N.M.K. holds a Wyeth/Canadian Institute for Health Research Chair in Transplantation Research and is a Senior Scholar of the Alberta Heritage Foundation for Medical Research.
Anti-HCV therapies in chimeric scid-Alb/uPA mice parallel outcomes in human clinical application†
Article first published online: 25 MAY 2006
Copyright © 2006 American Association for the Study of Liver Diseases
Volume 43, Issue 6, pages 1346–1353, June 2006
How to Cite
Kneteman, N. M., Weiner, A. J., O'Connell, J., Collett, M., Gao, T., Aukerman, L., Kovelsky, R., Ni, Z.-J., Hashash, A., Kline, J., Hsi, B., Schiller, D., Douglas, D., Tyrrell, D. L. J. and Mercer, D. F. (2006), Anti-HCV therapies in chimeric scid-Alb/uPA mice parallel outcomes in human clinical application. Hepatology, 43: 1346–1353. doi: 10.1002/hep.21209
Potential conflict of interest: Dr. Tyrrell is an officer and shareholder in Virexx. Dr. Mercer owns stock in KMT Hepatech.
- Issue published online: 25 MAY 2006
- Article first published online: 25 MAY 2006
- Manuscript Accepted: 22 MAR 2006
- Manuscript Received: 1 JUN 2005
- KMT Hepatech, Inc.
- Chiron, Inc.
Compounds with in vitro anti-hepatitis C virus (HCV) activity are often advanced directly into clinical trials with limited or no in vivo efficacy data. This limits prediction of clinical efficacy of compounds in the HCV drug pipeline, and may expose human subjects to unnecessary treatment effects. The scid-Alb-uPA mouse supports proliferation of transplanted human hepatocytes and subsequent HCV infection. Cohorts of genotype 1a HCV-infected mice were treated with interferon α-2b(IFN-α), BILN-2061 (anti-NS3 protease), or HCV371 (anti-NS5B polymerase). Mice treated with 1350IU/g/day IFN-α intramuscularly for 10 to 28 days demonstrated reduced viral titers compared with controls in all five experiments (P < .05, t test); viral titers rebounded after treatment withdrawal. A more pronounced antiviral effect with IFN-α was seen in genotype 3a–infected mice. Pilot studies with BILN2061 confirmed exposure to 10X replicon EC50 at trough and reduced viral titer over 2 log at 4 days. In a second 7-day study, mean HCV RNA titers dropped 1.1 log in BILN2061-treated animals, 0.6 log in IFN-treated mice, and rose 0.2 log in controls (P = .013, ANOVA). Pre-existing mutants with partial resistance to BILN2061 were identified by sequencing both the human inoculum and sera from treated mice. The polymerase inhibitor HCV371 yielded a decline in HCV titers of 0.3 log relative to vehicle-treated controls (P = NS). Performance of all three antiviral regimens in the chimeric mouse model paralleled responses in humans. In conclusion, this system may help selection of lead compounds for advancement into human trials with an increased likelihood of clinical success while broadening the tools available for study of the biology of HCV infection. (HEPATOLOGY 2006;43:1346–1353.)
Hepatitis C virus (HCV) is a pandemic viral disease affecting 170 million people worldwide.1 Despite viral RNA sequence identification in 1989,2 difficulties in developing cell-culture and small animal models have challenged the development of effective therapies. Standard interferon-based therapy remains imperfect with efficacy failure in 50%, including frequent treatment intolerance.3, 4 A validated small animal model would add significantly to the armamentarium available to researchers in developing novel antiviral compounds for advancement into clinical trials.
