key: cord-0776648-qvtbjkuz
authors: Stolovich-Rain, Miri; Kumari, Sujata; Friedman, Ahuva; Kirillov, Saveliy; Socol, Yakov; Billan, Maria; Pal, Ritesh Ranjan; Golding, Peretz; Oiknine-Djian, Esther; Sirhan, Salim; Sagie, Michal Bejerano; Cohen-Kfir, Einav; Elgrably-Weiss, Maya; Zhou, Bing; Ravins, Miriam; Gatt, Yair E; Das, Kathakali; Zelig, Orly; Wiener, Reuven; Wolf, Dana G; Elinav, Hila; Strahilevitz, Jacob; Padawer, Dan; Baraz, Leah; Rouvinski, Alexander
title: Intramuscular mRNA BNT162b2 vaccine against SARS-CoV-2 induces robust neutralizing salivary IgA
date: 2022-02-17
journal: bioRxiv
DOI: 10.1101/2022.02.17.480851
sha: 47586853d148250fb62266f3f25d420a47082fad
doc_id: 776648
cord_uid: qvtbjkuz

Intramuscularly administered vaccines stimulate robust serum neutralizing antibodies, yet they are often less competent in eliciting sustainable ‘sterilizing immunity’ at the mucosal level. Our study uncovers, strong neutralizing mucosal component (NT50 ≤ 50pM), emanating from intramuscular administration of an mRNA vaccine. We show that saliva of BNT162b2 vaccinees contains temporary IgA targeting the Receptor-Binding-Domain (RBD) of SARS-CoV-2 spike protein and demonstrate that these IgAs are key mediators of potent neutralization. RBD-targeting IgAs were found to associate with the Secretory Component, indicating their bona-fide transcytotic origin and their dimeric tetravalent nature. The mechanistic understanding of the exceptionally high neutralizing activity provided by mucosal IgA, acting at the first line of defence, will advance vaccination design and surveillance principles, pointing to novel treatment approaches, and to new routes of vaccine administration and boosting. Significance statement We unveiled powerful mucosal neutralization upon BNT162b2 vaccination, mediated by temporary polymeric IgA and explored its longitudinal properties. We present a model, whereby the molecular architecture of polymeric mucosal IgA and its spatial properties are responsible for the outstanding SARS-CoV-2 neutralization potential. We established a methodology for quantitative comparison of immunoreactivity and neutralization for IgG and IgAs in serum and saliva in molar equivalents for standardization in diagnostics, surveillance of protection and for vaccine evaluations.

Sterilizing immunity is defined as the ability of the immune system to prevent massive replication and subsequent transmission of a pathogen. Primary infection of some viral pathogens at mucosal surfaces is capable of eliciting sterilizing mucosal and systemic immunity, which is activated in a case of secondary exposure (e.g. enteric Polio-and Rota-viruses; as well as respiratory Influenza virus).

Mimicry of certain elements of viral infection by vaccination aims to train the immune system to be tuned for subsequent challenges with the actual pathogen (Holmgren and Czerkinsky, 2005 ). An ultimate goal of a vaccination campaign besides protection against the disease and death is to achieve a robust sterilizing effect, alleviate the carrier state and interrupt the transmission cycle in the population (Pollard and Bijker, 2021) . In this view, vaccine efficiency has several distinct, albeit interconnected aspects -(i) reduction of viral load at the entry site and preventing spread between individuals; (ii) preventing viral spread within the host and expediting virus clearance; (iii) protection from symptoms or reducing disease severity.

During the natural course of viral infections pre-symptomatic and asymptomatic individuals can transmit the virus. In an analogy, a vaccine that protects from the disease does not necessarily achieves the sterilizing effect. Pathogen-targeting IgA at mucosal surfaces is known to correlate with sterilizing immunity, thereby preventing transmission of respiratory and enteric viruses (Blutt et al., 2012; Donlan and Petri, 2020; Lavelle and Ward, 2021; Mostov and Deitcher, 1986; Renegar et al., 2004; Wright et al., 2014; Yu et al., 2021) .

Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2), the etiological agent of Coronavirus disease 2019 , is a highly contagious and difficult to contain respiratory virus, regardless of disease status and severity, mainly because both asymptomatic and pre-symptomatic individuals are responsible for substantial transmission events (Harrison et al., 2020; Zhou et al., 2020) . The two mRNA COVID-19 vaccines, BNT162b2 (Pfizer/BioNtech) and mRNA-1273 (Moderna), has successfully reduced the burden of symptomatic COVID-19, and its more serious outcomes, e.g. (Bruxvoort et al., 2021; Dagan et al., 2021; Haas et al., 2021) ). However, the emerging SARS-CoV-2 variants raise concerns as to its long-term protective capability. The immunoglobulin gamma (IgG) responses to natural SARS-CoV-2 infection and the role of IgG against the Spike protein and its

Receptor Binding Domain (RBD) in virus neutralization and disease prevention are well established, e.g. (Isho et al., 2020; Pullen et al., 2021; Röltgen et al., 2022; Yu et al., 2021) . The IgG response to the vaccine has been thoroughly reported, e.g. (Danese et al., 2021; Levin et al., 2021; Sokal et al., 2021) .

