Antibody binding to SARS-CoV-2 S glycoprotein correlates with, but does not predict neutralization

Convalescent plasma from SARS-CoV-2 infected individuals and monoclonal antibodies were shown to potently neutralize viral and pseudoviral particles carrying the S glycoprotein. However, a non-negligent proportion of plasma samples from infected individuals as well as S-specific monoclonal antibodies were reported to be non-neutralizing despite efficient interaction with the S glycoprotein in different biochemical assays using soluble recombinant forms of S or when expressed at the cell surface. How neutralization relates to binding of S glycoprotein in the context of viral particles remains to be established. Here we developed a pseudovirus capture assay (VCA) to measure the capacity of plasma samples or antibodies immobilized on ELISA plates to bind to membrane-bound S glycoproteins from SARS-CoV-2 expressed at the surface of lentiviral particles. By performing VCA and neutralization assays we observed a strong correlation between these two parameters. However, while we found that plasma samples unable to capture viral particles did not neutralize, capture did not guarantee neutralization, indicating that the capacity of antibodies to bind to the S glycoprotein at the surface of viral particles is required but not sufficient to mediate neutralization. Altogether, our results highlights the importance of better understanding the inactivation of S by plasma and neutralizing antibodies.

However, a non-negligent proportion of plasma samples from infected individuals as well as 23 S-specific monoclonal antibodies were reported to be non-neutralizing despite efficient 24 interaction with the S glycoprotein in different biochemical assays using soluble recombinant 25 forms of S or when expressed at the cell surface. How neutralization relates to binding of S 26 glycoprotein in the context of viral particles remains to be established. Here we developed a 27 pseudovirus capture assay (VCA) to measure the capacity of plasma samples or antibodies 28 immobilized on ELISA plates to bind to membrane-bound S glycoproteins from SARS-CoV-29 2 expressed at the surface of lentiviral particles. By performing VCA and neutralization 30 assays we observed a strong correlation between these two parameters. However, while we 31 found that plasma samples unable to capture viral particles did not neutralize, capture did not 32 guarantee neutralization, indicating that the capacity of antibodies to bind to the S 33 glycoprotein at the surface of viral particles is required but not sufficient to mediate 34

Introduction 42
The etiologic agent of the Coronavirus Disease , the severe acute respiratory 43 syndrome coronavirus 2 (SARS-CoV-2), spread from China to the rest of the world and is 44 the cause of the current pandemic [1]. In the absence of an effective vaccine to prevent SARS-45 CoV-2 infection, alternative approaches to treat or prevent acute COVID-19 are desperately 46 needed. A promising approach is the use of convalescent plasma containing anti-SARS-CoV-47

Virus neutralization assay 106
Target cells were infected with single-round luciferase-expressing lentiviral particles as 107 described previously [14]. Briefly, 293T cells were transfected by the calcium phosphate 108 method with the lentiviral vector pNL4.3 R-E-Luc (NIH AIDS Reagent Program) and a 109 plasmid encoding for SARS-CoV-2 Spike, SARS-CoV-1 Spike at a ratio of 5:4. Two days 110 post-transfection, cell supernatants were harvested and stored at -80°C until use. 293T-ACE2 111 target cells were seeded at a density of 1×10 4 cells/well in 96-well luminometer-compatible 112 tissue culture plates (Perkin Elmer) 24h before infection. Recombinant viruses in a final 113 volume of 100μl were incubated with the indicated sera dilutions (1/50; 1/250; 1/1250; 114 1/6250; 1/31250) for 1h at 37°C and were then added to the target cells followed by 115 incubation for 48h at 37°C; cells were lysed by the addition of 30μl of passive lysis buffer 116 (Promega) followed by one freeze-thaw cycle. An LB941 TriStar luminometer (Berthold 117 Technologies) was used to measure the luciferase activity of each well after the addition of 118 100μl of luciferin buffer (15mM MgSO4, 15mM KPO4 [pH 7.8], 1mM ATP, and 1mM 119 dithiothreitol) and 50μl of 1mM d-luciferin potassium salt (Prolume). The neutralization half-120 maximal inhibitory dilution (ID50) or the neutralization 80% inhibitory dilution (ID80) 121 represents the sera dilution to inhibit 50% or 80% of the infection of 293T-ACE2 cells by 122 recombinant viruses bearing the indicated surface glycoproteins. 123 124