The scid-Alb/uPA mouse model is the most robust small animal model of hepatitis C infection and replication currently available. T- and B-cell–deficient (scid) mice carrying a tandem array of murine urokinase genes under control of an albumin promoter (Alb/uPA)5 can be used as a platform to support engraftment and proliferation of transplanted human hepatocytes.6 Essentially, animals homozygous for both the scid trait and Alb/uPA transgene are transplanted intrasplenically with fresh or cryopreserved isolated human hepatocytes taken from noninfected human donors. The transplanted cells translocate to and engraft within the murine liver. Overexpression of urokinase within the murine portion of the liver is autotoxic and produces a strong stimulus for replication of nontransgenic (in this case, human) hepatocytes. The human hepatocyte population rapidly expands and replaces much of the diseased mouse parenchyma with nondiseased human cells. Mice with chimeric human livers can then be inoculated with human serum containing infectious HCV particles and support durable replication of the virus at levels equivalent to those seen in infected humans. These mice are currently being used to study both the basic biology of HCV infection as well as small molecule antivirals and biologicals.
We describe evidence of efficacy of interferon alpha2b (IFN-α) and a small molecule HCV NS3 protease inhibitor (BILN 2061) in an in vivo murine model. The antiviral effects of both drugs are shown to parallel human clinical outcomes, as did the negative outcomes with the polymerase inhibitor HCV371. A rebound in viral RNA levels seen during therapy with the protease inhibitor was found to be at least in part due to a natural mutation in the HCV sequence existing before selection pressure from anti-protease therapy; the mutation patterns previously described in the in vitro replicon system were not seen with these short-term in vivo treatments.
Materials and Methods
BILN was synthesized by Acme Bioscience, Ltd. (Belmont, CA), and solubilized in an aqueous suspension containing 0.2% carboxymethylcellulose and 0.5% Tween 80. It was active at 1 nmol/L in enzyme inhibition assays in vitro, confirming compound quality for the subsequent in vivo experiments.
HCV371 was obtained from Wyeth Research (Pearl River, NY) after formulation by the manufacturer. HCV371 is member of a novel class of HCV NS5B RNA-dedendent RNA polymerase inhibitors with a concentration that inhibits 50% (IC50) of 50 nmol/L against HCV NS5B enzyme.7
Interferon alpha 2b (Intron A, Schering-Plough, Kenilworth, NJ) was purchased from the University of Alberta Hospital Pharmacy, Edmonton, Canada.
Recipient animals (homozygous Alb-uPA/scid mice) were housed virus/antigen free, and cared for in accordance with Canadian Council on Animal Care (1993) guidelines. Experimental protocols were reviewed and approved by the University of Alberta Health Sciences Animal Welfare Committee.
Isolation and Transplantation of Human Hepatocytes.
Ethical approval for human tissue use was obtained from the University of Alberta Faculty of Medicine Research Ethics Board, and informed consent was obtained from all donors. Segments of human liver tissue (∼20 cm3) were flushed with cold phosphate-buffered saline and rapidly transported to the tissue isolation laboratory. Hepatocytes were isolated and purified using collagenase-based perfusion with 0.38 mg/mL Liberase CI solution (Boehringer Mannheim), using previously described techniques.6 Recipient mice (5-14 days old) were anesthetized with halothane/O2, and 1 × 106 viable hepatocytes were injected into the inferior pole of the spleen.
Human α1 Antitrypsin Analysis.
Samples of mouse serum (2 μLl) were diluted 1/100 in blocking buffer and analyzed by sandwich ELISA using a polyclonal goat anti–human alpha1-antitrypsin (hAAT) antibody (#81902, Diasorin, Stillwater MN) as the capturing antibody. A portion of the same antibody was cross-linked to horseradish peroxidase (#31489, Pierce, Rockford, IL) and used as the secondary antibody, with signal detection by 3.3′,5,5′-tetramethylbenzidine (Sigma, St. Louis, MO).
HCV RNA Quantitation.
Serum analysis was performed in blinded fashion by the Alberta Provincial Laboratory of Public Health (Edmonton, Canada) using the Cobas Amplicor HCV Monitor system (Roche Diagnostics, Laval, Canada).
HCV RNA Isolation and Sequencing.
