key: cord-0982431-8mmau2lz
authors: de Vries, Rory D.; Schmitz, Katharina S.; Bovier, Francesca T.; Predella, Camilla; Khao, Jonathan; Noack, Danny; Haagmans, Bart L.; Herfst, Sander; Stearns, Kyle N.; Drew-Bear, Jennifer; Biswas, Sudipta; Rockx, Barry; McGill, Gaël; Dorrello, N. Valerio; Gellman, Samuel H.; Alabi, Christopher A.; de Swart, Rik L.; Moscona, Anne; Porotto, Matteo
title: Intranasal fusion inhibitory lipopeptide prevents direct-contact SARS-CoV-2 transmission in ferrets
date: 2021-03-26
journal: Science
DOI: 10.1126/science.abf4896
sha: e00d29e4d3527fb29efef83506e5b5499fb0f639
doc_id: 982431
cord_uid: 8mmau2lz

Containment of the COVID-19 pandemic requires reducing viral transmission. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection is initiated by membrane fusion between the viral and host cell membranes, which is mediated by the viral spike protein. We have designed lipopeptide fusion inhibitors that block this critical first step of infection and, on the basis of in vitro efficacy and in vivo biodistribution, selected a dimeric form for evaluation in an animal model. Daily intranasal administration to ferrets completely prevented SARS-CoV-2 direct-contact transmission during 24-hour cohousing with infected animals, under stringent conditions that resulted in infection of 100% of untreated animals. These lipopeptides are highly stable and thus may readily translate into safe and effective intranasal prophylaxis to reduce transmission of SARS-CoV-2.

I nfection by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) requires membrane fusion between the viral envelope and the host cell, at either the cell surface or the endosomal membrane. The fusion process is mediated by the viral transmembrane spike glycoprotein (S). Upon viral attachment or uptake, host factors trigger large-scale conformational rearrangements in S, including a refolding step that leads directly to membrane fusion and viral entry (1) (2) (3) . Peptides corresponding to the highly conserved heptad repeat (HR) (Fig. 1A) domain at the C terminus of the S protein (HRC peptides) (Fig. 1B) can prevent this refolding and inhibit fusion, thereby preventing infection (4) (5) (6) (7) (8) . The HRC peptides form six-helix bundle-like assemblies with the extended in-termediate form of the S protein trimer, disrupting the structural rearrangement of S that drives membrane fusion (4) (see also movie S1).

Our approach in designing SARS-CoV-2 S-specific fusion inhibitors builds on our previous work that demonstrated that lipid conjugation of HRC-derived inhibitory peptides markedly increases antiviral potency and in vivo half-life (9, 10) . These peptides successfully inhibit human parainfluenza virus type 3 (HPIV-3), measles virus, influenza virus, and Nipah virus infection (9, (11) (12) (13) . Furthermore, the combination of dimerization and lipopeptide integration into cell membranes protects the respiratory tract and prevents systemic lipopeptide dissemination (14) . Lipidconjugated peptides administered intranasally to animals reached high concentrations both in the upper and lower respiratory tract, and the lipid could be designed to modulate the extent of transit from the lung to the blood and organs (9, 14) . We propose a SARS-CoV-2-specific lipopeptide as a candidate antiviral for preexposure and early postexposure prophylaxis for SARS-CoV-2 transmission in humans.

