Changes in serum-neutralizing antibody potency and breadth post-SARS-CoV-2 mRNA vaccine boost

Summary A better understanding of the durability and breadth of serum-neutralizing antibody responses against multiple severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants elicited by COVID-19 vaccines is crucial in addressing the current pandemic. In this study, we quantified the decay of serum neutralization antibodies (nAbs) after second and third doses of the original COVID-19 mRNA vaccine. Using an authentic virus-neutralization assay, we found that decay half-lives of WA1- and Delta-nAbs were both ∼60 days after second and third vaccine dose. Unexpectedly, the durability of serum antibodies that neutralize three different Omicron subvariants (BA.1.1, BA.5, BA.2.12.1) was substantially better, with half-lives of ≥6 months. A booster dose of the original COVID-19 vaccine was also found to broaden antibody responses against SARS-CoV and four other sarbecoviruses, in addition to multiple SARS-CoV-2 strains. These findings suggest that repeated vaccinations with the COVID-19 vaccine may confer a degree of protection against future spillover of sarbecoviruses from animal reservoirs.


INTRODUCTION
As of December 2022, more than 640 million people have been infected by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) with over 6.6 million deaths, according to the World Health Organization (''database:covid19.who.int''). At the same time, about 5.0 billion individuals (>63% of the global population; ourworldindata.org) have been vaccinated with various COVID-19 vaccines, most of which target the parental isolate of the virus that emerged in late 2019. Vaccination efforts, while successful, have slowed primarily due to hesitancy of select populations and the emergence of viral variants of concern (VOCs) that compromised vaccine efficacy. 1,2 More recently, multiple lineages of Omicron subvariants have resulted in breakthrough infections 3,4 in vaccinated individuals even though vaccines continue to protect against severe disease. 5,6 The increased risk of infection observed results from antigenic variation of the virus as well as declining neutralization antibody (nAb) titers in vaccinated individuals. In this study, we quantified the decay of vaccine-induced nAb titers against multiple SARS-CoV-2 variants after the second and third dose of vaccinations (referred to as pre-boost and post-boost, respectively). We also expanded our study to assess the breadth of nAbs against several SARS-CoV-2 variants and related sarbecoviruses that pose as future pandemic threats.

