Antiviral Properties of HIV-1 Capsid Inhibitor GSK878

Abstract

GSK878 is a newly described HIV-1 inhibitor that binds to the mature capsid (CA) hexamer in a pocket originally identified as the binding site of the well-studied CA inhibitor PF-74. Here, we show that GSK878 is highly potent, inhibiting an HIV-1 reporter virus in MT-2 cells with a mean 50% effective concentration (EC₅₀) of 39 pM and inhibiting a panel of 48 chimeric viruses containing diverse CA sequences with a mean EC₅₀ of 94 pM. CA mutations associated with reduced susceptibility to other inhibitors that bind to PF-74 binding site (L56I, M66I, Q67H, N74D, T107N, and Q67H/N74D) also reduced susceptibility to GSK878, with M66I, Q67H/N74D, and L56I having the greatest impact on antiviral activity. Amino acid substitutions in the CA cyclophilin A (CypA) binding loop (H87P and P90A), distal from the inhibitor binding site and associated with reduced CA-CypA binding, subtly, but reproducibly, also decreased GSK878 potency. Mechanism-of-action studies showed that GSK878 blocked both early (preintegration) and late (postintegration) steps in HIV-1 replication, with the early inhibition primarily determining the compound's antiviral activity. The early inhibition results from blocks to HIV-1 nuclear import and proviral integration and is associated with altered stability of the HIV-1 CA core.

Keywords: antiviral activity; capsid core stability; capsid inhibitor.

FIG 1

(A) Chemical structure of GSK878. (B) Susceptibility of chimeric reporter viruses to GSK878 and RAL. Shown is a plot of EC₅₀ values calculated from antiviral assays performed in MT-2 cells with replication-competent reporter viruses (NLRepRluc) with the Gag-PR region derived from HIV-1 clinical isolates of the indicated strain and subtype. Values are means and SD from ≥4 experiments.

FIG 2

Effect of cyclophilin A on GSK878 antiviral activity. (A) Scatterplot of EC₅₀ and EC₉₀ values of GSK878 from antiviral assays with wild-type (WT) and CA H87P HIV-1 reporter viruses (NLRepRluc) in MT-2 cells (****, P < 0.0001; WT, n = 92; H87P, n = 72). (B and C) Comparison of GSK878 dose-response inhibition curves from antiviral assays in MT-2 cells with VSV-G-pseudotyped replication-defective HIV-1 (VSV-G:NLRepRlucΔENV) with the indicated CA missense mutations in the absence (B) or presence (C) of 0.8 μM CsA. Points are averages of two experiments, each performed in duplicate. (D) Dose-response curves with a reference NNRTI (EFV) under the same conditions as for panels B and C.

FIG 3

GSK878 antiviral activity. (A) Time-of-addition experiment. MT-2 cells were infected with a replication-defective virus pseudotyped with gp160 from HIV-1 strain LAI (LAI:NLRepRlucΔEnv). Inhibitors were added at 0, 2, 4, 5, 6, 7, and 8 h postinfection, and Renilla luciferase activity was measured after an additional 72 h. Percent inhibition relative to the value for DMSO-treated control cells is plotted for each time point (means and SD from 8 replicates). (B) Effect of GSK878 on HIV-1 replication intermediates. MT-2 cells infected with replication-defective VSV-G-pseudotyped HIV-1 (VSV-G:NLRepRlucΔENV) were treated with the indicated inhibitors and harvested at 12, 24, and 48 h for quantification of total reverse transcripts, 2-LTR circles, and integrated proviruses, respectively (means and SD from ≥3 experiments). (C) Quantification of 2-LTR circles from VSV-G:NLRepRlucΔENV-infected cells treated with the indicated concentration of GSK878 in the presence of 200 nM RAL.

FIG 4

GSK878 antiviral activity and CA core stability. (A) Susceptibility hyperstable (E45A) and hypostable (P38A) HIV-1 CA core mutants on GSK878 and EFV antiviral activity. Shown is a comparison of dose-response inhibition curves from antiviral assays in MT-2 cells with VSV-G-pseudotyped replication-defective HIV-1 (VSV-G:NLRepRlucΔENV), parental (WT) or with the indicated CA missense mutation (means ± SD from 3 experiments). (B) GSK878 stabilizes the CA core in infected cells. HeLa cells infected with VSV-G:NLRepRlucΔENV were treated with the indicated concentrations of GSK878 or DMSO. After 6 h, cells were harvested and analyzed by the fate-of-the-capsid assay as described in Materials and Methods. CA (p24) in the input, supernatant, and pellet was detected with the Simple Western system (Wes; Bio-Techne, Minneapolis, MN). Quantification of CA in the pellet is shown as the means ± SD from 4 experiments. The images on the left side of panel B were derived from the same blot and spliced for labeling purpose.