Direct-Acting Antivirals for Chronic Hepatitis C: Can Drug Properties Signal Potential for Liver Injury?

Abstract

Chronic hepatitis C is a major public health burden both in the United States and globally. Direct-acting antivirals (DAAs) have revolutionized the treatment landscape for patients with chronic hepatitis C virus (HCV) infection1Mishra P. Murray J. Birnkrant D. Direct-acting antiviral drug approvals for treatment of chronic hepatitis C virus infection: Scientific and regulatory approaches to clinical trial designs.Hepatology. 2015; 62: 1298-1303Crossref PubMed Scopus (23) Google Scholar and the majority of them can now be successfully cured. DAA-based regimens are highly efficacious, better tolerated, and easy to administer with shorter treatment durations. These interferon (IFN)-free regimens have provided treatment options to patients, such as those with decompensated liver disease, who had been previously ineligible to receive IFN-based regimens. However, postmarketing experience has shown that certain DAA regimens are associated with significant hepatotoxicity,2FDA Drug Safety Communication: FDA warns of serious liver injury risk with hepatitis C treatments Viekira Pak and Technivie. Available at: www.fda.gov/Drugs/DrugSafety/ucm468634.htm. Accessed July 1, 2016.Google Scholar particularly in patients with advanced liver disease, which has led to labeling revisions for some drug products after approval. This further indicates the need for additional tools to predict or identify signal of drug-induced liver injury (DILI) during premarketing evaluation of clinical trial data. We previously observed and reported that an oral medication which is highly lipophilic (ie, logP ≥ 3) and when administered at a high daily dose (≥100 mg/d) was significantly associated with DILI risk in humans,3Chen M. Borlak J. Tong W. High lipophilicity and high daily dose of oral medications are associated with significant risk for drug-induced liver injury.Hepatology. 2013; 58: 388-396Crossref PubMed Scopus (222) Google Scholar and this was termed as the “rule of two” (RO2). As described in our previous work, a drug’s lipophilicity is measured as the log value for octanol-water partition coefficient P (ie, logP) and high lipophilicity is defined as logP ≥ 3.3Chen M. Borlak J. Tong W. High lipophilicity and high daily dose of oral medications are associated with significant risk for drug-induced liver injury.Hepatology. 2013; 58: 388-396Crossref PubMed Scopus (222) Google Scholar Although the RO2 model shows the potential for compound selection during drug development,4Kaplowitz N. Avoiding idiosyncratic DILI: two is better than one.Hepatology. 2013; 58: 15-17Crossref PubMed Scopus (42) Google Scholar whether it could signal hepatotoxicity at an early stage of clinical drug development especially for drugs which will be indicated for use in patients with preexisting liver disease including those with liver dysfunction (eg, DAAs) is largely unknown. In this commentary, we applied the RO2 model to the HCV DAAs and assessed its potential utility to support research and development considerations for future therapies. The RO2 analysis presented here was performed by using recommended daily doses as per the drug label, and logP values were calculated from chemical structure using an online package (available at: www.vcclab.org/web/alogps/, developed by Virtual Computational Chemistry Laboratory) as previously described.5Chen M. Borlak J. Tong W. A model to predict severity of drug-induced liver injury in humans.Hepatology. 2016; 64: 931-940Crossref PubMed Scopus (56) Google Scholar The chemical structure of asunaprevir (ASV) was retrieved from www.sigmaaldrich.com, and structures for the other DAAs were collected from the Food and Drug Administration (FDA)-approved product labels (available at: www.accessdata.fda.gov/scripts/cder/drugsatfda/). The RO2 model was applied to the 12 available DAAs for the treatment of chronic hepatitis C at the time of this analysis (Table 1). Specifically, we analyzed 6 HCV NS3/4A protease inhibitors, including boceprevir, telaprevir, simeprevir (SMV), paritaprevir, ASV, and grazoprevir (GZR); 4 HCV NS5A inhibitors including ledipasvir (LDV), ombitasvir, daclatasvir (DCV), and elbasvir (EBR); 1 NS5B polymerase inhibitor (sofosbuvir [SOF]); and 