WO2011156545A1 - Viral dynamic model for hcv combination therapy - Google Patents

Abstract

A model of HCV dynamics to quantify responses in subjects dosed with a combination of PR and a protease inhibitor is provided. The model quantitatively predicts clinical outcomes of alternative durations of the protease inhibitor and PR, and these predictions correspond well to the empirically observed outcomes in clinical trials. This model accounts for multiple HCV variants with reduced replicative fitness and susceptibility to protease inhibitors, and examines how these properties of the variants contribute to the HCV dynamics in subjects dosed with the triple combination therapy. Furthermore, the model provides insight into the mechanisms of how protease inhibitor-based regimens increase the SVR rate and reduce total duration of treatment, and predicts that the protease inhibitor-mediated reduction in the overall fitness of the HCV quasispecies, along with potential synergy of protease inhibitors and PR, may enable previous PR nonresponders to achieve SVR on T/PR treatment.

Description

VIRAL DYNAMIC MODEL FOR HCV COMBINATION THERAPY

CROSS-REFERENCE

The present application claims priority to U.S. Application No. 61/353,068 filed on June 9, 2010, the contents of which are incorporated herein by reference in their entireties.

BACKGROUND OF THE INVENTION Approximately 170 million people worldwide are infected with hepatitis C virus (HCV). Current standard treatment for HCV genotype 1 (the most common genotype) is peginterferon-alfa and ribavirin (PR) for 48 weeks, which achieves a sustained viral response (SVR) rate of about 40% to 50%.

As a consequence of its high replicative rate and its error-prone polymerase, HCV exists as a quasispecies. Consistent with this, variants with reduced susceptibility to agents such as HCV protease inhibitors are predicted to exist prior to dosing and have been detected. Upon dosing, the composition of the viral population is altered. We developed a model to predict the viral dynamics during dosing with a protease inhibitor and PR treatment, and used this model to predict the clinical outcomes obtained with varying durations of treatment.

Models of HCV dynamics in response to PR have been used to optimize therapy. These models have assumed a homogeneous viral population in response to the nonspecifically-targeted agents, PR. The present invention provides a multi-variant model to quantify responses to dosing with a triple combination of a protease inhibitor, and PR, and is used to optimize therapy. The model assumes a synergy between the protease inhibitor and PR. Model predictions were consistent with clinical study outcomes involving dosing with a triple combination of telaprevir and PR (T/PR).

SUMMARY OF THE INVENTION

A model that quantifies viral dynamics in regimens of a protease inhibitor

co-administered with PR is provided. Model predictions generally were consistent with data observed in completed Phase 2 clinical studies. We used the model to simulate how varying protease inhibitor treatment durations, in the presence of varying PR treatment durations, affect HCV elimination in subjects with chronic HCV genotype 1 infection. Before the start of treatment with direct-acting antiviral compounds, including a protease inhibitor, the HCV population must be considered functionally to be composed of a mixed population, consisting predominantly of wild-type HCV (WT) and a small population of HCV variants with varying levels of resistance to the a protease inhibitor. The resistant variants exist at a low level prior to the start of treatment because they are less fit (have lower replicative capacity) than WT. The resistant variants retain sensitivities to antiviral inhibition by peginterferon and ribavirin (PR) in vitro and in subjects.

In one embodiment, the invention provides a method of modeling treatment of an HCV patient with a protease inhibitor, peginterferon and ribavirin, comprising the step of: quantifying the patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin with a viral dynamic model using at least one of Equations 1-17.

In some embodiments, the patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin is quantified with a viral dynamic model using all of Equations 1-17.

In one specific embodiment, the patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin is quantified with a viral dynamic model using at least Equations 5A, 6A and 7A.

In some embodiments, the patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin is quantified with a viral dynamic model using all of Equations 1-7. In some embodiments, the patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin is quantified with a viral dynamic model using all of Equations 109. In some embodiments, the method further involves quantifying the patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin with a viral dynamic model using Equation 17. In some embodiments, the method further involves quantifying the patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin with a viral dynamic model using at least one of Equations 8- 16. In some embodiments, the method further involves quantifying the patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin with a viral dynamic model using at least one of Equations 10-16.

In some embodiments, the quantified patient's response is at least one value selected from the group consisting of a breakthrough rate, a relapse rate and a sustained viral response (SVR) rate. In some embodiments, the dosing regimens include a treatment duration for each of the protease inhibitor, peginterferon and ribavirin.

In some embodiments, the method further involves the step of comparing the quantified SVR rate with an intent-to-treat SVR rate.

In some embodiments, the viral dynamic model includes parameters for genotype 1. In some embodiments, the viral dynamic model includes parameters for genotype la or l b.

In some embodiments, the peginterferon is peginterferon-alfa. In some embodiments, the peginterferon- alfa is peginterferon-alfa 2a. In some embodiments, the peginterferon-alfa is peginterferon-alfa 2b. In some embodiments, the protease inhibitor is an inhibitor of hepatitis C proteases NS2-NS3. In some embodiments, the protease inhibitor is an NS3/4A protease inhibitor. In some embodiments, the protease inhibitor is telaprevir. In some embodiments, 750 mg of telaprevir is administered three times a day. In other embodiments, 125 mg of telaprevir is administered twice a day. In some embodiments, the patient is a treatment nai^'ve patient. In some embodiments, the patient is a PR treatment failure patient.

In one embodiment, the invention provides a method of adjusting the dosing level of a composition comprising a protease inhibitor, peginterferon-alfa and ribavirin administered to a patient, the method comprising: measuring plasma HCV RNA levels from a patient; utilizing the measured HCV RNA levels in a multi-variant kinetic model using at least one of Equations 1-17 to calculate the responsiveness of the patient to the administered composition comprising the protease inhibitor, peginterferon-alfa and ribavirin; comparing the calculated responsiveness to a predetermined responsiveness to compositions comprising the protease inhibitor, peginterferon-alfa and ribavirin; and adjusting the dosing level.

In some embodiments, the measured HCV RNA levels are utilized in a multi-variant kinetic model using all of Equations 1 -17. In some embodiments, the measured HCV RNA levels are utilized in a multi -variant kinetic model using all of Equations 1-9. In some embodiments, the method further comprises utilizing the measured HCV RNA levels in a multi-variant kinetic model using Equation 17. In some embodiments, the method further comprises utilizing the measured HCV RNA levels in a multi-variant kinetic model using at least one of Equations 8-16. In some embodiments, the method further comprises utilizing the measured HCV RNA levels in a multi-variant kinetic model using at least one of Equations 10-16.

In some embodiments, the method further involves adjusting the dosing level of the composition comprising a protease inhibitor administered to a patient based upon the comparison of the calculated responsiveness to the predetermined responsiveness.

In some embodiments, the multi-variant kinetic model accounts for one or more of HCV genotype 1 resistant variants. In some embodiments, the HCV genotype 1 resistant variant contains a mutation at one or more of an amino acid position selected from position 155, 54, 36, 156 and 155. In some embodiments, the one or more of HCV genotype 1 resistant variant is selected from R155M, T54A, T54S, V36M, R155K, V36A, A156S, R155T, V36M/R155K, A156T, A156V, and V36M/T54S. In some embodiments, the measured HCV RNA levels are utilized in a multi-variant kinetic model to calculate the responsiveness of the patient to the administered composition comprising a protease inhibitor, peginterferon-alfa and ribavirin includes determining the fitness.

In some embodiments, the plasma HCV RNA levels from a patient are measured within the first 20 days of administration. In some embodiments, the measured HCV RNA levels are utilized in the multi- variant kinetic model to calculate the initial responsiveness of the patient to the administered composition comprising a protease inhibitor, peginterferon-alfa and ribavirin. In some embodiments, the initial responsiveness is compared to a predetermined responsiveness and based upon that comparison calculating a concentration of a protease inhibitor to be subsequently administered.

In one embodiment the invention provides a computer system for modeling treatment of an HCV patient with a protease inhibitor, peginterferon and ribavirin, comprising a computer-readable medium storing a computer program for quantifying a patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin with a viral dynamic model using at least one of Equations 1-17 to provide quantified patient's response to the dosing regimens.

In some embodiments, the patient's response is quantified using all of Equations 1-17. In some embodiments, the patient's response is quantified using all of Equations 1-7. In some embodiments, the patient's response is quantified using all of Equations 1-9. In some embodiments, the patient's response is quantified using at least Equations 5(A), 6(A) and 7(A). In some embodiments, the patient's response is quantified using Equation 17. In some embodiments, the patient's response is quantified using all of Equations 8-16. In some embodiments, the patient's response is quantified using all of Equations 10-16.

In some embodiments, the quantified patient's response is at least one value selected from the group consisting of a breakthrough rate, a relapse rate and a sustained viral response (SVR) rate. In some embodiments, the dosing regimens include a treatment duration for each of a protease inhibitor, peginterferon and ribavirin.

In some embodiments, the computer system further involves the step of comparing the quantified SVR rate with an intent-to-treat SVR rate. In some embodiments, the viral dynamic model includes parameters for genotype 1. In some embodiments, the viral dynamic model includes parameters for genotype la or lb.

In some embodiments, the protease inhibitor is an inhibitor of hepatitis C proteases NS2-NS3. In some embodiments, the protease inhibitor is an NS3/4A protease inhibitor. In some embodiments, the protease inhibitor is telaprevir. In some embodiments, 750 mg of telaprevir is administered three times a day. In other embodiments, 125 mg of telaprevir is administered twice a day.

In some embodiments, the interferon is interferon-alfa. In some embodiments, the interferon- alfa is interferon-alfa 2a. In some embodiments, the interferon- alfa is interferon-alfa 2b.

In some embodiments, the patient is a treatment naive patient. In some embodiments, the patient is a PR treatment failure patient.

In one embodiment, the HCV protease inhibitor is telaprevir and equations 1 -4, 5A, 6 A, and 7 A are employed. In a specific embodiment, equations 1 -4, 5 A, 6 A, 7A, 8 and 9 are employed. In yet another specific embodiment, equations 1 -4, 5 A, 6A, 7 A, 8 and 9-17 are employed.

Subjects with genotype 1 HCV infection have variable responses to PR treatment. At one end of the spectrum, subjects with good responsiveness to PR treatment (e.g., subjects with rapid viral response, defined as undetectable HCV RNA at Week 4 of treatment) may achieve a sustained viral response after 24 weeks of PR treatment. At the other end of the spectrum, subjects with a poor responsiveness to PR treatment (e.g., null responders, defined as having less than a 2-logio decline in HCV RNA at Week 12) are not likely to achieve an SVR even after 48 weeks of PR treatment. The duration of treatment of a given regimen required to achieve an SVR is based on the time required to eradicate all HCV. For PR treatment regimens, models of viral dynamics have been successful in predicting SVR rates by predicting the percentage of patients whose on-treatment HCV RNA levels reach the viral eradication limit. Similarly, the duration of treatment with a regimen of telaprevir in combination with PR (T/PR regimen) required to achieve an SVR is based on the time required to eradicate both WT and variants with different degrees of resistance to telaprevir.

Analysis of viral populations derived from subjects who stopped treatment prior to viral eradication have led to the conclusion that to achieve SVR, a successful regimen will have (1) a telaprevir treatment duration sufficient to eradicate WT and variants with a low level of resistance to telaprevir (LV), and (2) a PR treatment-duration sufficient to eradicate variants with a high level of resistance to telaprevir (HV). Once WT and LV have been eradicated, and HV variants are the dominant residual viral population, telaprevir adds no additional effect. The PR duration required to eradicate HV depends greatly on the PR responsiveness of a given subject and likely the number of residual HV. Therefore, subjects with lower PR responsiveness may require longer durations of PR treatment. Because HV pre-exist at a lower prevalence than WT and have reduced fitness, a greater percentage of subjects can be treated with a shorter duration of PR treatment in the T/PR regimen than in the PR regimen. A schematic of the viral dynamics of PR and T/PR regimens is illustrated in Figure 1.

To analyze and describe HCV dynamics in response to telaprevir-based treatment, a viral dynamic model was developed that incorporates the presence of viral variants of differing telaprevir resistance profiles and fitness, and the variability in subject responses to PR treatment. The objectives of the model previously developed were to (1) represent HCV viral dynamics in subjects dosed with telaprevir monotherapy, (2) estimate the fitness of variants resistant to telaprevir, and (3) investigate the importance of replication space dynamics, mutations during treatment, and preexisting variants on the overall response. The model produced a good fit to the observed HCV RNA levels both during and after telaprevir dosing. WT-infected cells were eliminated more rapidly in the presence of telaprevir than in peginterferon-based regimens. Variants were estimated to have reduced replicative fitness compared to WT, with LV having higher fitness than HV. These variants are likely to pre-exist before the start of treatment, but at a lower prevalence than WT because of their reduced fitness.

The viral dynamic model was used to evaluate the effect of various durations of telaprevir and PR on treatment outcomes. These analyses were first conducted to design the Phase 3 studies of telaprevir in treatment-naive subjects, with the model developed using a combination of data from Study C and from the control arms of Study A and Study B.

Modeling analyses that incorporate data from four completed Phase 2 studies of telaprevir and examine the robustness of the conclusions from previous modelinj analyses are provided.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 depicts a schematic of the relationship between treatment durations and viral eradication in a quasi-species consisting of wild-type HCV and resistant variants. SVR = 44%; null responders = 12%; partial responders = 16%; prior relapsers = 18%.

