR and SVR of IFN-� andRBV combination therapy.16 However, at the same RBVdose, the � value was much lower to fit the slight viral titerdecline observed during RBV monotherapy.10 This gapcan only be explained if synergy exists between IFN-� andRBV. Is ADAR1 the missing link for synergy proposedbetween IFN-� and RBV?

Future modeling should also consider the pharmaco-dynamic properties of IFN-� and RBV. Early viral kinet-ics predicted early viral responses well, but its predictivevalue for long-term response has been somewhat unreli-able.3,18-20 Modeling at later time points (during weeks 6to 24) is lacking. Late viral kinetics may be more useful topredict treatment success or to optimize therapy, al-though it may be complicated by patients who achieveearly viral responses with undetectable viral titers at latertime points. Nevertheless, simultaneous modeling of viralkinetics and drug levels as described by Powers et al.21 mayshed light on the temporal effect of dose reduction or doseincrease toward treatment outcomes. Clearly there is aneed to refine the modeling under real-world treatmentsettings, where dose reduction may be needed to avoidintolerable side effects or dose increase may be required toimprove treatment response. Interestingly, a recent studyhas suggested that SVR might be further improved whenRBV doses were increased up to 4,000 mg daily.22 Can �do a better job? The model of Dixit et al. predicts it will.

ZHI HONG, PH.D.Valeant Pharmaceutical InternationalCosta Mesa, CA

References1. Lindsay KL. Introduction to therapy of hepatitis C. HEPATOLOGY 2002;

36:S114-S120.2. Di Bisceglie AM, Hoofnagle JH. Optimal therapy of hepatitis C. HEPA-

TOLOGY 2002;36:S121-S127.3. Neumann AU, Lam NP, Dahari H, Gretch DR, Wiley TE, Layden TJ, et

al. Hepatitis C viral dynamics in vivo and the antiviral efficacy of inter-feron-alpha therapy. Science 1998;282:103-107.

4. Blight KJ, Kolykhalov AA, Rice CM. Efficient initiation of HCV RNAreplication in cell culture. Science 2000;290:1972-1974.

5. Blight KJ, McKeating JA, Rice CM. Highly permissive cell lines for sub-genomic and genomic hepatitis C virus RNA replication. J Virol 2002;76:13001-13014.

HEPATOLOGY, Vol. 41, No. 6, 2005 HEPATOLOGY ELSEWHERE 1401

6. Samuel CE. Antiviral actions of interferons. Clin Microbiol Rev 2001;14:778-809, table of contents.

7. Hong Z. The role of ribavirin-induced mutagenesis in HCV therapy: aconcept or a fact? HEPATOLOGY 2003;38:807-810.

8. Scott LJ, Perry CM. Interferon-alpha-2b plus ribavirin: a review of its usein the management of chronic hepatitis C. Drugs 2002;62:507-556.

9. McHutchison JG, Gordon SC, Schiff ER, Shiffman ML, Lee WM, RustgiVK, et al. Interferon alfa-2b alone or in combination with ribavirin asinitial treatment for chronic hepatitis C. Hepatitis Interventional TherapyGroup. N Engl J Med 1998;339:1485-1492.

10. Pawlotsky JM, Dahari H, Neumann AU, Hezode C, Germanidis G, Lon-jon I, et al. Antiviral action of ribavirin in chronic hepatitis C. Gastroen-terology 2004;126:703-714.

11. Lau JY, Tam RC, Liang TJ, Hong Z. Mechanism of action of ribavirin inthe combination treatment of chronic HCV infection. HEPATOLOGY 2002;35:1002-1009.

12. Hong Z, Cameron CE. Pleiotropic mechanisms of ribavirin antiviral ac-tivities. Prog Drug Res 2002;59:41-69.

13. Crotty S, Maag D, Arnold JJ, Zhong W, Lau JY, Hong Z, et al. Thebroad-spectrum antiviral ribonucleoside ribavirin is an RNA virus muta-gen. Nat Med 2000;6:1375-1379.

14. Herrmann E, Lee JH, Marinos G, Modi M, Zeuzem S. Effect of ribavirinon hepatitis C viral kinetics in patients treated with pegylated interferon.HEPATOLOGY 2003;37:1351-1358.

15. Ribeiro RM, Layden-Almer J, Powers KA, Layden TJ, Perelson AS. Dy-namics of alanine aminotransferase during hepatitis C virus treatment.HEPATOLOGY 2003;38:509-517.

16. Dixit NM, Layden-Almer JE, Layden TJ, Perelson AS. Modelling howribavirin improves interferon response rates in hepatitis C virus infection.Nature 2004;432:922-924.

17. Herrmann E, Neumann AU, Schmidt JM, Zeuzem S. Hepatitis C viruskinetics. Antivir Ther 2000;5:85-90.

18. Layden TJ, Layden JE, Reddy KR, Levy-Drummer RS, Poulakos J, Neu-mann AU. Induction therapy with consensus interferon (CIFN) does notimprove sustained virologic response in chronic hepatitis C. J Viral Hepat2002;9:334-339.

