This study was carried out following the Guidelines for Animal Experimentation of the Japanese Association for Laboratory Animal Science and the recommendations set out in the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health. All protocols were approved by the Tokyo Metropolitan Institute of Medical Science Animal Experimental Committee (Approval No. 22-078).

In all experiments, 5–6-week-old BALB/c mice were purchased from Japan SLC (Shizuoka, Japan). The mice were given free access to rodent chow and water, and were maintained on a 12 hr light/12 hr dark cycle. In our previous study, we administered the SARS-CoV-2 spike protein and carboxyvinyl polymer (CVP) (Toko Yakuhin Kogyo Co., Ltd., Toyama, Japan) (S-CVP) intranasally to mice (15). To ensure that the results were comparable with those of the previous study, only female mice were used in this study.

A purified His-tagged active trimer of the SARS-CoV-2 S protein (#SPN-C52H8, Wuhan-Hu-1[R683A, R685A], Newark, DE) was purchased from AcroBiosystems (Newark, DE).

Mice were administered with 5 μg of the SARS-CoV-2 spike protein and CVP (S-CVP). The immunized mice were immunized at 3, 6, and 24 hr. The non-immunized mice that had not been administered the S-CVP were considered the healthy control mice (i.e., 0-hr mice). The CVP was mixed with an equal volume of protein solution just before vaccination. Mice were then immunized intranasally by 10 μL intranasal instillation using a microsyringe (5 μL in each NP) (#MS-NE10, Ito Corporation, Shizuoka, Japan). Mice were immunized once, euthanized at 3, 6, and 24 hr after immunization, and their heads were collected for tissue analysis.

Mice under anesthesia were dissected, and each head was separated into right and left halves using a scalpel. Tissues were fixed by immersing in 4% paraformaldehyde (PFA) (#315-1; TAAB, Berkshire, UK)/PBS(-) pH 7.0–8.0 at RT for 24 hr. After washing with Milli-Q water (18.2 MΩ.cm) obtained from a Millipore Milli-Q Integral 5 water purification system (Merck, Darmstadt, Germany), the tissues were immersed in decalcification solution (10% EDTA-4Na (#347-01881; FUJIFILM Wako, Osaka, Japan)/PBS(-) with 1.5% citric acid monohydrate (#09106-15; Nacalai Tesque, Kyoto, Japan) pH 7.0–7.5) and subjected to continuous shaking at RT for 46–50 hr to promote decalcification. The tissues were placed on grates in a container to ensure that they did not touch the bottom. After washing with Milli-Q water, the decalcified heads were penetrated with paraffin for 20 hr and then embedded in paraffin. Tissue-Tek Paraffin Wax II (#No. 7810; Sakura Finetek Japan, Tokyo, Japan) and Paraffin (#164-13345; FUJIFILM Wako) were used for tissue penetration and embedding; they were mixed at 25:2.1 for penetration or 25:4.2 for embedding. The prepared FFPE blocks were stored at 4°C until thin sectioning was performed. A microtome (#REM-710; Yamato Kohki, Saitama, Japan) and blades (#S35, FEATHER Safety Razor, Osaka, Japan) were used to prepare tissue slices.

For RNA extraction from the FFPE blocks of mouse heads containing the NP and NALT, 8 slices (10 µm) from a region were pooled per specimen (approx. 5 mm × 5 mm). Genomic RNA was isolated using a RNeasy FFPE Kit (#73504, QIAGEN, Hilden, Germany) according to the manufacturer’s instructions. Total RNA was eluted with 20 μL of RNase-free water. RNA concentration was determined using a NanoPhotometer NP80 (Implen, Munich, Germany) and the average fragment size was estimated using an Agilent Bioanalyzer (RNA 6000 Nano Kit, #5067-1511, Agilent, Waldbronn, Germany). Instead of an RNA integrity number, Visium Spatial for FFPE (#100339, #100194, #100251, 10x Genomics Inc., Pleasanton, CA) uses a DV200 value as indicator of RNA quality. According to the manufacturer, DV200 values of > 50% are recommended as being suitable for analysis.

