The development of highly immunogenic and safe vaccines will be critical for controlling the COVID-19 pandemic, and no one type of vaccine will likely fill the global need. Here, we describe the selection of immunogens that are effective at generating high titers of neutralizing antibodies for sustained periods of time and their evaluation in a NHP model species.
While the COVID-19 pandemic is caused by SARS-CoV-2, prior research on related coronaviruses have revealed key features that shaped vaccine and therapeutic development efforts. A key immunological target is the viral surface protein that is required for virus entry and replication (34). Indeed, the spike protein has been the primary target for SARS-CoV vaccines and has been demonstrated to elicit a robust and protective immune response in nonhuman primate animal model systems(29). Thus SARS-CoV-2 vaccine development has benefited from experience gained in the SARS-CoV vaccine studies. These studies demonstrated that neutralizing antibodies inhibit the interaction of spike protein with its receptor ACE2 and therefore prevent viral entry and replication(17, 32, 35).
Neutralizing antibodies from recovered COVID-19 patients target epitopes of spike S1, more specifically the RBD and NTD(18, 36, 37) (15, 17, 38). Studies by Qi et al. have shown that an RBD-Fc fusion protein (residues 319–541 fused to a mouse IgG1 Fc domain) elicited potent neutralizing antibodies in mice(39). A conflicting report from Wang et al. showed in a murine model that immunization with the RBD expressed with a norovirus shell domain was not sufficient to elicit neutralizing antibodies and the entire S1 portion was needed(40). While the RBD-based constructs are similar to the one used in this study, the RBD-Fc fusion protein used here also contains an additional 50 residues (541–591) of the C-terminal S1 domain, which may aid in immunogenicity directly, by providing more potential epitopes, or indirectly, by enhancing stability of the RBD-Fc construct or by providing conformational flexibility. Our C-terminally extended RBD construct dimerizes through the disulfide bridge present in the Fc portion, which may also provide non-linear epitopes in its dimerized form. Most proteins benefit from the enhanced pharmacological stability and solubility provided by fusion to a Fc domain(41). In addition, an Fc fusion can increase uptake by antigen presenting cells (APC) that express the Fc receptor, which in turn can enhance immunogenicity(42, 43). Our findings in the macaque model system are consistent with the work of Qi et al., as we found that immunization with the extended RBD-Fc containing formulations resulted in the generation of high antibody titers (Fig. 3B) with the ability to inhibit RBD binding to ACE2 (Fig. 4) and robust neutralizing activity (Fig. 5B).
A prior study by Ren et al., which used whole spike S1-Fc as an immunogen in macaques and a complex immunization schedule with five doses and multiple adjuvants, reported the successful generation of neutralizing antibody titers(44). We sought to test an immunization schedule with fewer administrations. The study was originally designed to deliver a primary immunization followed by 2 boosts, 2 weeks apart if needed. However, it was determined that the first 2 doses elicited a sufficiently robust antibody response with endpoint titers exceeding 106 (Fig. 3C) in the RBD-Fc receiving group (2 and 3) and strong neutralizing activity (Fig. 4C, 5A).
While studies have shown SARS-CoV-2 infection of humans elicits antibodies that target the NTD(16, 18), immunization with the NTD-Fc elicited only a moderate antibody response with a maximum spike-binding titer that was several orders of magnitude less than the extended RBD-Fc. Moreover, no neutralization activity was detected at any time (Fig. 5A). While this does not rule out a possible role for the NTD as an immunogen, it indicates that our extended RBD-Fc construct is sufficient for elicitation of a strong neutralizing antibody response. Using the mix of the RBD-Fc and NTD-Fc (group 3) resulted in the generation of a similar level of neutralization activity as did the use of the RBD-Fc alone (group 2) (Fig. 5B), indicating that the NTD-Fc does not contribute to the neutralizing activity in this context. These data also suggest that a lower dose of the extended RBD-Fc may be as effective as a higher dose, because group 3 received 50% less RBD-Fc than group 2. In addition, the neutralizing antibody titers generated by the extended RBD-Fc immunogen was significantly higher than is typical of individuals who have recovered from a natural SARS-CoV-2 infection or have been immunized with an mRNA or adenovirus vaccine(19, 45–47). In preclinical studies in rhesus macaques immunized with the mRNA-1273 vaccine, a reduction in both spike binding IgG and neutralizing antibodies was observed at 8 weeks after the initial vaccination(48). This suggests that the RBD-containing C-terminal domain protein used in this study (RBD-Fc) may result in a stronger, more durable antibody response than the newly approved mRNA-1273 vaccine.
Clearly an immunogen that can elicit a robust and durable antibody response would be a preferred vaccine candidate. In this study, we demonstrated a robust and sustained humoral response to immunization but did not examine the cell mediated response, which may contribute to protection. Studies with recovered COVID-19 patients have identified additional factors beyond spike-mediated binding and entry that may be important for protection or clearance of virus(49–51). However, studies by Du et al. on SARS-CoV demonstrate that protection from viral challenge was largely due to neutralizing antibodies resulting from RBD immunization(52–54). Further studies using the extended RBD-Fc and NTD-Fc could examine additional aspects of immunity, alternative adjuvants, alternative dosing schedules including a single dose in macaques.