19. Jessner W, Stauber R, Hackl F, Datz C, Watkins-Riedel T, Hofer H, et al.Early viral kinetics on treatment with pegylated interferon-alpha-2a inchronic hepatitis C virus genotype 1 infection. J Viral Hepat 2003;10:37-42.

20. Layden-Almer JE, Ribeiro RM, Wiley T, Perelson AS, Layden TJ. Viraldynamics and response differences in HCV-infected African American andwhite patients treated with IFN and ribavirin. HEPATOLOGY 2003;37:1343-1350.

21. Powers KA, Dixit NM, Ribeiro RM, Golia P, Talal AH, Perelson AS.Modeling viral and drug kinetics: hepatitis C virus treatment with pegy-lated interferon alfa-2b. Semin Liver Dis 2003;23(Suppl 1):13-18.

22. Lindahl K, Stahle L, Bruchfeld A, Schvarcz R. High-dose ribavirin incombination with standard dose peginterferon for treatment of patientswith chronic hepatitis C. HEPATOLOGY 2005;41:275-279.

Copyright © 2005 by the American Association for the Study of Liver Diseases.Published online in Wiley InterScience (www.interscience.wiley.com).DOI 10.1002/hep.20730Potential conflict of interest: Nothing to report.

New Perspectives in the Treatment ofHBeAg-Positive and HBeAg-NegativeChronic Hepatitis B

Marcellin P, Lau GK, Bonino F, Farci P, Hadziyannis S,Jin R, et al; Peginterferon Alfa-2a HBeAg-Negative ChronicHepatitis B Study Group. Peginterferon alfa-2a alone,lamivudine alone, and the two in combination in patientswith HBeAg-negative chronic hepatitis B. N Engl J Med

2004;351:1206-1217. (Copyright © 2004 MassachusettsMedical Society. All rights reserved.)

AbstractBACKGROUND: Available treatments for hepatitis B e anti-

gen (HBeAg)-negative chronic hepatitis B are associated with poorsustained responses. As a result, nucleoside and nucleotide ana-logues are typically continued indefinitely, a strategy associatedwith the risk of resistance and unknown long-term safety implica-tions. METHODS: We compared the efficacy and safety of pegin-terferon alfa-2a (180 microg once weekly) plus placebo,peginterferon alfa-2a plus lamivudine (100 mg daily), and lami-vudine alone in 177, 179, and 181 patients with HBeAg-negativechronic hepatitis B, respectively. Patients were treated for 48weeks and followed for an additional 24 weeks. RESULTS: After24 weeks of follow-up, the percentage of patients with normaliza-tion of alanine aminotransferase levels or hepatitis B virus (HBV)DNA levels below 20,000 copies per milliliter was significantlyhigher with peginterferon alfa-2a monotherapy (59 percent and43 percent, respectively) and peginterferon alfa-2a plus lamivu-dine (60 percent and 44 percent) than with lamivudine mono-therapy (44 percent, P�0.004 and P�0.003, respectively; and 29percent, P�0.007 and P�0.003, respectively). Rates of sustainedsuppression of HBV DNA to below 400 copies per milliliter were19 percent with peginterferon alfa-2a monotherapy, 20 percentwith combination therapy, and 7 percent with lamivudine alone(P<0.001 for both comparisons with lamivudine alone). Loss ofhepatitis B surface antigen occurred in 12 patients in the pegin-terferon groups, as compared with 0 patients in the group givenlamivudine alone. Adverse events, including pyrexia, fatigue, my-algia, and headache, were less frequent with lamivudine mono-therapy than with peginterferon alfa-2a monotherapy orcombination therapy. CONCLUSIONS: Patients with HBeAg-negative chronic hepatitis B had significantly higher rates of re-sponse, sustained for 24 weeks after the cessation of therapy, withpeginterferon alfa-2a than with lamivudine. The addition of lami-vudine to peginterferon alfa-2a did not improve post-therapy re-sponse rates.

Janssen HL, van Zonneveld M, Senturk H, Zeuzem S,Akarca US, Cakaloglu Y, et al.; HBV 99-01 Study Group;Rotterdam Foundation for Liver Research. Pegylated inter-feron alfa-2b alone or in combination with lamivudine forHBeAg-positive chronic hepatitis B: a randomised trial.Lancet 2005;365:123-129. (Reprinted with permissionfrom Elsevier.)

AbstractBACKGROUND: Treatment of HBeAg-positive patients with

chronic hepatitis B is not effective in most. A combination ofimmunomodulatory pegylated interferon alfa-2b and antivirallamivudine might improve the rate of sustained response. METH-ODS: 307 HBeAg-positive patients with chronic hepatitis B wereassigned combination therapy (100 microg/week pegylated inter-feron alfa-2b and 100 mg/day lamivudine) or monotherapy (100microg/week pegylated interferon alfa-2b and placebo) for 52weeks. During weeks 32-52 the pegylated interferon dose was 50microg/week in both treatment groups. The analyses were based

1402 HEPATOLOGY ELSEWHERE HEPATOLOGY, June 2005