Abstract
Hepatitis C virus (HCV) rapidly acquires drug resistance, complicating the design of effective antivirals. We developed a stochastic model of subgenomic HCV replicon replication that accounts for the emergence and selection of drug-resistant mutants. By incorporating key NS3 protease mutations (A156T, D168V, R155Q, etc.) resistant to inhibitors like BILN-2061 and VX-950, the model describes long-term viral RNA suppression dynamics. Our results suggest that pre-existing mutants contribute significantly to resistance during treatment, and observed kinetic differences across inhibitor concentrations stem from variations in mutant replication rates and drug sensitivities.
Introduction: The Challenge of Viral Adaptation
Hepatitis C is a master of adaptation. Its replication machinery, the NS5B polymerase, lacks a proofreading function, leading to an exceptionally high mutation rate. This results in a "quasispecies" population—a diverse cloud of genetic variants within a single host. When Direct-Acting Antivirals (DAAs) such as protease inhibitors are introduced, the dominant "wild-type" virus is suppressed, but rare, drug-resistant mutants can rapidly proliferate.
Traditional deterministic models, which rely on population averages, often fail to capture the "noisy," random events occurring at low viral concentrations—precisely when a single resistant mutant might determine therapeutic success. This research bridges that gap using a stochastic approach to simulate individual molecular interactions.
Methods: A Stochastic Framework for Replication
The researchers constructed a stochastic mathematical model of the HCV subgenomic replicon, a laboratory tool that replicates without infecting new cells.
The Gillespie Algorithm
To manage the randomness of low-concentration molecular events, the model employs the Gillespie Algorithm's "Next Step" method. This allows the simulation to track discrete events, such as the translation of a single polyprotein or the formation of an individual replication vesicle.
Biological Integration
The model tracks the complete intracellular cycle: IRES-mediated translation, polyprotein processing by the NS3 protease, the formation of membrane-bound replication vesicles, and the synthesis of new RNA strands.
Key mutations in the NS3 protease (e.g., A156T, D168V, R155Q) known to confer resistance to early-generation inhibitors like Telaprevir and Boceprevir were explicitly incorporated. The model was calibrated using the DEEP (Differential Evolution Entirely Parallel) optimization technique against experimental datasets for BILN-2061, VX-950, and SCH-503034.
Key Insights and Results
The stochastic model provided several critical insights into the dynamics of drug resistance:
- Pre-existing Mutants: The model confirms that resistance is rarely "invented" during treatment. Instead, resistant mutants typically pre-exist at low levels (approx. 1%). Treatment acts as a selective filter, removing the wild-type and allowing these mutants to flourish.
- Biphasic Decline: Viral RNA levels typically show a sharp initial drop (Phase 1) followed by a slower decline or rebound (Phase 2). The model identifies Phase 1 as the clearance of the wild-type virus and Phase 2 as the growth of drug-resistant mutants.
- Fitness Trade-offs: Many drug-resistant mutants exhibit a "fitness cost," replicating slower than the wild-type in the absence of inhibitors. The model quantifies these trade-offs to predict which mutants will dominate at specific drug concentrations.
- Long-term Predictive Power: Unlike previous deterministic models that lost accuracy after 72 hours, this stochastic approach accurately simulated viral dynamics over 9 to 15 days. It successfully predicted the dominance of resistant species as early as 5 days into monotherapy.
Conclusion and Future Directions
This model serves as a robust tool for interpreting anti-HCV drug tests and predicting the efficacy of potential therapies. While it currently treats vesicle RNA synthesis as a simplified "black box" rate constant and does not fully account for the host's innate interferon response, future iterations aim to include multi-drug "cocktails" and more detailed cellular immune interactions. Such advancements will better reflect modern clinical regimens and help optimize treatment strategies against highly mutable viruses.
Summary updated