Hepatitis C virus (HCV) is the leading indication for liver transplantation (LT) in the developed world.1 Recurrence is universal after transplantation, and the impact of immunosuppression is believed to be central to the accelerated course of liver injury from recurrent HCV after transplantation.2, 3 Unlike the smoldering progression of HCV to cirrhosis within 25 to 30 years of the primary infection in roughly 20% of infected individuals, posttransplant HCV recurrence results in a rapid rise in HCV RNA that peaks within 1 to 3 months, and it results in acute lobular hepatitis in 60% to 80% of patients within a median of 4 to 6 months after transplantation and in accelerated fibrosis and cirrhosis in 20% of patients within 5 years after transplantation.4-7 Understanding the impact of immunosuppressive drugs on HCV infections after LT is critical for improving outcomes both through the mitigation of the accelerated course of liver injury (if this is possible) and through the facilitation of the ability of current or future therapeutics to clear the virus. Calcineurin inhibitors remain central to effective immunosuppression in the liver recipients (>90%) to which they are currently administered; therefore, their impact on HCV replication is a critical factor in posttransplant HCV disease.
Cyclosporine A (CSA) has a well-documented effect of reducing HCV replication in vitro8; tacrolimus (TAC) does not. The treatment of acute rejection with corticosteroid boluses is an important factor in HCV recurrence and graft injury.9 TAC is the calcineurin inhibitor administered to more than 80% of liver recipients in North America at present. The balance between effective immunosuppression and pressure on HCV replication is an important factor in the long-term outcomes of HCV patients. Clinical evaluations have yielded conflicting results about the relative impacts of CSA and TAC on HCV replication and disease recurrence.10, 11 CSA may result in higher sustained virological response (SVR) rates when it is administered with pegylated interferon (IFN)/ribavirin therapy for HCV recurrence.12 Well-controlled studies that can answer these critical questions have been prohibitively difficult to achieve in the clinic. We studied the impact of CSA and TAC (alone and with the concomitant administration of IFN) on HCV replication rates in our chimeric mouse model of HCV infection.
MATERIALS AND METHODS
Animals and Care
The recipient animals, which were homozygous albumin/urokinase plasminogen activator (uPA)/severe combined immunodeficient (SCID) mice (hereafter called SCID/uPA mice) were kept virus- and antigen-free and received care in accordance with the 1993 guidelines of the Canadian Council on Animal Care. Experimental protocols were reviewed and approved by the health sciences animal welfare committee of the University of Alberta.
Isolation and Transplantation of Human Hepatocytes
Ethical approval for human tissue use was obtained from the research ethics board of the University of Alberta Faculty of Medicine, and informed consent was obtained from all donors. Segments of human liver tissue (∼20 cm3) were flushed with cold phosphate-buffered saline and were rapidly transported to the tissue isolation laboratory. Hepatocytes were isolated and purified by collagenase-based perfusion with a 0.38 mg/mL Liberase CI solution (Boehringer, Mannheim, Germany) according to previously described techniques.13 The recipient mice (5-14 days old) were anesthetized with isoflurane/O2, and viable hepatocytes (1 × 106) were embolized to the liver by an injection into the inferior pole of the spleen.
Human Alpha-1-Antitrypsin (hAAT) Analysis
hAAT analysis was used to confirm the stable ongoing function of the human hepatocyte grafts and to determine whether any changes in HCV titers were due to hepatocyte death or injury. Samples of mouse serum (2 μL) were diluted in a blocking buffer (1/100) and were analyzed with a sandwich enzyme-linked immunosorbent assay; a polyclonal goat anti-hAAT antibody (#81902, DiaSorin, Stillwater, MN) was used as the capturing antibody. A portion of the same antibody was crosslinked to horseradish peroxidase (#31489, Pierce, Rockford IL) with signal detection by 3,3′,5,5′-tetramethylbenzidine (Sigma).