Figure 2 A. depicts model verification: comparison between observed and predicted SVR rates in Studies A, B, C and D. Assumptions: The predictions of Study D by prior PR responses (all PR-treatment failures, relapsers, nonresponders) were calculated from a subset of simulated treatment-na^'fve subjects who have a HCV RNA response of PR treatment that is in agreement with the following definitions: all PR- treatment failures, not reaching viral eradication by 48 weeks of PR treatment; relapsers, undetectable HCV RNA at the end of treatment but not reaching viral eradication by 48 weeks of PR treatment; nonresponders: never had undetectable HCV RNA during 48 weeks of PR treatment. Peginterferon alia concentrations were simulated from re-sampled estimates of Studies A, B, and D. Ribavirin concentrations were simulated from published estimates (because the ribavirin concentrations collected in these studies were too sparse to allow good estimates of the pharmacokinetic parameters). Ribavirin dose modifications were assumed to be the average observed doses at intervals of dosing time from Studies A and B, which included dose interruptions, dose modification, and ribavirin dose stoppage. Telaprevir concentrations in Study D were simulated from the model developed from Studies A and B by re-sampling because the model predicted the PK of Study D subjects well. Telaprevir concentrations in Study C were simulated from estimates of Study C because the observed concentrations are higher than those in Studies A and B. Predictions for the 2 telaprevir dosing schedules of Study C assumed perfect compliance. B. depicts model verification: comparison between observed and predicted SVR rates in Studies PROVE 1 , PROVE 2, PROVE 3, C, K, L and M.

Figure 3 A. depicts predicted SVR rates by telaprevir treatment duration in treatment naive and PR treatment-failure subjects. B. depicts predicted SVR rates by PR treatment duration in treatment naive and PR treatment-failure subjects. Figure 4 A. depicts predicted treatment outcomes in simulated treatment-naive subjects who completed treatment with 4, 8, 12 or 24 weeks of telaprevir in combination with 24 weeks of PR. B. depicts observed treatment outcomes in treatment-naive subjects who completed treatment with 0, 8, or 12 weeks of telaprevir in combination with 24 weeks of PR. Figure 5 depicts simulated HCV RNA dynamics from baseline through week 12 in a typical treatment-naive subject and typical prior PR treatment failure subject treated with telaprevir, in combination with PR.

Figure 6 depicts predicted SVR rates in simulated treatment-na^'ive subjects who completed regimens with 12, 24, and 48 weeks of PR treatment, in combination with 12 weeks of telaprevir treatment.

Figure 7 depicts predicted SVR rates in simulated treatment-na^'ive subjects when subjects with eRVR were treated with 24 weeks or 48 weeks of PR. A PR treatment duration of 24 weeks was used only for subjects who had undetectable HCV RNA at Weeks 4 and 12 of treatment (eRVR). A PR treatment duration of 48 weeks was used for subjects who did not have an eRVR (values in left column of each panel) or 48 weeks for all subjects (values in right column of each panel). Assumptions: No additional discontinuations during PR treatment between Weeks 24 and 48.

Figure 8 depicts predicted SVR rates in regimens with and without a 4-week delayed start of telaprevir in simulated subjects completing treatment. On the x-axis, N= regimen without a delayed start of telaprevir treatment; Y= regimen with a 4-week delayed start of telaprevir treatment. Assumptions: Same total PR duration for both regimens; regimens with delayed start begin telaprevir dosing after 4 weeks of PR treatment. Figure 9 depicts HCV RNA dynamics in simulated subjects with representative PR responses who failed T12/PR24 treatment. The column on the right is a closer view of the HCV RNA dynamics of the panels in the column on the left. PR responses in the simulated subjects assumed median values of the responses in simulated PR null responders, PR nonresponders, and PR48 treatment failures. The SVR rates by different PR responses are given in Figure 12. Simulated subjects were infected with HCV genotype 1 a.

Figure 10 depicts predicted SVR rates by telaprevir durations and by prior PR response population. Assumptions: Discontinuation rates: see supplementary materials. Concentrations: Telaprevir concentrations were simulated using

parameters re-sampled from estimates of Study A and B. Peg-IFN: from estimates of Studies A, B, C. Treatment adherence was 100% for all 3 drugs; Random effects were controlled in the simulation. PK parameters and the viral responsiveness to telaprevir and to PR was the same across compared durations. The genotype 1 a: 1 b ratio was 1 : 1. The number of subjects was 350 per group for treatment-naive and 120 per group for prior nonresponder. The definitions of prior PR responses in each population are (1) failure, simulated subjects who failed to achieve eradication with PR48 treatment; (2) relapser, simulated subjects with HCV RNA undetectable at the end of PR48 treatment but failed to achieve eradication; (3) nonresponder, simulated subjects with HCV RNA detectable during PR48 treatment. The percentage of subjects who failed to reach SVR for reasons of lost to follow-up, noncompliance, and withdrawal of consent were assumed to be zero.

Figure 1 1 depicts sensitivity analyses of viral dynamics in typical simulated treatment-naive subject to different telaprevir dose schedules. With less frequent telaprevir dosing schedules, only the viral dynamics of wild-type (WT) and low-level resistant variants (LV) changed to oscillate with higher amplitude according to the variations of the simulated telaprevir concentrations. The viral dynamics of variants with high-level resistant variants (HV: A156 and V36M/R155K) did not change with telaprevir dose schedules. Assumptions: Simulated subject was HCV genotype la; median PK parameters; median responses of treatment-naive population treated with T12/PR48. Major HCV variants assumed were wild-type (WT), variant with low-level resistance (R155K), and variants with high-level resistance (A156 and V36M/R155K). Concentrations were simulated using PK parameters obtained from population estimates of Studies A and B. Telaprevir concentration assumed a 1- compartmental PK model.

Figure 12 depicts discontinuation rates assumed in the model when comparing alternative durations of telaprevir treatment.

Figure 13 depicts observed and predicted concentrations of peginterferon-alfa-2a.

Figure 14 depicts observed and predicted concentrations of ribavirin. Red line, y=x. Figure 15 depicts goodness of fit plot of HCV RNA logio decline for subjects on Peg- IFN and RBV treatment.

Figure 16 depicts goodness of fit plot of HCV RNA logio decline for subtype la subjects on telaprevir, Peg-IFN, and RBV treatment.

Figure 17 depicts goodness of fit plot of HCV RNA logio decline for subtype lb subjects on telaprevir, Peg-IFN, and RBV treatment.

Figure 18 A depicts the Visual Predictive Check for subjects treated with Peg-IFN and RBV. Data are obtained from Studies A and B (treatment-naive population), up to the time when dose modification occurred. B depicts the Visual Predictive Check for subtype la subjects treated with telaprevir, Peg-IFN and RBV. C depicts the Visual Predictive Check for subtype lb subjects treated with telaprevir, Peg-IFN, and RBV.

Figure 19 depicts the Predictive Check for SVR rates from Studies A and B.

Figure 20 depicts a schematic of a Model to Represent Evolutionary Dynamics.

Figure 21 depicts the response to PR and to T/PR treatment from the PROVEl study. Figure 22 depicts the resistance and in vivo fitness of variants

Figure 23 depicts the sensitivity to eradication assumption: clinical outcomes for T12PR24 regimen.

] I DETAILED DESCRIPTION OF THE INVENTION

Table 1 : List of Abbreviations

Abbreviation Definition

CI confidence interval

eRVR extended rapid viral response (undetectable HCV RNA at Weeks 4 and 12 of

treatment)

HCV hepatitis C virus

HV HCV variants with high levels of telaprevir resistance

LV HCV variants with low levels of telaprevir resistance

MAA Marketing Authorization Application

NDA New Drug Application

PK pharmacokinetic

PR peginterferon alfa and ribavirin

PR# peginterferon alfa and ribavirin treatment duration of # weeks (e.g., PR24 =

peginterferon and ribavirin treatment duration of 24 weeks)

SVR sustained viral response

TVR telaprevir

T# telaprevir treatment duration of # weeks (e.g., T12 = telaprevir treatment duration of

12 weeks)

T/PR telaprevir, peginterferon alfa, and ribavirin

WT wild-type HCV

Table 2: Definition of Terms

Term Definition

null responder Week 1 2 HCV RNA decline of <2-log,₀

nonresponder subject who never had undetectable HCV RNA during 48 weeks of PR treatment

relapser subject who had undetectable HCV RNA levels at the end of treatment but not reaching viral eradication

treatment-na^'fve subject who has not received prior treatment for HCV infection

PR treatment failure subject who was treated with peginterferon alfa and ribavirin but did not achieve SVR

relapse return of detectable HCV RNA levels after being undetectable at the completion of treatment

SVR, RVR and EVR Generally, the terms "RVR," "SVR," "EVR" are well known in the art and within the meaning well accepted in the art. For example, "RVR" means rapid viral response, "SVR" means sustained viral response, and "EVR" means early viral response. Typically, "RVR" indicates an undetectable HCV RNA level at week 4; "SVR" indicates an undetectable HCV RNA level 48 weeks after the end of treatment; and "EVR" indicates > 2-logl O reduction from baseline in HCV RNA at week 12 or undetectable HCV RNA at week 1 2. As described above, HCV RNA being "undetectable" means that the HCV RNA is present in less than 10 [U/mL as determined by assays currently commercially available, and preferably as determined by the Roche COBAS TaqMan™ HCV/HPS assay.

This invention also relates to specific doses and dosage regimens for administering VX-950. VX-950 is described in PCT Publication Numbers WO 02/018369, WO 2006/050250, and WO 2008/144072, with reference to the following structural formula or a pharmaceutically acceptable salt thereof:

Other descriptions of VX-950 can be found in PCT Publication Numbers WO

07/098270 and WO 08/106151.

The invention also includes prodrugs and solvates of VX-950.

As used herein, the phrase "pharmaceutically acceptable salt(s)" refers to the salts that are safe and effective for treatment of HCV infections. Pharmaceutically acceptable acid addition salts include, but are not limited to, hydrochloride, hydrobromide, hydroiodide, nitrate, sulfate, bisulfate, phosphate, acid phosphate, isonicotinate, acetate, and lactate salts. Pharmaceutically acceptable salts with various amino acids can also be used, and use of these amino acid salts is also within the scope of this invention. Suitable base salts include, but are not limited to, aluminum, calcium, lithium, magnesium, potassium, sodium, zinc, and diethanolamine salts. For a review on pharmaceutically acceptable salts, see Berge et al., J. Pharm. Sci., 66, 1 -19 (1977), the contents of which are incorporated herein by reference.

As used herein, the phrase a "pharmaceutically acceptable prodrug" of VX-950 refers to a compound that may be converted under physiological conditions or by solvolysis to VX-950 or to a pharmaceutically acceptable salt of VX-950 prior to exhibiting its pharmacological effect in the treatment of HCV infections. Typically, the prodrug is formulated with the objectives of improved chemical stability, improved patient acceptance and compliance, improved bioavailability, prolonged duration of action, improved organ selectivity, improved formulation (e.g., increased hydrosolubility), or decreased side effects (e.g., toxicity). A pharmaceutically acceptable prodrug can be readily prepared using methods known in the art, such as those described in Burger's Medicinal Chemistry and Drug Chemistry, Vol. 1 , 172- 178 and 949-982, John Wiley & Sons (1995). See also Bertolini et al, J. Med. Chem., 40, 201 1-2016 (1997); Shan et al., J. Pharm. Sci. 86(7), 765-767 (1997); Bagshawe, Drug Dev. Res., 34, 220-230 (1995); Bodor, Advances in Drug Res., 13, 224-331 (1984); Bundgaard, Design of Prodrugs, Elsevier Press (1985); and Larsen, Design and Application of Prodrugs, Drug Design and Development (Krogsgaard-Larsen et al, eds.), Harwood Academic Publishers (1991).

It will further be appreciated by those skilled in the art that the compounds described herein can exist in different solvate forms, for example hydrates, and yet retain the biological effectiveness. Such solvates may also form when solvent molecules are incorporated into the crystalline lattice structure of the compound molecule during the crystallization process. As used herein, the phrase a "pharmaceutically acceptable solvate" of VX-950 refers to a pharmaceutically acceptable solvate form of VX-950 that contains solvent molecule(s) and retains the biological effectiveness of VX-950.

Compounds that differ from VX-950 only in the presence of one or more isotopically enriched atoms are covered in the invention. For example, compounds having the present structures except for the replacement of hydrogen by deuterium or tritium, or the replacement of a carbon by a 13C- or 14C-enriched carbon are within the scope of this invention. Certain examples of isotopically enriched VX-950 can be found in

WO 2007/109080 and Maltais et al, J. of Medicinal Chemistry, "In Vitro and In Vivo Isotope Effects with Hepatitis C Protease Inhibitors: Enhanced Plasma Exposure of Deuterated Telaprevir versus Telaprevir in Rats" 2009:52(24):7993-8001.

VX-950 may contain one or more asymmetric carbon atoms and thus may occur as racemates and racemic mixtures, single enantiomers, diastereomeric mixtures and individual diastereomers. All such isomeric forms of these compounds are expressly included in the present invention. Each stereogenic carbon may be of the R or S configuration. The D- and L-isomers at the N-propyl side chain of VX-950 are expressly included within the scope of this invention.

It will be appreciated by those skilled in the art that the compounds described herein can exist in different polymorphic forms. As known in the art, polymorphism is an ability of a compound to crystallize as more than one distinct crystalline or "polymorphic" species. A polymorph is a solid crystalline phase of a compound with at least two diiTerent arrangements or polymorphic forms of that compound molecule in the solid state. Polymorphic forms of any given compound are defined by the same chemical formula or composition and are as distinct in chemical structure as crystalline structures of two different chemical compounds.

The amounts of VX-950 according to this invention are administered in a single dosage form or in more than one dosage form. If in separate dosage forms, each dosage form is administered about simultaneously. For the avoidance of doubt, for dosing regimens calling for dosing more than once a day, one or more pill or dose maybe be given at each time per day (e.g., 1 pill, three times per day or 3 pills, three times per day). Most embodiments of this invention will employ at least 2 pills per dose.

Methods of this invention may also involve administration of another component comprising an additional agent selected from an immunomodulatory agent: an antiviral agent; an inhibitor of HCV protease; an inhibitor of another target in the

HCV life cycle (other than NS3/4A protease); an inhibitor of internal ribosome entry, a broad-spectrum viral inhibitor; or a cytochrome P-450 inhibitor; or combinations thereof. The additional agent is also selected from an inhibitor of viral cellular entry.

Accordingly, in another embodiment, this invention provides a method comprising administering VX-950 and another anti-viral agent, preferably an anti-HCV agent.

