Mucosal administration of ChAdOx1 nCoV-19 induces durable mucosal anti-S IgA in the URT

Despite the robust systemic immunity induced by i.m. administration of COVID-19 vaccines, breakthrough infections and viral transmission continue to occur. Studies have suggested that vaccination regimens capable of inducing strong mucosal immune responses, including local IgA and tissue-resident cellular immunity, may be more effective in preventing respiratory infections, including SARS-CoV-2 infection^24,25,26,27. To investigate the ability of different vaccine routes to induce a mucosal immune response, we compared i.m. and i.n. routes of vaccination. Four- to 6-week-old C57BL/6 J (B6) male and female mice were vaccinated i.m. with two doses of an mRNA vaccine (1 µg), administered 3 weeks apart. Six weeks after the second dose, the mice received either a third dose of mRNA vaccine via i.m. administration (designated as group IM) or a single dose of ChAdOx1 nCoV-19 vaccine (3.8 × 10⁷ virus particles in 25 µl PBS) via i.n. administration (designated as group IN) (Fig. 1A). Mice that received two doses of an mRNA vaccine (1 µg), administered 3 weeks apart, served as the baseline group, while mice that received three doses of PBS were used as controls. Samples from the baseline group were collected on D0. All vaccines encoded the ancestral SARS-CoV-2 S protein.

Mucosal ChAdOx1 nCoV-19 administration results in enhanced spike-specific IgA response in the upper respiratory tract. (A) Outline of the study. 4- to 6-week old B6 mice (male and female) were vaccinated i.m. with two doses of an mRNA vaccine (1 µg), administered 3 weeks apart. Six weeks after the second dose, the mice received either a third dose of mRNA vaccine via the i.m. route or a single dose of ChAdOx1 nCoV-19 vaccine (3.8 × 10⁷ virus particles in 25 µl PBS) via the i.n. route. Mice that received two doses of an mRNA vaccine (1 µg), administered 3 weeks apart, served as the baseline group. Sera, NALT media, and BALF were collected at baseline, D3, D14, and D84 to investigate humoral immune responses against nine different SARS-CoV-2 variants. Image created with BioRender.com. (B,C) Radar plots showing IgG (B) and IgA (C) levels within sera, NALT media, and BALF samples at baseline, D3, D14, and D84. The scale indicates mean fluorescence intensity values. The blue plot corresponds to the IM group, whereas the purple plot corresponds to the IN group. Statistics were performed using two-way ANOVA followed by Tukey’s multiple comparisons test. Significance between IM and IN groups are indicated as follows: * = p-value < 0.05; ** = p-value < 0.01; *** = p-value < 0.001; **** = p-value < 0.0001.

To investigate differences in humoral responses across different tissue compartments, we collected sera, bronchoalveolar lavage fluid (BALF), and nasal-associated lymphoid tissue (NALT), a secondary lymphoid organ that runs lengthwise along the base of the nasal passage which is functionally equivalent to human tonsils²⁸, from the URT at baseline, D3, D14, and D84. The collected NALTs were incubated in media for 3 days, after which the media containing secreted antibodies was harvested. Antibodies in sera represent the systemic response, those in BALF represent the lower respiratory tract response, and those in NALT media represent the URT response. Binding IgG and IgA antibody titers were measured against nine different SARS-CoV-2 variants: Wuhan (Ancestral), Alpha, Beta, Delta, and Omicron variants BA.1, BA.2, BA.2.12.1, BA.2.75, and BA.5 (Fig. 1B-C). Significant differences were determined using the two-way ANOVA followed by Tukey’s multiple comparisons test.

Since the vaccines were based on ancestral SARS-CoV-2 S protein, IgG and IgA responses across all samples were stronger against earlier variants of SARS-CoV-2 than the Omicron variants. Binding IgG titers in sera collected at baseline and on D3 were equivalent. However, binding IgG titers in sera were significantly higher in the IM group compared to the IN group for Omicron variants at D14, and for Wuhan and Alpha variants at D84. In NALT media, binding IgG titers were elevated at both D14 and D84 compared to baseline and D3. In BALF, IgG titers were highest at D3 and D14. IgG titers in NALT media and BALF were similar across both vaccine groups at all time points (Fig. 1B).