1 non-nucleoside NS5B polymerase inhibitor (dasabuvir). Ribavirin (RBV) and ritonavir were also analyzed because they are part of some DAA regimens. The RO2 analyses are described in the chronological order of the approval of these DAAs/regimens.Table 1RO2 Model Results for HCV DAAs Based on Daily Dose and Lipophilicity (LogP)Drug ClassGeneric NameaMany drugs are approved as fixed-dose combinations: ledipasvir (fixed-dose combination product with sofosbuvir, Harvoni); paritaprevir and ombitasvir (fixed-dose combination product with ritonavir - Technivie and copackaged with dasabuvir – Viekira Pak); dasabuvir (copackaged with fixed-dose combination product ombitasvir, paritaprevir, ritonavir); elbasvir and grazoprevir (fixed dose combination product - Zepatier).Daily Dose (mg/d)LogPRO2 Model ResultHCV NS3/4A protease inhibitorsBoceprevir24001.93NegativeTelaprevir22502.56NegativeSimeprevir1504.69PositiveParitaprevir1503.50PositiveAsunaprevirbAsunaprevir is not approved in the United States.2003.12PositiveGrazoprevir1002.94NegativeHCV NS5A inhibitorsLedipasvir905.57NegativeOmbitasvir255.60NegativeDaclatasvir604.57NegativeElbasvir505.60NegativeNS5B polymerase inhibitor - nucleotideSofosbuvir4001.63NegativeNS5B polymerase inhibitor - nonnucleosideDasabuvir5004.70PositiveCYP3A InhibitorRitonavir1004.24PositiveNucleoside inhibitorRibavirin1200/1000−1.92Negative/Negativea Many drugs are approved as fixed-dose combinations: ledipasvir (fixed-dose combination product with sofosbuvir, Harvoni); paritaprevir and ombitasvir (fixed-dose combination product with ritonavir - Technivie and copackaged with dasabuvir – Viekira Pak); dasabuvir (copackaged with fixed-dose combination product ombitasvir, paritaprevir, ritonavir); elbasvir and grazoprevir (fixed dose combination product - Zepatier).b Asunaprevir is not approved in the United States. Open table in a new tab Boceprevir and telaprevir were the 2 first-generation NS3/4A protease inhibitors approved in 2011 for use in combination with pegylated IFN (PEG-IFN) and RBV. Since the recent approvals and availability of IFN-free regimens, use of PEG-IFN/RBV based regimens are not recommended by the professional societies’ guidelines.6AASLD-IDSA. Recommendations for testing, managing, and treating hepatitis C. Available at: www.hcvguidelines.org. Accessed July 1, 2016.Google Scholar Interestingly, both boceprevir and telaprevir are noted to be RO2 negatives owing to their low lipophilicity (logP < 3) even though these drugs were used at doses much higher than 100 mg/d (Table 1). In 2013, SMV was initially approved in combination with PEG-IFN and RBV and the indication was later extended for use in combination with SOF in 2014. SMV is associated with isolated increases in bilirubin levels given its inhibition of transporters involved in bilirubin’s metabolism. Postmarketing cases of markedly increased levels of bilirubin associated with hepatic decompensation and hepatic failure were reported, mostly in patients with advanced liver disease receiving an SMV-containing regimen. This led to the labeling revisions that included a warning for hepatic decompensation and hepatic failure.7US Food and Drug Administration. [email protected] drug approval package – Labeling action letter for OLYSIO (simeprevir) capsules. Available at: www.accessdata.fda.gov/drugsatfda_docs/appletter/2015/205123Orig1s008ltr.pdf. Accessed July 1, 2016.Google Scholar Furthermore, SMV is not recommended in patients with moderate or severe hepatic impairment.8US Food and Drug Administration. [email protected] drug approval package – US Prescribing Information for OLYSIO (simeprevir) capsules. Available at: www.accessdata.fda.gov/drugsatfda_docs/label/2016/205123s011lbl.pdf. Accessed July 1, 2016.Google Scholar Mean SMV exposures (ie, area under the curve at 24 hours) were 2.4- and 5.2-fold higher in patients with moderate and severe hepatic insufficiency compared with that of the pooled phase III population.9US Food and Drug Administration. [email protected] drug approval package – Clinical Pharmacology Review for OLYSIO (simeprevir) capsules. Available at: www.accessdata.fda.gov/drugsatfda_docs/nda/2013/205123Orig1s000ClinPharmR.pdf. Accessed July 1, 2016.Google Scholar Notably, SMV is a high lipophilic compound (logP = 4.69) administered at a daily dose of 150 mg and, therefore, is categorized as an RO2 positive drug. SOF in combination with PEG-IFN and RBV or with RBV was initially approved in 2013 followed by the approval of a fixed dose combination product of LDV and SOF in 2014. There have been few case reports in literature about hepatotoxicity associated with the LDV/SOF.10Tseng A. Wong D.K. Hepatotoxicity and potential drug interaction with ledipasvir/sofosbuvir in HIV/HCV infected patients.J Hepatol. 