Such anti-viral agents include, but are not limited to, immunomodulatory agents, such as α-, β-, and γ-interferons or thymosin, pegylated derivatized interferon -a compounds, and tymosin; other anti-viral agents, such as ribavirin, amantadine, and telbivudine; other inhibitors of hepatitis C proteases (NS2-NS3 inhibitors and NS3- NS4 inhibitors); inhibitors of other targets in the HCV lie cycle, including helicase, polymerase, and metalloprotease inhibitors; inhibitors of internal ribosome entry; broad-spectrum viral inhibitors, such as IMPDH inhibitors (e.g., compounds described in U.S. Pat. No. 5,807,876, 6,498,178, 6,344,465 and 6,054,472; and PCT publications WO 97/40028, WO 98/40381 , and WO 00/56331 ; and mycophenolic acid and derivatives thereof, and including, but not limited to, VX-497, VX-148, and VX-944); or any of their combinations. Examples of HCV protease inhibitors include those compounds described in WO 99/07733 (Boehringer Ingelheim), WO 99/07734 (Boehringer Ingelheim), WO 00/09558 (Boehringer Ingelheim), WO 00/09543 (Boehringer Ingelheim), WO 00/59929 (Boehringer Ingelheim), WO 02/060926 (BMS), WO 2006039488 (Vertex), WO 2005077969 (Vertex), WO 2005035525 (Vertex), WO 2005028502 (Vertex) WO 2005007681 (Vertex), WO 2004092162 (Vertex), WO 2004092161 (Vertex), WO 2003035060 (Vertex), of WO 03/087092 (Vertex), WO 02/18369 (Vertex), WO 98/17679 (Vertex), WO 07/025307 (Vertex) or WO 08/106139 (Vertex). Specific examples include VX-950 (WO 07/098270; WO 08/106151), MK-7009, SCH- 503034 (boceprevir, Victrelis™) and BILN-2061 (see, e.g., Liverton et al.,

Antimicrobial Agents and Chemotherapy, 54(1), 305-31 1 (2010)). Additional specific examples include ITMN-191 (R7227) (see, e.g., Seiwert et al., Antimicrobial Agents and Chemotherapy, 52(12), 4432-4441 (2008)). Other specific examples include TMC-435, GS-9451 , BI-201335, TMC-435, ABT-450, GS-9256, IDX-320, MK-5712, and ABT-267.

Other agents (e.g., non-immunomodulatory or immunomodulatory compounds) may be used in combination with a compound of this invention include, but are not limited to, those specified in WO 02/18369, which is incorporated herein by reference (see, e.g., page 273, lines 9-22 and page 274, line 4 to page 276, line 1 1 this disclosure being specifically incorporated herein by reference.

Still other agents include those described in various published U.S. Patent

Applications. These publications provide additional teachings of compounds and methods that could be used in combination with VX-950 in the methods of this invention, particularly for the treatment of hepatitis. It is contemplated that any such methods and compositions may be used in combination with the methods and compositions of the present invention. For brevity, the disclosures from those publications is referred to be reference to the publication number but it should be noted that the disclosure of the compounds in particular is specifically incorporated herein by reference. Examples of such publications include U.S. Patent Application Publication Nos.: US 20040058982, US 20050192212, US 20050080005, US

20050062522, US 20050020503, US 2004022981 8, US 20040229817, US

20040224900, US 20040186125, US 20040171626, US 200400110747, US

20040072788, US 20040067901 , US 20030191067, US 20030187018, US 20030186895, US 20030181363, US 20020147160, US 20040082574, US

20050192212, US 20050187192, US 20050187165, US 20050049220 and US 20050222236.

Still other agents include, but are not limited to Albuferon™ (albumin-Interferon alpha) available from Human Genome Sciences; PEG-INTRON® (peginterferon alfa- 2b, available from Schering Corporation, Kenilworth, NJ); INTRON-A®,

(VIRAFERON®, interferon alfa-2b available from Schering Corporation, Kenilworth, NJ); ribavirin (l-beta-D-ribfuranosyl-lH-l ,2,4-triazole-3-carboxamide, available from ICN Pharmaceuticals, Inc., Costa Mesa, CA; described in the Merck Index, entry 8365, Twelfth Edition); REBETROL® (Schering Corporation, Kenilworth, NJ); COPEGUS® (Hoffmann-La Roche, Nutley, NJ); PEGASYS® (peginterferon alfa-2a available Hoffmann-La Roche, Nutley, NJ); ROFERON® (recombinant interferon alfa-2a available from Hoffmann-La Roche, Nutley, NJ); BEREFOR® (interferon alfa 2 available from Boehringer Ingelheim Pharmaceutical, Inc., Ridgefield, CT);

SUMIFERON® (a purified blend of natural alpha interferons such as Sumiferon available from Sumitomo, Japan); WELLFERON® (interferon alpha nl available from Glaxo Wellcome Ltd., Great Britain); ALFERON® (a mixture of natural alpha interferons made by Interferon Sciences, and available from Purdue Frederic Co., CT); a-interferon; natural alpha interferon 2a; natural alpha interferon 2b; pegylated alpha interferon 2a or 2b; consensus alpha interferon (Amgen, Inc., Newbury Park, CA); REBETRON® (Schering Plough, Interferon-alpha 2B + Ribavirin); pegylated interferon alpha (Reddy, K.R., et al, "Efficacy and Safety of Pegylated (40-kd) Interferon alpha-2a Compared with interferon alpha-2a in Noncirrhotic Patients with Chronic Flepatitis C," Hepatology, 33, 433-438 (2001); consensus interferon

(I FERGEN®)(Kao, J.H., et al., "Efficacy of Consensus Interferon in the Treatment of Chronic Hepatitis," J. Gastroenterol, hepatol. 15, 1418-1423 (2000);

lymphoblastoid or "natural" interferon; interferon tau (Clayette, P., et al, "IFN-tau, A New Inteferon Type I with Antiretroviral activity", Pathol. Biol. (Paris) 47, 553-559 (1999); interleukin-2 (Davis, G.L., et al., "Future Options for the Management of Hepatitis C." Seminars in Liver Disease, 19, 103-1 12 (1999); Interleukin-6 (Davis et al., "Future Options for the Management of Hepatitis C," Seminars in Liver Disease, 19, 103-1 12 (1999); interleukin-12 (Davis, G.L., et al., "Future Options for the Management of Hepatitis C," Seminars in Liver Disease, 19, 103- 1 12 (1999); and compounds that enhance the development of type 1 helper T cell response (Davis et al, "Future Options for the Management of Hepatitis C," Seminars in Liver Disease, 19, 103-1 12 (1999)). Also included are compounds that stimulate the synthesis of interferon in cells (Tazulakhova, E.B., et al., "Russian Experience in Screening, Analysis, and Clinical Application of Novel Interferon Inducers" J. Interferon

Cytokine Res., 21 , 65-73, including, but are not limited to, doubled stranded RNA, alone or in combination with tobramycin, and Imiquimod (3M Pharmaceuticals; Sauder, D.N., "Immunomodulatory and Pharmacologic Properties of Imiquimod," J. Am. Acad. Dermatol., 43, S6-11 (2000). See also, WO 02/18369, particularly page 272, line 15 to page 273, line 8, this disclosure being specifically incorporated herein by reference.

Interferon is not typically administered orally, although orally administered forms are in development. Nevertheless, nothing herein limits the methods or combinations of this invention to any specific dosage forms or regime, thus, each component of a combination according to this invention may be administered separately, together, or in any combination thereof. As recognized by skilled practitioners, dosages of interferon are typically measured in IU (e.g., about 4 million IU to about 12 million IU). Interferon may also be dosed by micrograms. For example, a standard dose of Peg-Intron is 1.0-1.5 μg/kg/wk and of Pegasys is 180 μg/wk.

Ribavirin is typically administered orally, and tablet forms of ribavirin are currently commercially available. General standard, daily dose of ribavirin tables (e.g., about 200 mg tablets administered twice a day) is about 800 mg to about 1200 mg

(according to the dosage regimens described in its commercial product label).

In some aspects, the method includes the administration of agents over two phases, an initial phase and a secondary phase. For instance, the initial phase can be a period of less than about 12 or 24 weeks and the secondary phase can be greater or equal to about 12 weeks, e.g., the secondary phase can be between about 12-36 weeks. In certain embodiments, the secondary phase is 12 weeks. In still other embodiments, the secondary phase is 36 weeks. In certain embodiments, the sum of the initial and secondary phase is about 24 to 48 weeks (such as 24, 36 or 48 weeks). In some embodiments, the initial and secondary phases can be identical in duration. Objectives of Analyses

The objectives of the analyses are to support the duration rationale for the final regimen choices in the telaprevir ND A/ MAA; to assist in the interpretation of Phase 2 clinical study results, specifically as regards to predictions of SVR rates for the telaprevir and PR durations in the treatment regimens used in the Phase 3 clinical studies of telaprevir; to evaluate the model by comparing the observed and predicted SVR rates in Study C and Study D, which were not used in the model development, demonstrating the robustness of the predictions made from an earlier version of the model, after incorporating HCV RNA and drug concentration data available from Studies A and B.

Methods

Modeling analyses were conducted to predict the efficacy of alternative durations of telaprevir treatment and of PR treatment. Model-predicted telaprevir, peginterferon alfa, and ribavirin concentrations (based on the dosing records and population PK estimates of corresponding subjects in selected clinical studies) were entered into the viral dynamic model. Viral dynamic model parameters were estimated using HCV RNA data from T/PR and PR regimens in treatment-naive subjects (Studies A and B). SVR rates were predicted by entering the observed discontinuation rates into the model. Predicted SVR rates for treatment-naive and different categories of PR treatment-failure subjects (nonresponder, and relapser) were generated by simulations using re-sampled parameters from the distributions of individual Bayesian estimates from the population analysis approach.

Equation 1 represents the viral dynamics among HCV variants in subjects dosed with T/PR. Variant V._\. represents a virion with characterized amino-acid substitution(s) in the NS3/4A protease and different sensitivities to telaprevir in vitro. Variant Vi infects target cells (7) to form j-infected cells ( ,.) at rate βΤΥ:,.. Each infected cell (/j.) is assumed to be infected by only one variant, and each variant competes for the same target cells T. Target cells T also represent limited replication "space" shared by all variants. Each infected cell (/;.) produces a population of variants at production rate pf_\J._\., with a OT.j_j-fraction of this production mutating to produce variant j (V. . The production rate ratio ( .) quantifies variant replication advantage (or

disadvantage) in the absence of any drug. Different production rates (pf_\), but the same infection ( ?) and clearance (c) rates, are assumed for different variants. This assumption is consistent with the function of the NS3/4A protease to cleave a precursor polyprotein as a crucial step in the HCV replication cycle. .

Each drug (telaprevir, Peg-IFN, RBV) assumes a dual role in clearing HCV. First, each drug blocks the production rates p by a factor (1 - ε._\). In vitro Replicon data suggest that telaprevir antiviral blockage £ . depends on the identity of variant i; blockage is most effective on WT virus and least effective on variants with high resistance to telaprevir (HV). Blockage by Peg-IFN and RBV are assumed to be equal among variants, consistent with inhibition data in Replicon cells. The blockage factors were calculated as a function of plasma concentrations of each drug (multiplied by a factor K to convert plasma to effective concentrations), and the sensitivities of each valiant as measured in Replicon cells (represented by parameters IC₅o, and hill-power values h). Overall blockage in the T/PR regimen was calculated by assuming additive blockages of each drug e._i:r, £-p_, £-R- The second role of each drug is to enhance the infected cell clearance <S. WT <5.WT- values were up to 10-times higher in subjects dosed with telaprevir than in subjects treated with Peg-IFN/RBV. These observations were incorporated into the model by assuming that δ.\. increased proportionally to the logarithmic of blockage factor (l -e.j). Because the initial estimates of variant fitness assumed a constant background clearance <5._nodrug- obtained in subjects dosed in monotherapy with telaprevir (in the absence of Peg-IFN and RBV), the model of response on T/PR regimens constrained to the same _nodrug- value. The variability of infected-cell clearance on PR treatment was represented by different enhancement factors for Peg-IFN <5._P. and for RBV (¾.. Moreover, the δ.ρ. and <¾. values were assumed to be the same (Equations 1 -1 1) because the majority of the data were obtained from subjects treated simultaneously with Peg-IFN and RBV and therefore the antiviral contribution of RBV is difficult to separate from the antiviral

contribution of Peg-IFN. Equations for Multi-variant HCV RNA dynamic models:

J /(mutation nut*, nucleotide chaiit»¾f if j

Π tj A 7 i(i

EG]}^/i

v'.f ί ^ !

(10)

'/3

(11)

fv.i6A 11567^" = (^""V'''^'"' i subtype^; 11.) (:i6) a = ) +€log ,„ (HCVRNA decline } 17)

Soodvns - &Γ l g₁₀( I -

<*/> l°g|o(^J - ff) - ' i l g₁₀( I - i) (5) A

(6) A

Table 3 : Definition of variables

Data preparation were implemented in SAS (SAS Institute, Inc., Cary, NC).

Population pharmacokinetic estimations were implemented in NONMEM version 6. Simulation and estimation involving viral dynamic model were implemented in Jacobian 4.0 (RES group, Inc., Cambridge, MA). Calculations of SVRITT- were implemented in Perl. Figures were generated with JMP version 8.0.2 (SAS Institute, Inc., Cary, NC).

Peginterferon-alfa-2a concentrations were obtained from pooled data of Studies A, B, and D. Subjects with measured Peg-IFN concentrations with values < BQL in more than 50% of the scheduled measurements were excluded from the peginterferon pharmacokinetic model development. In the viral dynamic simulations, the Peg-IFN concentrations of these subjects were simulated by assigning clearance rates that correspond to a steady-state Peg-IFN concentration of half of BQL values.

For Study A, observed concentrations were limited to data from the first 6 weeks because of unexplained declines in the concentrations between Weeks 6 and 12 in this study. These declines were not consistent with the rest of the data in other studies (Studies B and D). For Studies B and D, all observed concentrations above quantification (BQL) values were included.