S-specific binding IgA titers in sera, NALT media, and BALF peaked at D14. In serum, there were no significant differences between the two vaccine groups at all time points tested. In stark contrast, IgA titers were significantly higher in the IN group in both NALT media and BALF. At D14, this was observed for all variants, whereas at D84, IgA titers in NALT samples remained significantly higher against Wuhan, Alpha, Beta, Delta, and Omicron BA.1, but not against later Omicron variants. IgA titers in BALF were low at D84. (Fig. 1C).

We then performed an ACE-2 competitive binding assay to quantify antibodies that block the binding of the S protein to ACE2, represented as the percentage of binding inhibition (Supplementary Fig. 1). This ACE-2 neutralization assay serves as an alternative to live virus or pseudovirus neutralization assays²⁹. In line with the binding antibody profile, we observed higher inhibition of earlier SARS-CoV-2 variants compared to Omicron variants in sera. Blocking antibody titers exceeded 75% for the Wuhan, Alpha, Beta, and Delta variants at all time points including baseline. As expected, at D14, blocking antibodies against Omicron variants BA.1, BA.2, and BA.2.12.1 were significantly higher in the IM group compared to the IN group. Blocking antibody titers in NALT samples and BALF were mostly below the limit of detection (20%).

Taken together, our results underscore that mucosal administration of the ChAdOx1 nCoV-19 vaccine induces durable IgA responses in the respiratory tract, whereas i.m. administration of the mRNA vaccine alone does not. In contrast, systemic neutralizing antibody responses against earlier Omicron variants were significantly higher in the IM group compared to the IN group.

Mucosal administration of ChAdOx1 nCoV-19 induces spike-specific tissue-resident CD8 + T cells in the URT

We hypothesized that myeloid cell populations in lung tissue at D3 would be influenced by where a vaccine was administered, and which vaccine platform/route was utilized. Thus, we collected lung tissue at D3 and performed high dimensional flow cytometry. Three minutes before tissue collection, an intravenous injection of CD45 antibody (CD45 IV) was performed. This allows enough time for the antibody to circulate throughout the body and stain CD45+ cells in circulation, but not to penetrate tissues and stain resident cells. In the IN group, we observed significantly elevated numbers of CD11b+ dendritic cells (DCs, CD45 IV−, CD11b+, CD11c+, Ly6C+, F4/80+) and interstitial macrophages (IMs, CD45 IV−, CD11b+, CD11c+, Ly6C−, F4/80+, MHC Class II+) compared to the IM group as well as PBS-injected control animals. We also detected increased counts of neutrophils (CD45 IV−, CD11b+, CD11c−, Ly6G+), plasmacytoid DCs (CD45 IV−, CD11b−, CD11c+, Ly6C+, F4/80−), Ly6C+ monocytes/macrophages (CD45 IV−, CD11b+, CD11c−, Ly6G−, Ly6c+, F4/80+) and Ly6C- monocytes/macrophages (CD45 IV−, CD11b+, CD11c+, Ly6C, F4/80+, MHC Class II−) in both vaccine groups compared to controls. No differences in alveolar macrophages (CD45 IV−, CD11b−, CD11c+, Ly6C−, F4/80+) were observed between groups. Thus, we observed an influx of CD11b+ DCs and IMs into the lungs of animals that received a mucosal vaccination (Fig. 2A).

