A high-throughput screening assay to identify inhibitory antibodies targeting alphavirus release

Abstract

Background: Several studies have demonstrated neutralizing antibodies to be highly effective against alphavirus infection in animal models, both prophylactically and remedially. In most studies, neutralizing antibodies have been evaluated for their ability to block viral entry in vitro but recent evidence suggests that antibody inhibition through other mechanisms, including viral budding/release, significantly contributes to viral control in vivo for a number of alphaviruses.

Results: We describe a BSL-2, cell-based, high-throughput screening system that specifically screens for inhibitors of alphavirus egress using chikungunya virus (CHIKV) and Mayaro virus (MAYV) novel replication competent nano-luciferase (nLuc) reporter viruses. Screening of both polyclonal sera and memory B-cell clones from CHIKV immune individuals using the optimized assay detected several antibodies that display potent anti-budding activity.

Conclusions: We describe an "anti-budding assay" to specifically screen for inhibitors of viral egress using novel CHIKV and MAYV nLuc reporter viruses. This BSL-2 safe, high-throughput system can be utilized to explore neutralizing "anti-budding" antibodies to yield potent candidates for CHIKV and MAYV therapeutics and prophylaxis.

Keywords: Anti-budding; HTS assay; Mayaro virus; Neutralizing antibodies; chikungunya virus.

Fig. 1

181/25 E2^nLuc virus particles carry nano-luciferase activity. A Illustration of CHIKV wildtype and nLuc reporter virus genomes. DNA encoding the nLuc reporter (highlighted) was inserted into viral cDNA clones immediately downstream of the furin cleavage site between E3 and E2. B Gradient-purified 181/25 and 181/25 E2^nLuc viral particles underwent three-fold serial dilutions and nLuc activity was measured. Gradient-purified particles were also separated on a SDS-PAGE followed by C Coomassie staining and immuno-blot with D anti-CHIKV and E anti-nLuc

Fig. 2

Generation of CHIKV 181/25 E2^nLuc and MAYV E2^nLuc reporter viruses. BHK-21 cells were infected with CHIKV/MAYV wild-type (WT) or nLuc reporter virus. Supernatant was harvested 6, 12, 18, and 24 hpi (hours post infection) and used to determine A infectious titres via plaque forming unit assay in U2OS cells, B relative viral RNA levels via qRT-PCR and C nLuc activity (represented by RLU) associated with released viral particles. Graphs represent the arithmetic mean ± SD of three independent experiments. Plaque size was compared for WT and nLuc CHIKV/MAYV with examples of plaques (D) and by quantitating 53 plaques from each virus as described in the methods (E)

Fig. 3

Optimizing parameters for CHIKV and MAYV anti-budding assays. Cells were bulk infected for 3.5 h with CHIKV or MAYV nLuc reporter virus, excess virus washed off, cells plated and incubated for 18 h with or without the addition of control inhibitor, and nLuc activity measured in 50 µL supernatant. A Different cell lines were tested for susceptibility to infection by CHIKV nLuc. Cells were infected in bulk and then removed from the plastic surface and counted. Thus, MOIs were calculated retrospectively and ranged from 0.2 (TZM) to 1.3 (U2OS). RD cells were used in subsequent assays to assess MOI ranges for CHIKV (B) and MAYV (C) reporter viruses using control inhibitors mAb C9 (0.5 µg/mL) or pooled sera (diluted 1:50) respectively. D RD cells were incubated with CHIKV nLuc virus for 2–3.5 h at 30 min increments, after which the assay was continued as described. E Post bulk infection with CHIKV nLuc, excess virus was washed off and 50 µL supernatant was removed from plated cells at 1 h intervals from 17 to 20 h to measure nLuc activity. F Pilot screen demonstrating suitability for HTS screening. Results are shown as percentage of nano-luciferase reading of wells containing no antibody. Infected cells with and without C9 on 10 test plates were subjected to Z′ analysis to determine variation and suitability for HTS systems. A Z′ factor of 0.5–1.0 is deemed sufficiently robust for HTS assays. All results shown carry a CV score of ≤ 15%

Fig. 4

Optimized CHIKV anti-budding assay. The assay uses a novel CHIKV nLuc reporter virus to screen specifically for antibodies that inhibit viral budding. RD cells were first bulk infected with reporter virus, excess virus washed off after 2 h, the monolayer disrupted, and infected cells seeded onto a 96-well plate at 2.5 × 10⁴ cells/well in 100 µL/well of complete growth medium. 100 µL diluted CHIKV-positive human sera or B-cell supernatant was then transferred to assay wells. All wells, including those with positive inhibition control mAb (2 µg/µL C9) and negative control (media only), were supplemented with 20 mM NH₄Cl. Plates were then incubated for 18 h at 37 °C/5% CO₂, following which 50 µL of supernatant was harvested from each well and transferred to white half-well plates (Greiner Bio-One) for luminescence readings. For all assays, spectrophotometer gain was adjusted to negative control wells

Fig. 5

Dose responses for varying dilutions of human polyclonal sera. Samples from three volunteers with previous exposure to CHIKV (TT018, TT002, and TT024) at different times post symptom onset (4, 195, and 240 dpo respectively) were screened using the anti-budding assay system. All volunteers were previously (TT002 and TT024) or at enrollment (TT018) confirmed by RT-PCR. Serum from one volunteer with an ongoing ZIKV infection (5 dpo) and no evidence of any exposure to CHIKV was included as a specificity control (TT101). Different serum dilutions were tested to determine an appropriate dilution range for the anti-budding assay. Positive control mAb C9 (2 mg/mL) was included for comparison. Nano-luciferase readings from control wells containing no antibody were set at 100%. Results were calculated as a percentage of these control wells and represent the mean ± S.D of triplicate measurements. All results shown carry a CV score of ≤ 15%

Fig. 6

Screening of B-cell supernatants. 800 test wells generated from TT052, an individual one year post onset of CHIKV infection, were screened for anti-budding activity. Results are represented as the logarithm of relative light units (nano-luciferase readings) from ten 96-well plates (inter-plate CV = 14.4%; Z′ = 0.57). Different symbols are used to represent individual plates and associated C9 control wells are shown in green. Lines representing mean readings (red) as well as those 2–3 SD lower than the mean are shown (dotted grey and black respectively). A reading of at least 3SD below the mean was considered a “hit”