IgA is the most abundant immunoglobulin isotype in humans with daily secreted amounts reaching 60 mg/kg/day (Kutteh et al., 1982; Monteiro and Winkel, 2003) . IgA plays a key role in the interaction between the immune system and environmental insults to provide mucosal protection, often serving as the first line of defence (Kerr, 1990; Woof and Russell, 2011) . Beyond its documented role at mucosal surfaces, IgA is the second most abundant isotype in the blood circulation following IgG, with about 20% of total circulatory immunoglobulin content. Serum IgA is predominantly a monomer, whereas secreted IgA at mucosal surfaces appears in a dimer/polymer form (Kerr, 1990) . The IgA dimer, joined through the J chain via disulfide bridges, forms a secretory component (SC-IgA) together with a portion of the polymeric immunoglobulin receptor (pIgR), that is necessary for trans-epithelial secretion (Brandtzaeg, 1981 (Brandtzaeg, , 2013 Brandtzaeg and Prydz, 1984; Mostov and Deitcher, 1986) . Still, in some cases, monomeric IgA can be found at mucosa, and traces of multimeric IgA have been reported in serum as well (Kutteh et al., 1982) . Importantly, the mechanistic relationship between mucosal and systemic immunoglobulin responses are not fully resolved (Iversen et al., 2017) . Both monomeric and dimeric RBD-targeting IgA elicited by SARS-CoV-2 infection were shown to possess strong neutralizing potential in biological fluids and when tested in a monoclonal Ab setup (Cervia et al., 2021; Sterlin et al., 2021; Wang et al., 2021b; Zeng et al., 2021) . However, the immunological characteristics and kinetics of the IgA response, particularly of its mucosal component, upon mRNA vaccination have not yet been deeply investigated, e.g. (Bleier et al., 2021; Danese et al., 2021; Gray et al., 2021; Juncker et al., 2021; Ketas et al., 2021; Matuchansky, 2021; Russell et al., 2020; Wisnewski et al., 2021) .

Here we analysed the humoral immune response to BNT162b2 vaccine and detected transitory secretory dimeric IgA, which targets the RBD of SARS-CoV-2 spike in the saliva of vaccinees. We unveiled the powerful neutralizing activity of this humoral component of the mucosal defence and explored its kinetic profile. Furthermore, we established a methodology for quantitative comparison of immunoreactivity and neutralization for humoral IgG and IgA response in serum and saliva in molar equivalents. We submit a universal approach for assessment of antibody response that can be applied globally and will ease standardization in diagnostics and surveillance practices, in decision making in patients' care, and in comparative vaccine evaluations.

In the course of monitoring the kinetics of the serological response in a BNT162b2 vaccinated cohort, we noticed that along with a well characterized IgG response toward RBD, a substantial proportion of vaccinees developed a time-dependent accrual of RBD-targeting IgA. To further understand functional aspects of BNT162b2 protection, we studied serum immunoglobulin responses to the vaccine and their kinetic properties. Serum samples were taken from 90 participants (Table S1 , cohort details), including pre-COVID-19 cohort, COVID-19 convalescents and vaccinees aged 24 to 75, who received two BNT162b2 doses at a three-week interval (time points included pre-vaccination and follow up till six-month after the first vaccine dose). A more detailed longitudinal follow-up cohort included serum samples (N=76) collected from 18 participants (Table   S2 , cohort details).

We focused on IgG and IgA directed against RBD region of the viral Spike, as many studies have shown that anti-RBD Abs hamper SARS-CoV-2 entry into host cells by competing with the binding to the host-cell receptor Angiotensin Converting Enzyme 2 (ACE2) (Hoffmann et al., 2020; Ju et al., 2020; Lan et al., 2020; Letko et al., 2020; Li et al., 2005) . To measure the antibody response, we produced fully glycosylated recombinant SARS-CoV-2 RBD in a mammalian expression system and used a custom ELISA amenable to quantitative measurements (see Figures S1A-C, methods and supplementary section for details and validation).

Robust anti-RBD IgG and IgA activity was evident in all vaccinees at 10-30 days after the second vaccine dose, versus naïve (pre-COVID-19) individuals ( Figure 1A ). Overall, vaccine-induced anti-RBD IgG was stronger than in convalescents, while the IgA levels were comparable between the two groups, in agreement with recent reports (Cho et al., 2021; Jalkanen et al., 2021; Wang et al., 2021a) . This suggests that at least from the quantitative perspective BNT162b2 vaccine prime/boost regimen initiates anti-RBD humoral immune response in circulation comparable to or even stronger than the one observed upon recovery from the natural infection.

Next, we carried out a detailed time course analysis of the serological response among the vaccinees. Notably, the fact that the majority of commercially available SARS-CoV-2 antibody assays, as well as our results presented in Figure 1B , use arbitrary unit values, impedes the capability to directly compare anti-RBD IgG and IgA levels in terms of molecular stoichiometry. Therefore, we used two different ordinate axes to represent IgG and IgA arbitrary values and only relatively superpose the respective shape and durability of the two isotypes. The magnitude of the serological IgA response among vaccinees was significantly more scattered and overall showed less steep increase than that of the IgG (Figure 1B) , suggesting a higher variability of the vaccine in induction of IgA isotype in the circulation. Monitoring of the circulatory levels for six months post-vaccination in a cohort subset revealed a decline of anti-RBD IgG and IgA ( Figure 1B , see also Figure S1D for violin plot representation of categorized periods).