HCV RNA Isolation and Quantitation
A murine serum analysis was performed in a blinded fashion by KMT Hepatech, Inc. (Edmonton, Canada). Viral RNA was extracted from 30-μL aliquots of mouse serum by a guanidine extraction method [purchased in a kit format from Qiagen (Buffer AVL, catalog no. 19073)] and was used according to the manufacturer's directions. The RNA was transcribed to complementary DNA by an HCV-specific primer (5′-aggtttaggattcgtgctcat) with a high-capacity RNA to complementary DNA kit (catalog no. 4387406, Applied Biosystems) according to the manufacturer's directions. Real-time polymerase chain reaction was performed with a real-time polymerase chain reaction system (model 7300, Applied Biosystems) and TaqMan chemistry; all measurements were performed in duplicate. The HCV-specific detection probe was 6-FAM-CACCCTATCAGGCAGTACCACAAGGCC-TAMRA, and it was used with a primer set that detected the conserved 5′-untranslated region of HCV (5′-TGCGGAACCGGTGAGTACA, 5′-aggtttaggattcgtgctcat). For absolute quantitation, a standard curve of known dilutions of a plasmid containing the sequence for HCV variant H77c (pCV-H77c) was created. Internal assay controls, which consisted of samples from a panel containing known amounts of HCV calibrated against the World Health Organization international standard for HCV RNA (OptiQuant HCV RNA panel, catalog no. 94-2011, AcroMetrix), were included in each assay. The lower limit of quantification was 300 IU/mL. HCV titers are reported as base-10 logarithms.
Drugs and Administration
CSA (Sandimmune IV, Novartis, Montreal, Canada), TAC (Prograf, Astellas Pharma Canada, Inc., Mississauga, Canada), and IFN-α2b (Intron A, Schering, Pointe-Clare, Canada) were purchased from the outpatient pharmacy of the University of Alberta Hospital (Edmonton, Canada). CSA and TAC were diluted in olive oil and were administered by oral gavage twice daily at dosages of 6.25 and 0.5 mg/kg, respectively. IFN was administered daily by subcutaneous injection (1350 IU/kg).
Six weeks after hepatocyte transplantation, the mice were screened for serum hAAT, and animals whose levels were greater than the cutoff of 100 μg/mL were inoculated with a 100-μL intraperitoneal injection of genotype 1a HCV–laden human serum (2 × 106 IU/mL). All mice received sera from the same donor sample. The baseline HCV levels and hAAT levels were obtained 1 and 2 weeks after the inoculation. The mice were allocated to treatment groups according to the week 1/baseline values. The allocation was designed to balance the groups with respect to the HCV titers, hAAT levels, sexes, and weights (in decreasing priority). A second blood draw and an HCV/hAAT assay were performed 2 weeks after the inoculation, and these measurements served as the baseline for assessing the impact of therapy. Only mice with HCV values greater than 3000 IU/mL and with hAAT levels greater than 100 ng/mL were allocated to active-intervention experimental groups; this allowed a minimum 1-log drop in HCV titers to be accurately assessed (10-fold greater than the assay cutoff for the quantitation of HCV RNA). Because no changes in HCV titers were expected in the vehicle-treated control mice, the cutoff was 600 copies in this group (2-fold greater than the sensitivity of the assay).
Pilot Pharmacokinetic/Toxicity Studies
CSA and TAC were administered to uninfected SCID/uPA mice with low hAAT levels to establish the dosages for the studies. The target blood levels were 800 to 1200 ng/mL for CSA (the 2-hour postdose level) and 8 to 10 ng/mL for TAC (the 12-hour trough whole blood level). These levels represent our center's target whole blood values in the early weeks after LT.
The blood levels of the immunosuppressant drugs were determined by liquid chromatography/tandem mass spectrometry at the clinical laboratory of the University of Alberta Hospital (Alberta Health Services Regional Client Response Services, Department of Laboratory Medicine and Pathology, University of Alberta Hospital).
