Uncovering strain- and age-dependent innate immune responses to SARS-CoV-2 infection in air-liquid-interface cultured nasal epithelia

Summary Continuous assessment of the impact of SARS-CoV-2 on the host at the cell-type level is crucial for understanding key mechanisms involved in host defense responses to viral infection. We investigated host response to ancestral-strain and Alpha-variant SARS-CoV-2 infections within air-liquid-interface human nasal epithelial cells from younger adults (26–32 Y) and older children (12–14 Y) using single-cell RNA-sequencing. Ciliated and secretory-ciliated cells formed the majority of highly infected cell-types, with the latter derived from ciliated lineages. Strong innate immune responses were observed across lowly infected and uninfected bystander cells and heightened in Alpha-infection. Alpha highly infected cells showed increased expression of protein-refolding genes compared with ancestral-strain-infected cells in children. Furthermore, oxidative phosphorylation-related genes were down-regulated in bystander cells versus infected and mock-control cells, underscoring the importance of these biological functions for viral replication. Overall, this study highlights the complexity of cell-type-, age- and viral strain-dependent host epithelial responses to SARS-CoV-2.


Figure S2 .
Figure S2.Immunofluorescent confocal microscopy staining at 40X magnification of ALI-HNECs reveals strain-and age-dependent loss of cilia, related to Figure 1b.A) Child ALI-cultures infected with Alpha-variant harvested and imaged at 72 hpi.(B-G) Individual channels shown with all donors with final combined channels.(B-D) Stains in adult donor cells infected with WT and Alpha SARS-CoV-2 in B) adult 1, C) adult 2, D) adult 3. (E-G) Stains in child donors infected with Alpha SARS-CoV-2 in E) child 1, F) child 2 and G) child 3. Stained for α-tubulin (AcTub, green), nucleoprotein (NP, red) and nuclei (DAPI, blue).WT-infected, Alpha-infected, and mock-control cells are shown for adults and only Alpha-infected cells are shown with children due to lack of spare ALIs available for children.Mock-control cells were harvested at 7 days post-infection (dpi).Scale bar: 50 μm.

Figure S3 .
Figure S3.Percentages of each cluster in infected compared with mock-control datasets in adult and child ALI-cultures, related to Figures 2f-g.X-axis shows the number representing the various cell-type clusters in the data and the Y-axis shows the percentage contribution of these cell-types.Bars with black outlines indicate the percentage of cluster in the mock-control dataset, and the bars without an outline indicate the percentage of cell-type cluster in the infected dataset (i.e.WT 48 hpi, WT 72 hpi, Alpha 72 hpi).Data are represented as mean, n=3, where n is each donor within an age-group (adult/child, p = * ≤ 0.05, ** ≤ 0.01, *** ≤ 0.001, **** ≤ 0.0001, moderated T-test).

Figure S4 .
Figure S4.Dot plot showing the expression of major cell-type-specific markers in each cell-type, related to Figures 2a & f.X-axis shows the individual cell-types within the data, and the Y-axis presents the cell-type marker genes.Bubble size indicates proportion of cells in cell cluster expressing the gene, and color indicates the average expression.

Figure S5 .
Figure S5.RNA velocity directionality reveals transcriptional dynamics in cells with secretory and/or ciliated properties, related to Figure 2b.A) RNA velocities (dynamical model) projected onto pre-computed UMAP embeddings.B) Partition-based graph abstraction (PAGA) graph shows directionality from Secretory/Ciliated → Secretory-Ciliated cells C) Phase portraits of marker genes for ciliated (FOXJ1, SNTN, CAPS) and secretory (SCGB1A1, MUC5AC, MUC5B) cells.X-axis represents the expression of spliced mRNA and Y-axis represents the expression of unspliced mRNA.D) Proportion of spliced (blue) and unspliced (red) transcripts in each cluster.

Figure S6 .
Figure S6.Significantly enriched reactome pathways analyzed using multiGO using significant DE results, related to Figures 3a-b & 4a.A) Infected vs mock-control cells, B) bystander vs mock-control cells and C) infected vs bystander cells.Columns with no matching DE data available are denoted with 'N'.Bubble size indicates -log10 enrichment p-values, and the color of the bubble indicates the proportion of upregulated genes involved in the pathway (i.e.fracUp).A subset of the results from top 35 terms are shown except for Figure S6c which shows the subset of results from top 100 terms.^ abbreviation of Respiratory electron transport, ATP synthesis by chemiosmotic coupling, and heat production by uncoupling proteins.The same numbering of cell clusters shown in the legend are used across the panels.Thresholds of pv_thresh=0.05,enrichment pv_thresh=0.005and logFC_thresh=1 were used.A full list of enriched reactome pathways is available via the links in TableS3.

Figure S7 .
Figure S7.Dot plot showing average expression of cytokine-related genes involved in GO biological term (cytokine production involved in immune response) separated into bystander, low and high/very high infection categories in Alpha 72 hpi adult datasets, related to Figures 3a-b.Xaxis shows the number representative of each cluster (shown in legend) and Y-axis shows the genes involved in the GO biological term (cytokine production involved in immune response).The color of the bubble represents the average expression level, and the size of the bubble represents the percentage of cells in each cluster expressing the gene.

Figure S8 .
Figure S8.Dot plot showing average expression of cytokine-related genes involved in GO biological term (cytokine production involved in immune response) separated into bystander, low and high/very high infection categories in WT 72 hpi adult datasets, related to Figures 3a-b.X-axis shows the number representative of each cell-type cluster (shown in legend) and Y-axis shows the genes involved in the GO biological term (cytokine production involved in immune response).The color of the bubble represents the average expression level, and the size of the bubble represents the percentage of cells in each cluster expressing the gene.

Figure S9 .
Figure S9.Significantly enriched reactome pathways analyzed using multiGO using significant DE results from comparing Alpha-vs WT-infected ALI-cultures, related to Figure 5. Columns with no matching DE data available are denoted with 'N'.Bubble size indicates -log10 enrichment p-values, and the color of the bubble indicates the proportion of upregulated genes involved in the pathway (i.e.fracUp).A subset of the results from the top 35 terms are shown.Same numbering of cell clusters shown in the legend in Figure S6 are used.Thresholds of pv_thresh=0.05,enrichment pv_thresh=0.005and logFC_thresh=1 were used.A full list of enriched reactome pathways is available via the link in TableS3.

Figure S10 .
Figure S10.Expression of VIM and immune profiles within mock-control cells across donors, related to STAR Methods.(A-B) DE genes comparing mock-control cells from donor 6 vs A) all other donors and B) other child donors.VIM is upregulated in donor 6 vs donors 1-5, but not compared with only the child donors.X-axis shows the log2FC and Y-axis shows the -log10padj, with cut-offs at padj=0.05.Dots in blue show the genes which did not meet the threshold of padj = 0.05, and dots in pink show the genes which met the threshold.C) MultiGO output of enriched GO biological terms in mock-control donor 6 against all other donors and against other child donors.Thresholds of padj < 0.05, enrichment p-value < 0.005 and |log2FC| > 1 were used.Bubble size indicates -log10 enrichment p-values, and the color of the bubble indicates the proportion of upregulated genes involved in the term (i.e.fracUp).