Viral Qiamp or Qiagen Qiamp Ultrasense kits (Qiagen, Valencia, CA) were used to extract plasma RNA from the initial human inoculum and from animals treated with BILN. Approximately 10% of the RNA was transcribed by Stratascript (Strategene, La Jolla, CA) reverse transcriptase (RT) in a 50-μL cDNA reaction according to the manufacturer's directions. The following set of nested polymerase chain reaction (PCR) primers were used to amplify the full-length NS3/4A (2.109 Kb) fragment from plasma RNA: external primers—5′GAG ATA CTG CTC GGG CCA GCC GA 3′ (1AN34A5′ and 5′ CCT TCT GCT TGA ACT GCT CGG CGA 3′ (1AN34A3′), followed by internal primers 5′ ATG GTC TCC AAG GGG TGG AGG T 3′ (1ANS3-5′) and 5′ TGC TCG ATG TAC GGT AAG TGC TGA 3′ (1ANS4B-3′). Each round of PCR was performed with the KOD Hotstart (Novagen, Madison WI) proof reading enzyme as directed by the manufacturer using the following amplification regimen: 94 °C 2 minutes, 94°C 15 seconds, 66°C 30 seconds, 68°C 2 minutes 50 seconds for 30 cycles followed by one cycle of 70°C 5 minutes and held at 4°C. RT-PCR products were either sequenced directly or cloned into a TOPO TA cloning kit vector (Invitrogen, Carlsbad CA) after A tailing with taq polymerase. PCR products from several cDNA reactions were sequenced to obtain an accurate view of the viral genomes in each sample.
IC50 Determination of HCV Replicons Containing Q80K Mutation.
Ten micrograms of in vitro transcribed replicon RNA was electroporated into 2 × 106 Huh7 cells as previously described.8 Cells were suspended into 26 mL Dulbecco's minimum essential medium immediately after electroporation and aliquoted 1 mL/well onto 24-well plates. Five to six hours later the medium was removed, and BILN 2061 serially diluted in complete Dulbecco's minimum essential medium was added. The concentrations of BILN2061 ranged from 0.01 nmol/L to 10 nmol/L for wild-type and Q80K replicons, 10 nmol/L to 3.3 μmol/L for D168A replicon. After 48 hours' incubation with compounds, cells were lysed for luciferase assays as previously described.8 All luciferase assays were done in duplicate; IC50 values for each replicon were determined by four independent experiments.
Cohorts of mice with HCV titers greater than 104 copies/mL were selected for use in antiviral studies. Over the course of these experiments, consistency of human hepatocyte graft function in study animals was monitored by serial assay of hAAT levels.
To study the efficacy of IFN-α in the mouse model, HCV genotype 1a–infected mice were treated with 1,350 IU/g/d IFN-α intramuscularly and compared with control animals given saline injections. This dose (10-fold the typical clinical dose) of human interferon has repeatedly demonstrated acceptable antiviral efficacy and minimal toxicity.
As a result of variable access to both human tissue and animals of appropriate age for transplantation, experimental cohorts were selected and assigned to treatment groups on availability. The first two vehicle-controlled experimental cohorts contained n = 5 and n = 4 animals produced from two different human donors and were treated with IFN-α by intramuscular injection for 14 days. They were compared with n = 7 and n = 4 controls, respectively. The results of these experiments are shown in Fig. 1A-B. Data are presented as change in log titers to normalize the experimental groups to a common baseline for graphical comparison; the starting titers and maximal drop in titers for each group are shown in Table 1. At the first two measured times (after 7 and 14 days of therapy), the IFN-α–treated animals had significantly lower viral titers than controls. In the first experiment, titers remained low after cessation of therapy, whereas in the second group, a rebound was observed. Some degree of rebound in titers after stopping IFN treatment has been seen in all subsequent experimental groups.