In order to assess the specific functional contribution of IgA and IgG in serum we performed neutralization analyses using a reporter assay in Vero E6 cells, based on SARS-CoV-2 spikepseudotyped Vesicular Stomatitis virus (VSV). First, we selected a pool of sera from four vaccinated individuals with significant anti-RBD IgA and IgG levels and depleted total IgG molecules. Figure S1E shows the complete drop in total IgG levels upon depletion, as measured by sandwich ELISA (depicted in Figure 2B ). While the original sera pool showed the half neutralization capacity (NT50) at the dilution of 1:360, the IgG depleted pool resulted in a complete loss of neutralization ( Figure   1C ). This indicates that the vaccine-elicited IgG is the functionally predominant neutralizing isotype in blood circulation in accordance with previous findings (Cho et al., 2021; Turner et al., 2021; Wang et al., 2021c) .

BNT162b2 vaccine-elicited circulatory anti-RBD IgA has drawn our attention, particularly due to the crucial importance of IgA in providing the ultimately desired mucosal defence.

As experience with intramuscular RNA vaccination is limited, especially with respect to IgA response overall, and in particular at the mucosal interface, we decided to explore the role of secretory IgA.

One notable obstacle in the functional assessment of the role of IgA in both circulation and mucosal surfaces is the inability to quantify and compare circulatory and mucosal IgG versus IgA. In our view, this is of utmost importance for SARS-CoV-2 studies, in particular due to the current need for a universal absolute measure of humoral response at different physiological sites (WHO/BS/2020.2403). We tackled this obstacle at three levels: (1) Serial dilutions of the serologically evaluated biological fluids to empirically determine the linear confidence range in immunoassays. (2) Implementation of pure human IgG and IgA fractions to create a calibration curve to convert Optical Density (OD) of secondary detection into absolute quantitative units (e.g. moles) independent of the secondary Ab conjugates, e.g. (Kannenberg et al., 2022; Wang et al., 2021a) . (3) Evaluation of the specific contribution of anti-RBD IgA or IgG by determining their proportion out of total immunoglobulins of the same isotype in a given biological fluid and subsequent functional assessment of the isotype-specific depletion. We used serum samples of vaccinated individuals, described in Figure 1B to establish and validate such measurements (see table S1 for the sub-cohort details).

In a typical ELISA used to measure RBD antibodies, we coat the plate with RBD and subsequently react it with the relevant biological fluids. RBD-reacting Abs of all isotypes are captured, while IgAs and IgGs are then differentially revealed by the corresponding isotype-specific secondary Abs ( Figure   2A ). Using commercial pure human IgG and IgA standards with defined concentrations, we established 'OD-to-mole' transformation ( Figure 2B ). To this end we use an ELISA setup, measuring the total IgG and IgA populations rather than antigen-specific subsets. In this case, the plate is first coated by capturing isotype-specific Abs and then the defined amounts of the reference isotypes are entrapped and revealed by the isotype-specific Ab-HRP (Horseradish Peroxidase) secondary conjugates. By introducing such a standard in our experimental routine, we could quantitatively relate OD to the absolute amount of captured immunoglobulins ( Figure S2A -C). Figure S2D shows the specificity of isotype capturing and detection, with no apparent cross-reaction.

We next applied this approach to evaluate molar concentrations of RBD-targeting IgG and IgA in the serum of vaccinees. Figure 2C quantitatively shows that the majority of individuals produced 200-1000 pmol/ml (nM) of RBD-targeting IgG versus 30-200 pmol/ml (nM) of RBD-targeting IgA. Our approach allowed the determination of the proportion of RBD-specific Abs out of the total amount of immunoglobulins of the given isotype in serum, (normalised, proportional formula

) ( Figure 2D ). While absolute quantities of RBD-specific IgG strongly dominate over the corresponding IgA in serum, the normalised fractional quantities were comparable and comprise 0.5 % of IgG and 0.4 % of IgA in circulation ( Figure 2D , graphical representation and Table   S4 , sub-cohort data). This near equal proportional representation of both isotypes toward the RBD antigen upon intramuscular mRNA vaccination, suggests a similar frequency of class-switch events.

In the next series of experiments, we monitored IgA and IgG kinetics expressed in pmol/ml values of anti-RBD in the serum of six individual vaccinees, allowing stoichiometric comparison ( Figure 2E ). All the individuals exhibited predominant IgG response that peaked around 40d post vaccination and gradually declined during six months. In contrast, IgA responses were more variable ( Figure 2E ). In all our measurements, the circulatory IgA picomole values were lower and with shorter duration than those of IgG, similar to other vaccine instances and upon natural immune responses to infections (Salonen et al., 1985) . Of note, we considered the IgA-monomer in serum for our molar transformations, as circulatory IgA is most commonly monomeric lacking the secretory component, while dimeric IgA is mostly found at mucosal surfaces and in mucosal secretions (see also Figure 4 ).

Given the substantial IgA amounts in serum elicited by the vaccine, and the well-established role of secreted IgA in providing mucosal immunity, we asked whether RBD-reactive IgA can be detected in resting saliva of vaccinees. Saliva samples (N=82) were obtained from 33 participants, aged 20-75 (see Table S5 , cohort details).

