INTRODUCTION
Globally, mRNA vaccines have prevailed to mitigate the coronavirus disease 2019 (COVID-19) pandemic. Given the prompt progress in the development of vaccines and their fast rollout at a global scale, population immunity against the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) will largely depend on vaccine-induced rather than the infection-induced immunity. In this start of acquiring vaccine immunity as a society against COVID-19, the de novo repertoire of vaccine-elicited antibodies in SARS-CoV-2 infection-naive individuals will be the first step to build an optimal host defense system toward vaccine-based population immunity.
Currently, the efficacy of vaccine-induced immunity against SARS-CoV-2 in an individual is evaluated by potential surrogate markers, such as half-maximal neutralization titers (NT50s) using live or pseudotyped viruses and total antibodies titers against the receptor binding domain (RBD) of the spike protein of the virus (
1–4). Understanding the epitope profile of both vaccine recipients and naturally infected individuals can readily help elucidate the molecular basis of these markers as a surrogate. Moreover, the coevolution of vaccine-induced host immunity and virus escape will be one of the most important elements to consider in the way of achieving herd immunity against COVID-19.
The RBD of the spike protein of SARS-CoV-2 is widely considered the key protein target for designing vaccines and developing neutralizing antibodies as therapeutic agents (
5,
6). Epitope profiles of sera from individuals naturally infected with COVID-19 have enabled the identification of several immunodominant regions in the spike protein (
7–9). While most immunodominant epitopes are located outside the RBD, the minor proportion targeting specifically the neutralizing RBD epitopes explain the majority of viral neutralizability and protection against reexposures (
10,
11). In fact, neutralizing monoclonal antibodies (NAbs) developed as potential therapeutics also target mainly the epitopes located in the RBD (
6,
10,
12–15). While a growing number of individuals acquire vaccine immunity, the detailed epitope profile of the humoral immune response to the mRNA vaccine is not fully understood (
1,
16,
17).
In this study, high-resolution linear epitope profiling targeting the RBD was performed using sera of both mRNA vaccine recipients and COVID-19 patients. By comparing the epitope profiles, we sought to describe the similarities and differences between the humoral immune responses induced by BNT162b2 mRNA (Pfizer/BioNTech) vaccination and natural infection. Information provided by this study will be crucial in this postvaccine era of the COVID-19 pandemic.
DISCUSSION
This study revealed the linear epitope profiles targeting RBD elicited by BNT162b2 mRNA vaccination and natural infection of SARS-CoV-2. Our principal finding was that the variation of linear epitopes was broader in vaccine-elicited antibodies than that in infection-elicited antibodies, which may contribute to potent neutralization and thus resistance of the vaccine-elicited antibodies against the SARS-CoV-2 variants of concern.
Now, four categories of NAb classes are proposed to characterize the mode of recognition and epitope specificity (
18). Class 1 NAbs block several proximal sites in the receptor binding motif (RBM) of the RBD and directly block ACE2 binding (
18); class 2 NAbs recognize both up and down formations of the RBDs and epitope overlapping or close to ACE2-binding site (
18); class 3 NAbs recognize both up and down RBD and bind outside the ACE2-binding site (
18,
22); and class 4 NAbs bind only to up RBDs and do not directly block ACE2 binding, but destabilize the virus prefusion spike conformation (
24,
25). Many of the human-isolated NAbs target RBD, while some target the N-terminal domain of the subunit 1 spike protein (
14,
27). In our study, two classes of NAbs exclusively relevant in vaccine-elicited sera were found to be of specific note, namely, peptide no. 6 (
Fig. 5a and
b) targeted by the class 3 NAb (
22) and peptide no. 19 to 20 (
Fig. 5c) targeted by the class 4 NAb (
24,
25). These epitopes locate outside the ACE2-binding RBM (
Fig. 5), while epitopes detected commonly in both vaccine- and infection-elicited repertoires clustered adjacent to the ACE2-binding site (
Fig. 4).
The majority of NAbs targeting the RBM, which correspond to class 1 and 2, have been shown to exhibit decreased neutralization against the virus variants (
15,
17,
28). For example, antibodies recognizing the linear epitope, here included in peptide no. 33, would possibly fail in neutralizing variants with the K417N mutation as described previously (
15,
17,
28,
29). To the contrary, the linear peptides no. 6 and no. 19 and 20 (
Fig. 5a, b, and
c), corresponding to epitopes found exclusively in vaccine-elicited sera (
Fig. 2b and
3c), revealed corresponding epitopes targeted by human NAbs isolated from SAR-CoV convalescent cases (S309 and CR3022, respectively) (
22,
24,
25). These cross-neutralizing antibodies, belonging to class 3 and 4, recognize linear epitopes highly conserved among different CoV species. The epitopes recognized by these class 3 and class 4 NAbs are major contributors broadening the repertoire of vaccine-induced immunity. Located remotely from the RBM, such neutralizing epitopes stay rather free from variants (
15,
17), and explain the resistance of vaccine-elicited sera toward viral mutational escapes (
30,
31). The vaccine recipients’ broader epitope profile spanning across the RBD may give immunological flexibility and resilience against this evolving virus. Our mutation peptide panels have also presented a rather optimistic view on discussing the efficacy of vaccine-induced immunity to efficiently recognize the SARS-CoV-2 variants (
Fig. 6). However, considering that the linear epitope profiles harboring the mutation loci were not dominant in either vaccine sera or patient sera (
Fig. 2b,
Fig. 3b and
c) and an abundance of conformational epitopes were found adjacent to the RBM, the extent to which these specific linear epitopes contribute in net neutralizability remains to be determined (
10,
11,
14,
16).
