Mucosal antibody responses following Vaxzevria vaccination

Abstract Mucosal antibodies play a key role in protection against breakthrough COVID‐19 infections and emerging viral variants. Intramuscular adenovirus‐based vaccination (Vaxzevria) only weakly induces nasal IgG and IgA responses, unless vaccinees have been previously infected. However, little is known about how Vaxzevria vaccination impacts the ability of mucosal antibodies to induce Fc responses, particularly against SARS‐CoV‐2 variants of concern (VoCs). Here, we profiled paired mucosal (saliva, tears) and plasma antibodies from COVID‐19 vaccinated only vaccinees (uninfected, vaccinated) and COVID‐19 recovered vaccinees (COVID‐19 recovered, vaccinated) who both received Vaxzevria vaccines. SARS‐CoV‐2 ancestral‐specific IgG antibodies capable of engaging FcγR3a were significantly higher in the mucosal samples of COVID‐19 recovered Vaxzevria vaccinees in comparison with vaccinated only vaccinees. However, when IgG and FcγR3a engaging antibodies were tested against a panel of SARS‐CoV‐2 VoCs, the responses were ancestral‐centric with weaker recognition of Omicron strains observed. In contrast, salivary IgA, but not plasma IgA, from Vaxzevria vaccinees displayed broad cross‐reactivity across all SARS‐CoV‐2 VoCs tested. Our data highlight that while intramuscular Vaxzevria vaccination can enhance mucosal antibodies responses in COVID‐19 recovered vaccinees, restrictions by ancestral‐centric bias may have implications for COVID‐19 protection. However, highly cross‐reactive mucosal IgA could be key in addressing these gaps in mucosal immunity and may be an important focus of future SARS‐CoV‐2 vaccine development.


INTRODUCTION
Mucosal immunity, particularly SARS-CoV-2 spike-specific mucosal IgA, has been suggested to be protective against breakthrough COVID-19 infections. 1,2Unfortunately, intramuscularly (IM) administered COVID-19 vaccines do not induce optimal mucosal antibody responses in uninfected COVID-19 vaccinees. 3Vaxzevria is an IM chimpanzee adenoviral vector COVID-19 vaccine used globally and is effective in reducing severe disease and death after two primary doses. 4,5ecently, it has been shown that Vaxzevria induces better mucosal IgG and IgA responses in the nasal fluid of COVID-19 recovered vaccinees than in vaccinated only vaccinees. 6In addition, while SARS-CoV-2 variants of concern (VoC) may escape neutralization, non-neutralizing antibodies remain robust in inducing Fc effector functionswhich have been suggested to contribute to protective immunity. 7,8Nonetheless, little is known about how prior COVID-19 infection impacts Fc-receptor engagement by mucosal antibodies, particularly against the SARS-CoV-2 VoCs in individuals receiving Vaxzevria vaccines.Here, we compare antibody isotypes, subclasses, as well as Fc-receptor engagement and cross-reactivity across VoCs in mucosal secretions (saliva, tears) with plasma responses from Vaxzevria vaccinees.
