Ad26.COV2.S priming provided a solid immunological base for mRNA-based COVID-19 booster vaccination

Summary The emergence of novel SARS-CoV-2 variants led to the recommendation of booster vaccinations after Ad26.COV2.S priming. It was previously shown that heterologous booster vaccination induces high antibody levels, but how heterologous boosters affect other functional aspects of the immune response remained unknown. Here, we performed immunological profiling of Ad26.COV2.S-primed individuals before and after homologous or heterologous (mRNA-1273 or BNT162b2) booster. Booster vaccinations increased functional antibodies targeting ancestral SARS-CoV-2 and emerging variants. Especially heterologous booster vaccinations induced high levels of functional antibodies. In contrast, T-cell responses were similar in magnitude following homologous or heterologous booster vaccination and retained cross-reactivity towards variants. Booster vaccination led to a minimal expansion of SARS-CoV-2-specific T-cell clones and no increase in the breadth of the T-cell repertoire. In conclusion, we show that Ad26.COV2.S priming vaccination provided a solid immunological base for heterologous boosting, increasing humoral and cellular responses targeting emerging variants of concern.


INTRODUCTION
The emergence of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) variants that are antigenically distinct and can evade vaccine-induced antibody responses 1,2 resulted in the recommendation of COVID-19 booster vaccinations. 3,4 Currently circulating variants are predominantly viruses from the Omicron sub-lineage. These variants harbor several mutations in the spike (S) protein that allow for partial immune escape at the antibody level. Previous studies have shown that mRNA-based booster vaccinations increase both S-specific antibodies and to a lesser extent T-cell responses, and restore clinical protection against severe disease after infection with antigenically distinct variants. [5][6][7][8] According to the final evaluation of the phase 3 clinical trial, vaccination with a single dose of Ad26.COV2.S induces protection against moderate to severe-critical COVID-19, to varying degrees between different SARS-CoV-2 variants and the ancestral virus. 2 This is explained by the fact that vaccination-induced antibodies have reduced reactivity with SARS-CoV-2 Omicron sub-lineages. In contrast, CD4 and CD8 T-cell responses do cross-react with emerging variants. 9 Compared to the mRNA-based vaccines, primary Ad26.-COV2.S vaccination yielded lower levels of S-specific antibodies, but these antibody levels remained stable for at least 6 months. 8,10 Since S-specific neutralizing antibodies were originally identified as a correlate of protection against COVID- 19 8,11,12 booster vaccinations of Ad26.COV2.S-primed individuals were recommended to increase protection against emerging variants. Boosting Ad26.COV-2.S-primed individuals with Ad26.COV2.S, BNT162b2, or mRNA-1273 proved safe and effective, 10,13,14 and SARS-CoV-2-specific antibody and T-cell responses are higher after heterologous boosting with an mRNA-based vaccine. 15 SARS-CoV-2 neutralization by antibodies is predominantly dependent on targeting the receptor binding domain (RBD) or N-terminal domain (NTD) of the S protein. 16 Mutations in these regions can lead to escape, Binding antibodies cross-react with the Delta and Omicron BA.1 variant Binding antibodies to ancestral SARS-CoV-2, and the Delta or Omicron BA.1 variant S proteins were assessed by ELISA ( Figure 1A). A significant increase in binding antibody levels was observed 28 days after both homologous and heterologous booster vaccination ( Figures 1B, S2, and S3A). We found the lowest binding antibody titer in the no-boost group (GMT of 1192). The binding antibody titers were higher after homologous (Ad26.COV2.S; GMT of 3774) and particularly after heterologous booster with mRNA-1273 (GMT of 117660) or BNT162b2 (GMT of 58747) ( Figure 1B). These patterns were compared with previously reported S1-specific binding antibodies as measured by commercial assay. 10 We found that binding antibodies were in general cross-reactive with both the Delta and Omicron BA.1 variant S proteins, although significantly lower antibody titers were found against Omicron BA.1 S across all groups and timepoints ( Figures 1C and S3B). No significant differences were observed between the ancestral S protein and Delta variant. To further analyze these responses at the cellular level, we determined the percentage of total RBD-specific B cells in peripheral blood mononuclear cells (PBMC) by flow cytometry. Ancestral RBD-specific B cells were detected in the pre-booster samples of all participants and no differences were observed at baseline between the groups. Interestingly, booster vaccination did not increase the frequency, nor did it change the phenotype, of RBD-specific B cells. Similar frequencies of RBD-specific B cells, RBD-specific memory B cells as well as RBD-specific IgG memory B cells were observed pre-and post-booster with all vaccination regimens ( Figure S4).
Antibodies with Fc-mediated functions cross-react with the Delta and Omicron BA.1 variant Two different Fc-mediated antibody effector functions were assessed: ADCC and ADCP. ADCC-mediating antibodies were measured in a functional NK cell degranulation assay performed on S protein-coated plates ( Figure 2A). Similar to the binding antibodies, higher levels of ADCC-mediating antibodies were observed after Ad26.COV2.S booster vaccination (median of 16% degranulating cells) compared to no boost (median of 9.5%). The highest levels of ADCC-mediating antibodies were observed after mRNA-1273 (median 20%) or BNT162b2 (median of 20%) booster vaccination ( Figures 2B and S5A). Although ADCC-mediating antibodies cross-reactive with the Delta variant S protein were detected in all groups at all timepoints, these were significantly lower compared to antibodies against the ancestral S protein ( Figures 2C and S5B). In contrast to what was observed with binding antibodies, ADCC-mediating antibodies cross-reactive with the Omicron BA.1 S protein were only detected after mRNA-1273 (median of 11%) or BNT162b2 (median of 11%) booster vaccination ( Figures 2C and S5B). Additionally, we measured ADCP-mediating antibodies in a functional THP-1 phagocytosis assay with ancestral S protein-coated beads ( Figure 2D). Similarly to ADCC-mediating antibodies, Fc-mediated phagocytosis was boosted by both homologous or heterologous vaccination and highest after mRNA-1273 (GMT of 41438) or BNT162b2 (GMT of 45788) booster vaccination as compared to Ad26.-COV2.S (GMT of 3373) vaccination ( Figure 2E). Flow cytometric analyses and individual dilution series per vaccination regimen are shown in Figures S6A and S6B, respectively.