We recently described a monomeric SARS-CoV-2 HRC-lipopeptide fusion inhibitor (4) against SARS-CoV-2 with in vitro and ex vivo efficacy superior to previously described HRCderived fusion inhibitory peptides (6, 7) . To further improve antiviral potency, we compared monomeric and dimeric HRC-peptide derivatives (Fig. 1C) . Figure 1D shows antiviral potency in a quantitative cell-cell fusion assay of four monomeric and two dimeric SARS-CoV-2 S-derived 36-amino acid HRC peptides ( Fig. 1B ; see also figs. S1A and S3 for structure and characterization), with or without appended cholesterol. Dimerization increased the peptide potency for both nonlipidated peptides and their lipidated counterparts (Fig. 1D) . A dimeric cholesterol-conjugated lipopeptide based on the HPIV-3 F protein HRC domain, used as a negative control, did not inhibit fusion at any concentration tested (black line in Fig. 1D; see fig. S1 , B and C, for additional negative controls). Among the monomeric lipopeptides, the peptide bearing 24 molecules of polyethylene glycol (PEG 24 ) was most potent. The dimeric cholesterolconjugated peptide ([SARS HRC -PEG 4 ] 2 -chol; red line in Fig. 1D ) was the most potent lipopeptide against SARS-CoV-2 among our panel. This peptide also robustly inhibited fusion mediated by the S proteins of several emerging SARS-CoV-2 variants [including D614G, the variant bearing the Asp 614 →Gly mutation (15) ], the recent variants of concern B.1.1.7 and B.1.351 (16, 17) , and the S protein of SARS-CoV and Middle East respiratory syndrome coronavirus (MERS-CoV) (Fig. 1E ). Proposed anchoring of the dimeric lipopeptide in the host cell membrane and interactions with the viral S protein are shown in fig. S2 and movie S1. Collectively, these data suggest that the [SARS HRC -PEG 4 ] 2chol lipopeptide is equipped to combat an evolving pandemic.

For other enveloped respiratory viruses, we previously showed that both ex vivo and in vivo dimeric lipopeptides (administered intranasally) displayed increased retention in the respiratory tract compared with monomeric compounds (14) . In this study, we compared local and systemic biodistribution of our most potent monomeric and dimeric lipopeptides (SARS HRC -PEG 24 -chol and [SARS HRC -PEG 4 ] 2 -chol) at 1, 8, and 24 hours after intranasal inoculation or subcutaneous injection in humanized K18 hACE2 mice ( Fig. 2  and fig. S4 ). The two lipopeptides reached a similar lung concentration at 1 hour after intranasal administration (~1 to 2 mM). At 8 and 24 hours, the dimeric [SARS HRC -PEG 4 ] 2chol lipopeptide remained at high concentrations in the lung with minimal entry into the blood, but the monomeric peptide entered the circulation and the lung concentration decreased ( Fig. 2A) . The dimeric [SARS HRC -PEG 4 ] 2 -chol lipopeptide was distributed throughout the lung after intranasal administration (Fig. 2B) . A cellular toxicity (MTT) assay in primary human airway epithelial cells showed minimal toxicity even after 6 days at the highest concentrations tested (<20% at 100 mM) and no toxicity at its 90% maximal inhibitory concentration (IC 90 ) (~35 nM) ( fig. S5 ). The longer respiratory tract persistence of [SARS HRC -PEG 4 ] 2 -chol, in concert with its in vitro efficacy, led us to advance this dimeric lipopeptide.

The lead peptide, [SARS HRC -PEG 4 ] 2 -chol, was assessed for its ability to block entry of SARS-CoV-2 in VeroE6 cells or VeroE6 cells overexpressing transmembrane serine protease 2 (TMPRSS2), one of the host factors thought to facilitate viral entry at the cell membrane (2) . Whereas viral fusion in VeroE6 cells predominantly occurs after endocytosis, the virus enters TMPRSS2-overexpressing cells by fusion at the cell surface, reflecting the entry route in airway cells (18) . This difference is highlighted by chloroquine's effectiveness against SARS-CoV-2 infection in Vero cells but its failure in TMPRSS2-expressing Vero cells and the human lung (19) . The [SARS HRC -PEG 4 ] 2 -chol peptide dissolved in an aqueous buffer containing 2% dimethyl sulfoxide (DMSO) inhibited virus entry after 8 hours with a half-maximal inhibitory concentration (IC 50 ) of~300 nM in VeroE6 and~5 nM in VeroE6-TMPRSS2 cells (Fig. 3A ). To strengthen translational potential toward human use, the lipopeptide was reformulated in sucrose instead of DMSO, resulting in equivalent in vitro potency (Fig. 3B) . A control dimeric fusion-inhibitory lipopeptide directed against HPIV-3 blocked infection by HPIV-3 but did not inhibit SARS-CoV-2 infection. The in vitro efficacy data are summarized in table S1.

