Bivalent COVID-19 Vaccines: Can the Original Antigenic Sin Be Forgiven?

On 31 August 2022, the United States (US) Food and Drug Administration authorized bivalent formulations of the Moderna and Pfizer-BioNTech coronavirus disease 2019 (COVID-19) vaccines [1]. These modified vaccines contain messenger RNA (mRNA) encoding for both the ancestral WA1/2020 and the Omicron BA.4/BA.5 spike proteins. The hope was that they would provide immunity to the BA.5 virus, which differs from WA1/2020 by >30 mutations in the spike protein and was the predominant variant in circulation at the time. Unfortunately, studies have shown that the levels of neutralizing antibodies to the BA.5 variant were not significantly higher in patients who received the bivalent vaccine than those who received the monovalent vaccine with WA1/2020 mRNA [2, 3]. In other words, vaccination with BA.5 spike protein did not lead to an appreciably better antibody response. The reason for this disappointing result has not yet been determined, but antigenic imprinting, also known as the “original antigenic sin,” has been invoked as a potential cause for these results.

To fully explain this tenant of immunotheology, one has to review some basic principles. While there are some notable exceptions [4], the process of extensive gene segment recombination leads to the production of diverse T- and B-cell receptors that should be capable of recognizing almost any conceivable pathogen we will ever encounter. The drawback is that there are very few cells with any given receptor and so when we first see a pathogen, naive cells with receptors specific for the pathogen's antigens have to proliferate extensively before we can mount an effective primary adaptive immune response. Some of these naive cells will become memory cells that circulate at higher levels so that if the same antigen is encountered again, a faster, more effective secondary adaptive immune response will occur.

The original antigenic sin occurs when we encounter antigens on a pathogen that is similar to one we have previously encountered [5, 6]. It is thought that rather than initiating a primary immune response where rare naive lymphocytes that have high-affinity receptors for the new antigens proliferate, a secondary response occurs where memory cells with receptors that are cross-reactive for both the original and new antigens are stimulated. This could potentially be advantageous if the targeted epitopes of the 2 pathogens are identical or very similar because then the 2 epitopes will be equally recognized. However, in many cases, because the cross-reactive memory cells were primed to respond to the antigen on the first pathogen, the cross-reactive receptors will have a higher affinity for the first pathogen, leading to a suboptimal response to the newer pathogen (Figure 1).

Figure 1.

Three broad categories of severe acute respiratory syndrome coronavirus 2–specific cells will exist after completing the primary vaccine series: cells with monoreactive high-affinity receptors for WA1/2020 spike protein (blue), cells with lower-affinity receptors that cross-recognize WA1/2020 and variant spike proteins (green), and cells with monoreactive high-affinity receptors for variant spike proteins (yellow). Before immunization with variant spike messenger RNA, patients will have a high frequency of WA1/2020-monoreactive memory cells as well as some cross-reactive memory cells. There will also be a very low frequency of naive cells with monoreactive receptors for every variant spike protein. The optimal result of vaccination would be a selective amplification of the naive T cells. The original antigenic sin occurs when there is an amplification of cross-reactive cells at the expense of naive cells with high-affinity receptors for the variant spike protein. Figure created with BioRender.com. –, no affinity; ++, low affinity; +++, moderate affinity; ++++, high affinity.

Does the original antigenic sin explain the failure of the bivalent vaccines to induce higher BA.5-specific responses than monovalent vaccines? What we know for sure is that monovalent vaccines containing just the WA1/2020 spike protein were able to induce BA.5-specific responses, suggesting that cells with cross-reactive receptors exist and were amplified by the ancestral spike protein. However, the presence of cross-reactive receptors alone does not necessarily lead to the original antigenic sin. We have previously shown that there are T cells with cross-reactive receptors that recognize spike proteins from both the common cold coronavirus HCoV-NL63 and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) WA1/2020 in patients who have recovered from COVID-19 [7]. These T cells were most likely generated during prior HCoV-NL63 infections as their receptors have higher affinity for the common cold virus than for SARS-CoV-2 [7]. SARS-CoV-2 mRNA vaccination resulted in an expansion of T cells that recognized HCoV-NL63, but these individuals also developed a very potent T-cell response to SARS-CoV-2 [8], and T-cell receptor analysis demonstrated that SARS-CoV-2–monoreactive T cells were generated [9]. A similar analysis should be done for recipients of the monovalent and bivalent vaccine to determine whether in the latter case, cells with cross-reactive receptors that co-recognized WA1/2020 and BA.5 spike proteins were amplified at the expense of naive cells with higher-affinity receptors for BA.5 spike proteins.

We also should consider the fact that unless one has had a natural infection with the BA.5 variant, the BA.5 protein produced by the bivalent vaccine will be eliciting a primary immune response. This will be the first time that naive cells with high-affinity receptors for this spike protein will encounter the antigen. In contrast, due to completion of the primary vaccine series, cells with high-affinity receptors for the WA1/2020 spike protein will exist in a memory state. Thus, the lack of a more potent response to BA.5 following bivalent vaccination in some cases may reflect the fact that we are looking at a primary immune response. If that is the case, then there is a chance that subsequent exposure to BA.5 spike protein, either by vaccination or natural infection, will lead to an improved response. Unfortunately, by the time 2 bivalent booster shots are given to a significant part of the population—an unlikely prospect given the limited uptake of the bivalent vaccine and the vaccine weariness of the US population—the variant in question will probably no longer be the dominant variant in circulation. The question then becomes should we periodically modify the bivalent vaccines to target the spike protein from the most prevalent circulating variant, or can we just rely on a monovalent vaccine to expand preexisting memory cells with cross-reactive receptors?

It is likely that we have low frequencies of naive B and T cells with high-affinity receptors specific for every SARS-CoV-2 variant that will emerge. So, if we continue to vaccinate with different spike proteins and generate more cells with cross-reactive receptors, will we be able to prime the rare naive cells with the high-affinity receptors we want, or will we just amplify cells with less effective cross-reactive receptors that are circulating at much higher levels? The good news is that if new variants escape to a point where there is little cross-recognition by preexisting memory B and T cells, then it should be possible to prime an effective primary immune response against the emerging spike protein. This phenomenon is routinely seen in untreated human immunodeficiency virus infection where there is a much higher degree of virologic escape [10]. At this point, hopefully the original antigenic sin will be forgiven.

Notes

Financial support. This work was supported by NIH grant R21AI67705.