2016; 66: 651-653Abstract Full Text Full Text PDF Scopus (4) Google Scholar However, these case reports have been confounded by the use of concomitant medications with hepatotoxic potential or concurrent medical conditions. Routine postmarketing pharmacovigilance surveillance is ongoing to detect any signal of hepatotoxicity associated with these drugs. SOF is RO2 negative owing to its low logP value (1.63) even with a high daily dose (400 mg/d). Meanwhile, LDV is a highly lipophilic drug (logP = 5.57) but its daily dose (90 mg/d) is a little lower than the threshold of 100 mg/d used in the RO2 model. Viekira Pak is a fixed-dose combined package of 1 tablet that includes ombitasvir, paritaprevir, and ritonavir, and a tablet of dasabuvir. Technivie is a combination of ombitasvir, paritaprevir and ritonavir. On October 22, 2015, the US FDA issued a Drug Safety Communication warning of serious liver injury risk associated with hepatitis C treatments Viekira Pak and Technivie, which were approved by the agency in December, 2014 and July, 2015, respectively.2FDA Drug Safety Communication: FDA warns of serious liver injury risk with hepatitis C treatments Viekira Pak and Technivie. Available at: www.fda.gov/Drugs/DrugSafety/ucm468634.htm. Accessed July 1, 2016.Google Scholar Viekira Pak or Technivie underwent comprehensive clinical evaluation during the FDA review process. Although, elevations in serum alanine transaminases (ALT) were identified as a potential hepatotoxicity signal that led to a warning in the label, no serious hepatotoxicity events were observed in the clinical trial database.11US Food and Drug Administration. [email protected] drug approval package – Medical Review for Viekira Pak (ombitasvir, paritaprevir, and ritonavir tablets; dasabuvir tablets), co-packaged for oral use. Available at: www.accessdata.fda.gov/drugsatfda_docs/nda/2014/206619Orig1s000MedR.pdf. Accessed July 1, 2016.Google Scholar During clinical development, paritaprevir doses of ≥200 mg were associated with an increased risk of grade 3 increases in ALT.12US Food and Drug Administration. [email protected] drug approval package – Clinical Pharmacology and BioPharmaceutics Review for Viekira Pak (ombitasvir, paritaprevir, and ritonavir tablets; dasabuvir tablets), co-packaged for oral use. Available at: www.accessdata.fda.gov/drugsatfda_docs/nda/2014/206619Orig1s000ClinPharmR.pdf. Accessed July 1, 2016.Google Scholar Paritaprevir is known to cause liver enzyme elevations13Poordad F. Lawitz E. Kowdley K.V. et al.Exploratory study of oral combination antiviral therapy for hepatitis C.N Engl J Med. 2013; 368: 45-53Crossref PubMed Scopus (264) Google Scholar in an exposure-dependent manner. A pharmacometric analysis has demonstrated that higher paritaprevir exposures were significantly associated with ALT elevations. Multivariate logistic regression analysis predicted a 2-fold increase in paritaprevir exposure increased the odds of ALT elevations (ALT >5 times the upper limit of normal) by 1.6-fold. In moderate (Child-Pugh class B) hepatic impairment paritaprevir area under the curve values increased by 62%.12US Food and Drug Administration. [email protected] drug approval package – Clinical Pharmacology and BioPharmaceutics Review for Viekira Pak (ombitasvir, paritaprevir, and ritonavir tablets; dasabuvir tablets), co-packaged for oral use. Available at: www.accessdata.fda.gov/drugsatfda_docs/nda/2014/206619Orig1s000ClinPharmR.pdf. Accessed July 1, 2016.Google Scholar However, in subjects with severe hepatic impairment, paritaprevir and dasabuvir area under the curve values increased by approximately 945% and 325%, respectively.12US Food and Drug Administration. [email protected] drug approval package – Clinical Pharmacology and BioPharmaceutics Review for Viekira Pak (ombitasvir, paritaprevir, and ritonavir tablets; dasabuvir