Antiviral Properties of the NSAID Drug Naproxen Targeting the Nucleoprotein of SARS-CoV-2 Coronavirus

Abstract

There is an urgent need for specific antiviral treatments directed against SARS-CoV-2 to prevent the most severe forms of COVID-19. By drug repurposing, affordable therapeutics could be supplied worldwide in the present pandemic context. Targeting the nucleoprotein N of the SARS-CoV-2 coronavirus could be a strategy to impede viral replication and possibly other essential functions associated with viral N. The antiviral properties of naproxen, a non-steroidal anti-inflammatory drug (NSAID) that was previously demonstrated to be active against Influenza A virus, were evaluated against SARS-CoV-2. Intrinsic fluorescence spectroscopy, fluorescence anisotropy, and dynamic light scattering assays demonstrated naproxen binding to the nucleoprotein of SARS-Cov-2 as predicted by molecular modeling. Naproxen impeded recombinant N oligomerization and inhibited viral replication in infected cells. In VeroE6 cells and reconstituted human primary respiratory epithelium models of SARS-CoV-2 infection, naproxen specifically inhibited viral replication and protected the bronchial epithelia against SARS-CoV-2-induced damage. No inhibition of viral replication was observed with paracetamol or the COX-2 inhibitor celecoxib. Thus, among the NSAID tested, only naproxen combined antiviral and anti-inflammatory properties. Naproxen addition to the standard of care could be beneficial in a clinical setting, as tested in an ongoing clinical study.

Keywords: SARS-CoV-2; antiviral; drug repurposing; inflammation; influenza; nucleoprotein; oligomerization; structure-based drug design.

Figure 1

Structural comparison of viral nucleoproteins N. (A) Influenza A 2IQH.pdb and of amino-terminal domains of (B) SARS-CoV-1 1SSK.pdb (C) MERS-CoV 6KL2.pdb (D) SARS-CoV-2 6VYO.pdb. Electrostatic surface potentials are computed according to Adaptive Poisson-Boltzmann Solver Electrostatics plugin from PyMOL in standard parameters. Electronegative potentials are colored red while electropositive potentials are colored blue. A large electropositive cleft is visible on each protein, pointed with a green arrow; it maps the putative RNA interaction sites. (E–H) Main chains flexibilities are approximated as crystallographic b-factors, with main chain ribbon diameter proportional to b-factor value; red is highest b-factor while blue is lowest b-factor. Most mobile elements are identified by a red asterisk (*); they group on one external side of the protein, facing solvent. Mobile elements of highest b-factors are proximal to electropositive surfaces, supporting an interaction scenario where (1) RNA docks onto an electro-complementary surface before (2) the protein conformation is induced into the N-RNA complex of highest stability. (I) Pairwise structural alignment of SARS-CoV-1 (blue), MERS-CoV (green) and SARS-CoV-2 (orange) nucleoproteins. The red arrow emphasizes the conserved RWYFYY sequence. Images were ray traced with PyMOL version 2.0.6. Orientations were selected to highlight discussed properties.

Figure 2

Binding sites of naproxen to N NTD monomer and dimer; comparison with naproxen derivatives. Naproxen is shown in (A), with details on its functional moieties as discussed in the text. Panels (B,C) show the binding sites of naproxen on monomeric N NTD (PDB 7ACT) [28], the main (more frequent) binding site represented in cyan, is associated with the dimeric interface, highlighted in (C) in yellow; the minor site of naproxen on monomeric N NTD is represented in green. In panels (D,E) is shown the main binding site of naproxen on dimeric N NTD (PDB 6VYO) [27]. For details, see also Supplementary Figure S2. The structures of naproxen A and C0 are shown in panels (F) and (G), respectively. Panel (H) summarizes the RNA binding site (shown in green) deduced from RMN studies [28], the main binding site of naproxen on the left and the main binding site of naproxen C0, located within the RNA binding groove. (See also Table 1 and Supplementary Figure S2).

Figure 3

Naproxen binds to recombinant N-terminal domain of SARS-CoV-2 N. (A): Close-up view of naproxen interacting with W52 (PDB 6VYO). (B): Normalized intrinsic fluorescence of N, corrected for the inner filter effect, as a function of naproxen concentration using 2.1 µM N in 20 mM Tris buffer pH = 7.9, 100 mM NaCl. (C): evaluation of the solvent-accessible surface area in the N dimer without and with naproxen.

Figure 4

Competition on N NTD of binding of RNA with naproxen or naproxen C0 deduced from fluorescence anisotropy r.

Figure 5

Naproxen competes with DNA-induced N oligomerization. Size distribution of N NTD alone (red) or in the presence of 48 mer DNA (green), demonstrating the DNA-induced oligomerization of N NTD. Naproxen (blue) addition to N largely impeded N NTD oligomerization while naproxen C0 (black) only somewhat decreased the size of the N-DNA complex. The sizes are described in the text. T = 20 °C, [NTD] = 60 µM, [48mer DNA] = 29.4 µM and naproxen or naproxen C0 were used at 37.5 µM. In these conditions, naproxen A or acetaminophen could not impede N oligomerization (Table 2).

Figure 6

Naproxen inhibits SARS-CoV-2 infection in Vero E6 cells and in HAE. (A,B) Dose–response curves of naproxen at 48 and 72 hpi in infected Vero E6 cells. Cells were infected at MOI 0.01 with SARS-CoV-2 and then treated 1 h postinfection with a large range of concentration of naproxen; (C) Relative intracellular viral genome quantification and (D) trans-epithelial resistance (TEER in Ohms/cm²) between the apical and basal poles in nasal and bronchial HAE at 48 hpi. Results are expressed in relative viral production compared to the infected untreated control and relative TEER compared to t = 0 (before infection). *** p < 0.001 and * p < 0.05 compared to the infected untreated (viral titers) or uninfected (TEER) groups by Student’s t-test. Data are representative of three independent experiment.

Figure 7

Antiviral effect of naproxen C0, as compared to the naproxen A derivative, acetaminophen and the COX2 inhibitor celecoxib in infected Vero E6 cells. Dose–response curves of naproxen C0 (A,B) or naproxen A (C,D) at 48 and 72 hpi in infected Vero E6 cells. Using the same experimental protocol as in Figure 4, Vero E6 cells were infected at MOI 0.01 with SARS-CoV-2 and then treated 1h postinfection with a large range of concentration of naproxen C0 or A or paracetamol or celecoxib. Panels (E,F) and (G,H) show the lack effect of acetaminophen (paracetamol) and the mainly proviral effect of celecoxib (COX2 inhibitor) on viral replication, demonstrating the specificity of naproxen among the COX inhibitors tested (Table 2).