Ferrets are an ideal model for assessing respiratory virus transmission, either by direct contact or by aerosol transmission (20, 21) . Mustelids are highly susceptible to infection with SARS-CoV-2, as also illustrated by frequent COVID-19 outbreaks at mink farms. Direct-contact transmission of SARS-CoV in ferrets was demonstrated in 2003 (22) , and both direct-contact and airborne transmission of SARS-CoV-2 have been shown in ferrets (20, 23) . Direct-contact transmission in the ferret model is highly reproducible (100% transmission from donor to acceptor animals), but ferrets display limited clinical signs. After infection via direct inoculation or transmission, SARS-CoV-2 can readily be detected in and isolated from the throat and nose, and viral replication leads to seroconversion.

To assess the efficacy of [SARS HRC -PEG 4 ] 2chol in preventing SARS-CoV-2 transmission, naïve ferrets were dosed prophylactically with the lipopeptide before being cohoused with SARS-CoV-2-infected ferrets. In this setup, transmission via multiple routes can theoretically occur (aerosol, orofecal, and scratching or biting), and ferrets are continuously exposed to infectious virus during the period of cohousing, providing a stringent test for antiviral efficacy. The study design is shown in fig. S6 . Three donor ferrets (gray in fig. S6 ) were inoculated intranasally with 5 × 10 5 50% tissue culture infective dose (TCID 50 ) SARS-CoV-2 on day 0. Twelve recipient ferrets ] 2 -chol peptides for intranasal administration were dissolved to a concentration of 6 mg/ml in an aqueous buffer containing 2% DMSO, and ferrets were administered a final dose of 2.7 mg/kg of body weight (450 ml, divided equally in both nostrils). Peptide stocks and working dilutions had similar IC 50 values, confirming that peptide-treated ferrets were dosed daily with comparable amounts (fig. S7, A and B) . Six hours after the second treatment on 2 dpi, one infected donor ferret [highly positive for SARS-CoV-2, as indicated by reverse transcription-quantitative polymerase chain reaction (RT-qPCR)] was cohoused with four naïve recipient ferrets (two mock-treated and two peptide-treated). After a 24-hour transmission period in three separate, negatively pressurized HEPA-filtered Animal Biosafety Level 3 (ABSL3)-isolator cages, cohousing was stopped and donor, mock-treated, and peptidetreated ferrets were rehoused as separate groups. Additional [SARS HRC -PEG 4 ] 2 -chol peptide treatments were given to recipient animals on 3 and 4 dpi.

The viral loads (detection of viral genomes using RT-qPCR) for directly inoculated donor animals (gray), mock-treated recipient animals (red), and lipopeptide-treated recipient animals (green) are shown in Fig. 4, A (Fig. 4, A and B, red curves) . Productive SARS-CoV-2 infection was not detected in the throat or nose of any of the peptide-treated recipient animals (Fig. 4, A and B, green curves) . A slight rise in viral loads in samples collected at 3 dpi was detected (at the end of the cohousing), confirming that peptide-treated animals were exposed to SARS-CoV-2. In Fig. 4C , the area under the curve (AUC) shows the notable difference between the mock-treated and the peptide-treated animals. No infectious virus was isolated from lipopeptide-treated ferrets, whereas infectious virus was detected in all mock-treated ferrets (Fig. 4D) . Virus isolation data correlated with genome detection (Fig. 4E) .