Results and Discussion 125
Virus capture assay 126 To measure the capacity of plasma samples and monoclonal antibodies immobilized on 127 ELISA plates to bind to membrane-bound S glycoproteins from SARS-CoV-2 expressed at 128 the surface of lentiviral particles we adapted a previously-described virus capture assay 129 (VCA) [19]. The pseudoviral particles used in this assay are generated by transfecting 130 HEK293T cell with the pNL4.3 Nef-Luc Env-construct [21][22][23][24]. This construct is co-131 transfected with a plasmid encoding the S glycoprotein from SARS-CoV-2 and a plasmid 132 encoding the G glycoprotein from vesicular stomatitis virus (VSV-G), resulting in a virus 133 capable of a single round of infection. Pseudovirus-containing supernatants are added to 134 plasma or antibody-coated ELISA plates and unbound pseudoviruses are washed away. 135 Retention of pseudoviruses on the surface of the plate by S-specific antibodies is visualized 136 by the addition of Cf2Th cells. These cells are refractory to SARS-CoV-2 S-mediated entry 137 (not shown) but are infected by the bound pseudoparticles via the incorporated G 138 glycoprotein from VSV. Cf2Th infection is measured as a function of luciferase activity two 139 days later. A scheme of the assay is depicted in Figure 1A. This assay was used with plasma 140 samples from SARS-CoV-2-convalescent individuals and, as expected, showed a robust 141 capture of pseudoviral particles whereas plasma obtained from SARS-CoV-2 uninfected 142 individuals failed to do so ( Figure 1B). Similarly, previously-described SARS-CoV-2 S-143 specific monoclonal antibodies isolated from convalescent donors [12] were able to capture 144 pseudoviruses expressing the S glycoprotein from SARS-CoV-2. Of note, the capture was 145 specific since plasma from convalescent individuals and monoclonal antibodies failed to 146 capture similar pseudoviruses expressing the S glycoprotein from SARS-CoV-1 ( Figure 1B). 147

148
Recognition of S glycoproteins at the surface of pseudoviral particles is required but no 149 sufficient to neutralize. 150 It is presently unclear whether the capacity of plasma from convalescent donors or 151 monoclonal antibodies to neutralize pseudoviral particles correlates with their capture 152 efficiency. Therefore, we first measured the capacity of plasma samples recovered six 153 weeks after symptoms onset from twenty-five convalescent individuals to neutralize 154 pseudoparticles bearing the SARS-CoV-2 S glycoprotein using 293T cells stably expressing 155 ACE2 as target cells, as described [14]. Neutralizing activity, as defined by the 156 neutralization half-maximum inhibitory dilution (ID50), is shown in Figure 2A. These results 157 illustrate the variable capacity of different convalescent plasma samples to neutralize. We 158 then measured the capacity of the same plasma samples to capture viral particles. As 159 mentioned above, HEK293T cells were co-transfected with pNL4.3 Nef-Luc Env-together 160 with plasmids expressing the SARS-CoV-2 S glycoprotein and VSV-G; released viral 161 particles were collected two days after transfection. For the VCA, 96-well microplates were 162 coated with plasma recovered from SARS-CoV-2 uninfected individuals (control) or plasma 163 from convalescent donors or with the receptor-binding domain (RBD)-specific CR3022 164 monoclonal antibody. This antibody was added to each plate for normalization purposes. 165 Pseudoviral particles were added to the plates and incubated for 4 hours at 37 ºC; the plates 166 were then washed to remove unbound pseudoviruses. Cf2th cells were added to the wells, 167 incubated at 37°C and lysed 48 hours later to measure luciferase activity. As reported in 168 Figure 2B, while a few plasma samples from convalescent donors failed to capture viral 169 particles, similar to plasma from uninfected donors, most did. Interestingly, even though we 170 observed a significant correlation between virus capture and neutralization ( Figure 2C), virus 171 capture did not always translate into neutralization. Indeed, we observed several 172 convalescent plasma samples that were able to efficiently capture pseudoviral particles but 173 did not neutralize, thus indicating that while interaction of SARS-CoV-2 S glycoprotein is 174 required for virus neutralization, it is not sufficient per se. 175

176
The capacity of convalescent plasma to bind to the S glycoprotein of SARS-CoV-2 and 177 neutralize pseudoviral particles decreases over time. 178 Several reports described a significant decrease in the neutralization capacity of plasma from 179 convalescent individuals starting six weeks after symptoms onset [13,14,25,26]. To 180 evaluate whether this was related to the capacity of plasma to capture viral particles, we 181 analyzed the neutralization and viral capture capacity of serological samples obtained from 182 fifteen convalescent donors at six and ten weeks after symptoms onset. As previously 183 reported [13], we observed a significant decrease in the neutralizing capacity of samples over 184 time ( Figure 3A). A similar decrease in their capacity to capture viral particles was observed 185 in the present study ( Figure 3B). Accordingly, we noted a strong correlation between 186 neutralization and virus capture ( Figure 3C C P 3 C P 7 C P 8 C P 1 4 C P 2 3 C P 2 4 C P 1 6 C P 1 9 C P 2 5 C P 1 C P 4 C P 1 0 C P 9 C P 5 C P 2 2 C P 1 2 C P 2 0 C P 2 C P 1 8 C P 1 5 C P 2 1 C P 1 3 C P 1 1 C P 1 7 C P 6 C P 3 C P 7 C P 8 C P 1 4 C P 2 3 C P 2 4 C P 1 6 C P 1 9 C P 2 5 C P 1 C P 4 C P 1 0 C P 9 C P 5 C P 2 2 C P 1 2 C P 2 0 C P 2 C P 1 8 C P 1 5 C P 2 1 C P 1 3 C P 1 1 C P 1 7 C P 6