First, we confirmed that the vast majority of total immunoglobulins in saliva detected by our quantitative ELISA were of the IgA isotype, in agreement with the well-characterized humoral repertoire of the salivary milieu ( Figure S3 ). Next, we turned to quantify the RBD-specific IgA in saliva samples collected at different time points post-vaccination, as indicated ( Figure 3A ). Anti-RBD salivary IgA response was rather variable between individuals, akin to its variability in serum. There The presence of anti-RBD immunoglobulins in saliva is rather encouraging, though the question remains as to its ability to prevent virus entry. Using the VSV-GFP-SARS-CoV-2-Spike pseudotype neutralization assay, a strong concentration-dependent neutralizing activity of saliva from vaccinees was discovered (NT50 1:60) ( Figure 3B , squares, five vaccinees samples). This value is significantly higher than the basal background neutralizing activity of saliva from naïve individuals ( Figure 3B , circles). The background neutralizing activity of saliva from naïve individuals may stem from basal innate antiviral properties of naïve saliva (e.g. proteolytic digestion and lectin properties). For the sake of sterility in neutralization assay, we used pre-diluted saliva samples that were cleared by centrifugation (12,000g, 5min) and subsequent size filtration (0.22µm) (see Figure 3C for the ELISA of clarified saliva samples used in Figure 3B ). The solubility of IgA molecules in saliva is often a matter of concern due to the viscous-colloid, mucus state. We confirmed quantitative recovery of solubilized saliva IgA by comparing pre-and post-centrifugation and filtration samples by quantitative ELISA ( Figure S3B and Tables S6).

Given the IgA prevalence in saliva, we turned to evaluate its functional contribution to neutralization ( Figure 4A ). We generated a pool of five saliva samples from vaccinees, and subsequently depleted either the IgA or IgG molecules ( Figure S4 ). Depletion of IgA, but not IgG, resulted in the loss of neutralization ( Figure 4A ). This result confirms that the strong neutralizing activity of vaccinees is attributed to IgA.

Salivary IgA, similar to all mucosal IgA forms, is produced in a dimeric, tetravalent form, as opposed to bivalent IgA and IgG monomers found in the circulation. To verify whether this is the case in the saliva samples of vaccinees, we employed anti-Secretory Component (SC) quantitative ELISA, measuring molar values of total and anti-RBD secretory dimeric IgA (experimental flow is depicted in Figures 4B and C, see details in Supplementary). We compared a pool of four saliva samples from naïve individuals to the two pools collected from vaccinees at two and three months' post vaccination. In parallel, we analysed the corresponding serological samples. Figure 4D demonstrates that anti-SC reveals RBD-targeting reactivity solely in the saliva of vaccinees and not in their sera. At 60d post vaccination overall anti-RBD IgA reactivity dominated in serum, while the SC-associated anti-RBD form predominated in saliva, in accordance with the strikingly eminent specific neutralization potency of salivary IgA ( Figures 3C and 4C) . In contrast to a sharp decline of anti-RBD IgA in serum at day 108 (3.5 months) post vaccination, the salivary IgA presence and its association with the SC were mounting, in line with data shown in Figures 1B and 3A . We conclude that anti-RBD IgA in saliva of vaccinees originates from bona fide transcytotic secretory pathway, validating its dimeric nature ( Figure 4D ).

To establish the specific neutralization potency of the anti-RBD immunoglobulins, we normalised their relative NT50 values to their actual [nM] concentration in the respective fluids ( Figure 5 -I). This pointed to a two orders of magnitude advantage of saliva anti-RBD IgA (NT50 = 0.02-0.05nM), vs serum anti-RBD IgG (NT50 1nM). We hypothesize that the remarkable neutralization potency of polymeric salivary IgA (relative to monomeric serum IgG) may stem from a combination of (i) the increased avidity of multivalent binding and (ii) a geometrical fit between dimeric IgA and the SARS-CoV-2 spike protein trimer as presented on the surface of virions.

Whereas the avidity components of multivalent binding and neutralization are well studied (Laursen et al., 2018) in viral infections, the subject of complementarity between a virion lattice and the immunoglobulin isotype is less explored. The simplified view of the molecular dimensions of SARS-CoV2 spike and of the studied immunoglobulin isotypes are presented in Figure 5 -IIA (molecules are drawn schematically with respect to their proportional scale). Since the surface glycoprotein lattice is sparse (e.g. majority of trimeric spike vertices are 20-25nm apart) (Klein et al., 2020; Yao et al., 2020) , circulatory IgGs and IgAs might bind to only single glycoprotein spike restricted by their Fab arm spread of (10-14nm). In contrast, dimeric SC-IgA can concomitantly capture two glycoprotein spikes due to its 25nm-longitudinal extension, thereby more efficiently covering -'mantling' the protection, whereby SC-IgA induced inter-particle oligomerization is promoted by the extended and elastic tetravalent branches. Such a situation, fortified by a plethora of polyclonal species, may appear as highly prominent in vivo. In oral immunity, the aggregation of exogenous particles is known to promote mechanical clearance of invaders from mucosal surfaces (Bustamante-Marin and Ostrowski, 2017; Nail et al., 1969) . The consequences might become even more significant in the case of a higher degree of multimerization that was recently observed for human intranasal IgAs.