One compartmental model is implemented using ADVAN2 in NONMEM 6. The NONMEM input file of the final run is provided. The final model is a one- compartmental model with no time delay. Steps for model selections were summarized in Table 4. Correspondence between data and individual model prediction is provided in Figure 13 (DV vs. IPRED). The final model did not show any evidence of bias. Final parameter estimates are provided in Table 8.

Table 4 Iterations of critical path peginterferon-alfa-2a pharmacokinetic model developments.

Run 6 is chosen as the final model based on the criterion that it has the lowest objective function value and that it has the least number of degree of freedom.

The model structure chosen was a three-compartment population pharmacokinetic model. The model was used to interpolate pharmacokinetics of ribavirin data in Studies A, B in the estimation step. Model was implemented as Empirical Bayesian Feedbacks, with parameters obtained from literature. No model selections were performed. The goodness-of-fit plots are provided in Figure 14. The observed concentrations were higher than the predicted concentrations (by up to 20%) at high concentrations (left panel). The impact of this over-prediction has been examined. No bias was observed as a function of time. The viral dynamic model is given in Equations 1-17. In one embodiment, the protease inhibitor is telaprevir. Complete concentration time profiles of telaprevir, peginterferon, and ribavirin were used as inputs to the viral dynamic model. To reduce computational expense in the estimation, the concentrations were

approximated as time-exponential function to the steady-state level. The

approximation was implemented only in the estimation of viral dynamic model parameters. For the simulation of SVR, the full time course of concentrations of three drugs were used. These approximations have been verified to predict the same SVR as the complete concentrations in 10⁴ simulation runs. Estimation used an Empirical Bayesian feedback approach, approximated by a weighted least-square objective function. The weights were obtained as the inverse of the expected variance, with values obtained from estimations from individual subjects. The initial estimates of parameter distributions were obtained from Study E (telaprevir monotherapy). The residual error estimate of HCV RNA log io- decline was obtained from the estimates of a published biphasic, one-variant viral dynamic model applied to the first 3 -day HCV RNA of Study E (telaprevir monotherapy).

The observation used in the estimation is the HCV RNA log.io decline. The following criteria were used to prepare the data: The data were obtained from the PR and T/PR arms of Studies A and B. The data was limited to HCV RNA up to the time when two or more consecutive HCV RNA levels were lower than LOQ values (30 IU/mL). Additional limitations were implemented: PR arms: up to the first dose modifications, or end of the treatment if the subject did not get dose adjusted. T/PR arms: Data up to the time when HCV RNA < LOQ for 2 consecutive times or end of treatment. Only subjects with WT-dominant quasispecies were included in the estimation. The subjects were limited to those whose pharmacokinetic estimates were available.

The predictability of the model was examined using two levels of verifications:

Internal visual predictive checks on HCV RNA dynamics were performed with the training set. The predictive checks were used to confirm whether the developed viral dynamic model described the time course of HCV RNA profiles reasonably well. The predictive checks were with all the detectable HCV RNA values. To verify the simulation results of long-term outcome (SVR) based on viral dynamics, predicted SVR rates were compared to the observed SVR rates from the studies used in model building and the other studies not used in the model building, separately.

Details of the Modeling Analyses

The predictions of SVR rates assumed discontinuation rates of telaprevir and of PR that are similar to those in the peginterferon-alfa-2a arms of Study C, in which the most updated rash management and study drug adverse event management were implemented. For the prediction of the effect of a regimen with delayed start of telaprevir, SVR rates were reported only for subjects who completed treatment because the discontinuation rate in regimens with a 4-week delayed start of telaprevir has not been characterized (the first study of telaprevir with a delayed-start regimen is ongoing).

The list of HCV variants used in this model was based on the major variants detected in clinical studies. All subjects with the same genotype were assumed to have the same set of major variants. The intermediate-resistant variants R155T/I and other minority variants observed in a few subjects in Studies A and B were not included because of lack of data to estimate their fitness. Including these variants in the model is expected to result in only small changes in the outcome, because these variants appear to be less fit than the variants used in the current model.

The model assumed that treatment-nai^'ve subjects have WT-dominant quasi-species at baseline. This was based on population sequence analysis showing WT-dominant quasi-species at baseline in at least 98% of the treatment-nai^'ve population. The SVR rates in the 2% of the treatment-nai^'ve population with variant-dominant quasi-species were ignored in the model prediction. Covariates of viral dynamic responses examined here were limited to viral genotypes la and lb. Genotypes 1 a and lb were modeled separately because when telaprevir was administered in monotherapy, different sets of resistant variants emerged. Other covariates have not been examined. The current model assumed a diagonal structure of the variance-covariance matrix (zero correlations between parameters). Despite this approximation, the resulting simulations corresponded well to the observed SVR rates, suggesting that the contribution of the non-diagonal component of the variance-covariance matrix to the SVR is negligible. The variability of the predictions was based both on the individual variability in responsiveness and on the sample size. The uncertainty of the parameter estimates have not been accounted for in the predictions. Given that the current model predictions were comparable to the observed SVR rates, the contribution of the uncertainty of the parameter estimates may be negligible compared to the variability in responsiveness.

The prediction of the reversion of HCV quasi-species after the end of treatment assumed an off-treatment infected-cell clearance rate that was based on the average rate observed in peginterferon-based treatment. This assumed off-treatment rate may be much more rapid than the actual rate; however, no direct data are available to estimate the off-treatment rate. Because the reversion rate of HCV quasi-species depends on the rate of infected-cell turnover, the model-predicted reversion of the HCV quasi-species (post telaprevir-based regimens) was faster than the observed rate.

Results

The model was used to predict SVR rates for 4, 8, 12, and 24 weeks of telaprevir treatment in treatment-na^'ive and PR treatment-failure populations. To provide additional insights into the sensitivity of the predicted SVR rates to telaprevir treatment durations, HCV dynamics in selected simulated subjects treated with 4 or 12 weeks of telaprevir treatment in combination with PR are provided. Moreover, the model was used to predict SVR rates for alternative telaprevir-based regimens consisting of PR treatment durations of 12, 24, and 48 weeks, and to predict SVR rates with and without delayed start of telaprevir. Finally, to assist the comparison of the predicted and observed clinical outcomes, the evolutionary dynamics of the HCV quasi-species in selected simulated subjects who failed a T12/PR24 regimen are provided. The mode] was verified with efficacy results (SVR) by comparing the predictions and the observed data from telaprevir studies for which final data were available.

Studies A and B: The treatment-na^'ive population whose on-treatment HCV RNA dynamics data were above the lower limit of quantification was used in the model development; this study included 1 duration of telaprevir treatment (12 weeks) and 3 durations of PR treatment (12, 24, and 48 weeks).

Study D: PR treatment-failure population, including subjects with different prior responses to PR (nonresponder and relapser); this study included 2 durations of telaprevir treatment (12 and 24 weeks) and 2 durations of PR treatment (24 and 48 weeks).

Study C: treatment-na^'ive population; this study included 1 duration of telaprevir treatment (12 weeks); 2 dose schedules of telaprevir (750 mg every 8 hours or 1125 mg every 12 hours); and 2 durations of PR treatment (24 or 48 weeks, depending on the on-treatment HCV RNA response). This study also included 2 different types of peginterferon alfa.

Table 5: Data sources for model development.

Table 6: Observed outcomes in T/PR regimes for studies PROVE 1 , PROVE 2 and Study C.

Figure 2 shows the correspondence between observed and predicted SVR rates for Studies A and B (used to train the model) and Studies C and D (used to verify the model). Predicted SVR rates were generated by inserting into the model the actual number of subjects treated, the number of subjects who prematurely discontinued treatment, the number of subjects who failed to reach SVR because of other reasons (lost to follow-up, noncompliant, and withdrawal of consent), the timing of treatment discontinuations, and the distribution of genotype (la and lb) for each

regimen/subject population.

The predicted SVR values are comparable to the observed values. The majority of the observed SVR values (12/14 treatment groups) are within the 90% confidence intervals (CIs) of the predicted values. In the two groups where the observed SVR values lie outside the predicted 90% CI bounds (Study C T12 ql2h/2a arm and Study D T12/PR24 prior relapsers), one was inside the 95% CIs, and the other was 1.7% lower than the nearest 98% CIs. When the results of Study D were broken down by prior PR response (nonresponders, relapsers), the predictions of SVR rates were comparable to the observed values.

Despite being trained only for the treatment-na^'ive population (using data from Studies A and B), the model produced consistently predictive results for different subject populations— treatment-na^'ive (Study C) and prior PR nonresponders and prior PR relapsers (Study D). This finding supports a hypothesis that a treatment-naive population contains several types of subjects that subsequently can be categorized by their prior PR responses (nonresponders, relapsers, SVR), and suggesting that a viral dynamic model estimated from the treatment-naive population can be used to predict results for the populations with different prior PR responses.

After the predictability of the model was established, the model was used to predict SVR rates for treatment-naive and treatment-failure populations treated with telaprevir 750 mg q8h for 4, 8, 12, and 24 weeks, in combination with PR for a total treatment period of either 24 or 48 weeks; the proportion of subjects treated for 24 or 48 weeks was assumed, in all the simulation scenarios, to be the same as the one observed in the Pegasys arms of Study C (the predicted SVR rates were robust to the assumed proportion of subjects treated for 24 and 48 weeks).

The predictions of SVR rates were made with the following assumptions:

1. The overall number of subjects per arm followed the planned numbers of subjects in Phase 3 studies: N=350 treatment-naive subjects, as in Study G; and

N=140 prior relapse subjects and N=l 20 prior nonresponder subjects, as in Study

F.

2. The genotype la: lb ratio is 1 : 1 , approximately reflecting the observed ratios from US and European studies.

3. For treatment-naive subjects, the PR treatment duration is 24 weeks for subjects with undetectable HCV RNA at Weeks 4 and 12 of treatment (eRVR) and 48 weeks for subjects without eRVR. For all PR treatment-failure subjects, the PR treatment duration is 48 weeks, regardless of viral response.

4. The percentage of subjects who failed to reach SVR because of being lost to follow-up, noncompliance, or withdrawal of consent is zero. Therefore, to compare the observed and predicted SVR rates, any nonzero percentage of these reasons for failing to reach SVR should be subtracted from the predicted SVR rates.

5. Discontinuation rates of telaprevir or of all drugs in the regimen (T/PR or PR) followed the profiles of Study C pooled peginteferon-alfa-2a arms, which are expected to be closest to the rates in Phase 3 studies. 6. Study drug concentrations were simulated from re-sampled estimates. For telaprevir, the estimates were from Studies A and B; for peginterferon alfa, the estimates were from Studies A, B, and D; for ribavirin, the estimates were taken from the literature. Sampling for plasma ribavirin concentration measurement in Studies A, B, and D was too sparse to allow good estimates of pharmacokinetic parameters.

7. Dosing adherence was 100% for all 3 drugs.

8. The same parameters generated randomly were used across simulation scenarios to avoid the influence of selection bias by re-sampling.

The predicted SVRJTT rate of each group in Figure 2 was computed by entering the actual number of subjects per group, the number of subjects who failed to reach SVR for other reasons (lost to follow-up, noncompliance, and withdrawal of consent), viral subtype, and treatment (both telaprevir and PR) durations of each group.

Simulations of regimens in Studies A, B and D assumed telaprevir concentrations that were simulated from parameters re-sampled from population pharmacokinetic estimates obtained from pooled Studies A and B. The Peg-IFN concentrations were simulated from estimates of pooled Studies A, B, and D. The RBV concentrations were simulated from published estimates.

Simulations of regimens in Study C were generated with the following assumptions:

1. Predictions were limited to Pegasys arms only because previous data were not available to model the pharmacokinetics and antiviral activity of Peglntron.

2. Telaprevir concentrations were simulated from the parameters re-sampled from population pharmacokinetic estimates of Study C.

3. Peg-IFN concentrations were simulated from parameters re-sampled from

population pharmacokinetic estimates of Studies A, B, and D. The Peg-IFN population pharmacokinetic estimates from Study C data have not been completed when this report is written.

In the simulation comparing the predicted SVR rates for different duration of treatments of combination regimens with telaprevir, peginterferon and ribavirin, the following assumptions were made: 1. The same random effects associated with pharmacokinetic and viral dynamic parameters were used in the simulation of SVR rates of the compared regimens. Therefore, the differences in SVR rates were due to the regimens applied only.

2. Discontinuation rates were assumed to be the observed discontinuations in the pooled Peg-IFN-alfa-2a RBV groups in Study C.

3. Telaprevir concentrations were simulated using parameters re-sampled from

population estimates of Studies A and B. Peg-IFN concentrations were simulated using parameters re-sampled from population estimates of Studies A, B, and D.

4. RBV concentrations were simulated using parameters re-sampled from published estimates. The average modified dose of RBV in the pooled T/PR arms of Studies

A and B was used for simulations for both regimens.

5. Treatment adherence to all three drugs was assumed at 100%

6. Subtype 1 a: lb ratio was 1 : 1

7. The number of subjects was 350 per group or Naive and Failure populations; 140 per group for Relapser, and 120 per group for Nonresponder. This number was the same as the number of subjects per arm in the active Phase 3 study of telaprevir (Study G and Study F).

8. The simulation for various populations with different PR responsiveness followed the method described above.

Predicted and observed SVR rates by telaprevir treatment duration are shown in Figure 3. The analyses support a telaprevir treatment duration of 12 weeks in treatment-naive and in treatment-failure populations. SVR rates increase as the duration of telaprevir treatment increases up to 12 weeks but a plateau in SVR rates is reached for durations of telaprevir treatment longer than 12 weeks.

Compared to a telaprevir treatment duration of 12 weeks, a duration of 4 weeks is predicted to result in an SVR rate that is 25% lower in treatment-naive subjects and 50% lower in treatment- failure subjects.