Tissue-resident spike-specific CD8 + T cells and IgA + B cells are enriched in animals that received an IN vaccination. Nasal turbinates (NT) and lung tissues were collected at baseline, D3, D14, and D84 and compared to cell populations in animals that received PBS injections. (A) Myeloid populations were quantified via flow cytometry in lung tissue collected at D3. (B) The percentage of residential CD4 + T cells (left) and CD8 + T cells (right) in lung tissue from total CD4 + or CD8 + T cells. (C) The percentage of residential CD4 + T cells (left) and CD8 + T cells (right) in lung tissue producing cytokines defined on the X-axis upon ex-vivo stimulation with SARS-CoV-2 spike peptides. (D) The percentage of residential spike-specific tetramer + CD8 + T cells in NT (left) and lung (right) tissue. (E) The percentage of residential spike-specific tetramer + CD69 + CD103 + CD8 + T cells in NT (left) and lung (right) tissue. (F) The percentage of residential CD19 + cells in lung tissue. (G) The total percentage of IgA + CD19 + cells in lung tissue. (H) B cell ELISpot assay showing the number of spike-specific IgA producing B cells in the spleen, lung, and CLN. A Kruskal–Wallis test followed by a Mann–Whitney test was used to determine statistical significance. * = p-value < 0.05; ** = p-value < 0.01; C = naïve animals; DC = dendritic cells; IM = interstitial macrophages; pDC = plasmacytoid dendritic cells; m/m = monocytes/macrophages; IFN- γ = interferon-γ; TNF-α = tumor necrosis factor α; IL = interleukin; Tet = tetramer; CLN = cervical lymph node.

Next, we investigated the presence and functionality of CD4+ and CD8+ T cells in lung and nasal turbinate tissues. First, we determined the fraction of CD45 IV- cells within total CD4+ and CD8+ T cells in lung tissue (residential T cells, Fig. 2B). Residential T cells were similar for controls, baseline animals, and animals that received a third IM vaccination. However, lung resident CD4+ and CD8+ T cells were significantly increased at D3 and D14 in the IN group compared to the IM group, but not at D84. At D14, single cell suspensions obtained from lung tissue were stimulated with SARS-CoV-2 S peptides, and the expression of IFN-γ, TNF-α, IL-2, IL-4, and IL-10 was compared to mock-stimulated cells using flow cytometry. In resident CD4+ T cells obtained from the IN group, a significant increase in the frequency of cells producing IFN-γ and IL-10 was detected compared to cells from the IM group. Furthermore, polyfunctional CD4+ T cells, producing both IFN-γ and TNF-α, were also significantly elevated in the lungs of animals from the IN group. In contrast, only the production of TNF-α by resident CD8+ T cells was significantly elevated in the IN group compared to the IM group (Fig. 2C).

We next assessed the immune phenotype of the T cells in lung and nasal turbinate tissues collected at baseline, D3, D14, and D84 by probing tissue-resident memory (T_RM) T-cell-specific markers, including CD44, CD103, and CD69. S-specific CD8 + T cells were sorted further using major histocompatibility complex (MHC) class I tetramer S_539-546 (VNFNFNGL). Compared to the IM group, nasal turbinates of animals from the IN group contained a significantly increased frequency of Tet + CD8 + T cells (CD45 IV-, CD3 + , CD8 + , CD44 + , Tet +) at all three time points. Importantly, up to 21.5% of Tet + CD8 + T cells displayed co-expression of CD103 and CD69, T_RM signature markers, and co-expression was found to be highest at D84. In contrast, in lung tissue of animals from the IN group the frequency of Tet + CD8 + T cells (CD45 IV−, CD3 + , CD8 + , CD44 + , CD62L−, Tet +) was highest at D14, and not significantly different from those detected in animals from the IM group at D84. Accordingly, co-expression of CD103 and CD69 was highest on D14 but was reduced to below 5% at D84 (Fig. 2D,E).