Mice from 4 sequential human hepatocyte transplant cohorts were allocated to each of 6 experimental groups (7-9 mice in each group). The 6 experimental groups received saline (n = 7), CSA (n = 8), TAC (n = 8), IFN (n = 7), CSA and IFN (n = 9), or TAC and IFN (n = 9) for 4 weeks. Blood (100 μL) was drawn from the animals after 1, 2, and 4 weeks of drug administration and from the surviving animals 2 weeks after the cessation of drug administration for quantitative HCV real-time polymerase chain reaction and hAAT assays. After the fourth week of drug administration, exsanguination by cardiac puncture was performed in 12 mice (5 on CSA with or without IFN, 5 on TAC with or without IFN, and 2 on IFN alone) to provide sufficient blood for assays of CSA and TAC levels so that we could confirm that the target levels were achieved in the treatment groups.
The data are presented as means and standard errors of the mean unless specified otherwise. The decline in the viral load was the primary endpoint of the study. Viral load changes were measured in a continuous fashion, and a viral titer decrease of 1 log (10-fold) or more at 4 weeks of follow-up was used as the primary endpoint. This endpoint was selected because of the limitations of the sensitivity of the HCV quantitative assay (a 10-fold or 1-log change in this case). The statistical analysis was performed with the chi-square test; Fisher's exact test was used when it was appropriate. Differences in HCV titers between the groups were compared with Mann-Whitney U tests or Kruskall-Wallis tests when they were appropriate.
The experimental design of this study allowed the detection of a 20-fold (1.3-log) difference between groups, and the probability of detecting a treatment difference was 80% at a 2-sided significance level of 0.05. All statistical analyses were performed with SPSS 15.0 for Mac (SPSS, Inc., Chicago, IL)
CSA, cyclosporine A; hAAT, human alpha-1-antitrypsin; HCV, hepatitis C virus; IFN, interferon; LT, liver transplantation; SCID, severe combined immunodeficient; SVR, sustained virological response; TAC, tacrolimus; uPA, urokinase plasminogen activator.
All drug treatment regimens were well tolerated by the chimeric mice; no discontinuations were required. One mouse on CSA alone was found dead after the 2-week blood draw; the results for week 2 were carried forward to week 4 for the analysis of this animal.
CSA-treated mice that were sacrificed 2 hours after dosing in week 4 had a whole blood level of 951 ± 281 ng/mL, whereas the mice that were treated with CSA and IFN had a mean level of 829 ± 225 ng/mL. The mice that were treated with TAC (with or without IFN) had a whole blood level of 7.9 ± 2.7 ng/mL. The mice treated with IFN alone had undetectable levels of CSA and TAC. The health and function of the human hepatocyte grafts were monitored with human serum protein (ie, the hAAT level) throughout the treatment. The hAAT levels declined at an average rate of 4% (the TAC/IFN group) to 10% (the CSA/IFN group) per week during the 1 month of the study; this was well within the expectations for this model, and there were no significant differences between the groups.
The changes in the HCV titers are listed in Table 1 and are illustrated in Fig. 1. The vehicle-treated mice remained within ±0.4 log of the baseline HCV titer during the follow-up. Similarly to several previous studies using the chimeric mouse model,14, 15 the mice treated with IFN alone demonstrated a rapid first-phase decline in the HCV titer during the first week, and this was followed by a more gradual decline, which reached 1.38 log after 4 weeks of treatment.
Table 1. Changes in Log HCV Titers
NOTE: The presented values are the changes from the baseline log serum HCV titers and are presented as means and standard errors of the mean.
In the non–IFN-treated groups, the baseline HCV titers were as follows: 3.2 ± 0.2 for the vehicle group, 4.8 ± 0.4 for the CSA group, and 4.2 ± 0.3 for the TAC group. The baseline HCV titer was significantly lower in the vehicle group versus the CSA group (P = 0.02), but there was no significant difference between the HCV titers of the CSA and TAC groups or between the HCV titers of the vehicle and TAC groups (P = 0.6 for the CSA group versus the TAC group and P = 0.1 for the vehicle group versus the TAC group). There was no significant difference in the HCV titers after 4 weeks of treatment between the groups (log HCV titers: 3.5 ± 0.3 for the vehicle group, 4.4 ± 0.6 for the CSA group, and 4.3 ± 0.4 for the TAC group, P = 0.3).
