Fig 1: Leading Edge Heterogeneity Revealed by Spatial Transcriptomics(A) Hematoxylin and eosin (H&E) staining of tissue sections and unbiased clustering of ST spots. Scale bar, 500 µm.(B) Violin plots of TSK scores of individual spots derived from scRNA-seq data (sc-TSK score) for each cluster. Dotted boxes outline clusters with highest average sc-TSK score.(C) Spatial feature plots of TSK-high cluster, sc-TSK score, and TSK marker MMP10 expression in tissue sections.(D) Violin plots of TSK-proximal signature score (intersection of differentially expressed genes in patient 2 cluster 4 and patient 10 cluster 2, n = 34 genes) for cell types in scRNA-seq data.(E) Left, H&E staining with leading edge of tumor annotated (dotted lines) and isolated leading edge spots labeled by cluster. Right, bar plots of total number and percentage of spots at leading edge per cluster.(F) Violin plots of non-TSK leading edge signature score (intersection of differentially expressed genes of non-TSK spots at leading edge from patient 2 and patient 10, n = 6 genes) by tumor subpopulations.(G) Violin plots of COL17A1 expression by tumor subpopulation in scRNA-seq data.(H) Immunohistochemical staining of COL17A1 in patient tumors. Scale bar, 50 µm.(I) Projection of non-TSK leading edge-associated clusters with dot plot of select gene ontology (GO) terms of differentially expressed genes in each cluster (n = 200 genes for patient 2 clusters, n = 20 for patient 10 cluster).(J and K) H&E, sc-TSK score, and MMP10 expression feature plots from (J) patient 4 and (K) patient 6 with data generated using the Visium ST platform.Scale bar, 500 µm. KC, keratinocyte; TAM, tumor-associated macrophage. See also Figure S3 and Table S4.
Fig 2: Histological and immunofluorescent verification of multi-origin 3D organotypic raft cultures.(A) The top row represents foreskin epithelia, the middle row represents cervical epithelia, and the bottom row represents oral epithelia. The first column shows H&E micrographs of 3D organotypic epithelial raft cultures. The second column depicts immunofluorescent staining of the 3D raft cultures with COL17A1 (Collagen XVII; green), CRNN (Cornulin; red), and DNA (DAPI-stained; blue). Columns three and four contain immunohistochemical reference staining of COL17A1 and CRNN from The Human Protein Atlas. Colorimetric 3,3’-diaminobenzidine (DAB)-stained COL17A1 micrographs include: normal skin tissue (T-X0500, Female, age 66, Patient id: 4786), normal cervical tissue (T-83000, Female, age 37, Patient id: 4504), and normal oral mucosa tissue (T-51000, Male, age 62, Patient id: 3724) all stained with antibody HPA043673. CRNN micrographs include: normal skin tissue (T-80100, Female, age 66, Patient id: 3357), normal cervical tissue (T-83000, Female, age 55, Patient id: 2005), and normal oral mucosa tissue (T-51000, Male, age 62, Patient id: 1505) all stained with antibody CAB026182. Images are available from v21.proteinatlas.org. Scale bars represent 50 µm. (B) Schematic illustration of the apical, midzone, basal, and dermal layers in stratified squamous epithelia. CRNN-positive cells are typically present in the apical layer (highlighted in red), while COL17A1-positive cells are typically present in the basal layer (highlighted in green). The midzone gap represents the double-negative area from the inferior portion of CRNN to the superior portion of basal epithelial cells. (C) Midzone gap distances were quantified between skin, cervical, and oral epithelia in our 3D rafts and reference Protein Atlas images (patient id indicated for each tissue, v21.proteinatlas.org). Violin plots were made for each independent donor measured while points represent the ratio of each midzone gap distance measurement divided by the lowest donor’s average midzone gap measurement. All measurements were taken at 25 µm intervals spanning the length of visible tissue across non-overlapping fields of view using ImageJ/Fiji. Mean ratios for each donor are indicated by red bars.
Supplier Page from MilliporeSigma for Anti-COL17A1 antibody produced in rabbit