RESULTS & DISCUSSION
The antigenic variability generated by spike mutations found in VOCs has been extensively studied because of the significant impact that these mutations have on viral escape from polyclonal antibodies. [7][8][9][10][11] To underscore the benefits of a booster vaccination, we undertook a systematic examination of the quantitative and qualitative changes in serum antibodies in a cohort of vaccinees who were not infected by SARS-CoV-2 as determined by clinical history and serologic tests for antibodies to the viral nucleocapsid. We first studied the durability of vaccine-induced antibodies in neutralizing the authentic virus of the original Wuhan isolate and two VOCs, Delta and Omicron BA.1.1, by measuring the serum neutralization titers for a period of up to 200 days after the last vaccine dose. We first compared, cross-sectionally, the rate of decline in nAb titers against these variants in 149 serum samples obtained from individuals after their second dose of mRNA vaccination and in 86 serum samples obtained from these individuals after their third dose of mRNA vaccination. We found that, although boosting raised the nAb titers against all variants, it did not significantly alter their decay rates. Pre-and post-boost, the decay of neutralizing titers against WA1 and Delta was similar with mean half-lives (t 1/2 ) of $2 months ( Figure 1A). Very few samples possessed a measurable titer to the Omicron BA.1.1 after the second vaccine dose, making it difficult to determine a reliable half-life. The 86 post-boost serum samples had measurable titers to BA.1.1, with an estimated mean decay t 1/2 of >6.6 months, suggesting that the neutralization titers to Omicron, although significantly lower, decayed more slowly than those against parental and Delta strains.
To confirm the robustness of the above findings, we then studied the decay of serum nAbs longitudinally in a cohort of 16 individuals receiving either the Pfizer (N = 14) or the Moderna (N = 2) vaccine and booster doses but were never infected with SARS-CoV-2 based on clinical history and serologic testing (Table S1). Sera collected from these individuals for a period of up to 200 days after the second dose or  Article the booster dose were tested for neutralization against three authentic viruses (WA1, Delta, and Omicron BA.1.1). Consistent with the cross-sectional results, boosting led to higher titers but did not have a significant impact on the decay rates of antibodies that neutralized WA1 and Delta ( Figure 1B). After two vaccine shots, the neutralizing titers against Omicron BA.1.1 were generally low again, with 6 of 16 individuals with all measurements below detection and 10 of 16 with very low levels, thereby precluding any meaningful determination of the decay half-life. In contrast, the antibody titers against BA.1.1 post-boost were all above detection but surprisingly very stable over the period of follow-up, with a mean decay half-life R6.6 months ( Figure 1B).
Since the initial wave of Omicron infections, newer subvariants have gained global prominence, particularly BA.4 and BA.5, which are identical in spike, as well as BA.2.12.1. 11 To further gauge the protection against these subvariants, we again assessed the serum neutralization against authentic viruses in the aforementioned 16 longitudinal cases after their vaccine boost. The results showed that serum neutralizing titers against BA.5 and BA.2.12.1 decayed nearly identically, with mean half-lives of $6 months for both (Figure S1). This observation again suggested that nAbs against Omicron subvariants decay considerably slower than those against WA1 and Delta.
We next turned our attention to determine the changes in the breadth of the antibody response after vaccine boost. Pre-and post-boost sera from vaccinees (Table S1) were tested for neutralization against the aforementioned five authentic SARS-CoV-2 isolates as well as five pseudoviruses constructed from the spike genes of related sarbecoviruses: GD pangolin, SARS-CoV, Rs4084, Rs7327, and LYRa11 ( Figure S2). The results showed that the vaccine boost increased the neutralization potency not only against diverse SARS-CoV-2 isolates by > 5.8-to 16-fold but also against the sarbecoviruses tested by > 2.5-to 16-fold ( Figure 2A). The enhanced antibody potency post-boost was seemingly coupled to an expansion of antibody breadth, which could be graphically represented on an antigenic map, showing the compression of the relative antigenic distances between WA1 and other viruses following a third mRNA vaccine dose ( Figure S3). To confirm the observed expansion of antibody breadth, we also compared the neutralization of serum samples from a subset of 15 individuals from this cohort at 4-10 weeks before and after boost. Once again, substantial boosting of titers was observed, along with a remarkable widening of breadth against all viruses tested ( Figure 2B). In fact, the post-boost serum from all cases demonstrated capability of neutralizing all of the sarbecoviruses tested, again verifying the expansion of antibody neutralization breadth.
It has been previously shown that vaccines can generate B cell memory that initially declines quickly, 12,13 but the re-exposure of these B cell clones to the antigen can induce more persistent levels of the antibodies with greater potency and breadth. In this study, we observed a decay in humoral responses to SARS-CoV-2 WA1 and Delta strains with a half-life of approximately 2 months, as has been reported previously. 14,15 However, using sera from the same clinical cohort, we noticed a remarkably greater stability in the duration of the antibody neutralizing responses to Omicron BA.  16 ; therefore, the nAb decay against these subvariants could not be reliably measured.
A recent SARS-CoV-2 vaccine study has shown that increased antigen availability in germinal centers after booster vaccination activates pre-existing memory B cells producing high-affinity antibodies targeting subdominant but conserved spike epitopes by masking the dominant epitopes. 17 Other studies have observed that a two-dose SARS-CoV-2 vaccination regimen after natural infection increased antibody durability and enhanced breadth across multiple SARS-CoV-2 VOCs. 15 Cross-sectional analyses have also suggested that a third dose of the original Pfizer or Moderna vaccine increases serum neutralization titers to Beta, Gamma, Delta, and Omicron BA.1 variants. 15 A booster shot of a COVID-19 mRNA vaccine is known to generate a more diverse repertoire of memory B cells and to yield monoclonal antibodies targeting conserved regions of the receptor-binding domain (RBD) of the viral spike. 18,19 In addition, a vaccine boost has been found to proportionally decrease antibodies targeting strain-specific epitopes (RBD Class 1 and 2) and to gradually ll OPEN ACCESS iScience 26, 106345, April 21, 2023 3 iScience Article increase antibodies targeting conserved epitopes (RBD Class 3 and 4). 18 Our findings on a COVID-19 vaccine boost are consistent with these prior observations, indicating that broadening of serum antibody responses could be elicited without boosting with a new vaccine based on the spike of an emergent variant. What is new, and perhaps unexpected, is the expansion of the nAb breadth extends beyond SARS-CoV-2 variants to include SARS-CoV and other related sarbecoviruses found ubiquitously in animal reservoirs, particularly in bats. 20 It is possible that the growing herd immunity in the population from repeated COVID-19 vaccinations and infections may be conferring a degree of protection against future spillover of sarbecoviruses from other animals into humans.

Limitations of the study
Our analysis was restricted to 16 individuals in the longitudinal cohort, and few samples had a shorter follow-up period due to exclusion of volunteers when they turned COVID-positive by nucleocapsid antigen testing. With these limitations, it should be noted that we employed a single exponential model that provided the best statistical support for decay analysis over a power-law or biexponential decay models, especially for minimal decay of titers against Omicron variants.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

DECLARATION OF INTERESTS
The authors do not declare any competing interests for this study.  Vaccine sera An end-point-dilution microplate neutralization assay was performed to measure the neutralization activity of vaccinee sera. Triplicates of each five-fold dilution beginning with a 1:50 dilution of sera were incubated with SARS-CoV-2 at an MOI of 0.1 in EMEM with 7.5% inactivated fetal calf serum (FCS) for 1 hour at 37 C. Post incubation, the virus-antibody mixture was transferred onto a monolayer of Vero E6 cells grown overnight. The cells were incubated with the mixture for $70 hrs. CPE was visually scored for each well in a blinded fashion by two independent observers. The results were then converted into percentage neutralization at a given sample dilution, and the averages G SEM were plotted using a non-linear five-parameter dose-response curve to obtain the ID50 of each sample using GraphPad Prism v.9.3. Statistical analyses between groups was performed using Wilcoxon paired t-test using GraphPad Prism 9.3