Three of the 8 mice (37.5%) that were treated with CSA alone demonstrated a 1-log or greater decline in the HCV titer at 4 weeks of follow-up, whereas 1 of the 8 TAC-treated animals (12.5%) did (odds ratio = 0.234, 95% confidence interval = 0.02-3.1, P = 0.59). None of the animals treated with saline alone achieved the outcome cutoff of a 1-log drop in the HCV titer. This difference did not achieve statistical significance (Fig. 1A).
When the 25 IFN-treated animals were considered [IFN alone (7), CSA and IFN (9), and TAC and IFN (9)], the baseline HCV titers were as follows: 4.6 ± 0.4 for the IFN group, 6.1 ± 0.4 for the CSA/IFN group, and 5.0 ± 0.4 for the TAC/IFN group (P = 0.07). Although the HCV titers differed between the test groups, this difference did not achieve statistical significance (P = 0.1 for the difference between the CSA/IFN and TAC/IFN groups). The treatment with the combination of CSA and IFN led to an HCV titer drop pattern that was remarkably similar to the pattern achieved with IFN alone; −1.31 log was reached by week 4 (P = 0.66 at 1 week, P = 0.51 at 2 weeks, and P = 0.85 at 4 weeks for the CSA/IFN group versus the IFN-only group). Treatment with TAC and IFN resulted in declines that were similar to but slightly less than those achieved with IFN alone or CSA/IFN (P = 0.4 at 1 week, P = 0.3 at 2 weeks, and P = 0.2 at 4 weeks for the TAC/IFN group versus the IFN-only group; Fig. 1B). The HCV titers at 4 weeks of treatment were as follows: 3.2 ± 0.3 for the IFN group, 4.7 ± 0.4 for the CSA/IFN, and 4.0 ± 0.5 for the TAC/IFN group (P = 0.07). There was no significant difference in the 4-week HCV titers of the CSA/IFN and TAC/IFN groups (P = 0.6).
Six of the 7 animals in the IFN-only group (85.7%) demonstrated an HCV titer decline of 1 log or higher. There was no difference in the test groups: in both the CSA/IFN and TAC/IFN cohorts, 6 of 9 animals (66.7%) achieved the primary outcome (P = 1).
The recognition of deteriorating rates of graft and patient survival after LT for HCV in the 1990s by Berenguer et al.,7 despite improving outcomes for other indications, led to questions about the impact of changing immunosuppression practices on HCV recurrence. The impact of steroid boluses and potent anti-lymphocyte induction was confirmed,9 and a possible adverse impact of TAC versus CSA was hypothesized. These impressions were reinforced by the demonstration of the in vitro impact of CSA on HCV replication in the replicon model8 and the elucidation of the mechanism of action (the inhibition of the action of cyclophilin, which is an important host cell cofactor in HCV replication).16
The clinical evidence has been conflicting. Inoue et al.17 were the first to suggest a clinical benefit from CSA. One hundred twenty nontransplant patients who were infected with HCV received IFN monotherapy or IFN with CSA. The SVR rates were significantly higher in the combination therapy group (42/76 versus 14/44, P = 0.01). Firpi et al.12 reported significantly higher SVR rates after IFN/ribavirin treatment in a retrospective review of patients with histological progression of recurrent HCV after LT when the maintenance immunosuppression treatment involved a CSA-based regimen versus a TAC-based regimen (46% versus 27%, P = 0.03). In apparent contrast, Martin et al.10 reported from the University of California Los Angeles that patients who underwent transplantation for HCV cirrhosis had superior survival rates when they were treated with TAC instead of CSA. A subsequent meta-analysis of calcineurin immunosuppression after LT for HCV failed to support a statistically significant benefit of one agent versus the other.11
This confusing picture of clinical outcomes with CSA and TAC immunosuppression after LT for HCV and the potential for a selective impact on patients undergoing IFN-based therapy for HCV recurrence after transplantation led us to evaluate the impact of these calcineurin inhibitors on HCV titers in our chimeric mouse model of HCV infection. This model was developed to facilitate the study of the biology and therapeutics of HCV through the crossing of an immunodeficient SCID mouse with a mouse carrying the uPA activator transgene linked to an albumin promoter. The SCID background allows the xenografting of human hepatocytes without rejection, whereas the transgene is targeted to the liver by the albumin promoter; it results in a form of subacute liver failure that both creates room for the human hepatocytes to proliferate and stimulates high levels of hepatocyte growth factor to stimulate the proliferation. By 8 to 12 weeks of age, the mice develop livers that are up to 90% human according to an immunohistochemical evaluation, and they support HCV infections with reasonably stable titers for several months.13
The levels of human hepatocyte chimerism and HCV titers can vary considerably between chimeric mice, but they are reasonably stable across time in any individual mouse for at least 8 to 10 weeks after the maximum hepatocyte xenograft expansion. The peak of hepatocyte graft expansion typically occurs within 8 to 10 weeks of transplantation. Although our understanding of the factors influencing HCV titers in the mouse model is incomplete, we do know that the level of human hepatocyte chimerism in each mouse is a major influence (as we also believe for natural killer cell activity). This period of relative stability of the human hepatocyte graft (as reflected by serum hAAT levels) and HCV titers is the optimal time for intervention studies.