ST analysis using the Visium system was performed according to the manufacturer’s instructions. Briefly, mouse FFPE samples were sectioned to a thickness of 5 μm. The sections were placed on Visium slides and ordered as a time series within the capture area (6.5 × 6.5 mm). Images of hematoxylin and eosin-stained sections were obtained after staining with HE. HE staining and imaging for Visium analysis were performed according to the Visium Spatial Gene Expression for FFPE User Guide (CG000409 Rev A, 10x Genomics Inc.). Image data acquisition was performed using an inverted fluorescence phase contrast microscope (BZ-X710, KEYENCE, Tokyo, Japan). Libraries for Visium were prepared according to the Visium Spatial Gene Expression for FFPE User Guide (CG000407 Rev C, 10x Genomics Inc.). The quality and quantity of libraries were measured using an Agilent 4200 TapeStation Qbit Fluorometer (Thermo Fisher Scientific, Waltham, MA) and a KAPA Library Quantification kit (#KK4824, Roche, Basel, Switzerland). The average size and concentration of libraries was 233–239 bp and 2.1–4.42 nM, respectively. Libraries were sequenced on a NextSeq 550 System (Illumina, San Diego, CA) using a NextSeq 500/550 High Output Kit v2.5 (150 cycles) (#20024907, Illumina) at a depth of approximately 69–127 million reads per sample.

Hematoxylin and eosin (HE) staining and imaging for Visium analysis were performed according to the Visium Spatial Gene Expression for FFPE User Guide (CG000409 Rev A; 10x Genomics Inc.). Briefly, the slides with tissue sections were placed on a thermocycler adaptor with the active surface facing up and incubated for 2 hr at 60°C. The slides were removed from the thermal cycler and allowed to cool to room temperature (RT). The tissue sections were deparaffinized in xylene two times for 10 min and then rehydrated twice in 100% ethanol for 3 min. The slides were placed on the thermocycler adaptor with the active surface facing up and incubated at 37°C for 15 min. The slides were then immersed in 96% ethanol for 3 min, 70% ethanol for 3 min, and Milli-Q water for 20 sec. Thereafter, the slides were immersed in Mayer’s Hematoxylin (#S330930-2; Agilent) at RT for 3 min. They were then immersed in a bluing buffer (#CS70230-2; Agilent) at RT for 1 min and incubated with alcoholic eosin (#HT110116; Merck) at RT for 1 min. Glycerol (85%) was added on the tissue sections, which were then covered with coverslips.

To integrate the four data sets, raw FASTQ files and histology images were processed using the aggr pipeline implemented in Space Ranger-1.3.1 (10x Genomics Inc.). The number of spots under tissues in unvaccinated mice or in mice at 3, 6, 24 hr after vaccination was 2,069, 1,967, 2,235, and 1,889 spots respectively, where the total number of spots arrayed on a single slide of the Visium platform is 5,000. The mean number of reads per spot was 52,872 and 63,343, 56,836, and 36,628, respectively. The median of the number of genes per spot was 6,545 and 6,179, 6,686, and 5,597, respectively.

The results were processed using Loupe Browser 6.1.0 (10x Genomics Inc.). Spots corresponding to bone in the HE-stained slides were selected manually. The gene expression data for the graph-based clusters generated by the Space Ranger software were exported as CSV files. The data consisted of the median-normalized average of gene expression, log2 fold changes, and statistical significance (p values) computed by the “Globally Distinguishing” method for genes with a p value < 0.05 in any of the clusters. The 25 most expressed genes with p values < 0.05 in the csv file were sorted in order of increasing “Median Normalized Average”.

Spots in the HE slides that corresponded to bone were selected in Loupe Browser 6, colored red, and saved as “Bone” and as same as “Nasal epithelium and Lymphoid tissues” (this image will henceforth be referred to as the dot pattern). The “Significant Feature Comparison” setting was set to “Globally Distinguishing” and “Feature Type” was set to “Gene”. The list of genes expressed in the selected region was extracted and the “Displayed Numeric Value” in the “Feature Table Options” was set to the “Median-Normalized Average”. “CSV Export Count” was set to “All Gene Expression” and the.csv file was prepared. The 25 most expressed genes with p values < 0.05 in the csv file were sorted in order of increasing “Median Normalized Average”. It is important to note that, compared to other software such as SEURAT, the analyses performed using Space Ranger and Loupe Browser are limited to filtering, integration, and dimensionality reduction.