Cross-neutralization of omicron BA.1 is increased after heterologous booster
Neutralizing antibodies were assessed in an infectious virus neutralization assay with the ancestral SARS-CoV-2, and the Delta, and Omicron BA.1 variants ( Figure 3A). mRNA-based booster vaccination after  Figures 3C and S7B). Individual S-curves per vaccination regimen are shown in Figure S8.

Correlations between serological assays
We examined the correlations between S-specific binding antibodies and S1-binding antibodies (Figure 4A), and their functionalities including neutralization (PRNT50) ( Figure 4B), NK cell degranulation (ADCC) ( Figure 4C), and phagocytosis (ADCP) ( Figure 4D) against the ancestral SARS-CoV-2 and found all correlations to be positive and significant (p < 0.05). We additionally performed correlations for the ancestral-, Delta-and BA.1-specific responses per assay ( Figure S9). We observed a direct relationship between ancestral-and variant-specific antibody levels and found that reduced (or absent) variant-specific antibody responses were directly related to low total antibody levels. Next, we measured T-cell responses before and after homologous or heterologous booster vaccination. To directly assess T-cell responses in whole blood, we previously performed an interferon gamma (IFNg) release assay (IGRA), and found that T-cell responses were boosted by both homologous and heterologous booster vaccination. 10 To assess T-cell responses in depth, PBMCs were stimulated with overlapping peptide pools spanning the full-length ancestral S protein, and responses were measured via IFN-g ELISPOT ( Figure 5A). Here, we found that mRNA-1273 booster vaccination induced significantly higher numbers of IFN-g producing T cells to ancestral SARS-CoV-2 compared to homologous booster vaccination ( Figure 5B).
To measure variant-specific responses, PBMCs were stimulated with overlapping peptide pools representing the full-length S protein from the ancestral SARS-CoV-2, and the Delta and Omicron BA.1 variants ( Figure 5A). Following stimulation, CD4 (OX40 + CD137 + ) and CD8 (CD69 + CD137 + ) T cell activation-induced marker (AIM) expression was measured by flow cytometry ( Figure S10A). CD4 and CD8 T-cell responses were detected in 32/60 (53%) of participants pre-booster, and levels were comparable between groups. Booster vaccination with either Ad26.COV2.S or mRNA-1273 did not significantly increase CD4 T-cell responses. Interestingly, booster vaccination with BNT162b2 increased the number of participants with a measurable CD4 T-cell  Figure S10C). Similar to CD4 T cells, CD8 T cells equally reacted with all SARS-CoV-2 variants tested ( Figure S10C).