Compared to a telaprevir treatment duration of 12 weeks, a duration of 8 weeks is predicted to result in an SVR rate that is 9% lower (the 90% confidence interval widths are 6% to 8%) in treatment-naive subjects and 16% lower (the 90% confidence interval widths are 7% to 8%) in PR-treatment-failure subjects. This result is comparable to the predicted SVR rate difference between T8/PR and T12/PR of 4% in treatment-naive subjects in the previous report (the predicted difference in the PR48 treatment-failure subjects was not available in this earlier report). The small predicted difference in SVR rates for 8 weeks and 12 weeks of telaprevir treatment in treatment- naive subjects, in conjunction with the possible reduction in the occurrence of severe rash with the 8 week regimen, was the basis for testing these telaprevir treatment durations in Phase 3 Study G.

Predicted SVR rates for 12 weeks versus 24 weeks of telaprevir treatment di fer by only 1% in treatment-na^'ive subjects and by only 2% in treatment-failure subjects.

To further dissect the relationship between telaprevir treatment duration and SVR rates, simulated treatment-na^'ive subjects who completed treatments with telaprevir for

4 or 12 weeks in combination with PR for 24 weeks were compared (a single 24-week

PR duration was selected to allow direct comparison of the telaprevir durations).

Figure 4 shows the predicted treatment outcome in these subjects. Subjects completing a T12/PR24 regimen are predicted to have higher SVR rates than those completing a T4/PR24 regimen because of both lower viral breakthrough rates and lower relapse rates. The majority of viral breakthrough on the T4/PR24 regimen occurred after telaprevir dosing was completed.

To further compare telaprevir treatment durations, HCV dynamics were analyzed in a typical (median) simulated treatment-na^'ive subject and a typical (median) simulated subject with PR48 treatment-failure. In this analysis, simulated subjects were treated with 4 weeks of telaprevir in combination with PR (T4/PR) or 12 weeks of telaprevir in combination with PR (T12/PR). This analysis included the period from baseline through Week 12 of treatment and provides a virological perspective for why SVR rates are expected to be lower with 4 weeks of telaprevir treatment than with 12 weeks of telaprevir treatment. Because PR responsiveness is highly variable in a given population, a typical simulated subject with prior PR48 treatment-failure was selected to provide a clearer illustration of the relationship between telaprevir duration and outcome. These simulations were performed with typical telaprevir

responsiveness; the contribution of variability in telaprevir responsiveness to SVR rates is indicated by the predicted SVR rates by telaprevir duration, as shown in Figure 3. Figure 5 shows that WT, LV (e.g., R155K), and HV (e.g., A156 and V36M/R155K) have different viral dynamics in response to T/PR treatment. The rate of elimination of WT and LV is dependent on telaprevir, and viral eradication (shown in figure as "HCV RNA total") is predicted to be accomplished by approximately 6 weeks in a typical treatment-naive subject (with average telaprevir and PR responses). If telaprevir treatment ceased before 6 weeks, the remaining WT and LV virus would need to be eliminated by PR treatment, and the success of eradication would depend more on the subject's PR responsiveness. The rate of eradication of HV in all subjects at all times is governed by PR responsiveness, as HV are poorly inhibited by telaprevir.

In the typical simulated treatment-na^'ive subject (subject with median PR

responsiveness), T4/PR and T12/PR regimens that include 48 weeks of PR treatment are predicted to result in viral eradication, as shown by a continuing decline in HCV RNA at the end of the first 12 weeks of treatment. However, a longer duration of PR treatment is required to eradicate all HCV with a T4/PR regimen than with a T12/PR regimen.

In contrast, in a subject with a less ideal PR response (typical subject with median responses of simulated prior PR48 treatment failure), predicted treatment outcomes differ for the 4-week and 12-week treatment durations of telaprevir. In the 4-week telaprevir regimen, viral breakthrough with WT and LV is predicted after telaprevir dosing is complete, as WT is not predicted to be eliminated until after Week 6, and the PR response is too weak to prevent WT replication at the levels at which virus survives at Week 4. The 12-week telaprevir duration is predicted to eradicate WT and LV, although some HV variants may still remain at Week 12. Eradication of WT and LV by Week 6 for a typical simulated subject with median responses is consistent with the observed eventual plateau in predicted SVR rates when telaprevir treatment duration is prolonged.

In Phase 2 studies A, B, and D, telaprevir-based regimens with PR treatment durations of 12, 24, and 48 weeks were explored in treatment-na^'ive and PR treatment-failure populations. The comparisons between model-predicted and observed SVR rates in these regimens are provided in Figure 2. To further understand these data, predicted SVR rates for subjects completing regimens with 12 weeks of telaprevir in

combination with 12, 24, and 48 weeks of PR are provided in Figure 6. Because the simulations compared simulated subjects who have the same

responsiveness but who are treated with alternative PR durations, they allow direct comparison of the SVR rates for different durations of PR treatment. To allow estimation of the maximum difference in SVR rates for different durations of PR treatment, the predicted SVR rates were calculated only among subjects who completed treatment. (Predicted SVR rates that account for discontinuations would tend to diminish any differences seen in the SVR rates.) The results showed that a PR treatment duration of 12 weeks is not sufficient: the difference in SVR rates between 12-weeks and 48-weeks of PR treatment is 23% in treatment-na^'ive and 38% in prior PR treatment- failure populations. However, a PR treatment duration of 24 weeks is sufficient. Comparing PR24 and PR48 durations, the difference in SVR rates is 2% to 3% in the treatment-na^'ive population and 4% in the prior PR treatment-failure population. Therefore, only PR durations of 24 and 48 weeks were used in Phase 3 studies.

In Phase 2 studies in treatment-na^'ive subjects (Studies A and B), subjects with undetectable HCV RNA at Weeks 4 and 12 (eRVR) who completed treatment with 12 weeks of telaprevir and 24 weeks of PR (T12/PR24) had low (2% to 5%) relapse rates. To evaluate whether this low relapse rate in treatment-na^'ive subjects with eRVR could be improved by extending PR treatment, modeling analysis was conducted to compare SVR rates in the 8-week and 12-week telaprevir regimens where (a) simulated subjects with eRVR were treated with 24 weeks of PR and without eRVR were treated with 48 weeks of PR and (b) simulated subjects were treated with 48 weeks of PR regardless of eRVR status. This analysis assumed that there were no premature discontinuations of treatment after

Week 24; discontinuations occurring during the period between Weeks 24 and 48 would diminish any differences seen in SVR rates between 24 and 48 weeks of PR treatment.

As shown in Figure 7, the results predict that when a response-driven PR duration is used in combination with a telaprevir treatment duration of 8 or 12 weeks, SVR rates are comparable for both 24- and 48-week PR treatment durations: the difference in SVR rates is 1% to 2% comparing T8/PR24 to T8/PR48, and 1% comparing

T12/PR24 to T12/PR48. Therefore, these results predict no advantage in extending the 24-week PR treatment in treatment-na^'ive subjects who achieve an eRVR. During the development of protease inhibitors, regimens with and without a 4-week delayed start of protease inhibitor treatment have been proposed. The rationale for the delayed start of protease inhibitor treatment is to allow PR concentrations to reach levels closer to steady-state and decrease HCV RNA levels before the start of protease inhibitor dosing, which might then maximize SVR rates and minimize viral breakthrough rates.

Modeling analyses were used to predict SVR rates for telaprevir-based regimens with and without 4 weeks delayed start. Regimens without a delayed start begin telaprevir dosing at Week 0; those with a delayed start begin telaprevir dosing at Week 4. The total durations of telaprevir treatment (12 weeks) and PR treatment (24 or 48 weeks) were controlled to be the same in the comparison. Because the contribution of a delayed start to discontinuation rates is not known, the modeling analyses were performed only in simulated subjects completing assigned treatment (The no discontinuation assumption during TVR and PR treatment periods resulted in higher SVR rates than are seen in Figure 3 and Figure 7, which were simulated with discontinuation rates considered). Results are shown in Figure 8. Predicted SVR rates with and without a delayed start of telaprevir treatment were predicted to be the same (within 1%) for both simulated treatment-nai^'ve and prior PR48 treatment-failure populations.

In a Phase 2 study in the PR treatment-failure population (Study D), regimens with 12 weeks (T12/PR24) and 24 weeks (T24/PR48) of telaprevir treatment had similar SVR rates. However, relapse rates were 19% for the T12/PR24 arm (N = 22: 3 WT, 12 LV and 7 HV) and 2% for the T24/PR48 arm (N=2, 1 WT and 1 LV). The model- predicted SVR rates, shown in Figure 3, suggest no additional increase in SVR rate if telaprevir duration is prolonged from 12 weeks to 24 weeks. Here, the model was used to further explore the expected evolutionary dynamics in simulated subjects who failed a T12/PR24 regimen.

Figure 9 demonstrates different mechanisms of failure to eradicate HC V in simulated subjects treated with T12/PR24 regimens. Mechanisms are shown separately for 3 simulated subjects with varying prior PR responses, as described in Table 7. These simulations illustrate only representative examples with median responses; each respective group of prior PR responses has variable PR responses (the predicted SVR rates by groups of prior PR responses are provided in Figure 10). Table 7 Description of Simulated Subjects

Figure Name of PR Description of PR Response Definition of PR Response of the

Panels Response of the Simulated Sub ject Group

9A Weakest Median response of prior PR Week 12 HCV RNA decline of <2-log_l0 treatment null responders during PR treatment

9B Intermediate Median response of prior PR Never had undetectable HCV RNA levels treatment nonresponders during 48 weeks of PR treatment

9C Slightly Median response of prior PR Did not reach eradication (SVR) with

better treatment failures PR48 treatment

In the simulated subject with the weakest PR response (Panel A of Figure 9), the PR response is too weak to overcome the replication of HV. High levels of HV lead to an increased incidence of back mutations to LV and to WT. This results in relapse with

WT or LV when PR is stopped at Week 24. If PR treatment is continued for 24 more weeks, the higher-fitness WT or LV replicate faster than HV during Weeks 24 to 48, which results in viral breakthrough with WT or LV during PR treatment (The

T12/PR48 results can be deduced by extrapolating the on-treatment Week 24 slope of the HCV RNA total).

In the simulated subject with an intermediate PR response (Panel B of Figure 9), the PR response is sufficient to prevent replication of HV but not eliminate it, and the simulated subject relapses when treated with either 24 or 48 weeks of PR. At the time of relapse, the quasi-species are predicted to be predominately HV if sample was taken at the time of relapse, or LV or WT if taken later because LV and WT have higher fitness than HV.

In the simulated subject with a slightly better PR response (Panel C of Figure 9), the PR response is sufficient to prevent replication of HV, but the HV elimination is slow. The simulated subject will relapse if treated with 24 weeks of PR but will reach SVR if treated with 48 weeks of PR.

Figure 9 also suggests that the type of viral sequence at the time of viral breakthrough will depend on the timing of the sample taken. If samples are taken immediately when relapse or viral breakthrough occurs, the chance of observing HV is higher. However, if samples are taken at a much later time, the chance of WT or LV being the dominant variant in the quasi-species is higher because the fitness of WT and LV are higher than the fitness of HV. When compared to the observed data, the predicted rate of back mutation to LV and particularly to WT may be more rapid.

Table 8 provides the final pharmacokinetic and viral dynamic parameter estimates that were used to resample parameters in the SVR rate simulations. Subtype la and lb were simulated separately because of differences in the

composition of variants in each viral quasispecies. All the parameters except the composition of the quasispecies were assumed the same between subtype la and lb. - Subtype l a: WT, R155K, A156T, and V36M/R155K

- Subtype lb: WT, V36A, A156T

Table 8: Final parameter estimates of viral dynamic and pharmacokinetic models.

Each parameter assumed lognormal distribution, of which log O of mean and variance were rovided.

PeglFN Ka_P 3.94E+00

Clearance of

PeglFN CI_P L h-¹ -2.77E+00 8.17E-01 -4.45E+00 1.38E+00 Studies A, B, D

Volume of Peg-IFN

Vp L 3.13E-01 3.42E-01 -2.50E+00 3.68E+00 Studies A, B, D

Absorption of RBV -1 .37E-

6

Ka_R h-¹ -2.66E-01 3.40E-04 -3.37E-01 01

Clearance (central

compartment) of

6

RBV CIR L h^"1 2.88E+00 1.64E-02 2.42E+00 3.22E+00

Volume (central

compartment) of

6

RBV V2_R L 6.16E+00 4.22E-03 5.82E+00 6.36E+00

Intercompartmental

clearance

(compartment 3) of

6

RBV Q3_R h-¹ 3.54E+00 4.65E-05 3.51 E+00 3.58E+00

Volume

(compartment 3) of

6

RBV V3_R L 8.38E+00 2.68E-02 7.82E+00 8.75E+00

Intercompartmental

clearance

(compartment 4) of

6

RBV Q4_R h-¹ 4.58E+00 3.14E-04 4.45E+00 4.63E+00

Volume

(compartment 4) of

6

RBV V4_R L 6.77E+00 2.74E-05 6.73E+00 6.79E+00

Bioavailability of

central

compartment of

6

RBV FI R unitless 2.85E-02 2.33E-02 -3.80Έ-0 6.67E-01

Absorption of -5.94E- telaprevir Ka_T h-1 -7.03E-01 8.64E-03 -7.82E-01 01 Studies A and B

Accumulation rate

constant of

absorption of

telaprevir ΑΚτ -3.07E-01 2.84E-01 -2.26E+00 2.03E+00 Studies A and B

Hill factor of

telaprevir GMT 1 .25E+00 1 .01 E-01 4.76E-01 2.34E+00 Studies A and B

Maximum

absorption of

telaprevir KAM_T 6.18E-01 2.97E-02 4.70E-01 8.19E-01 Studies A and B

Clearance of

telaprevir CIT h-¹ 3.50E+00 6.32E-02 2.69E+00 4.59E+00 Studies A and B

Volume of

telaprevir Vr L 5.93E+00 1.09E-01 4.99E+00 7.20E+00 Studies A and B

Bioavailability of

central

compartment of

telaprevir F1 T unitless 1.00E+00 O.OOE+00 na na Studies A and B

Clearance of

telaprevir CIT h^'1 4.05E+00 8.86E-02 3.14E+00 4.99E+00 Study C

Volume of

telaprevir VT L 5.28E+00 1 .08E-01 3.70E+00 6.35E+00 Study C

Bioavailability of

central

compartment of

telaprevir F1j unitless 2.15E+00 O.OOE+00 na na Study C

'Bounds were used to truncate distribution Table 9: Parameters obtained from literature or assumed.