tablets), co-packaged for oral use. Available at: www.accessdata.fda.gov/drugsatfda_docs/nda/2014/206619Orig1s000ClinPharmR.pdf. Accessed July 1, 2016.Google Scholar With the observed increased exposures of paritaprevir, and the demonstrated exposure–response relationship of increased paritaprevir exposures and risk of ALT elevations, Viekira Pak was contraindicated in patients with severe hepatic impairment (Child-Pugh class C) owing to risk of potential toxicity, at the time of initial approval. In addition, treatment of patients with Child-Pugh class B hepatic impairment was not recommended owing to lack of safety and efficacy data in this population to support its use. Since approval, postmarketing cases of hepatic decompensation and hyperbilirubinemia, including liver transplantation or death, in patients with cirrhosis have been associated with the use of Viekira Pak or Technivie. The serious outcomes were mostly reported in patients who had evidence of advanced cirrhosis before initiating therapy with Viekira Pak.2FDA Drug Safety Communication: FDA warns of serious liver injury risk with hepatitis C treatments Viekira Pak and Technivie. Available at: www.fda.gov/Drugs/DrugSafety/ucm468634.htm. Accessed July 1, 2016.Google Scholar The postmarketing assessment led to labeling revisions, including warnings about the risk of hepatic decompensation and hepatic failure in patients with cirrhosis, as well as contraindication for use in patients with moderate and severe hepatic impairment.14US Food and Drug Administration. [email protected] drug approval package – Labeling Revision Action Letter for Viekira Pak (ombitasvir, paritaprevir, and ritonavir tablets; dasabuvir tablets), co-packaged for oral use. Available at: www.accessdata.fda.gov/drugsatfda_docs/appletter/2015/206619Orig1s008ltr.pdf. Accessed July 1, 2016.Google Scholar All 4 drugs in Viekira Pak showed a high lipophilicity with a logP ≥ 3, and 3 of them (all except ombitasvir) were given at a high daily dose of ≥100 mg/d. As shown in Table 1, the analysis reveals that paritaprevir, dasabuvir, and ritonavir were RO2 positives, suggesting a high potential of DILI risk associated with their use in patients. Ritonavir (an HIV protease inhibitor) was used at 100 mg/d dose as a CYP3A inhibitor to increase paritaprevir exposures. Ritonavir is labeled for adverse hepatic reactions, including an increased risk of transaminase elevations.15US Food and Drug Administration. [email protected] drug approval package – US Prescribing Information for ritonavir. Available at: www.accessdata.fda.gov/drugsatfda_docs/label/2015/020659s062,022417s014lbl.pdf. Accessed July 1, 2016.Google Scholar An ASV and DCV combination regimen is approved in Japan. Severe cases of hepatotoxicity were reported after marketing approval in Japan.16Fujii Y. Uchida Y. Mochida S. Drug-induced immunoallergic hepatitis during combination therapy with daclatasvir and asunaprevir.Hepatology. 2015; 61: 400-401Crossref PubMed Scopus (28) Google Scholar ASV seems to be more likely associated with hepatotoxicity findings, and it was further reported that the discontinuation or reduction of ASV dosing could improve ALT levels.17Akuta N. Sezaki H. Suzuki F. et al.Relationships between serum asunaprevir concentration and alanine aminotransferase elevation during daclatasvir plus asunaprevir for chronic HCV genotype 1b infection.J Med Virol. 2016; 88: 506-511Crossref PubMed Scopus (19) Google Scholar The marketing application for ASV was withdrawn during the US FDA review process in 2014. Clinical trial data evaluating DCV in combination with SOF with or without RBV did not reveal any evidence of hepatotoxicity. Subsequently, DCV was approved by the FDA in combination with SOF in 2015. Of note, our analysis shows ASV is RO2 positive whereas DCV is an RO2 negative drug (Table 1). EBR and GZR combination regimen is available as a fixed dose combination product (Zepatier) approved by the FDA in 2016. Dose-related increases in hepatic transaminases were observed during early phase clinical development program when higher doses of GZR were evaluated in combination with PEG-IFN/RBV. This led to the regulatory decision that the GZR dose should not exceed 100 mg/d in HCV-infected subjects to be evaluated in late phase clinical trials.18US Food and Drug