T cell responses were also investigated in spleen, cervical lymph nodes (CLN, the draining lymph nodes of the URT), and inguinal lymph nodes (ILN). In CLN, Tet + CD8 + T cells numbers were relatively low at all time points throughout all groups except from the IN group at D14, and surface co-expression of CD69 and CD103 was significantly higher at this time point. Tet + CD8 + T cells in ILN were highest at D3 in the IM group, and expression of CD69 and CD103 was minimal. In spleen, the number of Tet + CD8 + T cells was significantly higher in the IM group at D3, increased to similar levels in both vaccine groups at D14, and then reduced to undetectable levels in both vaccine groups at D84. The median percentage of Tet + CD8 T cells in spleen at D14 hovered around 20% for both groups, slightly higher than that observed in lung tissue of the IN vaccinated group at this time point (14.7%). Expression of CD103 and CD69 was low, but slightly increased in the IM group compared to the IN group at D3 (Supplementary Fig. 2A).

Single cells suspensions obtained from spleen and CLN at D14 were then stimulated with SARS-CoV-2 S peptides, and release of IFN-γ or IL-2 was detected via ELISpot. In spleen, IFN-γ and IL-2 production was not significantly different between vaccine groups, and the number of cells producing IFN-γ was approximately 10–30 × higher than the number of cells producing IL-2. In CLN, only IFN-γ producing cells obtained from the IN group were above background levels (Supplementary Fig. 2B,C).

Resident B cell populations in lung tissue (CD45 + , CD19 + , CD45 IV−) were increased at D14 in the IN group compared to the IM group (Fig. 2F). Additionally, a higher percentage of these cells also expressed IgA on their surface (Fig. 2G). No differences in IgA + CD19 + CD45 IV− cells were seen in the nasal turbinates (Supplementary Fig. 2D).

The specificity and functionality of B cells was investigated in single cell suspensions of spleen, lung, and CLN tissue using a B cell ELISpot, in which the plate was coated with S protein (Ancestral variant), and IgA secretion was assessed using a 5-h incubation protocol. The number of IgA-producing plasma cells was significantly higher in the lung and CLN of the IN group compared to the IM group but was similar in the spleen (Fig. 2H).

Mucosal vaccine administration enhances immune cell expansion and proliferation in the URT

Since an increase in immune cell populations in nasal turbinate and lung tissues was observed in the IN group via flow cytometry, we aimed to investigate whether mucosal vaccination induced similar changes in the NALT. To assess this, we primed 4- to 6-week-old B6 mice with i.m. mRNA vaccination. Three weeks later, animals were boosted with a single dose of i.n. ChAdOx1 nCoV-19 vaccine. At 3-, 7-, 14-, and 28-days post vaccination (DPV), entire heads were collected. Coronal sections at approximately the level of the first molar were prepared to visualize the NALT, and sections from vaccinated animals were compared to those from unvaccinated animals (Supplementary Fig. 3A,B). Sections were stained for CD3 (T cells), PAX5 (B cells), IBA1 (macrophages), S protein, or Ki67 (proliferation).

Compared to unvaccinated controls, mucosal vaccination resulted in an increase in NALT size, with the largest size observed at 7 DPV, followed by a gradual decrease at 14 and 28 DPV (Fig. 3A, Supplementary Fig. 3C). Triplex immunostaining was performed to identify the localization of T cells, B cells, and any S protein which may be present in the vaccinated animals. Increased T and B cell staining was observed in the NALT tissues of animals that received a mucosal vaccination. In control animals, T cells were primarily localized at the periphery of the NALT. At 3 DPV, a robust increase in T cell number was observed, still predominantly located around the periphery. By 7 DPV, T cells had infiltrated the center of the NALT, where they were surrounded by B cells, and this distribution persisted at 14 and 28 DPV. Despite triplex staining showing no S protein presence in NALT, individual IHC revealed immunoreactivity in the olfactory epithelial mucosa of three out of four samples at 3 DPV and one out of three samples at 7 DPV, with no staining at 14 and 28 DPV (Fig. 3B-C). Ki67 staining showed gradual increases in cell proliferation from 3 to 28 DPV, with the highest levels at 14 DPV (Fig. 3D).