mRNA-based booster vaccination led to the expansion of S-specific T-cell clones
We further evaluated the expansion, breadth, and depth of the SARS-CoV-2-specific T-cell response after different booster regimens. TCRb sequencing was performed to define the repertoires of N = 30 participants (N = 7 no boost, N = 7 Ad26.COV2.S boost, N = 10 mRNA-1273 boost, and N = 6 BNT162b2 boost) pre-and post-booster vaccination. 42 Initially, we compared clones pre-and post-booster vaccination within donors to identify expanding clones after booster vaccination (representative example shown in Figure 6A). Expanding clones were detected in two donors that did not receive a boost, but in a time period of 28 day background expansion of G5-10 clones can be expected ( Figure 6B). More expanding clones were observed in the Ad26.COV2.S-boosted individuals as compared to no boost (dominated by 73 expanding clones in 1 individual), but especially in the mRNA-1273 and BNT162b2-boosted individuals the number of expanding clones was often >20 ( Figure 6B).
To identify SARS-CoV-2-specific T-cell clones, the TCR sequences were compared to a sequence dataset (the ImmunoCODE MIRA dataset) enriched in COVID-19 cases versus controls. 43 This method identifies clones that are specific to SARS-CoV-2 and reduces noise associated with clones that are very frequent or potentially cross-reactive. Breadth (number of unique SARS-CoV-2-specific TCRs) and depth (frequency of SARS-CoV-2-specific TCRs) were calculated for S-and ORF1ab-, ORF3a-, MÀ and N-specific T cells. As expected, a dominant S-specific T-cell response was detected, as SARS-CoV-2-infected donors were excluded from this study (Figures S11A and S11B). Interestingly, booster vaccinations did not lead to a significant increase in the breadth of the S-specific T-cell response ( Figure 6C). However, booster vaccination with mRNA-1273 led to a significant increase in the depth/frequency ( Figure 6D) of the SARS-CoV-2-specific T-cell response, which was not observed after Ad26.COV2.S booster vaccination.