Administration. [email protected] drug approval package – Cross Discipline Team Leader Review for ZEPATIER (elbasvir and grazoprevir). Available at: www.accessdata.fda.gov/drugsatfda_docs/nda/2016/208261Orig1s000CrossR.pdf. Accessed July 1, 2016.Google Scholar GZR exposures (area under the curve at 24 hours) were increased by 1.7-fold, 4.8-fold, and 11.7-fold, in non–HCV-infected subjects with mild, moderate, and severe hepatic impairment, respectively, as compared with matched healthy volunteers.19US Food and Drug Administration. [email protected] drug approval package – Clinical Pharmacology Review for ZEPATIER (elbasvir and grazoprevir). Available at: www.accessdata.fda.gov/drugsatfda_docs/nda/2016/208261Orig1s000PharmR.pdf. Accessed July 1, 2016.Google Scholar The exposure increases noted in subjects with moderate or severe hepatic impairment are considered clinically relevant as GZR-associated hepatotoxicity is exposure dependent.19US Food and Drug Administration. [email protected] drug approval package – Clinical Pharmacology Review for ZEPATIER (elbasvir and grazoprevir). Available at: www.accessdata.fda.gov/drugsatfda_docs/nda/2016/208261Orig1s000PharmR.pdf. Accessed July 1, 2016.Google Scholar The EBR and GZR combination regimen is contraindicated in patients with moderate or severe hepatic impairment (Child-Pugh class B or C) owing to the expected significantly increased GZR plasma concentration and the increased risk of ALT elevations.18US Food and Drug Administration. [email protected] drug approval package – Cross Discipline Team Leader Review for ZEPATIER (elbasvir and grazoprevir). Available at: www.accessdata.fda.gov/drugsatfda_docs/nda/2016/208261Orig1s000CrossR.pdf. Accessed July 1, 2016.Google Scholar EBR is an RO2 negative drug with a low daily dose (50 mg/d) but a high logP (5.60), and GZR is also an RO2 negative drug because of its logP < 3. However, GZR has a daily dose of 100 mg/d and a logP of 2.94 (borderline), suggesting its hepatotoxicity potential could be higher than other RO2 negative drugs. RBV (a synthetic nucleoside analogue) is part of many DAA-based regimens and its use is associated with hemolytic anemia resulting in hyperbilirubinemia, mainly driven by increase in indirect bilirubin, which is generally benign and resolves after treatment completion.20LiverTox – RIBAVIRIN. Available at: https://livertox.nlm.nih.gov/Ribavirin.htm. Accessed July 1, 2016.Google Scholar RBV is an RO2 negative owing to its low lipophilicity (logP = -1.92). In this commentary, we reported retrospective analyses of HCV DAAs using the RO2 model to predict the hepatotoxic potential of drugs for treating a serious and life threatening chronic liver disease owing to chronic HCV infection. The RO2 model’s analysis of DAAs is generally in agreement with the hepatotoxic risk observed in the postmarketing surveillance. Notably, in a recent seminal review,21Sarges P. Steinberg J.M. Lewis J.H. Drug-induced liver injury: highlights from a review of the 2015 literature.Drug Safety. 2016; : 1-21PubMed Google Scholar SMV, ASV, and Viekira Pak were cited as hepatotoxic DAAs/regimens, whereas SOF, LDV, and DCV were not reported to be associated with hepatic injury. It is important to note that understanding the interplay of drug properties with host factors and drug–host interactions is challenging and crucial to the understanding of DILI.22Chen M. Suzuki A. Borlak J. et al.Drug-Induced liver injury: interactions between drug properties and host factors.J Hepatol. 2015; 63: 503-514Abstract Full Text Full Text PDF PubMed Scopus (253) Google Scholar In our analysis of RO2-positive DAAs (eg, paritaprevir, dasabuvir, and SMV), we also observed that host factors such as advanced liver disease (Child-Pugh class B or C) may lead to increased exposure of the drugs, which is potentially associated with an increased risk of liver injury. In addition, other host factors, such as genetics, may affect drug metabolism and hepatotoxicity. DILI continues to be a concern for the FDA owing to the lack of effective tools for identifying the risk associated with drug candidates during clinical trials, before their widespread use in the large patient populations after marketing approval. Animal studies are required by regulatory agencies for investigational new drug applications; however, a retrospective analysis revealed such tests fail predicting risk for DILI in about 45% of clinical trials.23Olson H. Betton G. Robinson D. et al.Concordance of the toxicity of pharmaceuticals in humans and in animals.Regul Toxicol Pharmacol. 