Predicted SVR rates were calculated by simulating concentrations and HCV RNA dynamics using parameters re-sampled from estimates of population approaches summarized in Table 8. The re-sampled parameters were truncated by lower and upper bounds; with the bounds obtained from the extreme values of the observed individual parameter estimates.

SVR rate among completers were calculated from 10⁴ simulated treatment-na^'ive subjects treated with various durations of telaprevir and PR. Treatment durations simulated for telaprevir is in 2-week increments and for Peg-IFN and RBV in 2-week increments for the first 12 weeks and in 4- week increments subsequently.

Compliance to telaprevir, Peg-IFN, and RBV dosing were assumed to be 100%. The RBV dose was adjusted at the same amount of that observed in the pooled PR and T/PR regimens of Studies A and B. Peg-IFN dose was not adjusted during the simulations, because few Peg-IFN dose adjustments were observed in Studies A and B. In each simulated subject in a given duration of telaprevir and PR, viral eradication was assigned to the subject if his/her overall HCV RNA level by the end of treatment was below 1 copy in the body, or it reached a 12-log io decline from baseline in HCV RNA (assuming that the average baseline HCV RNA was 10⁷ IU/mL). If viral eradication was obtain during treatment, then the subject was assumed to obtain SVR. The HCV RNA level defined for viral eradication followed the number used in the literature.

Subjects with subtype la and lb were simulated separately because of the different compositions of variants emerging in subjects with different subtypes. In a given regimen, the predicted intent-to-treat SVRrrr rates, which include subjects who completing assigned regimens and who discontinuing earlier, were calculated by resampling randomly the SVR status with frequency equal to the cumulative number of subjects treated at a given durations of telaprevir and PR treatment. The proportion of subjects by different treatment duration were derived from the discontinuation profiles observed in the PR arms of pooled Studies A and B, and in the pooled telaprevir q8h and ql2h Pegasys and RBV arms of Study C.

The 90% confidence interval of the predictions were calculated by repeating the

S VRITT- sampling calculation 100 times and reported the 5^lh and 95 ^th percentiles of the results. The SVR_ITT rates for different populations were computed by limiting the random sampling of completer treatment-nai^'ve SVR simulations to subjects whose PR response follows the following criteria of each population: failure, if subject's viral load not reaching eradication by 48 week of PR treatment; relapser, if subject's viral load is undetectable by end of 48-week of PR treatment but not reaching eradication; and nonresponder, if subject' s viral load is detectable by end of 48-week of PR treatment.

Figure 12 shows the discontinuation rates assumed in the model when alternative durations of telaprevir treatment were compared. The assumptions about

discontinuation rates used in these analyses follow. For PR treatment, the discontinuations followed the rates observed from pooled PR arms from Studies A and B. Discontinuations because of virologic stopping rules were excluded. Discontinuation rates from Study D were not included because of large number of subjects that stopped treatment early because of virological stopping rules.

For T/PR treatment with a telaprevir duration of up to 12 weeks, discontinuations followed the rates obtained from pooled telaprevir 750 mg q8h and 1 125 mg ql2h PeglFN-alfa-2a/RBV arms in Study C. These treatment arms were chosen because Study C implemented 2 changes that are relevant to the Phase 3 studies:

(1) response-guided therapy in which subjects with eRVR treated with PR for 24 weeks and subjects without eRVR were treated with PR for 48 weeks); and (2) changes in the rash management plan to minimize treatment discontinuation.

For T/PR treatment with a telaprevir duration other thanl2 weeks, the discontinuation rates when telaprevir was present followed the rates described in the above point. When telaprevir was not present, the discontinuation rates followed the lower discontinuation rates of PR treatment. For T/PR treatment with a telaprevir duration of 24 weeks, the discontinuations from Weeks 12-24 were assumed to be the rates observed in the same period in the T24/PR48 arm in Study D.

For the first 20 weeks of treatment with alternative telaprevir durations, the assumed discontinuation profiles are provided in Figure 12.

Table 10: Estimates of viral dynamic parameters from Studies A and B PR and T/PR regimens.

Em irical Ba esian Feedback estimate were used. Values are given in loglO-scale.

Multiplier of plasma to effective 434 2.63E-02 7.07E-02 -6.61 E-01 2.49E+00 concentrations for ribavirin K_R

Multiplier of plasma to effective 318 1 .85E-01 1 .68E-01 -1 .88E+00 1 .45E+00 concentrations for telaprevir κ_Ί

Table 1 1 summarizes the fitness estimates obtained from subjects treated with telaprevir in monotherapy in Study E. The estimation method used here was refined from the method used earlier. In previous version, the method implemented was individual subject estimate; in this version, a population approach (Empirical

Bayesian Estimate) was implemented was implemented. The resulting estimates have similar trends to the previous estimates in Report D224; Sequence 0162, but with lower absolute median values and narrower distribution of fitness estimates.

Table 11 : Estimates of in-vivo fitness from Study E

Empirical Bayesian feedback estimate were used. Only variants observed in N>2 subjects were included.

'Fitness average when adjusted for 9 subtype lb patients with undetectable variants is 0.05.

2Fitness average when adjusted for 9 subtype lb patients with undetectable variants is 0.04.

Observed and simulated HCV RNA dynamics were compared in Figure 18. For subjects on PR treatment, results showed good correspondence between data and simulations; majority of the data were within the simulations. Few data lying outside the 5-95% percentiles of the simulations as expected. For subjects on T/PR treatment, data also corresponded well to predictions. For both subtypes, majority of the data were within 10th-90^th percentile of predictions. Few data lied outside the 5^th

percentile of prediction, as expected, and is consistent with low (about 5%)

breakthrough rates observed in T/PR regimens in treatment-naive population. The observed and the predicted (90% confidence interval) SVR rates for PR and T/PR regimens of Studies A and B are compared in Figure 19. T12PR12 regimen of Study A were excluded because of small number of subjects (N=17). The predicted SVRrrr rates were calculated using methods described here. Observed SVRnr- rates were within the 90% CI of the predicted rates.

Discussion

Modeling analyses suggest that in a T/PR regimen, the primary role of telaprevir is to eliminate WT and LV, and the primary role of PR is to eliminate HV. This is consistent with the observed sensitivities of HCV variants to telaprevir and to PR in vitro and in HCV-infected subjects. Furthermore, the model predicts that for T/PR regimens with 12 weeks of telaprevir, the rate-limiting step in HCV eradication is the elimination of HV by PR. This results in a dependency of SVR rates on PR

responsiveness. This prediction is consistent with the observed lower SVR rates in ■ Study D in subjects with lower PR responsiveness: in the T12/PR24 arm, SVR rates were 39.4% for prior PR nonresponders and 69.0% for prior PR48 relapsers; in the T24/PR48 arm, SVR rates were 37.5% for prior PR nonresponders and 75.6% for prior PR relapsers. Modeling analyses also suggest that the ranges of PR responsiveness observed in the PR treatment-failure populations are a subset of those observed in the treatment-na^'ive population. A model with PR responsiveness trained from data obtained in treatment- naive subjects, along with an assumption that PR responsiveness in the PR treatment- failure population is a subset of the PR responsiveness in the treatment-na^'ive population, predicted SVR rates that are comparable to the observed SVR rates from clinical studies in both treatment-na^'ive and different PR treatment-failure populations. The evolutionary dynamics of the HCV quasi-species for subjects who failed to reach eradication with telaprevir-based regimen showed reversion to WT- or LV-dominant quasi-species over time. In Study E, in which study subjects were dosed with 2 weeks of telaprevir monotherapy, the sequencing data showed replacement of HV by LV first (7 to 10 days after completing dosing) followed by replacement by WT in 3 to 7 months. In subjects who relapsed after completing a T12/PR regimen (with PR durations of 24 or 48 weeks, the sequences tended to be HV if the time of sample collection was soon after relapse, and tended to be LV or WT if the time of sample collection was later). A more complete picture of the evolutionary dynamics after telaprevir treatment is under investigation. These findings are consistent with the predictions of back mutations and reversion from HV to WT and LV in simulated subjects who did not achieve SVR with a T/PR regimen, because the fitness of WT and LV is higher than the fitness of HV.

Compared to observed data, the model predicted a faster rate of reversion to predominately WT. This discrepancy may arise from 2 sources. First, there is a lack of data available to estimate the rate of infected-cell elimination during off-treatment (because of the lack of data, the model assumed that the off-treatment rate was equal to the average rate during peginterferon-based treatment). Second, the model was based on a deterministic differential equation system, which tends to overestimate the rate of randomly generated back mutations near the limit of eradication. These 2 deficiencies are expected to affect only the predictions during off-treatment, because the on-treatment predictions, including the SVR rates, did not rely on these assumptions.

Modeling analyses predict that SVR rates increase with increasing telaprevir treatment durations of up to 12 weeks but that SVR rates plateau when the telaprevir treatment duration is increased from 12 weeks to 24 weeks, both for treatment-na^'ive and for overall prior PR treatment-failure populations.

Reducing the telaprevir treatment duration from 12 weeks to 4 weeks is predicted to increase viral breakthrough rates during subsequent PR treatment because the shorter duration of telaprevir is not sufficient to eliminate most WT and LV. A telaprevir treatment duration that is too short may result in a higher percentage of subjects returning back to detectable HCV RNA levels with WT or LV before the end of PR treatment, and a higher percentage of subjects relapsing with WT or LV after completing treatment. In Study I, in which subjects were treated with 4 weeks of telaprevir in combination with 48 weeks of PR, 2 of 12 subjects had viral

breakthrough with WT and LV during PR treatment, after telaprevir treatment was completed. In Studies A, B and C, in the subset of subjects treated with T/PR regimens who failed to achieve SVR because of early discontinuation of all study drug dosing and who were treated for less than 5 weeks, the HCV quasi-species were predominately WT (n=18; 13 with WT, 4 with LV, 1 with HV). Consequently, reducing the telaprevir treatment duration from 12 weeks to 4 weeks would reduce SVR rates, even after factoring in the lower treatment discontinuation rates anticipated for 4 weeks of telaprevir treatment. In the treatment-naive population, a telaprevir duration of 8 weeks is predicted to result in an SVR rate that is about 9% lower (with 90% CI bound widths of 6% to 8%) than the SVR obtained with a telaprevir duration of 12 weeks. This result is comparable to the predicted SVR rate difference of 4% between 8 weeks and

12 weeks of telaprevir in the previous modeling report. However, in Study A, an increase in the incidence of severe rash leading to discontinuation of study drug dosing occurred after 8 weeks of treatment with T/PR. The small predicted difference in SVR rates, combined with empiric tolerability data, suggest that a telaprevir treatment duration of 8 weeks may have a better benefi risk ratio than a telaprevir treatment duration of 12 weeks. Therefore, a treatment arm with a telaprevir duration of 8 weeks was included in Phase 3 Study G to evaluate this hypothesis.

In the treatment-na^'ive population, a regimen with 12 weeks of telaprevir is predicted to have an SVR rate within 1% of the SVR rate for a regimen with 24 weeks of telaprevir, suggesting that 12 weeks of telaprevir is sufficient. The conclusion that 12 weeks of telaprevir is sufficient is consistent with the low sum of viral breakthrough on PR and relapse rates with predominately WT or LV observed in studies in treatment-na^'ive subjects treated with a regimen containing 12 weeks of telaprevir (Studies A, B, and C). Subjects who complete their assigned treatment regimen but have viral breakthrough on PR or relapse with predominately WT or LV are the only subjects that have the potential to achieve an SVR with a longer telaprevir duration. In the combined Studies A, B, and C T12/PR24 or T12/PR48 arms, these subjects represented only 5% of the study population. Because this percentage provides the maximum potential increase in SVR rates with a longer telaprevir duration in these studies, its low value also suggests that 12 weeks of telaprevir is sufficient.

A duration of 12 weeks of telaprevir is also predicted to be sufficient for the prior PR-treatment-failure population. The predicted SVR rates for 12 weeks and 24 weeks of telaprevir were within 2% in the PR-treatment-failure population. In Study D, both 12-week and 24-week durations of telaprevir treatment were tested in this population, and the observed SVR rates were similar (51.3% for T12/PR24 and 53.1% for T24/PR48). The difference of the PR treatment durations in these arms is not expected to change the conclusion about the sufficiency of 12 weeks of telaprevir treatment, because the SVR rate of a T24/PR24 regimen should be less than or equal to the SVR rate in a T24/PR48 regimen. Similar conclusions were reached when the same type of analyses were repeated for each of the more refined categorization of prior PR response populations.

The sequencing analyses in subjects who failed to reach SVR in T/PR regimens are also consistent with the lack of evidence of virological benefit beyond 12 weeks of telaprevir treatment in the PR-treatment-failure population. In PR-treatment-failure subjects in Study D who were treated with a T12/PR24 regimen, 13% completed treatment and relapsed with WT or LV. From these relapse data alone, 2 possible explanations could be constructed: (1) the WT or LV was never eradicated during telaprevir treatment; or (2) the WT or LV were eradicated during telaprevir treatment but were re-generated from back mutations of HV and subsequently outgrew HV to dominate HCV quasi-species. The modeling analyses performed here, which accounted for different replicative fitness and elimination of these variants, predict that the second hypothesis is more likely. The lack of evidence of additional virologic benefit, and the observed poorer tolerance in the 24 weeks of telaprevir dosing (T24/PR48 arm of Study D) compared to the 12 weeks of telaprevir dosing

(T12/PR24 arm of Study D) support the 12-week telaprevir duration in the treatment- failure population.

