Mucosal vaccines that prevent SARS-CoV-2 infection might provide benefits beyond existing intramuscularly administered vaccines: through enhanced individual-level protection against disease, and population-level reduction of viral carriage and transmission. Secreted neutralising antibodies are most likely to be the crucial effectors for such vaccines generating a mucosal response.

¹

 However, it is unclear to what extent intramuscular administration of SARS-CoV-2 vaccines enhances neutralising antibody titres in the mucosal compartment, nor the breadth of variants that are neutralised.

In response, we describe an adapted version of our high-throughput live virus microneutralisation assay for mucosal samples to establish the effect of fourth dose intramuscular mRNA vaccination on neutralising antibodies against six SARS-CoV-2 variants (Omicron BA.1, BA.2, BA.5, BQ.1.1, XBB.1.5, and XBB.1.16) in paired serum and mucosal samples from 149 participants (appendix p 4)

²

 enrolled in the University College London Hospital and Francis Crick Institute Legacy study (NCT04750356).

²

^,  

³

^,  

⁴

^,  

⁵

^,  

⁶

 Mucosal samples were self-collected via nasopharyngeal swabs into viral transport media.

We found that intramuscular administration of a bivalent (ancestral + BA.1) mRNA vaccine enhanced mucosal neutralising activity (appendix pp 2,5), even when stratified by anti-nucleocapsid (anti-N) IgG negative and positive serostatus (previously uninfected or infected, respectively; appendix p 2). This boost appears restricted to variants closely related to the administered SARS-CoV-2 spike mRNA, with statistically significant enhancement for BA.1, irrespective of previous infection status (p=0·02 and p=0·032 in anti-N IgG negative and positive groups, respectively) and without statistically significantly improving neutralising capacity against more recently circulating (since summer 2023) and antigenically-distant variants in either anti-N IgG- (BQ.1.1 p=0·091; XBB.1.5 p=0·058; and XBB.1.16 p=0·67) or IgG+ groups (BQ.1.1 p=0·32; XBB.1.5 p=0·22; and XBB.1.16 p=0·25). Individuals with an equivalent number of exposures to the SARS-CoV-2 spike protein (through infection or vaccination) had similar quantities of mucosal neutralising antibodies (four exposures: [anti-N IgG- after dose four] vs [anti-N IgG+ before dose four]; BA.1 p=0·18; BA.2 p=0·32; BA.5 p=0·55; BQ.1.1 p=0·85; XBB.1.5 p=0·93; and XBB.1.16 p=0·18), reminiscent of serological findings

²

 in which the number of spike exposures was the crucial determinant of neutralising titres.

To establish which antibody isotype might contribute to the boosting of mucosal neutralisation capacity, we measured the concentrations of total IgA and IgG in viral transport medium samples (appendix p 2). We found that IgA contributed much more to the mucosal immunoglobulin compartment than IgG (median fold-difference [IQR]; pre 20·5 [9·8–53·3]; post 21·6 [12·5–81·2]), and that neither isotype was boosted after vaccination (appendix p 4, IgA p=0·66; IgG p=0·061). However, we did detect a post-vaccine increase in the concentration of immunoglobulins that bound ancestral receptor-binding domain (RBD), present in individuals with and without previous infection (appendix p 4, anti-N IgG- p=9·4 × 10^–7; anti-N IgG+ p=0·0064), by use of anti-N IgG serostatus to stratify previous infections. At a population level, anti-ancestral RBD antibody concentrations were positively correlated with variant-specific mucosal neutralisation (appendix p 4). These ancestral RBD binding per variant neutralisation relationships weaken with more antigenically distinct variants, and in those without previous infection.

Next, we established if there were relationships between serum and mucosal neutralisation (appendix p 7). We found positive correlations between serum and mucosal samples before dose four for BA.1, BA.2, BA.5, and BQ.1·1 (p=0·00025, p=0·00021, p=0·0025, and p=0·0015, respectively) but not the more recent variants XBB.1.5 (p=0·13) or XBB.1.16 (p=0·23). After dose four, there were no statistically significant correlations between serum and mucosal neutralisation, suggesting these compartments are differentially boosted. Within the pre-dose four comparisons, stratified by the presence of anti-N IgG, the correlations were predominantly driven by the individuals with previous infections (appendix p 7). Together, these observations suggest a weaker relationship between serum and mucosal neutralising immunity in uninfected individuals than in individuals with previous infections.

In summary, we have described a high-throughput method for live virus microneutralisation by use of nasopharyngeal sampling into viral transport medium and report several findings with broad applications. Nasopharyngeal self-swabbing into viral transport medium has the advantages of being an easy and economical method of sampling the upper respiratory compartment, avoiding proprietary sampling strips, and allowing repurposing of swabs used for viral testing and isolation. Importantly, parenteral vaccination boosts total mucosal neutralising capacity. Since this boost occurs in individuals both with and without previous mucosal challenge from infection, our data argue against a closed system of mucosal immunity only triggered by a mucosal challenge, such as infection.

⁷

 Similarly, we identified a positive correlation between serum and mucosal neutralisation that was most apparent in individuals with previous infections, thus showing that infection propagates antibodies in both serum and mucosal compartments, and arguing against large mucosal-only locally produced antibody after infection.

⁸

 Furthermore, similar quantities of mucosal neutralisation after equivalent numbers of exposures to SARS-CoV-2 spike proteins—via infection or vaccination—suggest that further boosts with intramuscular vaccines can enhance and potentially broaden mucosal neutralising capacity, as we have seen in the serum compartment. Identifying the sources for these mucosal neutralising antibodies will be the next steps (appendix p 8).

Ongoing large cohort studies will offer the opportunity to further explore these concepts and provide insights relevant to the adjudication of the best monovalent or multivalent formulations for vaccines targeting mucosal immunity. The viral transport medium sampling approach described here enables large-scale sample collection and testing for mucosal neutralisation required for vaccine evaluation,

⁹

 and will allow further exploration of cross-compartment neutralisation—for both present and future generation vaccines—including those directly generating a mucosal response.