Intranasal vaccination with Influenza virus in humans has revealed extreme potency and 'neutralization-breadth' toward variant strains. This powerful protection was attributed to nasopharyngeal SC-IgA with the elevating superiority of the multimeric states: dimers, trimers, tetramers and even higher order oligomers (Suzuki et al., 2015) .

Our findings, in conjunction with the 'GedankenExperiment' (Figure 5-IIB and C) of interaction modalities between surface SARS-CoV-2 lattice and mucosal dimeric IgA vs monomeric IgG and IgA in blood circulation, highlight the importance of implementing lattice design to improve the spatial surface-mimicry in the next-generation subunit vaccines. In this respect, the mRNA based vaccine may have had an unexpected benefit by enabling the host-cell to present the natural arrangement of SARS-CoV-2 spike membranal lattice upon its expression.

Overall, our results demonstrate that the BNT162b2 vaccine induces a five-month transient accrual of salivary anti-RBD IgA, extending beyond the time frame of detectable circulatory IgA, putting forward a basis for the establishment of mucosal memory. We suggest that the polymeric origin of the salivary IgA molecules forms the basis for the remarkably high specific neutralizing activity found in the BNT162b2 vaccinees' saliva, compared to serum IgG. Salivary anti-RBD IgA may represent a more general nasopharyngeal humoral component of mucosal protection.

Our study reveals a previously undescribed mucosal component resulting from the intramuscular administration of an mRNA vaccine. We unveil that saliva of vaccinees contains transitory anti-RBD dimeric secretory IgA (Figures 3A, 4D) with strong neutralizing activity ( Figure 3B , Figure 5 -I), possibly explained by its tetravalent nature. We show that this polyvalent IgA is the main mediator of potent neutralization activity in the vaccinees' saliva, remaining unchanged following IgG depletion.

Accordingly, vaccine-induced neutralization was completely abolished by depleting salivary IgA. In contrast, IgG was the predominant neutralizing isotype in serum since its removal resulted in loss of neutralization. Intriguingly, and contrary to the situation in saliva, residual serum IgA was devoid of measurable neutralization activity, despite its significantly higher concentration -about 30-fold higher content in serum vs saliva (when valence differences are accounted, see Methods). The unique feature of mucosal IgA is its association with the joining J-chain that bridges two iso-clonal immunoglobulin molecules upon their synthesis in secreting B-cell and with the SC that mediates trans-epithelial delivery and extends their lifespan in the highly hydrolytic mucosal environment. The functional and mechanistic impacts of such association in terms of avidity and stereochemical properties are discussed below. Epitopic repertoire of salivary and serum IgA may also differ due to affinity maturation driven somatic hypermutations of Nasopharynx Associated Lymphoid Tissue It is yet not clear whether BNT162b2 mRNA vaccination provides temporary sterilizing immunity, in addition to its proven capacity to ward off severe disease. 'Sterilizing immunity' -is crucial for the Comirnaty (mRNA vaccine) (Chan et al., 2021) . Intriguingly, Comirnaty induced anti-spike neutralizing IgA response detected in nasal epithelial lining fluid, while a similar response was not observed in CoronaVac vaccinees, highlighting the mucosal capabilities of mRNA based vs inactivated vaccine (Chan et al., 2021) .

In spite of the importance of IgA for protection against pathogens, a certain fraction of the human population is characterized by IgA deficiency (Brandtzaeg et al., 1999; Morawska et al., 2021) . Only 10-15% of IgA deficient individuals are susceptible to recurrent sino-pulmonary and gastrointestinal infections/disorders, while the vast majority remains asymptomatic and are often incidentally identified among healthy blood donors. In many cases, IgM appears to compensate the deficiency by

replacing IgA at mucosa, as it reacts with pIgR and can be transcytosed to mucosal surfaces (Brandtzaeg et al., 1999) . Whether mRNA vaccines boost mucosal IgM as well in such instances of IgA deficiencies remains to be explored.

The mechanism of eliciting the mucosal humoral component by an mRNA et al., 1996; Théry et al., 1999; Valadi et al., 2007; Zitvogel et al., 1998) . Whether the pre-existing immunological memory at mucosal sites to former instances of respiratory human common cold coronaviruses (e.g. OC43, NL63, HKU1, 229E) is stimulated by intramuscular mRNA boost, panning cross-reactive B-cells remains to be seen (Turner et al., 2021; Wang et al., 2021a) .

The approach we introduce here for the evaluation of anti-SARS-CoV-2 humoral response relies on molar units of antigen specific and of total immunoglobulins. Such molar expressions are welladopted in clinical diagnosis of autoimmune diseases, and provide universal international evaluation and decision making in patients' care (Monogioudi and Zegers, 2019; Tozzoli et al., 2002) . Beyond the obvious benefits of such universality for surveillance and comparative research, our work demonstrates the instrumental importance of absolute units and standardization for mechanistic understanding of functional neutralization. We suggest putting forward this methodological aspect for humoral diagnostics and assessment of vaccine efficiencies in comprehensive universal values.

Comparison of neutralizing activities in serum and saliva upon BNT162b2 vaccination provides the first in vivo evidence of augmented specific neutralization of polymeric IgA. At first glance, 'the multivalent state per se' is an obvious explanation following the orthodox proximity-based statistical models of association-dissociation shift -the 'avidity' component (Jendroszek and Kjaergaard, 2021) .

