Murine monoclonal antibodies against RBD of SARS-CoV-2 neutralize authentic wild type SARS-CoV-2 as well as B.1.1.7 and B.1.351 viruses and protect in vivo in a mouse model in a neutralization dependent manner
After first emerging in December 2019 in China, severe acute respiratory syndrome 2 (SARS-CoV-2) has since caused a pandemic leading to millions of infections and deaths worldwide. Vaccines have been developed and authorized but supply of these vaccines is currently limited. With new variants of the virus now emerging and spreading globally, it is essential to develop therapeutics that are broadly protective and bind conserved epitopes in the receptor binding domain (RBD) or the whole spike of SARS-CoV-2. In this study, we have generated mouse monoclonal antibodies (mAbs) against different epitopes on the RBD and assessed binding and neutralization against authentic SARS-CoV-2. We have demonstrated that antibodies with neutralizing activity, but not non-neutralizing antibodies, lower viral titers in the lungs when administered in a prophylactic setting in vivo in a mouse challenge model. In addition, most of the mAbs cross-neutralize the B.1.351 as well as the B.1.1.7 variants in vitro.
Importance Crossneutralization of SARS-CoV-2 variants by RBD-targeting antibodies is still not well understood and very little is known about the potential protective effect of non-neutralizing antibodies in vivo. Using a panel of mouse monoclonal antibodies, we investigate both of these aspects.
The RBD of the spike protein of SARS-CoV-2 is relatively plastic and can tolerate extensive mutations, at least in vitro. The plasticity of the RBD is alarming because extensive changes in the RBD could reduce the efficacy of current vaccines and additional booster vaccinations with updated vaccines may be needed for protection in the future (15, 16). We tested all 14 isolated mAbs for binding to a whole panel of mutant RBDs. While some mAbs lost binding for many mutant RBDs, other mAbs maintained binding well across the board. However, binding was not in all cases directly linked to neutralization. All of the neutralizing mAbs maintained binding and neutralizing activity to B.1.1.7 (N501Y) relatively well. However, two mAbs lost neutralizing activity against B.1.351 and one of these mAbs only showed a relatively low reduction in binding to E484K and B.1.351 RBDs. The second one showed a stronger reduction in binding which agrees better with loss of neutralizing activity. Other hotspots for loss of binding for neutralizing antibodies included amino acid positions 487 and 490.
For four of the neutralizing mAbs low resolution structures were solved using single particle EM. They included KL-S-1D2 which lost neutralizing activity to B.1.351. Our low resolution structural analysis precludes interpretation of molecular interactions but the reduction or loss of neutralization of B1.1.7 and B.1.351 by KL-S-1D2 suggests that N501 and E484 form critical interactions.
Protection in vivo by neutralizing mAbs could be a function of Fc-Fc receptor interaction. This has been shown for other mAbs developed against SARS-CoV-2 which showed less protection in vivo when the Fc was mutated (25). While the role of Fc-FcR interactions based effector functions for SARS-CoV-2 targeting antibodies is not fully understood yet, it is likely that these effector functions contribute to protection (26). This has also been demonstrated for influenza viruses as well as ebolaviruses (27, 28). We tested all isolated mAbs for their protective effect in a mouse model and found that the only correlation with protection was neutralizing activity while non-neutralizing antibodies had no effect. However, there is an important caveat that needs to be discussed for this experiment. All non-neutralizing antibodies that we isolated were of the IgG1 subtype, which in mice, is known to have low affinity for activating FcRs. This is in contrast to murine IgG2a and IgG2b which have high affinity for these FcRs. Therefore, we can only conclude that non-Fc-FcR based interactions do not contribute to protection by non-neutralizing antibodies. In fact, the two antibodies that provided the best protection, especially on day 3, KL-S-1D2 and KL-S-2C3, are both of the IgG2a subtype. While KL-S-1D2 showed the best in vitro neutralization of all isolated mAbs, which could cause this phenotype, KL-S-2C3’s in vitro activity was lower but still showed stronger activity in vivo than other mAbs. This could be seen as evidence that Fc-FcR interactions, especially engagement with activating FcRs, which is an important component of protection. Of note, the vast majority of antibodies induced in humans to SARS-CoV-2 spike by natural infection or vaccination are IgG1 and in humans – unlike in mice - IgG1 has strong affinity for activating FcRs (29).
In summary, we describe several antibodies to the SARS-CoV-2 RBD that maintain strong neutralizing activity against the B.1.1.7 as well as B.1.351 variant. These mAbs, if humanized, may be further developed into ‘variant resistant’ therapeutics.