|Figure||Groups||n||Duration of IFN (d)||Starting Log10HCV ± SD||Log10HCV Treatment End||Absolute Δ Log at Treatment End||Δ Log Relative to Control Treatment End|
|1a||IFN||5||14||5.91 ± 0.29||4.85 ± 0.61||−1.06 ± 0.68||−1.34|
|Control||7||14||5.85 ± 0.26||5.51 ± 0.35||−0.28 ± 0.30|
|1b||IFN||4||14||5.41 ± 1.14||4.32 ± 1.79||−1.09 ± 0.83||−1.54|
|Control||4||14||5.16 ± 1.35||5.61 ± 1.14||+0.45 ± 0.35|
|1c||IFN||7||18||6.25 ± 0.68||3.91 ± 0.55||−2.30 ± 0.46||−1.74|
|Control||5||18||6.48 ± 0.95||5.92 ± 0.71||−0.56 ± 0.47|
|1d||IFN||5||10||6.01 ± 0.97||4.90 ± 0.68||−1.11 ± 0.45||−1.04|
|Control||5||10||6.22 ± 0.65||6.15 ± 0.68||−0.07 ± 0.53|
|1e||IFN||4||28||6.16 ± 0.82||4.39 ± 1.25||−1.66 ± 0.62||−1.93|
|Control||4||28||6.80 ± 0.64||6.53 ± 0.63||−0.27 ± 0.53|
Animal welfare guidelines restrict the amount of blood that may be drawn from an animal to 100 μL per week. To “interpolate” data points between the standard weekly intervals, we offset the schedule of the next group of animals (new human donor, n = 7 IFN-treated and n = 5 controls) by 4 days, to have blood draws performed at 3, 11, and 18 days. We continued therapy up to the 18-day point in these mice. These data are presented in Fig. 1C and Table 1. A significant decrease in viral titers in the treated group was observed at 3 days after starting IFN-α therapy and remained significantly below controls over the duration of the experiment. A rebound toward baseline occurred after stopping IFN therapy. Controls remained constant over the experimental course, with a modest but non-significant drop (P = .055) at 39 days compared with baseline.
The fourth cohort (new human hepatocyte donor) was again designed to explore datapoints in the first 2 weeks of therapy. Animals treated with IFN (n = 5) for 10 days were sampled at 3, 8, and 11 days, with smaller volume blood draws at each point. A similar group of five controls was assayed on the same schedule. The data are presented in Fig. 1D and Table 1. The treatment group showed the most decline in viral titers after 3 days of IFN therapy, with a plateau over the next two times. Titers in treated animals were significantly below controls at all times.
To evaluate whether viral levels might “break through” therapy with longer treatment, a fifth group of mice (n = 4) were treated daily for 28 days and compared with four control mice (Fig. 1E, Table 1). A steady decline in viral levels was seen over the treatment period (6.2 ± 0.8 to 4.4 ± 1.3), with no evidence of breakthrough. After cessation of IFN treatment, levels again rose back toward baseline. The differences between IFN-treated animals and controls were significant at all times until cessation of therapy. The viral levels in control animals began to drift downwards toward the end of the experiment (P = .061 at 42 days compared with baseline); in two of the four control animals, hAAT levels fell off after the 28-day point, suggesting that diminished graft function may have contributed to reduction in viral titers. No change in hAAT was noted in the interferon-treated group.
The effect of IFN-α against another HCV genotype was studied in animals infected with genotype 3a. In pilot studies, two mice infected with genotype 3a were treated for 2 weeks with IFN-α. Pre-treatment viral titers were 3.23 and 4.48 logs, and in both cases fell to undetectable after 2 weeks of therapy. They remained below limits of detection (600 copies/mL) for an additional 2 weeks of observation, suggesting possible clearance (Fig. 2A, Table 2). A subsequent group of animals was treated intramuscularly with IFN-α once daily over a 5-day period, with sampling at the beginning and end of treatment and were compared with four saline-treated controls taken from the same cohort of genotype 3a–infected animals. Both treatment and control groups had similar viral levels before onset of treatment (4.7 ± 1.2 vs. 4.3 ± 1.1, P = .66, t test). After only 5 days of therapy, the treatment group had fallen by 2.1 logs to 2.6 ± 0.2, whereas the control group remained unchanged at 4.4 ± 1.3 (P = .035, t test). The data for individual mice are presented graphically in Fig. 2B.
