Halofuginone for COVID-19

The human transmembrane protease, serine 2 (TMPRSS2), essential for SARS-CoV-2 entry, is a key antiviral target. Here, we computationally profiled the TMPRSS2-binding affinities of 15 antiviral compounds. Molecular dynamics (MD) simulations for the docked complexes revealed that three compounds exited the substrate-binding cavity (SBC), suggesting noncompetitive inhibition. Of the remaining compounds, five charged ones exhibited reduced binding stability due to competing electrostatic interactions and increased solvent exposure, while seven neutral compounds showed stronger binding affinity driven by van der Waals (vdW) interactions compensating for unfavorable electrostatic effects (including electrostatic interactions and desolvation penalties). Positive and negative hotspot residues were identified as uncharged and charged, respectively, both lining the SBC. Despite forming diverse interactions with compounds, the burial of positive hotspots led to strong vdW interactions that overcompensated for unfavorable electrostatic effects, whereas negative hotspots incurred high desolvation penalties, negating any favorable contributions. Charged residues at the SBC’s outer rim can reduce binding affinity significantly when forming hydrogen bonds or salt bridges. These findings underscore the importance of enhancing vdW interactions with uncharged residues and minimizing the unfavorable electrostatic effects of charged residues, providing valuable insights for designing effective TMPRSS2 inhibitors.

AbstractSARS-CoV-2 (2019-nCoV) is the pathogenic coronavirus responsible for the global pandemic of COVID-19 disease. The Spike (S) protein of SARS-CoV-2 attaches to host lung epithelial cells through the cell surface receptor ACE2, a process dependent on host proteases including TMPRSS2. Here, we identify small molecules that reduce surface expression of TMPRSS2 using a library of 2,560 FDA-approved or current clinical trial compounds. We identify homoharringtonine and halofuginone as the most attractive agents, reducing endogenous TMPRSS2 expression at sub-micromolar concentrations. These effects appear to be mediated by a drug-induced alteration in TMPRSS2 protein stability. We further demonstrate that halofuginone modulates TMPRSS2 levels through proteasomal-mediated degradation that involves the E3 ubiquitin ligase component DDB1- and CUL4-associated factor 1 (DCAF1). Finally, cells exposed to homoharringtonine and halofuginone, at concentrations of drug known to be achievable in human plasma, demonstrate marked resistance to SARS-CoV-2 infection in both live and pseudoviral in vitro models. Given the safety and pharmacokinetic data already available for the compounds identified in our screen, these results should help expedite the rational design of human clinical trials designed to combat active COVID-19 infection.