The model predicted that in regimens with 8 and 12 weeks of telaprevir treatment, 24 weeks of PR treatment is sufficient for the majority of subjects. In combination with 12 weeks of telaprevir, a 12-week PR treatment duration was predicted to have 23% and 38% lower SVR rates than a 48-week PR treatment duration in treatment- naive and PR-treatment-failure populations, respectively. These results suggest that a 12-week PR treatment duration is not sufficient in the majority of subjects. However, 12 weeks of telaprevir treatment in combination with 24 weeks of PR treatment appear to be sufficient. The predicted SVR rate for the 24-week PR treatment duration is within 3% of that for the 48-week PR duration in the treatment-naive population. The similar SVR rates with 24-weeks and 48-weeks of PR treatment in combination with 12 weeks of telaprevir treatment in simulated subjects are also consistent with data from completed clinical trials. As mentioned above, low relapse rates were observed in subjects who achieved undetectable HCV RNA at Weeks 4 and 12 and were treated with 24 weeks of PR (in combination with 12 weeks of telaprevir).

Because the relapse rate provides the maximum limit of increase in SVR rates with longer treatment, the low relapse rate observed here suggest similar expected SVR rates with 24 and 48 weeks of PR treatment. Therefore, modeling analyses and observed data supported the evaluation of response-driven regimens in treatment- na^'ive subjects, and these regimens were included in Phase 3 Studies G and Study II. The model further predicts that SVR rates for 24-and 48-week PR treatment durations will be comparable for a regimen with an 8-week telaprevir treatment duration.

Regimens with and without a 4-week delayed start

treatment are predicted to have comparable SVR rates among subjects who complete treatment, both in treatment-na^'ive and in treatment-failure populations. The observed data in treatment-na^'ive- subjects in Studies A, B, and C, in which subjects were treated with 12 weeks of telaprevir without a delayed start in combination with PR treatment, showed low viral breakthrough and relapse rates (around 5% each). This finding supports that even without a delayed start of telaprevir, a considerably successful treatment result is expected for T/PR regimens. The prediction of any additional benefit of a delayed start of telaprevir will be re-evaluated when the results of Study F are available.

Modeling analyses support T/PR regimens that include a maximum telaprevir treatment duration of 12 weeks. Modeling analyses also support use of response- driven durations of 24 or 48 weeks of PR treatment in treatment-na^'ive subjects.

Regimens with these durations are being evaluated in Phase 3 studies of telaprevir. In addition, based on tolerability data from Phase 2 studies, a regimen with 8 weeks of telaprevir treatment is being evaluated in a Phase 3 study in treatment-na^'ive subjects (Study G).

Example 1

Modeling analyses were conducted to evaluate the efficacy of alternative treatment durations of telaprevir, Peg-IFN and RBV (T/PR) in subjects chronically infected with genotype 1 hepatitis C virus (HCV). The model was verified by comparing predicted SVR rates to the observed values in completed Phase 2 studies. The modeling analyses predictions follow.

SVR rates increase as the duration of telaprevir treatment increases up to 12 weeks but a plateau in SVR rates is reached for telaprevir treatment durations longer than 12 weeks, both for treatment-naive and for PR treatment-failure populations.

SVR rates are slightly lower for a treatment-na^'ive population treated with a T/PR regimen that includes 8 weeks of telaprevir treatment than for a T/PR regimen that includes 12 weeks of telaprevir treatment (9% lower SVR rates, with predicted 90% CI bound widths of 6% to 8%). SVR rates are comparable when subjects with undetectable HCV RNA at Weeks 4 and 12 of treatment are treated with 24 or 48 weeks of PR in combination with either 8 or 12 weeks of telaprevir: the difference in SVR rates is predicted to be 1% to 2% for T8/PR24 versus T8/PR48, and 1 % for T12/PR24 versus T12/PR48.

SVR rates are comparable for subjects who complete treatment with regimens with and without a 4- week delayed start of telaprevir, both in treatment-na^'ive and in PR treatment-failure populations.

Modeling analyses support regimens that include a maximum telaprevir treatment duration of 12 weeks in treatment-na^'ive and PR treatment-failure subjects and use of response-driven durations of 24 or 48 weeks of PR treatment in treatment-na^'ive subjects. Treatment regimens with these durations are being evaluated in Phase 3 studies of telaprevir.

Example 2

Control stream used in population pharmacokinetics of peginterferon-alfa-2a:

<?xml version- Ί .0"?>

<nonmem>

<text><![CDATA[

Z:\nonmem\pegVX950_ABC\control.model8jeroalagetaalag_10.ctl.2009101692331 .dir\control.mode etaalag_10.ctl.lst

2009.10.16.1511.21

; population PK of peginterferon

; 1 -compartment, estimated Ka, CI, V, ALAG; S2=V2; Eta=0, seed from results run 2 updated on 2009/01/16 ; used for Bayes estimates; optimal of study l-J-A applied to study B

SPROBLEM pegPK 1 comp, l-J, est Ka CI V ALAG, Eta=0 SINPUT ID DV AMT CMT EVID TIME STID TVR RBV BMI CRCL

SDATA Z:\Projects\vx950_ABC\data\data.nonmem.ABC.pegpoppk.csv IGNORE=@

SSUBROUTINE ADVAN2

$PK

KA =THETA(1 )*EXP(ETA(1 ))

CL =ΤΗΕΤΑ(2)ΈΧΡ(ΕΤΑ(2))

ALAG1 = THETA(3)*EXP(ETA(3))

V= THETA(4)*EXP(ETA(4))

S2 = V

K = CL/V

SERROR

Y=A(2)/V*(1 +EPS(1 ))+EPS(2)

DEL=0

IF(F.EQ.O) DEL=1

IPRED=F

W=lPRED+DEL

IRES=DV-IPRED

IWRES=IRES/W

STHETA

(0.01 ,0.331 )

(1 ,8.64)

O FIX

(0.05,0.912)

SOMEGA

(0.632)

(0.141 )

O FIX

(0.837)

SSIGMA

(0.102)

(0.285) SESTIMATION MAXEVAL=2000 SIG=3 PRINT=10 NOABORT METH0D=1 INTER

STABLE ID CMT TIME EVID AMT IPRED IRES IWRES NOPRINT ONEHEADER FILE=pred.tab STABLE ID ETA1 ETA2 ETA3 ETA4 KA CL V ALAG1 STID TVR RBV BMI CRCL FIRSTONLY NOAPPEND NOPRINT ONEHEADER FILE=eta.tab

1 NONLINEAR MIXED EFFECTS MODEL PROGRAM (NONMEM) DOUBLE PRECISION NONMEM VERSION VI LEVEL 1.1

DEVELOPED AND PROGRAMMED BY STUART BEAL AND LEWIS SHEINER

PROBLEM NO.: 1

pegPK 1 comp, l-J, est Ka CI V ALAG, Eta=0

ODATA CHECKOUT RUN: NO

DATA SET LOCATED ON UNIT NO.: 2

THIS UNIT TO BE REWOUND: NO

NO. OF DATA ITEMS IN DATA SET: 12

ID DATA ITEM IS DATA ITEM NO.: 1

DEP VARIABLE IS DATA ITEM NO.: 2

MDV DATA ITEM IS DATA ITEM NO.: 12

OINDICES PASSED TO SUBROUTINE PRED:

5 6 3 0 0 0 4 0 0

0 0

OLABELS FOR DATA ITEMS:

ID DV AMT CMT EVID TIME STID TVR RBV BMI CRCL MDV

O(NONBLANK) LABELS FOR PRED-DEFINED ITEMS:

KA CL ALAG V IPRE IRES IWRE OFORMAT FOR DATA:

(E5.0.E10.0, E4.0_J2E2.0,E13.0,3E2.0,2E13.0,1 F2.0)

TOT. NO. OF DATA RECS: 17605

TOT. NO. OF OBS RECS: 5063

TOT. NO. OF INDIVIDUALS: 877

0LENGTH OF THETA: 4

0OMEGA HAS BLOCK FORM:

1

0 2

0 0 3

0 0 0 4

0SIGMA HAS SIMPLE DIAGONAL FORM WITH DIMENSION: 2 OINITIAL ESTIMATE OF THETA:

LOWER BOUND INITIAL EST UPPER BOUND

0.1000E-01 0.3310E+00 0.1000E+07

0.1000E+01 0.8640E+01 0.1000E+07

O.OOOOE+00 O.OOOOE+00 O.OOOOE+00

0.5000E-01 0.9120E+00 0.1000E+07

OINITIAL ESTIMATE OF OMEGA:

BLOCK SET NO. BLOCK

1 NO 0.6320E+00

2 NO 0.1410E+00

3 YES O.OOOOE+00

4 NO

0.8370E+00

OINITIAL ESTIMATE OF SIGMA:

0.1020E+00

O.OOOOE+00 0.2850E+00

OESTIMATION STEP OMITTED: NO

CONDITIONAL ESTIMATES USED: YES

CENTERED ETA: NO

EPS-ETA INTERACTION: YES

LAPLACIAN OBJ. FUNC: NO

NO. OF FUNCT. EVALS. ALLOWED: 2000

NO. OF SIG. FIGURES REQUIRED: 3

INTERMEDIATE PRINTOUT: YES

ESTIMATE OUTPUT TO MSF: NO

ABORT WITH PRED EXIT CODE 1 : NO

IND. OBJ. FUNC. VALUES SORTED: NO

OTABLES STEP OMITTED: NO

NO. OF TABLES: 2

0-- TABLE 1 -

PRINTED: NO

HEADERS: ONE

FILE TO BE FORWARDED: NO

OUSER-CHOSEN ITEMS

IN THE ORDER THEY WILL APPEAR IN TH E TABLE: ID CMT TIME EVID AMT IPRE IRES IWRE

0-- TABLE 2 -- OFIRST RECORDS ONLY: YES

04 COLUMNS APPENDED: NO

PRINTED: NO

HEADERS: ONE

FILE TO BE FORWARDED: NO

0USER-CHOSEN ITEMS

IN THE ORDER THEY WILL APPEAR IN THE TABLE:

ID ETA1 ETA2 ETA3 ETA4 KA CL V ALAG STID TVR RBV BMI CRCL 1 DOUBLE PRECISION PREDPP VERSION V LEVEL 1.0

ONE COMPARTMENT MODEL WITH FIRST-ORDER ABSORPTION (ADVAN2)

OMAXIMUM NO. OF BASIC PK PARAMETERS: 3

OBASIC PK PARAMETERS (AFTER TRANSLATION):

ELIMINATION RATE (K) IS BASIC PK PARAMETER NO.: 1

ABSORPTION RATE (KA) IS BASIC PK PARAMETER NO.: 3

0COMPARTMENT ATTRIBUTES

COMPT. NO. FUNCTION INITIAL ON/OFF DOSE DEFAULT DEFAULT

STATUS ALLOWED ALLOWED FOR DOSE FOR OBS.

1 DEPOT OFF YES YES YES NO

2 CENTRAL ON NO YES NO YES

3 OUTPUT OFF YES NO NO NO

1

ADDITIONAL PK PARAMETERS - ASSIGNMENT OF ROWS IN GG

COMPT. NO. INDICES

SCALE BIOAVAIL. ZERO-ORDER ZERO-ORDER ABSORB

FRACTION RATE DURATION LAG

t * * * 4

2 ^ * * * *

3

- PARAMETER IS NOT ALLOWED FOR THIS MODEL

* PARAMETER IS NOT SUPPLIED BY PK SUBROUTINE;

WILL DEFAULT TO ONE IF APPLICABLE ODATA ITEM INDICES USED BY PRED ARE:

EVENT ID DATA ITEM IS DATA ITEM NO.: 5

TIME DATA ITEM IS DATA ITEM NO.: 6

DOSE AMOUNT DATA ITEM IS DATA ITEM NO.: 3

COMPT. NO. DATA ITEM IS DATA ITEM NO.: 4 OPK SUBROUTINE CALLED WITH EVERY EVENT RECORD. PK SUBROUTINE NOT CALLED AT NONEVENT (ADDITIONAL OR LAGGED) DOSE TIMES.

OERROR SUBROUTINE CALLED WITH EVERY EVENT RECORD.

OERROR SUBROUTINE INDICATES THAT DERIVATIVES OF COMPARTMENT AMOUNTS ARE USED. 1

MONITORING OF SEARCH:

OPERATION NO.: 0 OBJECTIVE VALUE: 0.21214E+05 NO. OF FUNC. EVALS.: 7

CUMULATIVE NO. OF FUNC. EVALS.: 7

PARAMETER: 0.1000E+00 0.1000E+00 0.1000E+00 0.1000E+00 0.1000E+00 0.1000E+00 0.1000E+00 0.1000E+00

GRADIENT: -0.7346E+02 -0.1387E+03 -0.1252E+03 0.1493E+02 0.4249E+02 0.1594E+02 0.1056E+04 - 0.3394E+02

OPERATION NO.: 10 OBJECTIVE VALUE: 0.21 121 E+05 NO. OF FUNC. EVALS.: 8

CUMULATIVE NO. OF FUNC. EVALS.: 93

PARAMETER: 0.3146E+00 0.1200E+00 0.6514E+00 0.1636E+00 0.8492E-01 0.2441 E+00 -0.3596E-01 0.4388E+00

GRADIENT: 0.1123E+03 -0.1180E+03 0.7693E+02 0.2990E+02 -0.4237E+02 -0.2372E+02 0.2204E+03 0.6745E+02

0ITERATION NO.: 20 OBJECTIVE VALUE: 0.21110E+05 NO. OF FUNC. EVALS.: 8

CUMULATIVE NO. OF FUNC. EVALS.: 221

PARAMETER: 0.2890E+00 0.1345E+00 0.6001 E+00 0.1282E+00 0.1206E+00 0.2405E+00 -0.5414E-01 0.4124E+00

GRADIENT: 0.7713E+02 0.1443E+01 0.6070E+02 0.1923E+00 -0.1438E+01 -0.1135E+02 0.5147E+00 0.3549E+02

OPERATION NO.: 27 OBJECTIVE VALUE: 0.211 10E+05 NO. OF FUNC. EVALS.:25

CUMULATIVE NO. OF FUNC. EVALS.: 336

PARAMETER: 0.2874E+00 0.1355E+00 0.5972E+00 0.1277E+00 0.1214E+00 0.2405E+00 -0.5406E-01 0.4094E+00

GRADIENT: 0.9041 E+04 0.2328E+01 -0.8572E+04 -0.2743E-01 -0.7282E-01 0.2143E+05 0.5154E+05 - 0.1256E+05

NUMSIGDIG: 3.3 2.3 3.3 3.9 3.5 3.3 3.3 3.3

0MINIMIZATION TERMINATED

DUE TO ROUNDING ERRORS (ERROR=134)

NO. OF FUNCTION EVALUATIONS USED: 336

NO. OF SIG. DIGITS IN FINAL EST.: 2.3

ETABAR IS THE ARITHMETIC MEAN OF THE ETA-ESTIMATES,

AND THE P-VALUE IS GIVEN FOR THE NULL HYPOTHESIS THAT THE TRUE MEAN IS 0.