Seroconversion occurred in donor ferrets and six of six mock-treated animals by 21 dpi but occurred in none of the peptide-treated recipient animals, as shown by S-and nucleocapsid (N)-specific immunoglobulin G (IgG) enzyme-linked immunosorbent assay (ELISA) and virus neutralization (Fig. 4 , F to H). Successful challenge infection confirmed that inhost virus replication had been completely blocked by the [SARS HRC -PEG 4 ] 2 -chol treatment ( Fig. 4I and fig. S8 ) and that none of the peptide-treated animals were protected, whereas the mock-treated animals (which had seroconverted) were all protected. Collectively, these data show that intranasal prophylactic administration of the [SARS HRC -PEG 4 ] 2 -chol peptide had protected six of six ferrets from transmission and productive infection.

In light of the persistence of the dimeric lipopeptide in the murine lung ( Fig. 2 and fig.  S4 ), we assessed the potential for a single administration of sucrose-formulated lipopeptide in a ferret transmission experiment 2 hours before cohousing to prevent or delay infection. In this experiment, we used a dimeric HPIV-3-specific lipopeptide as a mock control ( fig. S9 ). Although sucrose formulation had resulted in promising results in vitro at small scale (Fig. 3B) , formulation at larger scale resulted in incomplete dissolution. As a consequence, the sucrose-formulated [SARS HRC -PEG 4 ] 2 -chol lipopeptide was administered at a substantially lower concentration than in the experiment with the DMSO-formulated lipopeptide ( fig. S7, C and  D) . Nevertheless, the SARS-CoV-2 lipopeptide provided a significant level of protection as compared with the HPIV-3 control group, and four of six SARS-CoV-2 lipopeptidetreated animals were protected against infection. This experiment suggests that singleadministration preexposure prophylaxis is promising, although the optimal formulation and dosing regimen are an area of ongoing experimentation.

The intranasal [SARS HRC -PEG 4 ] 2 -chol peptide presented in this study is a successful prophylaxis that prevents SARS-CoV-2 transmission in a relevant animal model, providing complete protection during a 24-hour period of intense direct contact. Parallel approaches to prevent transmission that target the interaction between S and angiotensin-converting enzyme 2 have shown promise in vitro [e.g., the "miniprotein" approach (24) ]. The lipopeptide described here acts on the S2 domain after shedding of S1 ( fig. S2 and movie S1) and is complementary to strategies that target the functions of S1 or maintain S in its prefusion conformation, e.g., synthetic nanobodies (25) . Fusion-inhibitory lipopeptides could be used for pre-and postexposure prophylaxis in combination with these strategies and in conjunction with treatments [e.g., ribonucleoside analogs (26) ] that reduce replication in a treated infected individual. A combination of drugs that target different aspects of the viral life cycle is likely ideal for this rapidly evolving virus. Of note, the [SARS HRC -PEG 4 ] 2chol lipopeptide is equally active against several emerging SARS-CoV-2 variants, including D614G as well as the recent variants of concerns B.1.1.7 and B.1.351. The [SARS HRC -PEG 4 ] 2chol peptide has a long shelf life, does not require refrigeration, and can easily be administered, making it particularly suited to treating hard-to-reach populations. This is key in the context of COVID-19, which has affected every community, with the burden falling disproportionately on low-income and otherwise marginalized communities. This HRC lipopeptide fusion inhibitor is feasible for advancement to human use and should readily translate into a safe and effective nasal spray or inhalation-administered fusion inhibitor for SARS-CoV-2 prophylaxis, supporting containment of the ongoing COVID-19 pandemic.

are listed as inventors of [SARS HRC -PEG 4 ] 2 -chol on a provisional patent application covering findings reported in this manuscript. Data and materials availability: All data are available in the manuscript or the supplementary materials. Materials are available by MTA with the Trustees of Columbia University, NYC. Reagents are available from the corresponding authors under a material agreement with Columbia University