mRNA and protein expression of MECOM, PAX8, SOX17 and WT1 loci in normal tissues and cancer

(A) Unsupervised hierarchical clustering of MECOM, PAX8, SOX17 and WT1 mRNA expression in the pan-normal GTEx dataset. TCGA data was clustered based on the 5 main clusters of TF expression from GTEx. MTFlow = Cluster of tissues that did not or lowly expressed MTFs, MTFhigh = cluster of tissues that moderately or highly expressed all four MTFs, Mhigh = Cluster of tissues that highly expressed MECOM, M&Phigh = cluster of tissues that highly expressed both MECOM and PAX8, Phigh = cluster of tissues that expresses PAX8 and Shigh = cluster of tissues that expresses SOX17). (B) A boxplot of the average Pearson correlation values of MECOM, PAX8, SOX17 and WT1. Ranked from highest average Pearson correlation value to lowest. (C) A boxplot representing the mean positivity rate of MECOM, PAX8, SOX17 and WT1 expression in each histotype. In B and C, the limits of the boxes represent the interquartile range, and the limits of error bars represent the minimum and maximum value without outliers (D) Ratio of samples with number of co-stained TFs based on a 0.1 positivity rate threshold.

MECOM, PAX8, SOX17 and WT1 super-enhancer landscape in FTSECs and HGSCs

Landscapes of active chromatin and chromatin loops (based on Hi-C maps) in FTSECs and HGSCs. (A) MECOM, (B) PAX8, (C) SOX17 and (D) WT1.

MECOM, PAX8, SOX17 and WT1 are tumor-specific therapeutic targets

(A) TF knock-down followed by colony formation assays stained with crystal violet. Representative wells are shown. (B) Barplots representing the quantification of crystal violet staining in Figure 3A. Error bars represent biological replicates. (C) Dose response curves for FT246, FT282, KURAMOCHI and OVCAR4 cells treated with THZ1, THZ531 and JQ1. Error bars represent standard deviation of mean cell survival values from biological replicates. (D) RT-qPCR quantification of MECOM, PAX8, SOX17 and WT1 expression upon THZ1 and THZ531 treatment in FT282 and OVCAR4 cells. Data for MECOM, PAX8 and SOX17 expression upon THZ1 and THZ531 treatment of OVCAR4 cells are reproduced from Reddy et al., 2021.

MECOM, PAX8, SOX17 and WT1 form a core-regulatory circuit in FTSEC to HGSC but cross-relation between factors changes during tumorigenesis

(A) MECOM, PAX8, SOX17 and WT1 co-occupies its own and others genomic loci. (B) MECOM, PAX8, SOX17 and WT1 co-occupy active enhancer regions across the genome. CPM-normalized CUT&RUN reads were centered on 3 kilobase windows of FTSEC or HGSC-specific PAX8 peaks. Rows are the same across feature. (C) Metagene plot, MECOM, PAX8, SOX17, WT1, H3K27ac and H3K27me3 signal centered on FT of HGSC-specific PAX8 peaks. (D) Set analysis of CUT&RUN peaks from representative MECOM, PAX8, SOX17 and WT1 samples in FT282 and KURAMOCHI. (E) Chromatin state of TF peaks categorized by number of TF overlaps. (F) MECOM, PAX8, SOX17 and WT1 co-regulation based on TF knock-down followed by RNA-seq and differential expression analysis with DESEQ2. (G) Node and edge plot representing co-regulation of each TF based on gene expression measured by RNA-seq after TF knock-down.

MECOM, PAX8, SOX17 and WT1 cistromes are remodeled during tumorigenesis

(A) Set analysis of TF binding sites that were common or specific to FT282 or KURAMOCHI. (B) MECOM, PAX8, SOX17 and WT1 co-occupies regions that were bound in a common or context-specific manner. CPM-normalized CUT&RUN reads were centered on 3 kilobase windows of PAX8 peak start and stop positions. Rows are the same across samples. (C) Ratio of chromatin states associated with TF binding regions categorized by cellular context. (D) Ratio of FT282 and KURAMOCHI specific enhancers bound by one, two, three or four TFs. (E) BHLHE41 and PBX1 loci displaying the co-localization of MECOM, PAX8, SOX17 and WT1 at a KURAMOCHI specific super-enhancer.

Gene regulation by MECOM, PAX8, SOX17 and WT1 in FTSECs and HGSC

(A) Log2 fold-change of 28,158 genes following normalization informed by ERCC spike-in RNA content, (B) Number of differentially expressed genes for each TF knock-down based on absolute log2 fold-change ≥ 0.5. (C) Schematic to integrate differential expression, TF binding and topologically associated domain (TAD) maps. (D) Alluvial plot displaying the status of high confidence differentially expressed genes following TF depletion in FTSECs and HGSCs. (E) Heatmap and pathway analysis of genes displayed in D. (F) BRCA1 locus highlighting H3K27ac signal and TF binding at the BRCA1 promoter. (G) Log2 fold change of BRCA1 expression following MECOM, PAX8, SOX17 and WT1 knock-down relative to scrambled controls. (H) Chromatin landscape of RUNX3 locus in FTSECs and HGSCs. (I) Log2 fold change of RUNX3 expression following MECOM, PAX8, SOX17 and WT1 knock-down relative to scrambled controls (RNA-seq).