DISCUSSION
We performed immunological profiling of the SARS-CoV-2-specific immune response, including reactivity to the Delta and Omicron BA.1 variants, after homologous or heterologous booster vaccination of Ad26.COV2.S-primed individuals. We found that Ad26.COV2.S priming provided a solid immunological base for strong and broad SARS-CoV-2-specific immune responses upon subsequent mRNA-based booster vaccination. A limitation of this study was that the age was significantly different between the groups that received an mRNA-based booster vaccination, Ad26.COV2.S boost, or no boost. This potentially contributed to the fact that we concluded that mRNA-based vaccines are superior boosters, as older age was reported as a negative factor contributing to the induction of lower antibody titers following SARS-CoV-2 infection. 44 However, in the original SWITCH trial, with a larger number of participants, we also found the superior boosting capacity of mRNA-based vaccines, despite this difference. 10 Additionally, other studies reported similar superiority of mRNA-based over vector-based COVID-19 vaccines. 45,46 Samples were collected between August and September of 2021 when Omicron sub-lineages were not circulating in the Netherlands. To exclude recent infections, a nucleocapsid (N) ELISA was performed on all samples before participants received their booster vaccination. 10 Here, we compared four different booster regimens in a random selection of individuals from the larger SWITCH study. 10 Binding antibodies targeting the ancestral SARS-CoV-2, and the Delta and Omicron BA.1 variants, increased after booster vaccination and levels were highest in participants that received an mRNA-based booster. Strikingly, we found that the proportion of RBD-specific memory B cells in blood did not increase after booster vaccination. This indicates that the original Ad26.COV2.S priming induced a sustained RBD-specific memory B cell response and we speculate that final maturation had already occurred in the 3 months after the initial vaccination. This is in line with the slow increase of S1-specific antibodies after priming vaccination with Ad26.COV2.S and the stable levels of these antibodies. 8 Therefore, booster vaccination led to the rapid induction of antibody production by memory B cells rather than Antibodies can have a multitude of effector functions, ranging from direct neutralization to Fc-mediated triggering of cytotoxicity or phagocytosis targeting infected cells and/or cell-free virions, depending on the antibody isotype, glycosylation pattern, and Fc receptor bound. 47 The majority of the participants in this study developed neutralizing antibodies against the ancestral SARS-CoV-2 (independent of the vaccination regimen). Neutralizing antibodies targeting the Delta variant were readily detected at slightly lower levels, but neutralizing antibodies targeting Omicron BA.1 could only be detected after mRNA-based booster vaccination, at considerably lower levels compared to the ancestral SARS-CoV-2. 8 Importantly, mRNA-based booster vaccination resulted in significantly higher neutralizing antibody titers as compared to Ad26.COV2.S booster vaccination.
We assessed Fc-mediated effector functions of antibodies. It was previously hypothesized that these functions might play a role in contributing to protection against COVID-19, 48,49 but relatively little is known iScience Article about the impact of Fc-mediated antibody effector functions. 34 Novel antigenically distinct SARS-CoV-2 variants, like the Delta variant and Omicron sub-lineages, are partly capable of evading neutralizing antibodies by accumulating mutations in the RBD. 1,50-52 Functional non-neutralizing antibodies are speculated to be less susceptible to immune escape by emerging variants, as they are not dependent on the recognition of specific epitopes in the RBD and they can target the entire S protein. 25,34 Here, we show an increase in ADCC-and ADCP-mediating antibodies against the ancestral SARS-CoV-2, Delta, and Omicron BA.1 variants following both homologous and heterologous booster vaccination. Similar to the neutralizing antibody responses, Fc-mediated antibodies were higher following mRNA-based booster vaccination. Although effector functions mediated by non-neutralizing antibodies were also reduced towards the Delta and Omicron BA.1 variant, ADCC-mediating antibodies were still clearly detected after mRNA-based booster vaccination. We speculate that escape from antibodies with the potential to target the entire S protein is caused by the numerous mutations in Omicron BA.1, even outside the RBD (and NTD). For ADCP we were not able to measure variant-specific responses due to a lack of the required reagents. However, based on the observed correlation between binding, ADCC-mediating, and ADCP-mediating antibodies, we expect similar patterns of cross-reactivity.