Increased size, accompanied with increased T and B cell numbers and Ki67 staining in NALT tissue of i.n. vaccinated mice. (A) H&E staining of NALT tissues. (B) IHC staining of NALT tissues with CD3 (yellow), PAX5 (teal), and SARS-CoV-2 S protein (purple). (C) IHC staining of nasal turbinate tissues with S protein (purple). (D) IHC staining of NALT tissues with Ki67 (purple). Images are representative of n = 4 mice per group. Magnification, × 200; scale bars, 100 µm.

Robust T cell infiltration into the lamina propria of the nasal turbinates and sinuses was visible at 3 DPV and peaked at 7 DPV. Intriguingly, T cell staining showed aggregate formation at 14 DPV which persisted through 28 DPV (Supplementary Fig. 4A). Macrophage staining indicated increased antigen-presenting cells in the NALT starting at 7 DPV. Macrophage numbers in the nasal turbinates peaked at 7 DPV and were significantly reduced at 28 DPV (Supplementary Fig. 4B).

Intranasally administered ChAdOx1 nCoV-19 vaccine provides long-term protection against SARS-CoV-2 infection in the URT

We demonstrate a clear distinction in mucosal immunological outcomes between the two vaccine regimens investigated in this study. To evaluate whether these differing immunological outcomes translate into differences in protection against infection, we assessed vaccine efficacy in K18-hACE2 transgenic mice. These mice express the human ACE2 gene under the control of the human epithelial cell cytokeratin 18 promoter, which is critical because SARS-CoV-2 utilizes ACE2 for cellular entry.

Due to the unavailability of an mRNA vaccine encoding the ancestral SARS-CoV-2 S protein, this part of the study was performed using an mRNA vaccine that encodes the S protein of the XBB.1.5 subvariant of Omicron (SpikeVax, Moderna). Mice were vaccinated using the same regimen as outlined previously (Fig. 4A). Eight animals were assigned to each group, with unvaccinated animals (administered PBS via the i.m. route) serving as the control group. At 14 and 84 DPV, all animals were exposed to aerosols containing approximately 3 × 10⁴ tissue culture infectious dose 50% (TCID₅₀) of Omicron EG.5.1 variant, which is antigenically closely related to XBB.1.5³⁰. Mice were euthanized at D2 and D4 post exposure, with four animals per group per time point.

Intranasal ChAdOx1 nCoV-19 administration protects the upper respiratory tract up to 12 weeks post vaccination. (A) Experimental schedule of the study. Two doses of mRNA vaccination i.m. were followed by either one dose of mRNA vaccine i.m. or one dose of ChAdOx1 nCoV-19 vaccine i.n.. Two weeks (schedule on top) and 12 weeks (schedule on bottom) post-third vaccination, animals were exposed to Omicron EG.5.1 variant via aerosol. Four mice per group were euthanized at D2 and D4. Animals that received PBS served as the control group. Image created with BioRender.com. (B) Relative weights of the animals from D0 to D4. Significance indicated in blue and grey asterisks represents comparisons between IN vs. IM and IN vs. control, respectively C. Viral sgRNA present in oral swabs obtained on D1 to D4. (D) Viral load in NT at D2 and 4. (E) Viral load in lung tissues at D2 and 4. Significance was calculated using two-way ANOVA mixed-effects model followed by Tukey’s multiple comparisons test for B and C, and unpaired t-test followed by Mann–Whitney U test for figures D & E. Dotted lines indicate the limit of detection.* = p-value < 0.05; ** = p-value < 0.01; *** = p-value < 0.001.

A significant difference in relative body weight between the IN and IM groups was observed on Day 1 in the 14 DPV group. In contrast, no differences in relative weight were observed between study groups exposed to virus at 84 DPV (Fig. 4B). High viral loads were detected in oral swabs obtained from control animals, but viral RNA was absent from swabs obtained from vaccinated animals challenged at 14 DPV. In the animals exposed at 84 DPV, viral RNA was detected in swabs from two animals in the IM group but remained undetectable in the IN group (Fig. 4C).