2000; 32: 56-67Crossref PubMed Scopus (1467) Google Scholar Currently, the agencies still largely rely on clinical trials to identify DILI risk in humans. Solely relying on clinical trials to identify DILI risk could be challenging especially in the cases with preexisting liver disease owing to the presence of multiple confounding factors, such as elevated baseline liver enzymes in chronic HCV patients, elevated indirect bilirubin level owing to enzyme inhibition, and inhibition of hepatic transporter associated with certain DAAs. For example, one of the components of Viekira Pak and Technivie, paritaprevir, is a known inhibitor of bilirubin transporters (OATP1B1 and OATP1B3) that could cause increases of total and indirect bilirubin levels observed in clinical trials. Similarly, SMV also inhibits hepatic transporters. The RO2 model is a simple tool only requiring a daily dose, which can be estimated from animal studies or determined at the early stage of clinical trial, and logP value, which can be predicted based on chemical structure. Importantly, the RO2 model has a high specificity (∼95%) but low sensitivity (∼40%),3Chen M. Borlak J. Tong W. High lipophilicity and high daily dose of oral medications are associated with significant risk for drug-induced liver injury.Hepatology. 2013; 58: 388-396Crossref PubMed Scopus (222) Google Scholar suggesting its potential use as an alert signal for screening drug candidates with DILI risk for humans at an early stage of clinical development and the results could further inform dose selection in patients. More important, the application of predictive models such as the RO2 model opens the door for rational design of clinical trials to help to capture those rare, but severe adverse events in the early phases of drug development. The following limitations associated with the use of this model are worth noting. First, the RO2 model remains a scientific research tool that is largely descriptive and retrospective. Although our current study demonstrates its promising performance, further validation or refinement is warranted in prospective clinical trials. Second, the model has been tested in the dataset with limited numbers of historical drugs; therefore, independent evaluation in larger databases is needed. Third, the RO2 model uses 2 parameters with a hard threshold (ie, daily dose ≥ 100 mg/d and logP ≥ 3) and generates a binary outcome (ie, positive versus negative). This may pose challenges during interpretation for some drugs, which are very close to the cutoff values used to report outcome. In conclusion, these analyses for HCV DAAs across multiple drug classes further demonstrate the potential role of RO2 model in identifying DILI risk in humans. A model with continuous variables might serve the field better and is being currently explored.5Chen M. Borlak J. Tong W. A model to predict severity of drug-induced liver injury in humans.Hepatology. 2016; 64: 931-940Crossref PubMed Scopus (56) Google Scholar These models need to be evaluated prospectively in a systematic way and more data are needed to refine them. Collaborative research with pharmaceutical companies and academia may provide much-needed tools for DILI assessment in future clinical trials. This article reflects the views of the authors and should not be construed to represent FDA’s views or policies. This report is original unpublished work and was performed as part of US. Government work. The authors would like to acknowledge the FDA review teams for their excellent work during the drug approval review process and postmarketing assessments. We appreciate the discussions amongst the FDA Liver Toxicity Working Group members and thank them for their contributions. We acknowledge Drs Ayako Suzuki and Weida Tong for their comments and/or discussions. The data for these drug products were submitted to the United States Food and Drug Administration by AbbVie Inc, Bristol-Myers Squibb Company, Gilead Sciences Inc., Janssen Products, Merck Sharp & Dohme Corp. and Vertex Pharmaceuticals as part of a new drug application.