Lack of antiviral activity of probenecid in vitro and in Syrian golden hamsters

Abstract

Objectives: Antiviral interventions are required to complement vaccination programmes and reduce the global burden of COVID-19. Prior to initiation of large-scale clinical trials, robust preclinical data to support candidate plausibility are required. This work sought to further investigate the putative antiviral activity of probenecid against SARS-CoV-2.

Methods: Vero E6 cells were preincubated with probenecid, or control media for 2 h before infection (SARS-CoV-2/Human/Liverpool/REMRQ0001/2020). Probenecid or control media was reapplied, plates reincubated and cytopathic activity quantified by spectrophotometry after 48 h. In vitro human airway epithelial cell (HAEC) assays were performed for probenecid against SARS-CoV-2-VoC-B.1.1.7 (hCoV-19/Belgium/rega-12211513/2020; EPI_ISL_791333, 2020-12-21) using an optimized cell model for antiviral testing. Syrian golden hamsters were intranasally inoculated (SARS-CoV-2 Delta B.1.617.2) 24 h prior to treatment with probenecid or vehicle for four twice-daily doses.

Results: No observable antiviral activity for probenecid was evident in Vero E6 or HAEC assays. No reduction in total or subgenomic RNA was observed in terminal lung samples (P > 0.05) from hamsters. Body weight of uninfected hamsters remained stable whereas both probenecid- and vehicle-treated infected hamsters lost body weight (P > 0.5).

Conclusions: These data do not support probenecid as a SARS-CoV-2 antiviral drug.

Figure 1.

(a) Concentration–effect relationship for the inhibition (%) of SARS-CoV-2 cytopathic activity for remdesivir and probenecid. Non-linear regression using an E_max model was performed on data taken from three independent biological replicates to generate concentration–effect predictions (solid black lines). EC₅₀ values, hillslope and replicate number (n) are shown. Dashed lines represent the EC₅₀. Squares, diamonds and circles represent individual biological replicates and error bars represent standard deviation calculated from technical triplicates. (b) GS-441524, the parental form of remdesivir, but not probenecid, inhibits SARS-CoV-2 replication in HAEC cultures. Viral RNA in apical washes was determined and mean viral load, converted to equivalent TCID₅₀ values based on a standard curve using a virus stock with known titre, and standard deviation is shown. LLOQ, lower limit of quantification in the RT–qPCR as derived from the standard curve.

Figure 2.

Average body weight data p.i. Error bars represent the standard deviation between individual animal weights.

Figure 3.

Lung viral RNA normalized to the total RNA (a) and 18 S subunit (b) in untreated controls and SARS-CoV-2-infected animals, treated with vehicle or probenecid. Error bars represent standard deviation between samples obtained from individual animals. Limit of detection (LOD) is set to y = 5 (SgE assay) and y = 2 (N assay) as lowest possible standards in assays.

Figure 4.

Concentration (pfu/mL) in terminal lung samples from SARS-CoV-2-infected animals, treated with vehicle (n = 5) or probenecid (n = 5). Error bars represent standard deviation of values obtained for individual animals.