The model has been validated for the prediction of outcomes of anti-HCV therapeutics (including combinations with IFN) in a clinical study14, 15 and has been used for the evaluation of cyclophilin inhibitor candidates during clinical development18 as well as other putative HCV therapeutics with a wide range of different action mechanisms. Therefore, we felt that this mouse model could provide insights into the potential differential impacts of calcineurin inhibitors on HCV in vivo.
Both CSA and TAC, which were administered by oral gavage alone and in combination with subcutaneous IFN, were well tolerated by the study animals; this allowed a full 4-week course of both drugs at dosages that achieved target blood levels equivalent to maintenance immunosuppression levels achievable in clinical practice. Because of the limitations of the quantitation assay (which had a 1-log sensitivity limit), it was unclear whether a smaller difference (<1-log or 10-fold) in continuous tests would be statistically or clinically meaningful. We sought to better understand our results by measuring the number of subjects that had achieved at least a 10-fold (1-log) decline in the viral titer. We saw no consistent beneficial impact of CSA on HCV levels in the chimeric mice. A trend of lower HCV titers was observed after CSA treatment, but this was not significantly different from what was observed for the mice receiving TAC or for the vehicle-treated controls. Similarly, the HCV titers were nearly identical for the IFN-only and CSA/IFN groups at all time points. No evidence was seen for an additive impact of CSA when it was administered at dosages that led to clinically appropriate blood levels of the drug.
The HCV titer decline with the TAC/IFN combination showed a minor trend of less impact with IFN alone (P = 0.2 for the TAC/IFN group versus the IFN-only group at 4 weeks and P = 0.4 for the TAC/IFN group versus the CSA/IFN group at 4 weeks). Although the differences did not reach statistical significance, this trend in a study with limited numbers raises the possibility that TAC may have a modest negative impact on the response to IFN. This would agree with the possibility that the lower IFN-induced SVR rate with concomitant TAC immunosuppression in select clinical studies has been due to a lesser negative effect of CSA rather than a positive impact. However, Pan et al.19 recently reported that TAC did not interfere with the IFN-induced suppression of HCV in the Huh7 subgenomic replicon or with the de novo production of viral particles in the infectious HCV model.19
Our results support more recent clinical evaluations that have not demonstrated a significant clinical benefit of CSA-based immunosuppression for LT patients with HCV recurrence. Berenguer et al.20 reported a prospective, pseudorandomized series of 253 HCV-positive patients who underwent transplantation at a single center between 2001 and 2007 and were treated with CSA- or TAC-based immunosuppression with routine protocol biopsy at 1 year. Patients receiving CSA (136) and patients receiving TAC (117) demonstrated no significant differences in a range of virological and clinical outcomes. Equivalent outcomes were seen for severe diseases (ie, bridging fibrosis, cirrhosis, fibrosing cholestatic hepatitis, or allograft loss or death within 1 year; 27% versus 26%), fibrosis on 1-year protocol biopsy samples (75% versus 70%), advanced fibrosis (30% versus 25%), and survival at 1 and 7 years (83%/67% versus 78%/64%); all P values were nonsignificant. These investigators commented that their study was underpowered despite its substantial size. Their power calculations demonstrate how difficult (and thus unlikely) it would be to achieve an adequately powered study in this area: they stated that 1192 patients would be needed in each arm to achieve a statistical power of 0.80 for a 5% difference in the primary outcome.