The valence influence is often more prominent for weak affinities, in agreement with the expected In conclusion, our data reveal the existence and the unprecedented specific neutralization potency of spike-targeting temporary mucosal secretory IgA in saliva of BNT162b2 vaccinees. Moreover, our approach of molar quantification of SARS-CoV-2 immunoglobulins in various body fluids may have practical implications for basic research, as well as for accurate assessment of humoral immunity in diagnostics and in epidemiological surveillance studies. Surveying salivary IgA is non-invasive and easily accessible and as such may be beneficial in the search for correlates of protection. Therefore, if predictive, it can be used for large scale, or individual screening, and for determining the need of an additional boost.

Vero E6 and HEK293T cell lines were obtained from the American Type Culture Collection (ATCC).

Vero E6 and 293T cells were grown in DMEM medium supplemented with 10% (v/v) Foetal Bovine Serum (FBS), 100 IU/ml of penicillin and 100 mg/ml of streptomycin sulfate, and the cells were grown in 5% CO2 and 95%air. Cells were passaged at 80% confluence and seeded as indicated for the individual assays. Proteins were produced in Expi293 or ExpiCHO that were obtained from Thermo Fisher and grown according to the manufacturer's instructions.

From December, 2020 to May, 2021 we enrolled medical and research personal at our university medical center, without previous documented COVID-19, to participate in our study. Eligible participants were both male and female adults prior to or after receiving the BNT162b2 vaccine. The vaccine has been provided as a part of ongoing national vaccination campaign. This study was part of an ongoing study and was reviewed by the Institutional Review Board (0278-18-HMO). All the participants provided written informed consent. Serum samples were obtained from each participant; saliva samples were obtained from a portion of the participants (saliva cohort). Multiple serum samples were obtained from some of our participants in order to investigate the kinetics of immunoglobulins response to the vaccine (longitudinal cohort). These samples were collected in the following time frames indicated in the cohort tables (supplementary) starting from day 0 till day 180 after the first vaccine. Convalescent plasma and pre-COVID-19 era serum samples were provided from the sample bank and used as a reference for the serum studies; saliva from unvaccinated persons was used as reference for saliva studies. The collected samples were kept at -70⁰C (sera), -20⁰C (saliva). They were heat inactivated and filtered prior to ELISA or neutralization assays.

The following plasmids were used for VSV-pseudovirus production: pVSV-ΔG-GFP, pCAGGS-G or pBS-N-Tϕ, pBS-P-Tϕ, pBS-L-Tϕ and pBS-G (Whitt, 2010) . 

Receptor Binding Domain (RBD) of SARS-CoV-2 spike was expressed in mammalian expression system (Expi293 or ExpiCHO) as a secretory protein with sreptag and subsequently purified using To expand the dynamic range of the assays, the samples (sera or saliva) were always applied as a series of 2-or 3-fold dilutions, and the data was obtained from the dilutions, resided in the linear OD range ("Normalised OD"). 

Streptavidin magnetic beads were washed 4x with PBS and then incubated for 45 minutes with biotinylated capturing antibody diluted 1:10 in PBS. Following incubation, the beads were washed again 4x with PBS before the addition of the sera or saliva samples. The samples were incubated while rotating at 4⁰C for 45 minutes with the beads before being removed. The samples were then checked via ELISA to determine whether complete and selective depletion had been achieved.

VSVdeltaG-GFP single round infectious particles were first generated by pseudotyping with VSV-G- 

Serum and saliva samples were heat-inactivated (60°C, 30 min) prior to their use in neutralization assays. Saliva samples were pre-diluted 1:1 with PBS before inactivation to avoid coagulation. Next, the samples were diluted in the cell culture growth medium and filtered by using 0.2 m, cellulose acetate. Vero E6 cells were seeded in 96-well plate and used for neutralization assay at 75-80%

confluency. The sera and saliva samples were serially diluted and subsequently incubated with constant amount SARS-CoV-2 spike pseudotyped virus for 1 hr, 37°C. Next, the mixture was transferred to the monolayer and incubated for 16-24hrs. The green fluorescence signal was observed under the microscope at 18-24h.p.i. The reduction in amount of green-fluorescent cells due to neutralization was calculated in percentage of un-inhibited control infection. The images were captured from several fields of each well and the green cells were calculated by using automated image analysis by ImageJ (NIH). The graphs were plotted to get the 50% neutralization titer (NT50), in GraphPad Prism. Table S1 for cohort and sampling details. Independent ordinate axes for IgG (left, blue) and IgA (right, red) highlight the restricted, relative nature of the comparison between isotypes in this experiment, as discussed in the text, see also 

Intramuscular mRNA BNT162b2 vaccine against SARS-CoV-2 induces production of robust neutralizing salivary IgA

To evaluate the magnitude and the composition of anti-SARS-CoV-2 acquired humoral immunity, we have developed quantitative ELISA using the recombinant RBD derived from SARS-CoV-2 spike protein ( Figure S1A, B) . Figure S1A shows Figure S1B ). The RBD glycoconjugate analysis demonstrates that approximately 10% of the MW is attributed to N-linked glycans, highlighting their significance in the RBD surface antigenic properties ( Figure S1B ). The Pearson correlation coefficient of 0.967 between our quantitative test and the ARCHITECT (Abbott, Illinois, U.S.A.) anti-RBD IgG test, confirmed the reliability of our assay, see Figure S1C , for correlation analysis and Table S3 , for cohort details.