In contrast, saliva and plasma responses against IgA1, total IgG as well as FccR3a after COVID-19 recovered individuals received their second Vaxzevria dose (1 9 prior COVID-19 infection + 2 9 Vaxzevria) did not rise to the levels seen 2 weeks after the first dose (Figure 1c, d; Supplementary figure 2b, d).However, it should be noted that despite the decrease, total IgG and FccR3a responses in saliva remained significantly elevated (P < 0.05) over pre-vaccination responses particularly against the more novel Spike 1 protein (Figure 1e).Salivary IgA1 responses in COVID-19 recovered individuals were comparable to pre-vaccination levels after their second Vaxzevria dose.A similar trend was observed with IgA1, total IgG and FccR3a responses in tear fluid from COVID-19 recovered individuals after their Vaxzevria vaccination, with significant increases in antibody responses detected after the first Vaxzevria dose (P < 0.05; Figure 1f).
Taken together, our findings support previous observations that salivary IgA responses dip after the second IM vaccine dose in COVID-19 recovered subjects,  (c, d), the median of each cohort/ timepoint's antigen-specific MFI was divided by the antigen-specific MFI in the 99th percentile for that detector (99th percentile was chosen to minimize the impact of outliers on the data transformation).IgA, total IgG and FccR3a saliva (e) and tear (f) antibody features from COVID-19 recovered individuals after their second Vaxzevria vaccination are also illustrated in respective bar graphs.Statistical significance was calculated using the Friedman test followed by Dunn's test for multiple comparisons and where significant or trending significance, P-values are reported (*P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001).and suggest that repeated mucosal antigen exposures (either mucosal infection or potentially mucosal vaccination) may be required to maintain robust mucosal IgA and antibody-mediated Fc-responses. 9cosal IgG and Fc-antibodies target more ancestralcentric SARS-CoV-2 antigens Mucosal neutralizing antibodies are key in preventing active SARS-CoV-2 infection and is the ongoing goal of nasal vaccine design. 10,11Here, we explored whether IM Vaxzevria vaccines neutralizing activity at the mucosa, particularly among COVID-19 recovered individuals, using a surrogate neutralization ACE2-inhibition assay. 12We found only a few individuals with salivary antibody responses that could inhibit RBD-ACE2 binding.Most individuals from both cohorts and after either one or two vaccines did not induce responses in saliva above the assay detection threshold (dotted line) (Figure 2a-c).
Given the poor neutralizing activity in saliva following IM vaccination, mucosal Fc-responses could play a larger role in protecting against active infections.As COVID-19 recovered individuals induced better antibody-mediated Fc-engagement against the ancestral spike in their mucosal secretions, we next studied whether their mucosal antibodies could also induce effective Fc-receptor  engagement to the range of SARS-CoV-2 VoCs spikes.While we detected salivary antibody-mediated Fc-receptor engagement in COVID-19 recovered vaccinees (1 9 prior COVID-19 infection + 2 9 Vaxzevria) across a broad range of SARS-CoV-2 VoCs, these responses were ancestralcentric with the largest responses against variants (Alpha, 7.9-fold; Delta 14.2-fold) more similar to the ancestral wildtype (Figure 2g).Whilst more modest, Fc-engagement responses in tear fluid against the SARS-CoV-2 VoCs also showed a similar ancestral-centric trend (Figure 2h).