|Mouse Number||Log10HCV day 0||Log10HCV day14||Log10HCV day28|
To study the effect of small molecule antivirals in the mouse model, two molecules were accessed: an inhibitor of the HCV NS3 protease and an inhibitor of the NS5B polymerase. A limited quantity of the protease inhibitor BILN 2061 was obtained from a commercial supplier (Acme Bioscience, Ltd.) and was formulated as an aqueous suspension for oral administration. The first experimental series was conducted to evaluate tolerability and pK and to provide initial efficacy evaluation. Three HCV-infected mice (Log10 HCV titers 4.3, 4.2, and 4.0) received 10 mg/kg twice daily dosing by oral gavage. Quantitative viral RNA levels were drawn before initiating therapy and after 4 days of treatment, at which point a terminal bleed was conducted to confirm adequacy of anti-viral exposure. All three mice tolerated the twice-daily dosing well, with trough plasma concentration at 4 days averaging 45 nmol/L, approximately 10 times the median effective concentration in replicons and similar to that seen in previously published human trials9; in each case HCV RNA titers dropped below the limit of detectability of the quantitative Roche Amplicor assay (∼600 copies/mL) after 4 days' treatment.
Having established tolerability and achievement of target plasma concentrations, and with initial evidence of efficacy, a phase II confirmation study was initiated. A group of genotype 1a–infected mice were allocated to 3 groups to achieve balance in mean HCV titers, hAAT levels, and sex, and then received 14 days' treatment with either BILN 2061 (n = 6), IFN-α (1,350 IU/g/d given once daily, n = 4), or oral gavage vehicle (n = 4). Animals were monitored daily for health status.
All animals tolerated the study interventions without adverse outcome. The raw data for the experiment are presented in Table 3, and the comparison between BILN-2061 and negative controls is represented graphically in Fig. 3. Mean HCV RNA dropped by 1.05 log in BILN 2061–treated animals after 7 days' treatment (6.55 ± 0.49 to 5.50 ± 0.0.68), in comparison with a 0.18 log rise in vehicle controls (6.80 ± 0.64 to 6.98 ± 0.67); the difference between groups was significant (P = .013, ANOVA), and BILN 2061 was better than negative control (P < .05, SNK post-hoc analysis). A positive control group of four mice that received 1,350 IU/g/day IFN (once daily) demonstrated a decrease in HCV RNA titer of 0.58 log (6.16 ± 0.82 to 5.58 ± 0.68; P < .05, SNK post-hoc analysis) over the same period. A slight rise in titers was seen in the BILN-treated group after the second week despite remaining on therapy; after cessation of treatment, levels rose slightly. Negative control animals demonstrated a slight downward trend over the course of the experiment, although not statistically different from pre-treatment baseline at any point. The downward trend in HCV levels persisted in the four positive control animals that were maintained on IFN-α therapy over the duration of the 4-week experiment.
|Group||ID||Pre-Treatment||Week 1||Week 2||Week 3||Week 4|
|Negative control (vehicle)|
|Positive control (IFN)*||D849||206||6.46||181||5.40||116||5.36||170||4.04||218||4.04|
Despite the initial decrease in HCV titer in all mice treated with BILN 2061, four of six mice had varying degrees of rebound in viral titers at the 14-day point, although they remained on therapy. Because a series of mutations in NS3 (A156T, R155Q, D168V) were previously reported to result in a protease-resistant phenotype in replicons exposed to BILN2061,10, 11 we sequenced the plasma RNA from the two most pronounced rebounders in an attempt to detect these mutations in vivo. The HCV sequence in all samples (including the inoculum) had the expected wild-type residues at all three positions. A partial sequence from the liver of one rebounder showed the expected wild-type D at amino acid 168.