Nasal turbinate and lung tissues collected at D2 and D4 revealed high viral load in control animals. In contrast, vaccinated animals exhibited full protection in the lower respiratory tract, regardless of regimen or time since vaccination. Nasal turbinates were also protected in both vaccine groups at 14 DPV. However, 84 DPV, viral RNA levels in the nasal turbinates of animals in the IM group were comparable to those of control animals at D2. Conversely, animals in the IN group maintained full protection. At D4, viral RNA was undetectable in the nasal turbinates of both vaccine groups, whereas it was high in the nasal turbinates of control animals (Fig. 4D-E).

Thus, regardless of vaccine regimen, all vaccinated animals were equivalently protected from infection at 14 DPV, as indicated by minimal weight loss and lack of viral RNA in the respiratory tract. In contrast, at 84 DPV, protection of the URT at D2 was observed only in animals in the IN group.

Differentially expressed genes between tissues collected from the IN and IM groups

Using bulk RNA-sequencing, we identified genes that were differentially expressed between vaccine groups at each timepoint post-virus challenge in nasal turbinates, NALT, and lung tissue. Numerous genes showed differential expression between the groups across timepoints (Supplementary Fig. 5). However, when clustering samples using principal component analysis (PCA) based on these differentially expressed genes, we found limited evidence that these genes fully captured group differences at each timepoint. Notably, distinct clustering by treatment was observed only in nasal turbinate samples collected at D4 (Fig. 5A). In other cases, control samples either separated from the vaccine groups (lung tissue and NALT at D4), or IN group samples clustered separately, whereas IM and control samples clustered together (nasal turbinates and NALT at D2).

Intranasal ChAdOx1 nCoV-19 administration results in upregulation of adaptive immune responses. (A) Principal component analysis of samples using differentially expressed genes between vaccine groups, separated by the tissue sequenced and the day of collection post-exposure. Sample points are shaped based on vaccine type (PBS, IM, or IN). Samples were clustered based on expression similarity using the k-means clustering method, assuming three clusters. Some clusters are not encircled due to the size and dimension of the cluster. (B) Heatmap of expression (shown as fragments per kilobase of exon per million reads, FPKM) of B-cell receptor signaling pathway genes that are differentially expressed between IN and IM samples, collected on D2 from NT. Sample names are colored based on vaccine type, with IN (left, red) and IM (right, blue). (C) KEGG pathway enrichments among differentially expressed genes in the comparison between IN and IM samples, collected on D2 from NT. Pathways are separated by enrichment in upregulated or downregulated genes. Pathway color indicates the degree of significance (following FDR multiple testing correction), while the X-axis shows the fold enrichment of significant genes within that pathway compared to expectation.

Given the observed differences in virus replication in the nasal turbinates at D2, we focused our comparisons between the two vaccine groups on this time point and tissue. Among the differentially expressed genes, we found significant enrichment of B and T cell receptor signaling pathways, with a majority of B cell receptor signaling genes being upregulated in the IN group compared to the IM group. These genes included those related to antibody secretion, such as immunoglobulin light and heavy chains, and the J chain (Fig. 5B). We also identified other enriched functional KEGG pathways among the significantly differentially expressed genes between the IN and IM groups. In the IN group, several immune-associated pathways were upregulated, including genes related to Th1, Th2, and Th17 cell differentiation. Additionally, we found that multiple other pathways, including the PI3K-Akt, MAPK, and JAK-STAT signaling pathways, were downregulated compared to the IM group (Fig. 5C).

Notably, we found the peroxisome proliferator-activated receptor (PPAR), which is associated with enhanced antibody production and B cell differentiation³¹ amongst other immunoregulatory functions³², and the T-cell receptor signaling pathways were the most significantly enriched pathways in the IN group. Additionally, we observed a significant increase in ITGAE expression, which encodes CD103, in nasal turbinates of the IN group at D2.

Thus, we demonstrate that the IN group exhibits enhanced adaptive immune responses in the nasal turbinate at D2 compared to the IM group. These enhanced responses likely account for the observed differences in viral load in the nasal turbinate at this time point.