In Vitro Evaluation and Mitigation of Niclosamide's Liabilities as a COVID-19 Treatment

Abstract

Niclosamide, an FDA-approved oral anthelmintic drug, has broad biological activity including anticancer, antibacterial, and antiviral properties. Niclosamide has also been identified as a potent inhibitor of SARS-CoV-2 infection in vitro, generating interest in its use for the treatment or prevention of COVID-19. Unfortunately, there are several potential issues with using niclosamide for COVID-19, including low bioavailability, significant polypharmacology, high cellular toxicity, and unknown efficacy against emerging SARS-CoV-2 variants of concern. In this study, we used high-content imaging-based immunofluorescence assays in two different cell models to assess these limitations and evaluate the potential for using niclosamide as a COVID-19 antiviral. We show that despite promising preliminary reports, the antiviral efficacy of niclosamide overlaps with its cytotoxicity giving it a poor in vitro selectivity index for anti-SARS-CoV-2 inhibition. We also show that niclosamide has significantly variable potency against the different SARS-CoV-2 variants of concern and is most potent against variants with enhanced cell-to-cell spread including the B.1.1.7 (alpha) variant. Finally, we report the activity of 33 niclosamide analogs, several of which have reduced cytotoxicity and increased potency relative to niclosamide. A preliminary structure-activity relationship analysis reveals dependence on a protonophore for antiviral efficacy, which implicates nonspecific endolysosomal neutralization as a dominant mechanism of action. Further single-cell morphological profiling suggests niclosamide also inhibits viral entry and cell-to-cell spread by syncytia. Altogether, our results suggest that niclosamide is not an ideal candidate for the treatment of COVID-19, but that there is potential for developing improved analogs with higher clinical translational potential in the future.

Keywords: COVID-19; SARS-CoV-2; drug repurposing; niclosamide; polypharmacology.

Figure 1

Niclosamide is toxic at antiviral concentrations after long-term exposure. (A) Workflow for high content anti-SARS-CoV-2 bioassay screening to determine infection inhibition and cytotoxicity. (B) 10-point, 2-fold dilution concentration–response curves for VeroE6 cells infected with B.1.1.7 variant at MOI = 0.1 for 48 h. (C) Representative images for mock, vehicle, and 10 µM niclosamide-treated (infected) VeroE6 cells. (D) Concentration–response curves for H1437 cells infected with B.1.1.7 variant at MOI = 1 for 48 h. (E) Representative images for mock, vehicle, and 10 µM niclosamide-treated (infected) H1437 cells. Data points in concentration–response curves represent mean ± SEM for n = 10 replicates per condition. Curve fitting was performed in GraphPad Prism 9 using a semi-log 4-parameter variable slope model. Percent infection is shown using red curves, while percent viability is shown in black. Images were captured at 10× magnification, and the overlays were generated in ImageJ such that cyan = nuclei and magenta = SARS-CoV-2 N protein (uniform scale barb = 80 micrometers).

Figure 2

Niclosamide potency is SARS-CoV-2 variant dependent. (A) Assay timeline for 24-h infection experiment. The assay window was shortened to reduce niclosamide toxicity. (B) 10-point 2-fold concentration–response curves for niclosamide against the different SARS-CoV-2 variants of concern (MOI = 0.1 for each variant) with a top concentration of 10 μM. Curves were fitted with GraphPad Prism 9 software using a semi-log 4-parameter variable slope model. Data for each variant were normalized to the average percent infected of its respective viral control. Data points represent mean ± SEM for n = 3 replicates. (C) IC₅₀ values for niclosamide potency against SARS-CoV-2 variants of concern. Values were extracted from curve fitting using GraphPad 9 and include SEM error bars (WA1: 1664 ± 149 nM, B.1.1.7: 298 ± 23 nM, B.1.351: 440 ± 21 nM, B.1.617.2: 774 ± 58 nM, P.1: 399 ± 34 nM). Significance was determined using Student’s t-tests (* = p < 0.05, ** = p < 0.01).

Figure 3

Morphological profiling of B.1.1.7 infection versus niclosamide treatment in VeroE6. Image analysis reveals mechanistic characteristics of niclosamide activity against SARS-CoV-2 infection. (A) The maximum area of viral objects decreases with increasing niclosamide concentration. Data are the max area for viral objects in each condition. Viral control: n = 17452, +156 nM niclosamide: n = 2425, +313 nM niclosamide: n = 1470, +625 nM niclosamide: n = 496. (B) Viral objects per well decrease with increasing niclosamide concentration. Viral objects include single infected cells and syncytia. Replicate values are indicated on the X axis. (C) Mean pixel intensity for viral objects in each condition. Pixel intensity increases with increasing niclosamide concentration. (D) Representative images for each condition including N-protein channel, nuclear channel, an overlayed image, and a fire lookup table (LUT) image of the N-protein channel. Images were taken on a CX5 high content microscope at 10× magnification. * = p < 0.05, *** = p < 0.001, **** = p < 0.0001. (Scale bar = 80 µM).

Figure 4

Niclosamide analogs have improved efficacy and reduced cytotoxicity in VeroE6. (A) Structure of niclosamide indicating the chlorosalicyl/nitroaniline rings, and analog scaffold with modified substituent positions labeled. (B) IC₅₀ vs. CC₅₀ plot highlighting efficacious compounds in VeroE6. Compounds with improved potency and cytotoxicity profiles are circled on the plot. (C) 10-point, 2-fold concentration–response curves for the top four niclosamide analogs with a starting concentration of 20 μM. Data are shown as the mean ± SEM of n = 3 replicate wells per condition. Curves for infection (in red) and cell viability (in black) are included. (D) Representative images of infected cells treated with indicated compounds and viral control (Vehicle). (10× magnification, cyan = nuclei, magenta = SARS-CoV-2 nucleocapsid protein, scale bar = 80 µm).

Figure 5

Efficacy of protonophores against SARS-CoV-2. (A) Removal of protonophore hydroxyl of niclosamide results in a complete loss of efficacy in both VeroE6 and H1437 cells. (B) Chemical structures for other protonophores evaluated against SARS-CoV-2 infection. (C,D) 10-point, 2-fold concentration–response curves for protonophores versus WA1 variant (C) and B.1.1.7 variant (D) with starting concentrations of 100 μM. (E) IC₅₀ plot for antiviral protonophores indicating significantly different potency against variants. ** = p < 0.01, *** = p < 0.001. (F) Table of IC₅₀ and CC₅₀ values for protonophores.

Figure 6

Diagram of niclosamide effect on SARS-CoV-2 entry and spike protein-mediated syncytia formation. (1) SARS-CoV-2 binds to the ACE2 receptor of the host cell and enters. Niclosamide has been shown to inhibit this entry step in vitro. (2) Viral replication generates many copies of the RNA genome. (3) Infection results in an increased expression of viral spike (S) protein and host cell TMEM16F at the plasma membrane. (4) The S protein at the surface of an infected cell binds to the ACE2 receptor of an adjacent uninfected cell. (5) Spike-dependent syncytia formation is mediated by the calcium-dependent lipid scramblase TMEM16F to generate multinucleated infected cell bodies. Niclosamide, an inhibitor of TMEM16F, has been shown to block spike-dependent syncytia formation.