Mucosal IgA broadly recognizes SARS-CoV variants
4][15] To investigate the impact of immunological imprinting on vaccine-induced responses, we compared the antibody signatures against the VoCs against those made against the ancestral wildtype (Figure 2i, j; Supplementary figure 3a, b).Ancestral-centric bias was largely detected for salivary and plasma total IgG and Fc-engagement responses (P < 0.01), with weaker responses against the more novel Omicron variants BA.1 and BA.2 compared with variants more similar to the ancestral wildtype, such as Alpha and Delta (Figure 2i, j).In contrast, salivary IgA, but not plasma IgA, showed cross-reactivity across all the variants tested (Alpha, Delta, Beta, Omicron BA.1, Omicron BA.2) (Figure 2i-l; Supplementary figure 4a, b), particularly after the first Vaxzevria dose in COVID-19 recovered individuals.As such, the induction and retention of localized crossreactive IgA at the mucosa could be vital in the protection against emerging SARS-CoV-2 variants.Larger cohort studies should be done to further explore these findings.

DISCUSSION
Mucosal immunity is a key barrier in preventing active infections by respiratory pathogens, such as SARS-CoV-2.Higher levels of salivary antibodies, particularly secretory IgA targeting SARS-CoV-2 RBD, has been associated with protection against breakthrough infections. 2 Our group and others have shown previously that IM COVID-19 mRNA vaccinations are insufficient for inducing robust mucosal humoral responses in vaccinated only individuals. 3,16ecently, it has been shown that COVID-19 recovered individuals receiving IM Vaxzevria vaccination produced more IgG and IgA in their nasal secretions, compared with vaccinated only individuals. 6re, we investigated whether COVID-19 recovered individuals also generate improved IgG and IgA responses in other important mucosal secretions, namely saliva and tear fluid.Furthermore, we explored whether these mucosal antibodies induced after Vaxzevria vaccination could also effectively engage Fc-receptors.Compared with vaccinated only individuals, COVID-19 recovered individuals did induce better mucosal (saliva, tears) IgG, IgA1 and FccR3a engagement responses after the first Vaxzevria dose.However, both systemic and mucosal antibody responses in COVID-19 recovered individuals dipped after the second Vaxzevria vaccination.
Local antigen stimulation at the mucosa has been shown to be key to producing mucosal IgA. 11,17As such, repeated IM vaccinations may be limited in the robust induction of mucosal IgA.Furthermore, antibody feedback from pre-existing high-affinity IgG antibodies has also been suggested to limit the humoral responses towards vaccination through epitope masking. 18Indeed, an extended dosing interval of 45 weeks between the first and second Vaxzevria dose has been shown to be four times more effective than a 12-week interval. 19In our study, the dosing interval between the first and second Vaxzevria doses for COVID-19 recovered individuals was only 70-86 days (10-12 weeks).This limited dosing interval may have influenced the humoral responses made after the second Vaxzevria dose.However, we also acknowledge that the week delay in our sampling process after the second Vaxzevria vaccination of COVID-19 recovered individuals may have also contributed to the observed differences in antibodies responses.
Unsurprisingly, while we did detect neutralizing activity against SARS-CoV-2 in plasma following IM COVID-19 vaccination, we did not detect robust salivary neutralizing activity even among COVID-19 recovered vaccinees.This finding highlights the urgent need for an effective COVID-19 mucosal vaccine as a tool to reduce breakthrough infections. 17In addition, while mucosal IgG antibodies could detect a broad range of SARS-CoV-2 VoCs and induce Fc-receptor engagement, responses against novel variants such as Omicron were limited by ancestral-centric bias.Immunological imprinting can not only influence mucosal humoral responses made after vaccination, but also responses following breakthrough infections. 15This emphasizes the need for future COVID-19 vaccines to move away from the ancestral SARS-CoV-2 wildtype template and instead be updated with emerging variants. 20inally, we observed that salivary IgA, but not plasma IgA, displayed cross-reactivity across SARS-CoV-2 variants, particularly after the first Vaxzevria dose in COVID-19 recovered individuals.Secretory IgA has been shown to be broadly cross-reactive for different strains of influenza A and B viruses respectively. 21,22Indeed, increased levels of ancestral SARS-CoV-2 wildtype spikespecific mucosal IgA has been associated with a decreased risk of acquiring Omicron breakthrough infection. 1As such, the induction of cross-reactive mucosal IgA could be key in preventing breakthrough infections by emerging SARS-CoV-2 variants.Given the modest size of our study, future studies involving larger cohorts of Vaxzevria vaccinees should further expand upon these findings.
Future work should focus on studying the effectiveness of updated mono-or bi-valent COVID-19 vaccines in reducing ancestral-centric bias.The impact of repeated mucosal exposures, through either receiving mucosal vaccines or acquiring infections, on the levels of mucosal IgA and Fc-responses could be further explored.Likewise, IgA and Fc-responses in secretions from other mucosal sites where SARS-CoV-2 has been detected, such as the intestinal and nasal tracts, could also be investigated.

CONCLUSION
IM Vaxzevria vaccination is not only effective in generating systemic humoral responses, but can also induce modest mucosal humoral responses, particularly among COVID-19 recovered individuals.However, these IgG-driven mucosal responses are largely non-neutralizing and ancestral-centric and may therefore have limited capacity to prevent breakthrough infection.Highly crossreactive mucosal IgA could be key in addressing these gaps in mucosal immunity and should be the emphasis of SARS-CoV-2 mucosal vaccines development.