The mice nasal samples were fixed in 10% neutral buffered formalin, embedded in paraffin, and sectioned at a thickness of 4 μm for histological analysis. For immunohistochemical (IHC) evaluation, deparaffinized tissue sections were treated by autoclaving sections in 10 mM citrate buffer (pH 6.0) for 10 minutes at 121°C for antigen retrieval. The sections were then immersed in 1% hydrogen peroxide in methanol at room temperature for 20 min to inactivate endogenous peroxidase. The sections for CD4 and CD8α staining were blocked with BlockAce (#UKB80, DS Pharma Biomedical, Osaka, Japan) for 20 min at room temperature. They were then incubated overnight at 4°C with CD4-specific antibody (#ab183685, [EPR19514], 1:500, Abcam, Cambridge, UK) or CD8 alpha-specific antibody (#98941, [D4W2Z], 1:500, Cell Signaling Technology, Danvers, USA) diluted in 0.25% Tween20 PBS. The sections for CD19 staining was blocked with Odyssey blocking buffer (#927-40000, LI-COR, Lincoln, NE) for 20 min at room temperature. Samples were then incubated overnight at 4°C with CD19 specific antibody (#90176, [D4V4B] 1:8000, Cell Signaling Technology, Danvers, USA) diluted with Can Get Signal Immunostain Immunoreaction enhancer solution B (#NKB-601, TOYOBO, Osaka, Japan). Subsequently, secondary labeling was performed by incubation with EnVision+ System-HRP-labeled Polymer Anti-Rabbit (#K4003, Dako Denmark A/S, Glostrup, Denmark) for 1 hr at room temperature. Detection with DAB chromogen was completed using an ImmPACT DAB Peroxidase Substrate kit (#SK−4105, Vector Laboratories, Newark, CA, USA). Nuclear staining was performed with hematoxylin solution.

NALT and NP cells from mice were collected under a microscope (Stemi 305 Cam, Carl Zeiss, Oberkochen, Germany). The cells were then stained with fluorescence-conjugated antibodies for cell surface markers for 20 min at 4°C. The following fluorescence-conjugated monoclonal antibodies were used for FACS analysis: APC-conjugated anti-mouse CD3e antibody (#17-0031-82, eBioscience, San Diego, California), APC-Cyanine (Cy)7-conjugated anti-mouse CD4 antibody (#100413, BioLegend, San Diego, California), Pacific Blue-conjugated anti-mouse CD8a antibody (#100728, BioLegend), Brilliant Violet (BV)650-conjugated anti-mouse CD11c antibody (#117339, BioLegend), phycoerythrin (PE)-CF594-conjugated anti-mouse CD19 antibody (#562291, BD Biosciences, San Jose, California), PE-Cy7-conjugated anti-mouse CD45 antibody (#103113, BioLegend), and BV605-conjugated anti-mouse F4/80 (#123133, BioLegend). Cells were fixed in 2.5% formalin/PBS(-) before being resuspended in 1 mL FACS buffer for analysis using a flow cytometer system (BD LSRFortessa™ X-20, BD Biosciences).

A volcano-plot was created using the R statistical software package (version 4.2.2, R Core Team (2022); Vienna, Austria) and RStudio (version 2022.12.0, Build 353, RStudio Inc., Boston, Massachusetts).

First, we attempted to establish a protocol for obtaining thin slices of the mouse nasal cavity that contained enough quality RNA for ST analysis. This is a challenging step because slicing the mouse nasal tissue requires pre-processing by decalcification, which typically damages RNA. Using a two-day decalcification step with a solution containing both EDTA and citric acid, we established a protocol that allowed us to obtain the sliced sections while limiting damage to the RNA (i.e., DV200 values of above 70%) ( Supplemental Figure 1A ).

By following the manufacturer’s instructions using FFPE blocks prepared from BALB/c mouse nasal tissue, we performed ST analysis using the Visium system. Briefly, 5 μm sections of FFPE samples were placed on Visium slides and ordered as a time series within the capture area (6.5 × 6.5 mm). Images of HE-stained sections were obtained after staining with HE. NGS libraries were subsequently prepared by synthesizing cDNA on the glass slides. The resulting NGS library was sequenced and analyzed using Space Ranger and Loupe Browser 6 software from 10x Genomics ( Supplemental Figure 1B ).

Based on morphological evaluation of the HE-stained slides, “bone” and “nasal epithelium and lymphoid tissue” were selected for analysis as these features were the easiest to evaluate morphologically. HE-stained tissue images showed the frontal bone in the upper cranium, the nasal passage, a hematoxylin-stained area in the nasal cavity, and NALT in the nasopharynx. We wanted to ascertain whether the genes in bone could be identified after demineralization, and whether the genes in lymphocytes could be identified in the lymphoid tissue, as we were interested in clarifying immune system dynamics in the lymphoid tissue.

