Fig 2: Lineage plasticity of basal and mesenchymal tumor cell populations isolated from mouse and human bladder tumors.a Cytokeratin 5 tomato reporter (KRT5ERT2-LSL) mice showing that CK5-positive basal cells are enriched for CD49f expression in normal bladders (top). Subsequent to OHBBN-induced carcinogenesis, Tdt cells show heterogeneity in CD49f and Epcam expression (bottom). Right bar graphs, distribution of Tdt-positive;Epcam-CD49f-positive cells in normal and OHBBN-induced bladder cancer. b In-vivo transplantation of CD49f-high and CD49f-low tumor populations. Sorted and transplanted basal (GFP+EpcamhighCD49fhigh) cells or mesenchymal cell populations (GFP+EpcamlowCD49flow) were implanted to the bladder muscle wall of FVB mice and evaluated for tumor formation. Formed tumors were dissociated to single cells and assessed for redistribution of Epcam-CD49f expression according to the defined quadrants (Q1, Q2, Q3, and Q4) after progression. c CD49f-low cells were isolated by flow cytometry from an early (P0) passage PDX model of MIBC (tumor 15934) followed by transplantation to immune compromised mice. Expression analysis of derived tumors under the same flow cytometry settings, shows marked redistribution and increased expression in CD49f (n = 5). Lower, representative histology of resulting tumor outgrowths derived from CD49flowEpcamlow cells (bar = 100 µM). Individual data points are represented by circles, the error bars represent mean ± SD of the three independent replicates. Each experiment was performed at least six times.

Fig 3: There are multiple age-related gene expression changes in lung epithelial cells. (a) Cell isolation strategy by FACS. CD31- CD45- EpCAM+ cells were sorted. (b) Volcano plot to show distribution of gene expression. Cutoffs of -log10 p-value > 1.3 and fold change > ±1.5 were applied (dotted lines). (c) Heatmap depicting clustering of samples from 3 month old (n = 6) and 22 month old (n = 6) lung epithelia according to gene expression profile. (d) RT-PCR confirmation of differential regulation of genes. Expression was normalised to UBC. Error bars are standard deviations. All genes depicted are significantly differentially expressed (p < 0.05, two-tailed T-test).

Fig 4: Identification of CD49f (Itga6) as a differentially expressed cell surface marker in OHBBN-induced bladder tumors.a Bulk RNA-seq analysis showing differentially expressed genes belonging to various lineage subtypes in the normal bladder (n = 5) and OHBNN induced bladder cancer (n = 5). The surface antigen, CD49f (Itga6, arrows), is overexpressed in bladder tumors and found within the basal mRNA cluster. b tSNE cell clusters 3, 5, 8, and 11 defining Epcam (left) and CD49f (Itga6) expression. c Strategy for discriminating GFP-positive, CD45-negative OHBBN-induced cancer cells from non-tumor host cells in FVB/NJ-transplanted mice. Tumor cells are evaluated for expression of Epcam and CD49f, and define lineage populations within defined quadrants (Q1, Q2, Q3, and Q4). d Differential gene expression for lineage subtypes using Epcam-CD49f expression in cell populations of GFP-positive;CD45-negative tumor cells sorted from quadrants Q1, Q2, Q3, and Q4. Individual data points are represented by circles, the error bars represent mean ± SD of the three independent replicates. Each experiment was performed at least in triplicate.

Fig 5: Lineage plasticity of mouse bladder cancer cell populations in vitro.a Strategy to evaluate the plasticity in Epcam and CD49f expression using sorted populations of OHBBN-induced primary bladder cancer cells cultured as organoids. b Flow cytometry plots showing isolated cell populations of tumor cells (left), grown in vitro for 10–15 days, and analyzed for changes in Epcam and CD49f expression. The effects of switching to adherent conditions are shown for gate two (G2) cultured cells (top right) and total CD45-negative cells grown continuously in suspension (bottom left). c Graphs showing redistribution of individual sorted cell populations before (input profile) and after in-vitro culture. Individual data points are represented by circles, the error bars represent mean ± SD of the three independent replicates. Each experiment was performed at least eight times.