One of the limitations of this study is the various levels of HCV titers at the baseline (ie, before the initiation of the drug treatment); they were significantly lower in the control group of the non–IFN-treated arm (P = 0.01) and were nonsignificantly lower in the IFN-treated arm (P = 0.07). However, the HCV titers did not significantly differ in our test groups at the baseline, although there was a nonsignificant trend toward higher baseline HCV levels in the CSA arm versus the TAC arm. Our chimeric mouse study results do not completely rule out the potential for higher SVR levels in clinical practice when CSA is used during IFN/ribavirin therapy. The chimeric mouse model lacks an adaptive immune system; this component of the immune response may have an impact on SVR rates due to therapeutic interventions in the clinical setting. The lack of any change in the rate of decline in HCV titers in the chimeric mouse model does not provide support for predicting such benefits, however, because of the strong correlation in clinical trials between rapid virological response rates and ultimate SVR rates. In parallel, Firpi et al.21 recently reported a prospective, randomized, single-center study of calcineurin inhibitor immunosuppression in IFN-naive patients with posttransplant HCV recurrence who were undergoing IFN/ribavirin therapy. After 72 weeks of treatment for 20 TAC patients and 18 CSA patients, no differences were seen in the early virological response rates (55% for TAC versus 72% for CSA, P = 0.2), the rates of viral clearance at the end of treatment (45% versus 50%), or the relapse rates (22% in both groups). The intention-to-treat SVR rates were 35% and 39%, respectively (P = not significant).
The limitations of our study include the modest numbers of chimeric mice, the limited duration of therapy, and the lack of a fully functional immune system in the chimeric mouse model. The study is, therefore, potentially underpowered to detect differences; this is a practical reality imposed by the substantial costs and time required to generate the chimeric mice required to support in vivo HCV infections. Multiple donor and recipient factors play roles in the progression of HCV disease after transplantation and in the impact of IFN therapy in the posttransplant clinical setting; this model simply assesses the direct impact of immunosuppressive drugs and IFN on HCV replication levels. Although human data are based on SVRs (after 6-12 months of therapy and 6 months follow-up) rather than rapid virological responses, there is a high correlation between the 2 outcomes. A study requiring daily IFN injections and twice daily gavage with immunosuppressive drugs in an immunodeficient transgenic mouse is also limited by the practicality of animal health over longer periods.
In summary, we have demonstrated that this SCID/uPA mouse model can achieve and maintain immunosuppressive drug levels (CSA and TAC) that are representative of clinical drug target levels during a 4-week period. The CSA cohort had a greater though nonsignificant decline in HCV viral titers during the study period in comparison with the TAC group; however, there was no difference between the 2 groups of IFN-treated animals. These animal data partially corroborate the well-documented findings suggesting that CSA inhibits HCV replication in a subgenomic replicon model; however, this did not translate into a better response to IFN in this study.
Our animal data have failed to provide any robust evidence supporting a change to CSA immunosuppression in HCV-infected patients undergoing LT or indeed a conversion to CSA in posttransplant patients with HCV recurrence who are going to undergo IFN-based therapy. The potential for a benefit from a change to CSA during IFN-based anti-HCV therapy remains debatable. We can hope that improved outcomes with the next generation of anti-HCV therapies will lessen the importance of this debate. Interestingly, in vitro outcomes,23 clinical trial evaluations,24 progress in clarifying the mechanism of action,25-27 and evidence for a higher barrier to resistance development28 all suggest that the cyclophilin inhibitor class of anti-HCV agents may have an important role to play.