Basolateral B-cells in the lamina propria generate J-joined dimers of IgA that are then luminally delivered by transcytosis to be secreted at mucosal surfaces. The part of the pIg receptor, cleaved during transcytosis (the secretory component, SC), remains permanently associated with the mucosal IgA and is important for its stabilization. To determine the origin of RBD-specific salivary IgA in vaccinees, we asked whether it contains SC. To this end, we employed anti-SC quantitative ELISA measuring total and anti-RBD secretory dimeric IgA (experimental flow is depicted in Figures 4B, C) .

The molar values were inferred using pure commercially available dimeric IgA as a reference standard ( Figure 4C ). Table S3 . Table S1 for cohort and sampling details. Independent ordinate axes for IgG (left, blue) and IgA (right, red) highlight the restricted, relative nature of the comparison between isotypes in this experiment, as discussed in the text, see also calculate these values we normalised NT50 expressed in dilutions of serum and saliva (Figures 1 and 3) to the molar concentration in saliva and serum, respectively [nM] . NT50 value for salivary IgA was calculated based on average NT50 dilution of 1:20, upon normalization to basal inhibitory activity of naïve saliva (see Figure 3B ). NT50 value for salivary IgA was calculated based on average NT50 dilution of 1:20 anti-RBD IgG, average dilution of 1:300), (Figure 5 -I). The plausible mechanisms behind such stark, two orders of magnitude difference in NT50 between salivary and serum immunoglobulins are addressed in the 'Thought experiment' described below and summarized in the model presented in 

Figures S1-S4

Tables S1-S6 Figure 2B (see Methods for details). Tested samples included: naïve saliva (N=13), vaccinated saliva (N=22), vaccinated sera (N=13). IgG is the predominant isotype in serum, while IgA is predominant in saliva in accordance with previously published data. (B) Isotype specific OD-to-mole transformation for saliva samples presented in Figure 3 (see methods for details). Shown are standard dilution curves used for extrapolation in capture ELISA format. Extrapolated molar concentration of used to evaluate recovery yields for soluble anti-RBD IgG and IgA from saliva samples upon (i) centrifugation and (ii) filtration are presented in Table S3 . Part II: Samples of participants included in the survey that stayed unvaccinated after recovery or recovered and further vaccinated: "rec" corresponds to recovered participants and "rec vac" to participants that were recovered from disease and further vaccinated. ii.

i. Table S6 

Secretory immunity with special reference to the oral cavity

Direct evidence for an integrated function of J chain and secretory component in epithelial transport of immunoglobulins

The B-cell system of human mucosae and exocrine glands

Real-world effectiveness of the mRNA-1273 vaccine against COVID-19: Interim results from a prospective observational cohort study. The Lancet Regional Health -Americas

Cilia and Mucociliary Clearance

Systemic and mucosal antibody responses specific to SARS-CoV-2 during mild versus severe COVID-19

The Mucosal and Serological Immune Responses to the Novel Coronavirus (SARS-CoV-2) Vaccines

Anti-SARS-CoV-2 receptor-binding domain antibody evolution after mRNA vaccination

Differential Kinetics of Immune Responses Elicited by Covid-19

Systemic and Mucosal Immune Responses in Mice after Rectal and Vaginal Immunization with HIV-DNA Vaccine

Mechanisms of SARS-CoV-2 Transmission and Pathogenesis

An intranasal vaccine durably protects against SARS-CoV-2 variants in mice

Distinguishing features of current COVID-19 vaccines: knowns and unknowns of antigen presentation and modes of action

Mucosal Vaccination via the Respiratory Tract

Systematic Review of Mucosal Immunity Induced by Oral and Inactivated Poliovirus Vaccines against Virus Shedding following Oral Poliovirus Challenge

SARS-CoV-2 Cell Entry Depends on ACE2 and TMPRSS2 and Is Blocked by a Clinically Proven Protease Inhibitor

Mucosal immunity and vaccines

Persistence of serum and saliva antibody responses to SARS-CoV-2 spike antigens in COVID-19 patients

Strong Clonal Relatedness between Serum and Gut IgA despite Different Plasma Cell Origins

COVID-19 mRNA vaccine induced antibody responses against three SARS-CoV-2 variants

Nanoscale spatial dependence of avidity in an IgG1 antibody

Cross-presentation by dendritic cells

Intranasal Protollin TM /F1-V vaccine elicits respiratory and serum antibody responses and protects mice against lethal aerosolized plague infection

Human neutralizing antibodies elicited by SARS-CoV-2 infection

The Levels of SARS-CoV-2 Specific Antibodies in Human Milk Following Vaccination

Antibody course and memory B-cell response in the first year after SARS-CoV-2 infection

Generation of antigen-specific immunity following systemic immunization with DNA vaccine encoding CCL25 chemokine immunoadjuvant

The structure and function of human IgA

Antibody Responses to SARS-CoV-2 mRNA Vaccines Are Detectable in Saliva

SARS-CoV-2 structure and replication natural and vaccine-induced defences

Certified Reference Materials and their need for the diagnosis of autoimmune diseases

