A dual drug regimen synergistically blocks human parainfluenza virus infection

Abstract

Human parainfluenza type-3 virus (hPIV-3) is one of the principal aetiological agents of acute respiratory illness in infants worldwide and also shows high disease severity in the elderly and immunocompromised, but neither therapies nor vaccines are available to treat or prevent infection, respectively. Using a multidisciplinary approach we report herein that the approved drug suramin acts as a non-competitive in vitro inhibitor of the hPIV-3 haemagglutinin-neuraminidase (HN). Furthermore, the drug inhibits viral replication in mammalian epithelial cells with an IC50 of 30 μM, when applied post-adsorption. Significantly, we show in cell-based drug-combination studies using virus infection blockade assays, that suramin acts synergistically with the anti-influenza virus drug zanamivir. Our data suggests that lower concentrations of both drugs can be used to yield high levels of inhibition. Finally, using NMR spectroscopy and in silico docking simulations we confirmed that suramin binds HN simultaneously with zanamivir. This binding event occurs most likely in the vicinity of the protein primary binding site, resulting in an enhancement of the inhibitory potential of the N-acetylneuraminic acid-based inhibitor. This study offers a potentially exciting avenue for the treatment of parainfluenza infection by a combinatorial repurposing approach of well-established approved drugs.

Figure 1. Structures of N-acylneuraminic acid-based inhibitors of hPIV-3 HN.

Chemical structures of N-acetylneuraminic acid (Neu5Ac, 1), Neu5Ac2en (2), zanamivir (3), BCX-2798 (4), the phenyltriazole derivative of BCX-2798 (5) and 2′-(4-methylumbelliferyl) α-D-N-acetylneuraminide (MUN, 6). Ph = phenyl, Ac = acetyl.

Figure 2. Enzymatic inhibition of HN by suramin (7).

(a) Chemical structure of suramin (7). (b) Dose-response of suramin (7) against hPIV-3 neuraminidase activity. Data points are the mean of duplicate values and are representative of a least 2 independent experiments. The error bars represent the standard error of the mean (SEM).

Figure 3. Determination of the inhibition mode of suramin (7) by enzyme kinetics.

The initial velocities v_i of the HN neuraminidase activity were determined at several concentrations of the substrate MUN (2, 4, 8, 16, 20 mM, 6) for each concentration of suramin [suramin]. The Lineweaver-Burke graph was created by plotting duplicate values of as a function of according to Equation (1), and is representative of 3 independent experiments. The straight lines are linear regressions calculated for each concentration of inhibitor.

Figure 4. In vitro antiviral effect of suramin (7) on hPIV-3-infected LLC-MK2 cells.

Dose-dependent inhibition of hPIV-3 infection by suramin (7) at different stages of infection (a). The antiviral potencies of the drugs were evaluated by focus reduction assay, and the drugs were added either during virus binding (4 °C for 1 h), at adsorption stage (37 °C for 1 h), or post-adsorption (37 °C for 72 h). Foci numbers (virus binding and adsorption at 37 °C) or foci size (post-adsorption) were used to determine viral replication. Immunostaining was carried out after 72 h of incubation. (b) Post-adsorption effect of suramin (7) on reduction of foci size at 30 μM as compared to an untreated control (mock). Top: scan of a focus reduction assay from a 24-well plate immunostained 72 h post-infection. Bottom: image of the same well after image conversion to a binary image and particle detection for automated foci counting and size measurements using Fiji. Each detected focus is outlined in black and numbered in red.

Figure 5. Evaluation of the synergism of suramin (7) in combination with competitive in vitro inhibitors of HN.

Data sets in red and blue correspond to suramin (7)—compound 5 or suramin (7)—zanamivir (3) combinations, respectively. (a) Dose-response curves of each individual compound. Suramin (7) was evaluated twice, for each of the combinations with suramin and compound 5. The antiviral effect was determined by measurement of foci size. (b) Median-effect representation of the dose-response curves for each individual compound, using Equation (4). m is the linear regression slope, f_a is the “fraction affected”, or (% effect) ÷100. (c) Normalised isobologram that represents, for each combination, the normalised dose of each compound individually required to reach the observed effect in combination (Equation (3)). D₁ is the dose of a compound 1 in combination required to achieve x% of inhibition, while D_x1 is the dose of a compound alone required to achieve x% of inhibition. Data points in zone a, b and c correspond to combinations with synergistic, additive or antagonistic effects, respectively. (d) CI-effect plot representing the combination index CI, calculated using Equation (2), of each combination as a function of their associated antiviral effect. The zones a, b and c are the same as the ones described in (c). (e) log(DRI)-effect plot representing the drug reduction index (DRI) of compounds as a function of their antiviral effect in combination. The DRI is calculated for each drug in each combination according to Equation (5), and represents the dilution factor required for a drug to reach the same level of inhibition individually compared with it when in combination. All combinations were tested in quadruplicate, post-adsorption, by a focus reduction assay. The results are representative of 3 independent experiments.

Figure 6. Competition STD-NMR of suramin (7) and zanamivir (3) in presence of purified hPIV-3.

(a) ¹H-NMR spectrum of suramin (7) and zanamivir (3). (b,c) Competition STD-NMR spectra of suramin (7) and zanamivir (3) in presence of virus where zanamivir (3) was added before (b) or after (c) suramin (7). Absolute intensities of STD NMR signals (b,c) are comparable. (d) STD-NMR spectra of suramin (7) alone in presence of virus (bottom), and after addition (top) of zanamivir (3). (e) STD-NMR spectra of zanamivir (3) alone in presence of virus (bottom), and after (top) addition of suramin (7). Drugs in combination were tested at an equimolar ratio.

Figure 7. Probing of suramin^h (8) in the hPIV-3 HN binding site.

(a) Chemical structure of Suramin^h (8), used for blind docking simulations. (b) Solvent-accessible surface representation of HN active site (1V3E), with zanamivir (3) docked. Zanamivir (3) from the structure 1V3E is represented in green, the 3 best conformations of zanamivir (3) from docking simulations are represented in magenta, orange and yellow. (c) Suramin^h (8) binding site probing strategy. A total of 36 overlapping grids (search spaces) were designed to cover the entire surface of an HN monomer (represented in grey), and simulations were run on each of them. Left: volume of the 36 grids combined. Right: volume of a single grid centred on the HN active site. (d) Solvent-accessible surface representation of HN with the best conformation of 8 docked to the active site of the apo-form (1V3B, suramin^h (8): in magenta, grey surface). (e) Solvent-accessible surface representations of HN with the 2 lowest energy conformations of 8 resulting from 2 overlapping search spaces positioned over the active site of zanamivir (3)-bound HN (1V3E; zanamivir (3): in green; suramin^h (8): in magenta). (f) Solvent-accessible surface representation of HN with a dominant cluster of suramin^h (8, left) and suramin (7, right) docked over the active site of zanamivir (3)-bound HN (1V3E; zanamivir (3): in green; suramin^h (8) and suramin (7): in orange. The HN surface in (b,d–f) is coloured from red (−20 kT/e) to blue (20 kT/e) according to the electrostatic potential (APBS).