The virus in all three of the tested sera, including the inoculum, was found to have a lysine codon at position 80 of the NS3 gene rather than the glutamine codon typical of genotype 1a. Based on computational modeling, amino acid 80 lies near the active site of the enzyme, and amino acid changes at this site may influence compound binding.12 We introduced a Q80K mutation into the NS3 contained within a transient HCV replicon harboring the reporter gene luciferase13 and found that the antiviral activity of BILN 2061 was reduced approximately threefold relative to the wild-type replicon (Fig. 4). The Asp168 mutation previously found to be associated with resistance in the replicon was run as a parallel control in this assay and resulted in a 3 log decrease in sensitivity at 2.36 μmol/L.
The small molecule NS5B polymerase inhibitor HCV 371 is a molecule under study through an alliance between Wyeth (Pearl River, NY) and Viropharma (Exton, PA). After formulation as guided by the manufacturer, a 21-day course of subcutaneous therapy was administered to 8 mice concomitantly, with a positive control group (n = 5) receiving IFN-α at 1,350 IU/g once daily by intramuscular injection, and a negative control group (n = 8) receiving vehicle. One animal each in the HCV371 and IFN-α control groups developed clinical scores reflecting poor condition and were sacrificed and removed from analysis at subsequent times (before days 12 and 21 outcomes, respectively). IFN-α–treated mice demonstrated a typical decrease of 1.5 logs in HCV titer over 3 weeks' treatment, which was significantly different from the vehicle control group at all times (P < .05 for all; ANOVA with SNK post hoc analysis). In comparison, HCV371-treated mice declined by a maximum of 0.6 logs at 12 days of therapy (a 0.3 log drop relative to vehicle control), which was not significantly better than control over the course of the experiment. Data are presented graphically in Fig. 5. HCV371 failed to reach efficacy targets in a subsequent clinical trial.14
These data suggest that the response to HCV antiviral therapy in the scid-Alb-uPA mouse models the treatment response in humans.3 In the case of therapy with IFN-α2b, viral levels drop after administration of IFN begins, and continue to fall over the duration of therapy. Furthermore, most of the decline appears to occur in the first few days, with a subsequent more gradual decrease. In all cases, over the same periods viral levels in control animals remain statistically unchanged. This biphasic pattern is reminiscent of what is observed in human populations.15 Because the presented experiments were performed on chimeric mice produced from a number of different human donors, the antiviral effect of IFN appears to be at least in part independent of the source of human hepatocytes. Additionally, given the immunodeficient nature of the scid/Alb-uPA mice, these data also argue for a substantial direct antiviral effect of IFN.
Scid/Alb-uPA mice in our colony have been infected with HCV of genotypes 1a, 1b, 2a, 3a, 4a, and 6a. Study of the impact of interferon on viral titers has been carried out for genotypes 1a and 3a. In the chimeric mouse model, genotype 3a viral infections appear to respond to interferon therapy in a fashion somewhat distinct from 1a infections. The virus appears to be found at slightly lower copy number in genotype 3a infection, with viral RNA levels typically averaging 104 to 105 copies/mL, versus 105 to 106 in genotype 1a infection. The 3a virus in the mouse model also may be more sensitive to IFN therapy, as we observed drops in viral titer of up to 3 logs after only 5 days of therapy. In point of fact, two of the four tested animals dropped to below detectability of the qualitative RT-PCR assay (300 RNA copies/mL) before being sacrificed (Fig. 2B). Some of our earliest pilot work suggested that durable clearance of 3a virus could be achieved with IFN therapy: two genotype 3a–infected mice treated with high-dose interferon demonstrated a decrease in viral titer to below assay sensitivity after 7 days' treatment and remained with viral titers at undetectable levels during 14 further days of follow-up (Fig. 2A, Table 2). This increased sensitivity of genotype 3a HCV to interferon therapy in the mouse model again recalls what is observed with genotype 3a–infected human patients undergoing interferon treatment16, 17 and provides additional validation of the model across viral genotypes.