We manually selected two sets of spots on HE-stained slides; each set corresponded to “bone” and “nasal epithelium and lymphoid tissue” (214 and 411 spots, respectively). We then checked the gene expression profiles of these spots in the naïve mouse sample ( Figure 1A ). Of the 25 most highly expressed genes showing significant up-regulation (p values < 0.05) in the bone-associated spots, 10 were specifically related to bone and muscle. The most expressed gene, Col1a2, encodes type I collagen, the main component of connective tissue, and mutations in the Col1a2 gene are known to cause osteogenesis imperfecta (18). The second most expressed gene, Bglap, encodes the osteogenic marker osteocalcin, a non-collagenous protein that is most abundant in the bone matrix and is a bone-derived endocrine hormone (19). Other genes related to bone and muscle were Col5a2, Col11a2, Serpinh1, Col2a1, Dmp1, Col5a1, Col11a1, and Sost.

The area circled in blue on the HE-stained image ( Figure 1B ) is the NP, and the area circled in red is the NALT. Of the 25 most highly expressed genes showing significant up-regulation (p values < 0.05) in the nasal epithelium and lymphoid tissue-associated spots, most were found to be related to nasal function (olfaction) or to the connective tissues that make up the nose. The gene for Odorant-binding protein 1a (Obp1a), the most abundant gene in the nasal epithelium and lymphoid tissue, encodes an odorant-binding protein and is considered to mediate olfactory transduction in mice (20). Among the olfactory system proteins that function to reversibly bind to odors and pheromones, Mup4 encodes Major urinary protein 4, an isoform of major urinary proteins (MUPs) (21–23). MUPs are expressed at pheromone excretion sites (liver and kidney), while MUP4 is the nasal isoform (24). Of the 25 most highly expressed genes, seven were related to nasal functions (red) ( Figure 1B ) and two were involved in immunity, i.e., Cd74 and Bpifa1 (blue) ( Figure 1B ). We then compared all 18,404 genes expressed in the nasal epithelium and lymphoid tissue-associated spots using log2 fold changes (vertical axis) and median normalized average (horizontal axis) ( Figure 1C ). The five most highly expressed genes (Obp1a, Mup4, Lcn11, Gfy, and Tmsb4x) and six nasal immunity-related genes (Cd74, Bpifa1, Ptprc (CD45), Cd19, Cd4, and Cd8a) (25) are shown in red and blue, respectively ( Figure 1C ). Visium effectively detected genes that are highly expressed in the nasal cavity, such as those in the supporting tissues of the nasal cavity, as well as genes that were weakly expressed (i.e., micro gene expression), such as Cd8a. Log2 fold changes in Ptprc, Cd19, Cd4, and Cd8a were all ≥ 1, indicating that expression was higher in the red spots of the selected area than the median normalized average for the overall region. The median normalized average gene expression levels were used to identify Ptprc, Cd19, Cd4, and Cd8a using the Visium system ( Figure 1C ). The immunohistochemical analysis also detected CD19, CD4, and CD8a positive cells in NALT and NP ( Supplemental Figure 2 ). The FACS analysis also detected CD45, CD19, CD4 and CD8 positive cells in the NALT and NP ( Supplemental Figure 3 ).

We then used ST analysis to evaluate the effect of S-CVP intranasal vaccination on mice. To identify the genes involved in the induction of immune responses by the S-CVP vaccination, BALB/c mice were administered the S-CVP intranasal vaccine. Mouse heads were sampled at 0 hr (unvaccinated) and at 3, 6, and 24 hr post-vaccination ( Figure 2A ). Changes in Ptprc, Cd4, Cd8a, and Cd19, which are genes that are representative of the immune response, were also examined ( Figure 2B ). The expression of Ptprc (Protein tyrosine phosphatase receptor type C, or CD45), which was originally considered to be a leukocyte common antigen, can be used as a marker for all differentiated hematopoietic cells. In this study, the expression of Ptprc was observed in the mouse nasal tissues throughout the experimental period (i.e., at 0, 3, 6 and 24 hr post-vaccination) ( Figure 2B ). A similar pattern of gene expression was observed in other specific lymphoid cell makers, such as Cd19 (B-lymphocyte surface antigen), Cd4 (a glycoprotein that serves as a co-receptor for the T-cell receptor), and Cd8a (a glycoprotein found on most cytotoxic T cells) ( Figure 2B ). ST analysis using the Visium system revealed that in non-vaccinated animals, cells associated with the immune response are present in the nasal cavity, including the NP and NALT. In addition, gene expression of Ptprc, Cd19, Cd4, and Cd8a was not significantly altered shortly after S-CVP vaccination.