ETABAR: -0.35E-01 -0.28E-02 O.OOE+00 -0.74E-01 SE: 0.22E-01 0.11 E-01 O.OOE+00 0.20E-01 P VAL: 0.12E+00 0.80E+00 0.10E+01 0.23E-03

1

*************************************

******************** ********************

******************** MINIMUM VALUE OF OBJECTIVE FUNCTION ********************

******************** ********************

************************************************************************************************************************

*_**_**_**_*******_**_********_***_***_**_{******************} 2^ Ί QQ 728 ************_**_**_**_************_********_**_***_*** 1

************************************************************************************************************************ ******************** ********************

******************** FINAL PARAMETER ESTIMATE ********************

******************** ********************

************************************************************************************************************************ THETA - VECTOR OF FIXED EFFECTS PARAMETERS *********

TH 1 TH 2 TH 3 TH 4

3.97E-01 8.92E+00 O.OOE+00 1.47E+00 OMEGA - COV MATRIX FOR RANDOM EFFECTS - ETAS

ETA1 ETA2 ETA3 ETA4

ETA1

+ 6.68E-01 ETA2

+ O.OOE+00 1.47E-01 ETA3

+ O.OOE+00 O.OOE+00 O.OOE+00

ETA4

+ O.OOE+00 O.OOE+00 O.OOE+00 1.11 E+00 SIGMA - COV MATRIX FOR RANDOM EFFECTS - EPSILONS

EPS1 EPS2 EPS1

+ 7.50E-02

EPS2

+ O.OOE+00 5.29E-01

2009.10.16.1520.33 Example 3

Control stream used in population pharmacokinetics of ribavirin:

<?xml version- ' 1.0"?>

<nonmem>

<text><! [CDATA[

Z:\nonmem\rbvVX950_ABC\contiOl.rtv.model2.ctl.20091019135514.dir\control.rbv.model2.ctl.lst 2009.10.19.1355.18

; population PK of rbv

; 3 -compartment, using bayesian feedbacks from Wade, et al., Br. J. Clin. Pharm. 2006,62,710 ; 2009.08.04 - created

$PROBLEM rbvPK 3comp bayes feedbacks with complete mdl

$IN PUT ID DV AMT CMT EVID TIME LBW

SDATA Z:\Projects\vx950_ABC\data\data.nonmem.ABC.rbvPK.csv IGNORE=@

; ADVAN 12= 3 compartments with first order absorption

SSUBROUTINE ADVAN 12

SPK

KA =THETA(1)*EXP(ETA(1 ))

CL =THETA(2)*EXP(ETA(2))*(l+THETA(8)*(LBW-67))

Q3 =THETA(3)*EXP(ETA(3))

Q4 =THETA(4)*EXP(ETA(4))

V2 =THETA(5)*EXP(ETA(5))

V3 =THETA(6)*EXP(ETA(6))*(l +THETA(9)*(LBW-67))

V4 =THETA(7)*EXP(ETA(7))

S2 =V2 Fl =EXP(ETA(8))

; need to define relationships with the basic rate constants K=CL/V2

23=Q3/V2

K32=Q3/V3

24=Q4/V2

K42=Q4/V4 SERROR

Y=A(2)/V2*( 1+EPS(1 ))+EPS(2) ; assume linear and additive error DEL=0

IF(F.EQ.O) DEL= 1

1PRED=F

W=IPRED+DEL

IRES=DV-1PRED

IWRES=IRES/W

$THETA

(0.767); Ka

(19.8); CL

(34.4); Q3

(97.7); Q4

(472); V2

(4910); V3

(871 ); V4

(0.00869) ; LBW on CL

(0.01 1 ) ; LBW on V3 SOMEGA

(0.090); Ka

(0.026); CL

(0.036); Q3

(0.123); Q4

(0.176); V2

(0.137); V3

(0.032); V4

(0.048); F l SS1GMA (0. 17); proportional error

(0.5); additive error

SESTIMATION MAXEVAL=0 SIG=3 PRINT=10 NOABORT METHOD= 1 INTER

STABLE ID CMT TIME EVID AMT IPRED IRES IWRES NOPRINT ONEHEADER FILE=pred.tab STABLE ID ETA 1 ETA2 ETA3 ETA4 ETA 5 ETA6 ETA7 ETAS F l KA CL Q3 Q4 V2 V3 V4 F1RSTONLY NO APPEND NOPRINT ONEHEADER FILE=eta.tab

OTHER EMBODIMENTS

While we have described a number of embodiments of this invention, it is apparent that our basic examples may be altered to provide other embodiments which utilize the compounds and methods of this invention. Therefore, it will be appreciated that the scope of this invention is to be defined by the appended claims rather than by the specific embodiments that have been represented by way of example above.

Claims

What is claimed is:

1. A method of modeling treatment of an HCV patient with a protease inhibitor, peginterferon and ribavirin, comprising the step of:

quantifying the patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin with a viral dynamic model using at least one of Equations 1-17.

2. The method of Claim 1, wherein the patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin is quantified with a viral dynamic model using all of Equations 1 - 17.

3. The method of Claim 1 , wherein the patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin is quantified with a viral dynamic model using all of Equations 1 -7.

4. The method of Claim 1, wherein the patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin is quantified with a viral dynamic model using all of Equations 1-9.

5. The method of Claim 1 , wherein the patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin is quantified with a viral dynamic model using at least Equations 5(A), 6(A) and 7(A).

6. The method of Claim 3 or 4, further comprising quantifying the patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin with a viral dynamic model using Equations 17.

7. The method of Claim 3, further comprising quantifying the patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin with a viral dynamic model using at least one of Equations 8- 16.

8. The method of Claim 1 , wherein the quantified patient's response is at least one value selected from the group consisting of a breakthrough rate, a relapse rate and a sustained viral response (SVR) rate.

9. The method of any one of Claims 1-8, wherein the dosing regimens include a treatment duration for each of the protease inhibitor, peginterferon and ribavirin.

10. The method of any one of Claims 1-9, further comprising the step of comparing the quantified SVR rate with an intent-to-treat SVR rate.

1 1. The method of any one of Claims 1 -10, wherein the viral dynamic model includes parameters for genotype 1.

12. The method of Claim 1 1, wherein the viral dynamic model includes parameters for genotype la or lb.

13. The method of Claim 1 , wherein the peginterferon is peginterferon-alfa.

14. The method of Claim 13, wherein the peginterferon-alfa is peginterferon-alfa 2a.

15. The method Claim 13, wherein the peginterferon-alfa is peginterferon-alfa 2b.

16. The method of Claim 1 , wherein the protease inhibitor is an NS2/4A3 protease inhibitor.

17. The method of Claim 1 , wherein the protease inhibitor is an NS3/4A protease inhibitor.

18. The method of Claim 17, wherein the NS3/4A protease inhibitor is telaprevir.

19. The method of Claim 18, wherein 750 mg of telaprevir is administered three times a day.

20. The method of Claim 18, wherein 1 125 mg of telaprevir is administered twice a day.

21 . The method of any one of Claims 1 -20, wherein the patient is a treatment naive patient.

22. The method of any one of Claims 1-20, wherein the patient is a PR treatment failure patient.

23. A method of adjusting the dosing level of a composition comprising a protease inhibitor, peginterferon-alfa and ribavirin administered to a patient, the method comprising:

measuring plasma HCV RNA levels from a patient;

utilizing the measured HCV RNA levels in a multi-variant kinetic model using at least one of Equations 1-17 to calculate the responsiveness of the patient to the administered composition comprising the protease inhibitor, peginterferon- alfa and ribavirin;

comparing the calculated responsiveness to a predetermined responsiveness to compositions comprising the protease inhibitor, peginterferon- alfa and ribavirin;

and adjusting the dosing level.

24. The method of Claim 23, wherein the measured HCV RNA levels are utilized in a multi-variant kinetic model using all of Equations 1-17.

25. The method of Claim 23, wherein the measured HCV RNA levels are utilized in a multi-variant kinetic model using all of Equations 1-7.

26. The method of Claim 23, wherein the measured HCV RNA levels are utilized in a multi-variant kinetic model using all of Equations 1 -9.

27. The method of Claim 23, wherein the measured HCV RNA levels are utilized in a multi-variant kinetic model using at least Equations 5(A), 6(A) and 7(A).

The method of Claim 25 or 26, further comprising utilizing the measured HCV RNA levels in a multi-variant kinetic model using Equation 17.

The method of Claim 25, further comprising utilizing the measured HCV RNA levels in a multi-variant kinetic model using at least one of Equations 8-16.

The method of Claim 23, wherein the protease inhibitor is an NS2/3 protease inhibitor.

The method of Claim 23, wherein the protease inhibitor is an NS3/4A protease inhibitor.

32. The method of Claim 31 , wherein the NS3/4A protease inhibitor is telaprevir.

33. The method of Claim 23, further comprising adjusting the dosing level of the composition comprising the protease inhibitor administered to a patient based upon the comparison o f the calculated responsiveness to the predetermined responsiveness.

34. The method of Claim 23, wherein the multi-variant kinetic model accounts for one or more of HCV genotype 1 resistant variants.

The method of Claim 34, wherein the HCV genotype 1 resistant variant contains a mutation at one or more of an amino acid position selected from position 155, 54, 36, 156 and 155.

The method of Claim 35, wherein the one or more of HCV genotype 1 resistant variant is selected from R155M, T54A, T54S, V36M, R155K, V36A, A156S, R155T, V36M/R155K, A156T, A156V, and V36M/T54S.

The method of Claim 23, wherein utilizing the measured HCV RNA levels in a multi-variant kinetic model to calculate the responsiveness of the patient to the administered composition comprising a protease inhibitor, peginterferon-alfa and ribavirin includes determining the fitness.

The method of Claim 23, wherein the plasma HCV RNA levels from a patient are measured within the first 20 days of administration.

The method of Claim 23, wherein the measured HCV RNA levels are utilized in the multi-variant kinetic model to calculate the initial responsiveness of the patient to the administered composition comprising a protease inhibitor, peginterferon- alfa and ribavirin.

The method of Claim 23, wherein the initial responsiveness is compared to a predetermined responsiveness and based upon that comparison calculating a concentration of the protease inhibitor to be subsequently administered.

A computer system for modeling treatment of an HCV patient with a protease inhibitor, peginterferon and ribavirin, comprising a computer-readable medium storing a computer program for quantifying a patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin with a viral dynamic model using at least one of Equations 1-17 to provide quantified patient's response to the dosing regimens.

The computer system of Claim 41 , wherein the patient's response is quantified using all of Equations 1 -17.

The computer system of Claim 41 , wherein the patient's response is quantified using all of Equations 1 -7.

The computer system of Claim 41 , wherein the patient's response is quantified using all of Equations 1 -9.

45. The computer system of Claim 41 , wherein the patient's response is quantified using at least Equations 5(A), 6(A) and 7(A).

46. The computer system of Claim 43 or 44, wherein the patient's response is quantified using Equation 17.

47. The computer system of Claim 43, wherein the patient's response is quantified using all of Equations 8-16.

48. The computer system of Claim 41 , wherein the quantified patient's response is at least one value selected from the group consisting of a breakthrough rate, a relapse rate and a sustained viral response (SVR) rate.

49. The computer system of any one of Claim 41 or 42, wherein the dosing regimens include a treatment duration for each of a STAT-C agent, peginterferon and ribavirin.

50. The computer system of any one of Claims 41-43, further comprising the step of comparing the quantified SVR rate with an intent-to-treat SVR rate.

51. The computer system of any one of Claims 41-43, wherein the viral dynamic model includes parameters for genotype 1 .

52. The computer system of Claim 51 , wherein the viral dynamic model includes parameters for genotype la or lb.

53. The computer system of Claim 41 , wherein the protease inhibitor is an NS2/3 protease inhibitor.

54. The computer system of Claim 41, wherein the protease inhibitor is an NS3/4A protease inhibitor.

55. The computer system of Claim 54, wherein the NS3/4A protease inhibitor is telaprevir.

56. The computer system of Claim 41, wherein the peginterferon is peginterferon-alfa.

57. The computer system of Claim 56, wherein the peginterferon-alfa is

peginterferon-alfa 2a.

58. The computer system of Claim 56, wherein the peginterferon-alfa is

peginterferon-afa 2b.

59. The computer system of Claim 55, wherein 750 mg of telaprevir is administered three times a day.

60. The computer system of Claim 55, wherein 1 125 mg of telaprevir is administered twice a day.

61. The computer system of Claim 41 , wherein the patient is a treatment naive patient.

62. The computer system of Claim 41 , wherein the patient is a PR treatment failure patient.

63. The method of Claim 4, further comprising quantifying the patient's response to one or more dosing regimens of the protease inhibitor, peginterferon and/or ribavirin with a viral dynamic model using at least one of Equations 10-16.

64. The method of claim 26, further comprising utilizing the measured HCV RNA levels in a multi-variant kinetic model using at least one of Equations 10-16.

65. The computer system of Claim 44, wherein the patient's response is quantified using all of Equations 10-16.