spread rapidly through human hepatocarcinoma cell lines and has been used to successfully develop virus growth assays and screens [19,20,21,22]. Next, chimeric viruses with genotype 1 structural protein coding sequences fused to JFH1 non-structural regions were produced [16,18], followed by chimeras with structural proteins from each HCV genotype [14,18,23,24,25,26,27]. Genotype 1 infections are the most common worldwide, and are most recalcitrant to interferon-containing therapy. Therefore, inhibitor activity against genotype 1 is a prerequisite for any novel DAA to enter clinical development. Novel HCV DAAs often exhibit selectivity for the genotype or subtype of the virus used for screening necessitating significant medicinal chemistry efforts to achieve broader genotype coverage. In addition, high-throughput screening (HTS) is often facilitated using viruses containing reporter gene proteins, such as luciferase. However, the intergenotypic HCV viruses, and those with reporter genes, often replicate to lower titers and with slower kinetics than those needed for extensive drug discovery. While a full-length genotype 1 clone with robust growth properties has yet to be developed [28], intergenotypic chimeras, where Core-NS2 of JFH1 is replaced with the corresponding region from genotype 1, are a potential source of viruses that can be adapted for comprehensive drug discovery activities. Despite their delayed growth kinetics relative to Jc1 [18], these viruses represent powerful tools for drug discovery since the entire early stage (i.e., virus entry and nucleocapsid uncoating) of the virus life cycle is mediated by genotype 1 proteins while virus assembly is orchestrated by a combination of genotype 1 and 2 proteins. Here, we report on the use of a genotype 1a/2a chimeric, reporter virus to develop a robust, homogeneous, high-throughput, multi-cycle virus replication assay and demonstrate its capability in HTS of a large-scale, small-molecule compound library. This novel screening approach was validated using a comprehensive array of secondary assays that classified hits according to potency, selectivity, life cycle stage targeted and genotype coverage.
EIPE (1.960.8) or LY411575 (2.060.6; Figs. 1B & C), suggesting that virus spread was not detectable at this time point. Expansion of virus foci was typically observed within 72 h (17.669.4) and large foci (88.0629) were observed at 96 h pi (Figs. 1B & C). The expansion of virus foci was blocked by both EIPE and LY411575 with .20 fold inhibition observed at the 96 h time point (Figs. 1B & C). Importantly, the expansion of virus foci correlated with an increase in Renilla luciferase expression (Fig. 1D). Consistent with the results above, EIPE and LY411575 exhibited only modest inhibition (,20%) of luciferase expression at 48 h pi but achieved .80% at 72 h and .90% inhibition at 96 pi (Fig. 1E). As expected, addition of the entry inhibitor or an NS3 protease inhibitor (BMS-339) at the time of infection inhibited luciferase expression at all time points (Fig. 1E). Taken together, these results suggested that a 96 h incubation period was necessary and sufficient for unbiased identification of inhibitors of all phases of viral replication using the gt 1a/2a-Rluc virus. Our goal was to develop an assay where .90% of the Luciferase signal was due to virus spread so as to avoid bias towards early or genome replication inhibitors. While it has been reported that earlier time points are optimal for spreading of fully genotype 2a viruses, Figure 1 shows that a 96 h incubation period was required to achieve this goal. These results further demonstrated that single and multi-cycle virus replication could be delineated by monitoring luciferase expression at either the 48 or 96 h pi time points, respectively.
Transition to a homogenous, 384 well format assay
Rapid screening of large-scale compound libraries requires homogenous, miniaturized platforms that can be automated. A first step towards assay simplification was to determine if virus and trypsinized cells could be mixed in suspension to initiate infection, and dispensed onto plates containing screen compounds, instead of adding virus to adherent cells plated 24 hours prior. It was observed that the effective titer of a virus stock (ffu per well) was typically reduced approximately 1.5-fold in the one-step procedure relative to a multi-step infection protocol (data not shown) and the expansion of viral foci (virus per foci) was only modestly (,2 fold) delayed (Fig. 2A). Importantly, the homogenous protocol had no impact on the potency of control inhibitors (data not shown). A low MOI is necessary to enable the multiple rounds of replication required to interrogate all replication stages. However, the MOI needs to be high enough to achieve a robust signal/ background with minimal well-to-well variability. This signal/ variability relationship is referred to as the Z factor [41] and is generally preferred to be $0.5 to enable statistically reliable identification of inhibitors during HTS. First, the optimal ratio of gt 1a/2a-Rluc virus to Huh-7.5 cells in a 384 well plate format was determined. While cell numbers influenced Z factor somewhat, the amount of virus used had a more dramatic effect, with $400 ffu of virus/well and 4,000 Huh7.5 cells providing an acceptable Z factor (Fig. 2B). With 400 ffu of virus/well and 4,000 Huh7.5 cells (MOI = 0.1), entry (EI), replication (BMS-339) and late-stage (LY411575) inhibitors all achieved .90% inhibition in a 96 h assay (Fig. 2C). As expected, increasing the input of virus (MOI.0.1) reduced the percent inhibition for the late-stage inhibitor (Fig. 2C) confirming that when using 4,000 Huh-7.5 cells/well an MOI of 0.1 or lower was critical to ensuring detection of late-stage inhibitors. Additional experiments were performed to enable HTS, including studies that showed that the gt 1a/2a-Rluc virus (and other HCVcc viruses) was amenable to large-scale preparation and storage at 280uC and that Huh-7.5 cells could be cultured on a preparative-scale using robotics without loss of HCVcc permissiveness (data not shown).