M cells, which are known to play an important role in initiating local immune responses in mucosal tissues(8), have been detected in the follicle-associated epithelium of Peyer’s patches in the villous epithelium of the small intestine and the nasal cavity. Gp2 (Glycoprotein 2), Sox8 (SRY-box transcription factor 8), and Spib (Ets transcription factor Spi-B) are M cell markers (26–29), and data obtained from ST analysis of the nasal cavity after S-CVP vaccination were used to examine changes in these M cell markers at 0, 3, 6, and 24 hr ( Figures 3A, B ). The expression of the M-cell marker genes was observed to increase in the NP at 3–6 hr after S-CVP vaccination ( Figure 3A ). In addition, gene expression patterns of Gp2 & Sox8 and Gp2 & Spib were altered shortly after administering the S-CVP vaccination, i.e., at 3–6 hr ( Figure 3B ).

Since no quantitative or positional changes were observed in key immune cells, such as Ptprc, Cd4, Cd8a, and Cd19 ( Figure 2B ), we examined other cell types through integration of the ST profiles across the time series. Visualization of the spots through dimension reduction by Uniform Manifold Approximation and Projection (UMAP) analysis showed that our integration was successful, as a substantial fraction of the spots agreed across the time series ( Figure 4A ).

We also performed clustering of the spots based on their expression profiles. Notable exceptions were Clusters 6 and 9, which appeared only at 3–6 hr after vaccination, and then disappeared again at 24 hr ( Figure 4B ). In particular, Cluster 9 was restricted to the nasal cavity surface within the NP region. Analysis of the difference between the spots in Cluster 9 and the others at 6 hr post-vaccination showed that of the 3344 expressed genes, 295 were significantly upregulated (p values < 0.05 and log2 fold values ≥ 1) and 100 were downregulated (p value < 0.05 and log2 fold values ≤ -1) ( Figure 5A ). The change in expression levels of the genes with p values < 0.05 at 6 hr in Cluster 9 were ordered from the highest to the lowest.

The results showed that Interleukin-6 (Il6) had the highest log2 fold values, so we checked the site and time at which Il6 was expressed using the Loupe Browser 6 program. The results showed a characteristic pattern in which Il6 had the highest log2 fold values, with expression peaking at 3–6 hr and then decreasing at 24 hr. The genes that showed the same gene expression pattern (i.e., a small change in expression level at 0 hr, an increase at 3–6 hr, and a decrease at 24 hr) were then selected using the search function in the Loupe Browser 6 program. The results for genes with p values < 0.05 and those with log2 fold values ≥ 1 were sorted in descending order of log2 values. Among the genes with p values < 0.05 and log2 fold values ≥ 1, Edn2 (Endothelin-2), Ccl20 (C-C motif chemokine 20), Hk2 (Hexokinase-2), Egln3 (Prolyl hydroxylase EGLN3), Fam189a2 (Endosomal transmembrane binding with Epsin1), Elf3 (ETS-related transcription factor Elf-3), and Tmem173 (Sting1) showed similar patterns to that observed for Il6 ( Figure 5B ). In this study, we focused on the four genes with the highest log2 fold changes in the NP, i.e., Il6, Edn2, Ccl20, and Hk2, and examined their dot distribution patterns ( Figure 5C ).

The four genes in Cluster 9 in DCs ( Figure 6A ; Supplemental Figure 4 ), which are immune early response cells, were examined. At 3–6 hr after vaccination, an increase in positive cell signal dots for co-expression of the dendritic cell markers Itgax and Il6 was observed in the NP region. However, Edn2, Ccl20 and Hk2 showed no change at 0, 3, 6, and 24 hr ( Figure 6B ).

A similar analysis was performed for macrophages ( Figure 6A ; Supplemental Figure 5 ). Positive cell signal dots indicating co-expression of the Il6, Edn2, Ccl20, and Hk2 genes, as well as expression of the macrophage marker gene Adgre1, were observed in the NP region at 3–6 hr after vaccination. However, no significant changes in the number of signal-positive dots co-expressing the macrophage marker Adgre1 and Il6, Edn2, Ccl20, and Hk2 were observed. The values indicated by the bars indicate co-expression of the Itgax or Adgre1 genes with Il6, Edn2, Ccl20, and Hk2 ( Figure 6B ).