Cohort and sample collection
We enrolled individuals with (COVID-19 positive between March to September 2020, thus likely infected with the ancestral virus or D614G strain) and without prior SARS-CoV-2 infection from a previously described cohort 23 to donate blood, saliva, and tear samples prior to and following vaccinations with Vaxzevria (Oxford-AstraZeneca) vaccines, as well as mRNA boosters (Table 1; Supplementary table 1).Whole blood was collected with sodium heparin anticoagulant (Becton Dickinson, Franklin Lakes, NJ USA) and plasma was collected and stored at À80°C until use.Saliva was collected by SalivaBio Oral Swabs (Salimetrics, Carlsbad, CA, USA) and processed following the manufacturer's instructions, before being stored at À80°C until use.Basal (non-stimulated) tear samples ( $ 7 lL per eye) were collected by capillary flow (Drummond Scientific, Broomall, PA, USA) from the inferior tear meniscus as previously reported, and also stored at À80°C until use. 24Plasma and saliva from COVID-19 unvaccinated, uninfected healthy controls were also collected on 16 March 2020, while tear controls were from pre-pandemic samples.All participants provided written informed consent, and the study was approved by the University of Melbourne Human Research Ethics Committee (2021-21198-15398-3, 2056689, 11507).
Bead-based multiplex against SARS-CoV-2 ancestral antigens SARS-CoV-2 specific antibody isotypes (IgG, IgA, IgM) and subclasses (IgG1-4, IgA1-2) in plasma (1:1600), saliva (1:12.5)and tear (1:25) from pre-pandemic and vaccinated cohorts were assessed using a customized multiplex bead-based array consisting of four ancestral SARs-CoV-2 proteins, including Spike 1 (S1; Sino Biological, Beijing, China), Spike 2 (S2; Acro Biosystems, Newark, NJ, USA), Receptor Binding Domain (RBD) and whole Spike Trimer (ST), as described previously. 3SIVgp120 protein (Sino Biological) and uncoupled BSAblocked beads were included as negative controls for background subtraction.Plasma and saliva concentrations used in the array were chosen based on a dilution series.Briefly, antigen-coupled beads were incubated with the respective samples on a shaker overnight at 4°C, before being washed, and incubated with Phycoerythrin (PE)-conjugated detection antibodies (Southern Biotech, Birmingham, AL, USA) on a shaker for 2 h at room temperature (RT).The beads were then washed and read on the Flexmap 3D.Assays were repeated in duplicate (Supplementary figure 1a, b).
Engagement of SARS-CoV-2 specific antibodies to Fc gamma receptors (FccR) were measured using surrogate Fc gamma receptor dimers (FccR2a, CD32; FccR3a, CD16) as described previously (kind gift from Mark Hogarth and Bruce Wines). 25After incubation with samples, the washed beads were first incubated with surrogate FccR dimers on a shaker for 2 h at RT, washed again, and then incubated with Streptavidin-R-Phycoerythrin (SAPE; Thermo Fisher Scientific, Waltham, MA, USA) on a shaker for a further 2 h at RT. Finally, the beads were washed and read on the Flexmap 3D.Assays were repeated in duplicate.

Bead-based multiplex against SARS-CoV-2 variant antigens
To compare plasma (1:25600), saliva (1:12.5)and tear (1:25) SARS-CoV-2 ancestral WT and variant antibody responses (Alpha, Beta, Delta, Omicron BA.1, Omicron BA.2), WT and VoC Spikes were used to form a customized bead array (Supplementary figure 3a, b).To measure SARS-CoV-2 specific total IgG and IgA responses, beads were first incubated with samples, washed and then incubated with the biotin-conjugated detection antibodies (MabTech, Nacka Strand, Sweden) on a shaker for 2 h at RT. Subsequently, the beads were washed and then incubated with SAPE for another 2 h at RT. Finally, the beads were washed again and then read on the Flexmap 3D.The ability of SARS-CoV-2 variant specific plasma (1:12800) and saliva (1:12.5)antibodies to mediate FccR engagements (FccR2a, CD32; FccR3a, CD16) were measured using the surrogate Fc-receptor dimers as described above.Assays were repeated in duplicate.

Surrogate RBD-ACE-2 inhibition assay
Neutralizing activity of plasma (1:800, 1:4000) and saliva (1:12.5)against the SARS-CoV-2 variants of concern (Alpha, Beta, Delta, Omicron BA.1, Omicron BA.2) (Sino biological) were accessed using a surrogate RBD-ACE-2 inhibition assay.As described previously, 3,12 ancestral or variant RBD-coupled beads were incubated with avi-tagged biotinylated ACE2 in the presence of the respective plasma and saliva samples on a shaker for 2 h at RT.The beads were washed and then incubated with SAPE on a shaker for 1 h at RT. R-Phycoerythrin Biotin-XX conjugate (Thermo Fisher Scientific) was then added to the beads and incubated on a shaker for a further hour at RT. Finally, the beads were washed and read on the Flexmap 3D.Assays were repeated in duplicate.Calculation of % ACE2 inhibition was done using the following formula: ([1 -(Average MFI of sample wells/Average MFI of ACE2 only wells)] 9 100).Graphs were plotted between À20% to 100% ACE2 inhibition.A nominal cutoff of 20% (depicted by dotted line) was also set as described previously 12 (Supplementary figure 1c, d).
To validate that the RBD-ACE2 inhibition assay could detect ACE2 inhibition in the presence of saliva, we spiked dilutions of either plasma from an mRNA boosted individual (1:200-1:25 600 final plasma dilutions) or a RBD neutralizing antibody SAD-S35 (40-0.002nM) in saliva collected from three COVID-19 unvaccinated, uninfected healthy individuals (Supplementary figure 1e, f).