We observed discrepancy between the neutralizability of sera obtained from vaccine recipients and patients, which could be partially explained by the difference in the time course of epitope selection and immune maturation. When comparing convalescent-phase sera and vaccine-elicited immunity, we found that the distribution of neutralizing epitopes was less generalized and focalized at specific peptides (
Fig. 3; individual epitope distribution can be found in Fig. S1). Among the two modes of acquired immunity, our results indicate that infection-induced humoral immunity had established a more mature, finely selected antibody repertoire. Our snapshot observations are in line with the ideas that the maturation of infection-provoked repertoires occurs as early as 10 to 20 days after onset, or even earlier in the case of COVID-19 beginning at 4 to 7 days after onset (
32,
33). The positive selection of relevant epitopes and the maturation of an antibody repertoire thus may lag behind in vaccine-induced immunity. Nevertheless, in this study, sera were sampled during the peak period of immune reaction in the host for both groups. Longitudinal evaluation of the epitope profiles and serological markers are needed for assessing host immune evolution and drawing conclusions to the above speculations.
In conclusion, we evaluated the similarity and difference in humoral immunity elicited by both the BNT162b2 mRNA vaccine and natural infection of SARS-CoV-2. High-resolution linear epitope profiles revealed the characteristic distribution of polyclonal antibodies spanning the RBD in vaccine recipient sera, which possibly accounted for the discrepancy observed in serological markers. Based on the multiplicity of neutralizing epitopes supporting the protectivity of vaccine-elicited antibodies, mRNA vaccine-elicited humoral immunity may harbor advantages in resisting the rapidly evolving pathogen.
There are several limitations in our study. The primary objective of our study was to compare the epitope profiles induced by vaccine and natural infection. However, the disease severity of the COVID-19 patients evaluated in the epitope analysis was skewed toward higher severity (one mild, two moderate, and seven critical patients), whereas the vaccine recipients were relatively healthy without major comorbidities. Indeed, the epitope profiles of critical patients (P01 to P07) seemed to differ from the ones of mild/moderate patients (P08 to P10) (
Fig. 2b), suggesting the potential discrepancy in COVID-19 immune response depending on patient clinical conditions. Their ages were equally distributed in both groups. This analysis was focused exclusively on the linear epitope profile targeting RBD. Moreover, the methodological limitation of the peptide binding assay using microarrays was that the results were semiquantitative. Therefore, as mentioned in the results, the reactive epitopes were characteristically selected by combining local peak detection and steric conformation in the structure with NAbs (
34,
35). In this manner, unreported and nondominant epitopes could be overlooked in this study. Experimental observations on compositional epitopes and epitopes outside the RBD region were not made in this study. Nonetheless, our results reporting the mRNA vaccine’s broad RBD epitope variety are in concordance with preceding reports (
30,
36).
ACKNOWLEDGMENTS
This work was funded by Japan Agency for Medical Research and Development (AMED) under grant no. JP20wm0125003 (Y.K.), JP20he1122001 (Y.K.), JP20nk0101627 (Y.K.), and JP20jk0110021 (Y. Nakagama). This work was also supported by JSPS KAKENHI grant no. JP21441824 (N.K.). We received support from Osaka City University’s “Special Reserves” fund for COVID-19. We also received the COVID-19 Private Fund (to the Shinya Yamanaka laboratory, CiRA, Kyoto University). Y. Nitahara received the BIKEN Taniguchi Scholarship.
We thank healthy volunteers and patients who participated in this study.
We are grateful to National Institute of Infectious Diseases, Tokyo, Japan, for providing for the virus and to James A. Rankin for his contribution in checking the manuscript.
We declare no conflicts of interest.
Y. Nitahara, Y. Nakagama. N.K., and Y.K. designed the study. Y. Nitahara, Y. Nakagama, N.K., H.Y., Y. Mizobata, H.K., and Y.K. selected patients and acquired clinical data. Y. Nitahara, Y. Nakagama, N.K., K.C., Y. Michimuko, E.T.-K., and M.Y. performed immunological assays. Y. Nitahara, Y. Nakagama, N.K., and Y.K. performed epitope mapping analysis. Y. Nakagama and M.Y. performed neutralization assays. Y. Nitahara, Y. Nakagama, N.K., and Y.K. wrote the manuscript and contributed to analysis and interpretation of the data. H.Y., Y. Mizobata, H.K., A.K., and M.Y. contributed to critical discussion of the manuscript. All authors have read and approved the manuscript.