Fig 2: Ectopic expression of GATA4 in human squamous esophageal epithelial NES-B10T cells alters the expression of squamous cell marker genes. Expression of four squamous epithelial cell marker genes (p63, KRT5, KRT13, KRT15) were examined by qRT-PCR and/or immunoblot in three pInducer20-GATA4 NES-B10T cell clones (G4_1, G4_2, G4_3) and one pInducer20 NES-B10T control cell clone (C) after 72 h of doxycycline treatment to induce GATA4. (A–D) qRT-PCR demonstrated reduced steady-state levels of p63, KRT5, KRT13, and KRT15 mRNA in NES-B10T cells expressing GATA4 compared with controls cells lacking GATA4 protein (n = 3 experiments; *P ≤ 0.05). (E–I) Quantitative infrared immunoblotting (LI-COR) was used to analyze p63, KRT5, and KRT13 protein expression in control and GATA4 expressing NES-B10T cell clones. Revert Total Protein Stain was used for normalization. Immunoblots were performed using protein extracts from two independent doxycycline induction experiments. Representative blots are shown in (E) and (F). Quantification of blots is shown in panels G-I with the abundance of protein detected in GATA-expressing clones expressed as the percent relative to the non-expressing control cells. Full immunoblots used to generate panels are shown in Supplementary Fig. S1.

Fig 3: Pan-cancer survey of TP63 expression and associations with survival: (a) TP63 expression shown for 8,519 TCGA patients from 29 disease cohorts where expression has been z-score normalized per patient for comparability. (b) Pan-cancer hazard ratios calculated using a Cox regression. Diseases with less than 100 patients or fewer than 10% patients events were excluded. (c) Bladder cancer patient TP63 expression distribution was bi-modal with a larger proportion (90%) having expression greater than 1 log2 TPM. Dashed line indicates 1 log2 TPM cut off. (d) Kaplan–Meier visualization comparing survival of bladder cancer patients with low TP63 expression versus patients with increased TP63 expression using a 10th percentile threshold. (e) Bladder cancer patients were stratified using bladder cancer molecular subtypes from Robertson et al. with luminal subtypes combined as a single class. Hazard ratios and maximum likelihood statistics calculated using Cox regression. Upper/lower quartiles and medians represented as boxplots with whiskers indicating +/– 1.5 IQR. Forest plot HR error bars represent 95% CI. (Robertson et al. [4]).

Fig 4: Core TFs regulate LSC function and identity by SE–P interaction networks.a Schematic of TF motif enrichment analysis. b Identified TF motifs in the indicated groups. c Metaplots and heatmaps showing the enrichment of the indicated ChIP-Seq signals at the CTCF and SMC1 sites. d Metaplots showing the enrichment of p63 and RPS at promoter-anchored enhancers and enhancer-anchored promoters. e Chromatin interactions between p63-bound enhancers and selected genes. Blue nodes represent p63-bound enhancers and orange nodes represent target genes. Edges represent chromatin interactions. f Identified chromatin interactions and tracks for the indicated ChIP-Seq signals at NOTCH1 locus. g Metaplot and heatmap showing the enrichment of p63 at SEs. h Identified chromatin interactions and tracks for the indicated ChIP-Seq signals at ITGB4 locus. i p63 and RPS mediated an intersected gene regulatory network via SE–P interactions.

Fig 5: ΔNp63α depletion results in EBV lytic reactivation in EBV+ carcinomas (CNE-2 and NPC43) and telomerase-immortalized oral keratinocyte (NOKs) cell lines.A) CNE-2-Akata cells were infected with lentiviruses expressing shRNAs against p63 or a control sequence, selected with puromycin, and harvested for immunoblot analyses to measure expression levels of the EBV Z, R, and BMRF1 lytic proteins, ΔNp63ɑ, and tubulin proteins as indicated. B) NPC43 cells were infected with lentiviruses expressing shRNAs that target p63, or a control sequence. Two days post-infection, cells were harvested for immunoblot analysis to assess the expression of Z, BMRF1, ΔNp63ɑ, and the loading control tubulin. C) NOKs-Akata cells were transfected with siRNAs against p63 or a control siRNA. Two days after transfection, the cells were harvested for immunoblot analyses and expression of the Z, R, ΔNp63ɑ and tubulin proteins was examined as indicated. The black line indicates where irrelevant lane(s) were removed. The original western blots used to construct this figure are shown in supplemental Fig 1. All figures shown are representative of eight independent biological replicates.