IgA Fc Receptors

The Epidemiology and Clinical Presentations of Atopic Diseases in Selective IgA Deficiency

Polymeric immunoglobulin receptor expressed in MDCK cells transcytoses IgA

Epipharyngeal receptors responding to mechanical stimulation

Mucosal Immunity Induced by Enhanced-Potency Inactivated and Oral Polio Vaccines

Intranasal Vaccination with 1918 Influenza Virus-Like Particles Protects Mice and Ferrets from Lethal 1918 and H5N1 Influenza Virus Challenge

Lung mucosal immunity: immunoglobulin-A revisited

A guide to vaccinology: from basic principles to new developments

Selective functional antibody transfer into the breastmilk after SARS-CoV-2 infection

Intramuscular vaccination targeting mucosal tumor draining lymph node enhances integrinsmediated CD8+ T cell infiltration to control mucosal tumor growth

B lymphocytes secrete antigen-presenting vesicles

Role of IgA versus IgG in the Control of Influenza Viral Infection in the Murine Respiratory Tract

Antigen Presentation of mRNA-Based and Virus-Vectored SARS

Immune imprinting, breadth of variant recognition, and germinal center response in human SARS-CoV-2 infection and vaccination

BNT162b2 mRNA vaccine elicited antibody response in blood and milk of breastfeeding women

Mucosal Immunity in COVID-19: A Neglected but Critical Aspect of SARS-CoV-2 Infection

Kinetics of specific IgA, IgD, IgE, IgG, and IgM antibody responses in rubella

mRNA vaccination of naive and COVID-19-recovered individuals elicits potent memory B cells that recognize SARS-CoV-2 variants

IgA dominates the early neutralizing antibody response to SARS-CoV-2

Multiplex Beads Assay and Volumetric Absorptive Microsampling to Generate Immune Repertoire Cartography

Enhanced SARS-CoV-2 neutralization by dimeric IgA

mRNA vaccine-elicited antibodies to SARS-CoV-2 and circulating variants

Generation of VSV pseudotypes using recombinant ΔG-VSV for studies on virus entry, identification of entry inhibitors, and immune responses to vaccines

WHO/BS/2020.2403 Establishment of the WHO International Standard and Reference Panel for anti-SARS-CoV-2 antibody. WHO/BS/2020.2403 EXPERT COMMITTEE ON BIOLOGICAL STANDARDIZATION

Human IgG and IgA responses to COVID-19 mRNA vaccines

Structure and function relationships in IgA

Intestinal Immunity Is a Determinant of Clearance of Poliovirus After Oral Vaccination

Mucosal vaccines against respiratory syncytial virus

Molecular Architecture of the SARS-CoV-2 Virus

Distinct kinetics of immunoglobulin isotypes reveal early diagnosis and disease severity of COVID-19: A 6-month follow-up

Characterization of SARS-CoV-2-specific antibodies in COVID-19 patients reveals highly potent neutralizing IgA

A pneumonia outbreak associated with a new coronavirus of probable bat origin

Eradication of established murine tumors using a novel cellfree vaccine: dendritic cell derived exosomes

We would like to thank for the generous support of the research from The Edmond and Benjamin de Rothschild Foundation. The research was also supported by grants from Panet for the help with reagents and plasmids used for the initial establishment of a variety pseudotype assays. We are thankful to the Hadassah blood bank members and the Hadassah Clinical virology lab members. We would like to especially thank to all our cohort participants and to blood bank donors for their readiness and collaboration.

Hebrew University of Jerusalem, Hadassah-Hebrew University Medical Center and Hadassah Academic College has filed a patent application "Compounds and methods for increasing antibody's neutralization properties and methods for assessing antibody response" on which MSR, SK, AF, PG, RW, DP, LB, and AR are listed as inventors.

 Figure 4A . Left-side bars (IgA-depleted) samples were measured by sandwich ELISA for total IgG (blue column on the left) and for IgA (red column on the left). Right-side bars (IgG depleted) samples were measured by sandwich ELISA for total IgG (blue column on the right) and IgA (red column on the right), see methods for further details. IgG and IgA values are represented on separate ordinate axes (indicated), as explained in the main text. Horizontal dashed line indicates saturation level of ELISA measurement. Horizontal dashed line indicates saturation level of ELISA measurement. Total immunoglobulins were measured in this experiment in saturated conditions to confirm the completeness of the depletion. (B) Specific detection of 'bona-fide' dimeric IgA of mucosal origin bound to Secretory Component (SC) using anti-SC antibody (see methods and diagrams in Figure 4B , C for further details). The experimental details are depicted in diagrams below the corresponding bars. Serial dilutions of commercial standards: (i) monomeric human IgA purified from serum and (ii) dimeric secretory human IgA from colostrum were used as indicated. Graphical legend explains the pictogram identity. (C) Standard curves used for OD to molar transformations in the experiment presented in Figure 4D . Dimeric secretory human IgA from colostrum was used as a standard for detection with anti-human-IgA and anti-human-SC antibodies as indicated. "rec" corresponds to recovered participants and "rec vac" to participants that were recovered from disease and further vaccinated. Sample numbering is coded and presented according to combined cohort table.