Organoid formation from embryonic mouse OE cells

(A-D) The dissection steps of mouse OE for organoid culture preparation. NE, nasal epithelium; R, rostral; C, caudal.

(E) Immunostaining of E13.5 mouse nasal epithelium for Tuj1. The scale bar represents 200 µm.

(F) Schematic model of the culture method for organoid formation from mouse developing OE tissues.

(G) Representative bright-field images of OE organoids. Cyst-like structures are visible after 3-6 days. The scale bar represents 500 µm.

(H) Temporal changes of the organoid size (n = 8). The organoids reach an area of approximately 80,000-160,000 µm2 within 10 days. **p < 0.01; *p < 0.05 Student’s t test.

(I) qPCR for Lgr5 in OE organoids from day 10 to 30 (mean ± s.e.m, n = 3 independent experiments).

(J) qPCR for Lgr5 in E13.5 mouse OE, OE organoids on day 10, and mESCs (mean ± s.e.m, n= 3 independent experiments).

(K) Time-dependent growth of a representative cyst-like structure generated from a single cell derived from E13.5 mouse OE tissues.

OE organoids form a polarized pseudostratified structure

(A-E) Time-course qPCR of OE organoids from day 10 to 30 (mean ± s.e.m, n = 3 independent experiments).

(F-L) Cryosections of OE organoids on day 10 (F and G), on day15 (H-J and L), and E13.5 mouse OE

(K). Immunostaining for Sp8 (F), PanCK (F-H), NCAM (G and H), Tuj1 (I-K), Sox2 (I-K), aPKC (I-K), p63 (L), and Ebf2 (L).

(M) qPCR for NCAM in E13.5 mouse OE, OE organoids on day 20, and mESCs (mean ± s.e.m, n = 3 independent experiments). ***p < 0.001; Student’s t test.

(N-P) The effect of Notch signaling on the expression of OMP. (N) untreated, (O) DAPT treatment (500 nM, day 10-20). (P) Quantification of the OMP+ cell percentages (n = 5). **p < 0.01; Student’s t test.

(Q-S) The effect of NFs and FBS on the expression of Foxj1. (Q) NFs and FBS treatment (day 0-10),

(R) FBS treatment (day 0-10). (S) qPCR for Foxj1 in E13.5 mouse OE, OE organoids on day 10 treated with NFs and FBS, and OE organoids on day 10 treated with FBS (mean ± s.e.m, n = 3 independent experiments). *p < 0.05; Student’s t test.

(T) Schematic of olfactory/respiratory epithelial differentiation from olfactory basal cells by treatment of NFs or FBS.

The scale bars represent 100 µm (A-D), 50 µm (E-G, N, O, Q, and R).

Analysis of cellular organization in OE organoids by scRNAseq

(A) Visualization of clustering results of day10 OE organoids cultured in NF- or FBS-medium on UMAP.

(B) Cell number ratio in each cluster shown in A

(C) Visualization of expressions of markers for NRE (Foxj1, Krt19), mesenchymal cells (Prrx1) and OSN (Ebf2) on UMAP.

(D) The expression of each subtype marker in OE organoids cultured under each condition is shown in the dot plots.

(E) DESC-assisted clustering result of day 20 OE organoid on UMAP. Each cluster was named according to the cell subtype identified by the gene expression pattern.

(F) Genes specifically expressed in each cluster (DEGs) are shown in heatmap.

(G) Heatmap of correlation coefficients between clusters using all gene expression patterns.

(H) The dot plot showing the expression of each subtype marker in OE organoids on day 20.

(I) The SUS cell population was re-clustered and embedded into UMAP to visualize the expression of specific marker for BG (Muc5B) and SUS (Cyp2g1, Cyp2a5).

(J) The MV cell population was re-clustered and expanded into UMAP to visualize the expression of specific marker for MVs (Trpm5, Cftr).

Lineage trajectory and RNA velocity of OE organoid on day 20

(A) Dimensionality reduction of scRNA-seq results using first and third principal components. The left panel shows color coding by DESC-assisted clustering result. The middle and right panel show the expression of Krt5 (red), Ebf2 (blue, in middle panel) and Cyp2a5 (blue, in right panel)

(B) Inferred lineage trajectory by Slingshot.

(C) Visualization of pseudotime of 3 different trajectories estimated by Slingshot on UMAP.

(D) The result of arranging the cells of each trajectory on a line along the pseudotime.

(E) Normalized gene expression along pseudotime of the top four clusters of highly variable genes extracted by Tradeseq. The bottom panel shows the results of enrichment analysis of the genes included in each cluster.

(F) RNA vectors of day20 OE organoids were calculated with scVelo and overlayed on the PCA plot (PC1/PC3).

(G) The phase portraits of marker genes.

(H) Calculated latent time by scVelo was overlayed on the PCA plot (PC1/PC3).

(I) The proposed model for cell lineage relationships during OE organoids formation.

Analysis of the functionality of mature OSNs and identification of molecular codes for classifying OSNs by OR expression

(A) Immunostaining for axons extended from OE organoids on day 20. The scale bars represent 200 µm (upper panel) and 50 µm (lower panel).

(B) Analysis of Ca2+ dynamics in individual axons on day 20. Numbers correspond to individual ROIs analyzed simultaneously by Ca2+ imaging.

(C) Number of OR genes expressed in OSNs of OE organoids on day 20.

(D) Heatmap showing the expression of each OR gene in OSNs that exclusively express one type of OR.

(E) The matured OSN was re-clustered and embedded into UMAP to visualize the expression of Ncam2 (blue) and Acsm4 (red).

(F) Scatter plots based on the expression of Ncam2 and Ascm4.

(G) Heatmap showing the correlation coefficients of expression of 249 axon guidance-related genes in mature OSNs.

(H) Discrimination of OSNs based on OR classes using a deep neural network. The learning curve using 249 genes is shown in the left panel, and the learning curve using the remaining 17 genes after feature reduction by random forest fitting is shown in the right panel.

(I) Heatmap showing the correlation coefficients of expression of the remaining 17 genes in mature OSNs

(J) Heatmap showing the expression of the remaining 17 genes in mature OSNs.

(K) Immunostaining for Cdh2 and Plxnb2 in mouse OE.