Pseudovirus production of sarbecoviruses
Spike gene for SARS-CoV and sarbecovirus S genes were codon-optimized for mammalian expression, synthesized by Twist Biosciences, and cloned into the same expression vectors as above by Gibson Assembly (New England Biolabs). Sarbecovirus sequences were retrieved from GenBank under the following accession numbers: GD Pangolin (MT799524), Rs7327 (KY417151), Rs4084 (KY417144), and LYRa11 (KF569996). Recombinant VSV pseudoviruses in which the native glycoprotein was replaced with sarbecovirus S proteins were generated as previously described. 24,25 Briefly, human embryonic kidney (HEK) 293T cells (ATCC), at a confluency of 80% were transfected with a S protein expression vector using PEI (1 mg/mL) and cultured overnight at 37 C under 5% CO 2 . Twenty-four hours later, cells were infected with VSV-Gpseudotyped DG-luciferase (G*DG-luciferase, Kerafast) at a multiplicity of infection (MOI) of 3 for 2 hours. Afterward, cells were washed three times with 13 PBS, changed to fresh medium, and cultured at 37 C for another 24 hours before supernatants were harvested and clarified by centrifugation at 300gfor 10 min.

Pseudovirus neutralization assay
Pseudoviruses were titrated to standardize the infectivity levels for target cells before setting up neutralization assays. Neutralization assays were then performed as described earlier 24  The model was implemented in two versions. In the first, for the cross-sectional data ( Figure 1A), we fitted all the data for the different variants before and post-boost simultaneously, using the software R, 26 with the package censReg, 27 which fits general linear models with censored data. This model included covariates for boost and strain, allowing for different initial values and decay rates for each variant and pre-post-boost. We used the log-likelihood ratio test to compare the different models (for example, the effect of boosting on the decay rates). In the second version, we used the same model with a mixed-effects population approach to fit the longitudinal data. We used Monolix 2021R2 (Lixoft SAS, a Simulations Plus company, Antony, France) to fit the data for all the individuals simultaneously, pre-boost and post-boost, for the Wuhan, Delta and Omicron BA.1 variants (we fitted the Omicron BA.5 and BA.2.12.1 longitudinal data separately, because there was no pre-boost data for these variants). In this model version, l 2 and l 3 were parameterized as a 1 l 1 and a 2 l 2 , respectively, which allowed us to test whether a 1 = 1 and/or a 2 = 1, and thus whether the decay rates were the same for the three different variants with pre-and post-boost. The model also included the effect of boost, both as a random effect and a covariate for the initial titers and the decay rates. We compared the fits using the corrected Bayesian Information Criterion (cBIC), as provided by Monolix. The cBIC is used for model selection, with the lowest cBIC indicating the preferred model. The cBIC takes into consideration the number of parameters used to fit the data and penalizes models with more parameters. We tested other model structures, for example biexponential decay or a power-law decay, but there was no statistical support for these over the simpler single exponential model. For the fits in Figure S1, for the BA.5 and BA.2.12.1 variants, we used a similar approach with parameterization l 5 = a 3 l 4 . But we found that a 3 = 1 provided the best fit, indicating no difference in the decay rates of antibody titers against these two variants.

Antigenic mapping of neutralization data
We utilized antigenic cartography to explore the impact of a third mRNA vaccine dose on antigenic distances between multiple SARS-CoV-1 and SARS-CoV-2 related viruses. Antigenic maps were generated separately using sera samples following two or three doses of an mRNA vaccine. The relative positions of viruses and sera on the maps were determined and optimized as described previously 21,28-30 and such that each antigenic distance unit (AU) corresponds to a two-fold change in ID50. Geometric uncertainty was assessed for all points on both maps using a stress limit of one. Maps were created using the Racmacs package (https://acorg.github.io/Racmacs/)in R. All maps were constructed using 1000 optimizations per map with a dilution step size of 0 and the minimum column basis parameter set to ''none.'' Serum positions are represented by grey squares, while virus positions are represented by colored circles. Geometric uncertainty was assessed for all points on both maps using a stress limit of one and is illustrated for virus positions as colored regions. iScience Article QUANTIFICATION AND STATISTICAL ANALYSIS Neutralization assays were quantified by using a non-linear five parameter regression analysis curve fit to determine the IC50 values that were used in generation of the decay and breadth plots as determined in GraphPad Prism 9.3. Quantification of the decay half-lives were calculated as described in detail in the analyses of ID50 decay section for both the cross-sectional and longitudinal cohorts. Racmacs package was used to define the antigenic cartography maps using the antigenic distance units as described in the antigenic mapping methods section.