The protease inhibitor BILN-2061 was the first small molecule advanced into clinical testing that demonstrated efficacy against HCV.9 The compound was shown to bind the NS3 protease with a Ki of 0.3-0.4 nmol/L, and in replicon studies was effective in inhibiting viral replication with an median effective concentration of 3 to 4 nmol/L. When administered to humans (with a steady-state trough of approximately 42 nmol/L), it induced a 2- to 3-log decrease in HCV RNA titers after 48 hours of therapy. This novel compound provided a small molecule that could be validated in the chimeric mouse model of HCV infection.
After 4 days of therapy with BILN2061 in the initial experiment, HCV titers decreased by approximately 2 log, similar to the impact seen in clinical trials, and with serum BILN levels roughly equal to those used in the earlier clinical studies. In the subsequent controlled experimental series, 1 week of therapy with BILN 2061 resulted in a geographic mean drop of over 1 log in treated animals, versus a slight rise in negative controls (a relative decrease of 1.3 log with therapy). Despite remaining on therapy, however, there was a slight rise in viral titers at the end of the second week. This rise suggested the possible emergence of resistant mutants. Sequencing, however, did not reveal the A156T, R155Q, or D168V mutations previously reported to result in resistance in replicons.10, 11 We did identify a Q80K mutation that existed in the inoculum before exposure to the drug, and that attenuated the response to BILN by a factor of 3 to 4 when inserted into a replicon system. This mutation may well have contributed to the rebound seen in some of the mice. Additionally, as data are limited to the 1-week and 2-week points, we cannot interpolate viral dynamics in the surrounding intervals. It is possible, as was seen in the human studies of BILN 2061 and as suggested in the pilot experiments, that the actual drop in initial viral titers was quite rapid, and perhaps even of greater magnitude than observed at 1 week. Additional BILN-2061 could not be obtained for further study.
The scid-Alb-uPA mouse model of HCV infection is a constructed model requiring transplantation of each individual mouse with human hepatocytes. The requirement for high-quality human hepatocytes to achieve high-level human chimerism remains a significant challenge and limitation to the model. Batch-to-batch variability in human cells can theoretically challenge reproducibility of the model and create difficulties with scale-up. Because of variable availability of fresh human tissue, a breeding protocol must be used wherein animals of a suitable age for transplantation (typically 7-14 days of age) are available on a weekly basis in the event that tissue is procured. As such, cohorts of animals available for study can be of variable size according to breeding numbers, but are typically 12 to 20 animals at any one time in our colony. Animals are typically allocated between groups according to HCV titer, hAAT levels, and sex and age of animals. In this fashion, the variability both within and between groups is minimized, and statistical power is enhanced. Based on the variances seen in Table 2, although an absolute 1.05 drop in log HCV titers was observed after 1 week with BILN therapy, the study would be capable of detecting a mean drop of only 0.25 log as significant compared with negative controls (ANOVA and SNK modeling).
From these experiments, we believe the scid-Alb-uPA mouse system can model the HCV antiviral treatment response in humans with reasonable accuracy. Treatment with both IFN and BILN 2061 appears to produce similar antiviral effect in HCV- infected human hepatocytes, whether in a human or a murine host. The model also appears valid for negative predictive value as demonstrated by parallel outcomes in the mouse model and clinical trials with the polymerase inhibitor HCV371. The HCV subgenomic replicon system has enabled evaluation of putative inhibitors of the viral nonstructural proteins and provided valuable insight into in vitro efficacy of pre-clinical and clinical drug candidates.18 Because there are differences between replicons and the in vivo replicating virus,19 uncertainty is added to the prediction of which compounds will be efficacious in human trials based on replicon data alone. This mouse model of HCV infection should provide a bridge between in vitro data and human clinical trials to enhance prediction of efficacy and avoid human exposure to compounds with low likelihood of clinical efficacy.
- 14Viropharma, Inc. Press Release, (www.viropharma.com) July 15, 2003.