SARS-CoV-2-specific T cells play an important role in reducing COVID-19 severity following re-or breakthrough infection. 53 T cells can clear virus-infected cells, contributing to the reduction of virus replication. 11
Virus-specific T cells are thought to be long-lived, as these have been detected up to six months after the completion of primary vaccination regimens, 8 and up to 17 years after SARS-CoV infection. 36 T cells can target epitopes dispersed throughout proteins, including conserved epitopes under functional constraints, and therefore retain cross-reactivity to SARS-CoV-2 variants, 9,30,40,41 including the Omicron sublineage. 7,8,54 Here, we show that T-cell responses are boosted in Ad26.COV2.S-primed individuals especially after mRNA-based booster vaccination, as measured by both IFN-g levels and expansion of S-specific T-cell clones. However, these T-cell responses were not significantly higher than after Ad26.-COV2.S homologous booster vaccination. Alternatively, the contraction phase after mRNA-based boost is reported to be rapid, 55 leaving a small window of opportunity to detect virus-specific T-cell increases after booster vaccination. Although based on TCRb sequencing the breadth of the s-specific response did not increase after heterologous booster vaccination, reactivity of both CD4 and CD8 T cells with the Delta and Omicron BA.1 variants was retained. No significant increase in CD8 T-cell responses was detected following any of the booster vaccinations, following the same pattern as CD4 T-cell responses. However, high variability in the AIM results and less sensitivity in detecting SARS-CoV-2-specific CD8 + T-cells makes it more complicated to interpret these data.
Currently, several sub-lineages of the Omicron variant are circulating. Although the BA.1 lineage quickly became dominant upon introduction, it was rapidly replaced by the BA.2 lineage. Both variants have shown significant escape from neutralizing antibodies. 17,20,56,57 Currently, other Omicron variants are rapidly establishing dominance in different geographical locations, [21][22][23] and escape has been demonstrated for the newer Omicron variants. 24 In our study, we have focused on cross-reactive immune responses to Omicron BA.1, since at the time of the experiments the newer variants were not yet circulating. Based on cross-reactivity with BA.1 and available literature, we expect that non-neutralizing antibodies and T-cell responses have at least equal potential for cross-reactivity with these novel immune-evasive variants, based on the targeting of conserved epitopes.
In conclusion, we showed that Ad26.COV2.S priming provided a solid immunological base for SARS-CoV-2-specific immune responses triggered by mRNA-based booster vaccination. Additionally, we show that heterologous mRNA-based boosters are more potent compared to homologous Ad26.COV2.S boosting. Neutralizing antibodies targeting immune-evasive variants were detectable after a mRNA-based booster, and non-neutralizing antibodies and T-cell responses to these variants were retained or even boosted. These findings are similar to previous findings in individuals primed with another vector-based vaccine (ChAdOx1-S), who received an mRNA-based booster vaccination. 46 It is crucial to further investigate how these responses to booster vaccination compare to individuals primed with ChAdOx1-S or an mRNA-based vaccine, and whether the initial priming vaccination still has an effect on ongoing booster campaigns with bivalent vaccines. Although there currently is a high prevalence of breakthrough infections with viruses from the newly emerging Omicron sub-lineages, the related disease has been reported to be relatively mild. 58  iScience Article reducing COVID-19 disease severity and boosting these could be crucial for vaccine effectiveness in the future. 53,59 Limitations of the study Because of the complexity of the used techniques, and the amount of different immunological parameters studied, it was not feasible to test large cohorts in a high-throughput setting. Therefore, we present a relatively small dataset. Although most immunological parameters show a clear pattern, the TCR-sequencing data are limited by the small dataset and should be interpreted carefully. We expanded the analysis of functional antibodies beyond neutralization and included experiments to measure Fc-mediated effector functions. We could not analyze all Fc-mediated effector functions against the variants of interest, as not all reagents were available at the time of the study. Additionally, we focused our analyses around cross-reactivity of immune responses with Omicron BA.

Materials availability
SARS-CoV-2 peptide pools used in this study are from Alessandro Sette (alex@lji.org), and are available upon reasonable request with a completed materials transfer agreement. Other unique/stable reagents generated in this study are available from the lead contact with a completed materials transfer agreement.
Data and code availability d All data reported in this paper will be shared by the lead contact upon request. This work is licensed under a Creative Commons Attribution 4.0 International (CC BY 4.0) license, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/. This license does not apply to figures/photos/artwork or other content included in the article that is credited to a third party; obtain authorization from the rights holder before using such material.
d This paper does not report original code.
d Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Study design
The SWITCH trial is a single-(participant)-blinded, multi-center, randomized controlled trial among HCWs without severe comorbidities performed in four academic hospitals in the Netherlands (Amsterdam University Medical Center, Erasmus University Medical Center, Leiden University Medical Center, and University Medical Center Groningen), according to the published protocol. 60 The trial adheres to the principles of the Declaration of Helsinki and was approved by the Medical Research Ethics Committee from Erasmus Medical Center (MEC 2021-0132) and the local review boards of participating centers. All participants provided written informed consent before enrollment.

Participants
For analysis of humoral and cellular immune responses, 60 donors were randomly selected, taking into account whether sufficient material was available. Participants randomly selected for immunological profiling received a priming vaccination with Ad26.COV2.S, followed by a booster vaccination with Ad26