Data transformation and statistical analysis
Statistical analysis was performed with GraphPad Prism 9 (GraphPad Software).Prior to analysis of the multiplex data (ancestral and variant antigens), the readings from negative control antigens (SIV gp120, BSA) were background subtracted for each individual antigen/detector.The multiplex data were then right shifted to remove negative values.To transform the multiplex data (ancestral antigens) into percentages for use in the radar plots (Figure 1c, d), the median of each cohort/timepoint's antigen-specific MFI was divided by the antigen-specific MFI in the 99th percentile for that detector (99th percentile was chosen to minimize the impact of outliers on the data transformation).Antibody levels between cohort/timepoints were compared using Mann-Whitney U-tests or Friedman tests, with corrections for multiple comparisons as required.

Figure 1 .
Figure 1.COVID-19 recovered Vaxzevria vaccinees displayed enhanced IgA and FccR responses in saliva and tear fluid.Paired saliva and plasma samples were collected pre-and post-Vaxzevria vaccination from vaccinated only (a) and COVID-19 recovered (b) individuals at the indicated time-points.Saliva (c) and plasma (d) antibody isotype and subclass responses from both cohorts against the various SARS-CoV-2 spike antigens were compiled into respective radar plots.To transform the data into percentages for use in the radar plots(c, d), the median of each cohort/ timepoint's antigen-specific MFI was divided by the antigen-specific MFI in the 99th percentile for that detector (99th percentile was chosen to minimize the impact of outliers on the data transformation).IgA, total IgG and FccR3a saliva (e) and tear (f) antibody features from COVID-19 recovered individuals after their second Vaxzevria vaccination are also illustrated in respective bar graphs.Statistical significance was calculated using the Friedman test followed by Dunn's test for multiple comparisons and where significant or trending significance, P-values are reported (*P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001).

Figure 2 .
Figure2.Cross-reactive salivary IgA avoids vaccination-induced ancestral-centric bias.Bar graphs describe salivary (a-c) and plasma (d-f) inhibition of RBD-ACE2 interactions against the ancestral wildtype (WT) SARS-CoV-2 or the VoCs (a, Alpha; d, Delta; b, Beta; ο BA.1, Omicron BA.1; ο BA.2, Omicron BA.2) by vaccinated only (a, d) and COVID-19 recovered individuals (b, c, e, f), respectively.The number of individuals with detectable responses above the arbitrary 20% assay threshold (dotted line) at either timepoint are listed under the bar graphs in their respective colors.Significant differences between both timepoints were calculated using the two-tailed Mann-Whitney U-test, followed by Bonferroni-Dunn's test for multiple comparisons.Bar graphs also show the salivary (g) and tear (h) FccR3a responses against WT SARS-CoV-2 or the VoCs in COVID-19 recovered individuals after their second Vaxzevria vaccination.Fold changes listed above the bar graphs were calculated for post dose 2 Vaxzevria responses (purple) over their respective pre-vaccination responses (gray) for each antigen.The number of individuals with detectable responses above the assay threshold (median responses from COVID-19 unvaccinated, uninfected healthy controls; dotted line) at either timepoint were listed under the bar graphs in their respective colors.Significant differences between both timepoints were calculated using the two-tailed Mann-Whitney U-test, followed by Bonferroni-Dunn's test for multiple comparisons.Heat maps illustrate the VoC-specific Spike Trimer salivary (i) and plasma (j) antibody responses post-mRNA booster and post-Vaxzevria vaccination (dose 1 or 2) for both vaccinated only and COVID-19 recovered cohorts respectively.The median antibody response for each VoC spike was described as a fold change to the wildtype spike.Statistical significance was calculated using Friedman's test followed by Dunn's test for multiple comparisons.Bar graphs show the salivary IgA responses against WT SARS-CoV-2 or the VoCs in COVID-19 recovered individuals after their first (k) and second (l) Vaxzevria vaccines.Fold changes listed above are calculated against that for the WT SARS-CoV-2.WT SARS-CoV-2 salivary IgA responses from COVID-19 unvaccinated and uninfected healthy controls are shown by the dotted line.Where significant, P-values are reported (*P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001)."