TP63 mediates the generation of tumour-specific chromatin loops that underlie MYC activation in radiation-induced tumorigenesis

Abstract

Alterations in 3D chromatin conformation may disrupt the interplay between promoters and distal enhancers. How gene regulatory circuits are reshaped during ionizing radiation-induced tumorigenesis remains unclear, and little is known about the mediators that drive these processes. To decipher the chromatin alterations in radiation-induced lung cancer, we performed ATAC-seq, RNA-seq and Hi-C analyses of human bronchial epithelial cells and corresponding radiation-induced malignantly transformed cell lines. We found that this malignant transformation is accompanied by chromatin switching from the inactive B compartment to the active A compartment, an increased number of TADs and gained ATAC-seq peaks that mediate new distal chromatin contacts. We identified tumour protein 63 (TP63) as a mediator of new chromatin-accessible sites that anchor tumour-specific chromatin contacts in radiation-induced tumour cells. A TP63-mediated accessible chromatin site anchors a tumour-specific TAD boundary and multiple tumour-specific chromatin loops, which might underlie MYC oncogene activation during malignant transformation.

Fig. 1. Ionizing radiation-induced tumour cells exhibit aberrant gene expression and pervasive accumulation of accessible chromatin sites.

A Volcano plots showing genes that are up- and downregulated in BERP35T1 and BERP35T4 tumour cells compared to BEP2D cells. B Heatmap showing significantly enriched GO terms of up-regulated genes in BERP35T1 and BERP35T4 tumour cells. C Heatmaps showing the read intensities of 2-kb regions centered on ATAC-seq peaks in normal BEP2D cells and tumour cells. All ATAC-seq peaks were categorized into 3 types: gained in tumour cells, lost in tumour cells, and unchanged (from top to bottom). All replicates were shown in one heatmap with the same order of all genes. The pie charts show the proportions of ATAC-seq peaks located at gene promoters of each type of ATAC-seq peak (orange: ATAC-seq peaks at promoters, ATAC-seq peaks not at promoters).

Fig. 2. A/B compartments switching and alterations in intra-TAD structures are associated with dysregulated gene expression in the ionizing radiation-induced tumour cells.

A Pie charts showing the proportions of each type of locus (stable compartment A, A-to-B switching, B-to-A switching and stable compartment B) comparing BEP2D with BERP35T1 and BERP35T4 respectively. B Overlap of the compartments from B-to-A in BERP35T1 and BERP35T4. C Violin plots showing the log2 fold changes in the expression of genes located within each type of locus (A-to-B compartment switching, unchanged compartment, and B-to-A compartment switching). P-values from a two-sided t-test between groups are shown. Box limits represent 25^th percentile, median and 75^th percentile. Whiskers represent 5^th and 95^th percentile. Source data are provided as a Source Data file. D Boxplots showing the log2 fold changes in the expression of genes located within each type of TAD (from left to right: decreased SDOC in tumour, unchanged SDOC in tumour, increased SDOC in tumour). P-values from a two-sided t-test between groups are shown. Box limits represent 25th percentile, median and 75th percentile. Whiskers represent 5th and 95th percentile. Source data are provided as a Source Data file. E Stacked bar plots showing the fractions of the 3 groups of TADs in regions with an A-to-B compartment switching and regions with a B-to-A compartment switching. F Examples of A/B compartment switching with chromatin accessibility and gene expression tracks, where the shadings correspond to genomic regions with compartment switching that associates with alterations in chromatin accessibility and transcriptional activity.

Fig. 3. Dynamic changes in different SDOC during the process of radiation-induced tumor formation.

A The contact frequency alterations within SDOC-decreased, SDOC-stable, and SDOC-increased TADs in BERP35T4 (left) and BERP35T1 (right) cells. Box limits represent 25th percentile, median and 75th percentile. Whiskers represent 5th and 95th percentile. Significance: n.s. not significant, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, two-sided t-test. Source data are provided as a Source Data file. B Boxplots showing the log2 fold changes in the expression of genes located at the 3 types of loci (contact-increased, contact-decreased and other) within the 3 groups of TADs (SDOC increased in tumour cells, SDOC unchanged in tumour cells and SDOC decreased in tumour cells). Box limits represent 25th percentile, median and 75th percentile. Whiskers represent 5^th and 95^th percentile. Significance: n.s. not significant, *P < 0.05, ***P < 0.001, two-sided t-test. Source data are provided as a Source Data file. C Boxplots showing the log2(FPKM + 1) in the expression of oncogenes located at the 3 groups of TADs (SDOC increased in tumour cells, SDOC unchanged in tumour cells and SDOC decreased in tumour cells). Box limits represent 25th percentile, median and 75th percentile. Whiskers represent 5th and 95th percentile. Significance: n.s. not significant, *P < 0.05, **P < 0.01, ***P < 0.001, two-sided t-test. Source data are provided as a Source Data file.

Fig. 4. TP63 is a mediator of tumour-specific chromatin contact.

A Scatter diagram of SDOC alteration correlated with TFBS in BERP35T1 and BERP35T4. The color and size represent log2 STD of gene expression alteration in BERP35T1 (left) and BERP35T4 (right). B Average chromatin accessibility levels at binding sites of each TF in normal BEP2D cells (x-axis) and tumour cells (y-axis) measured as reads per kilobase of transcript per million reads mapped (RPKM). The pink and green dashed lines correspond to 3-fold increases and 3-fold decreases, respectively, in ATAC-seq read abundances in tumor cells. The names of TFs that were significantly upregulated and showed an increase in the ATAC-seq read abundance of more than 3-fold in either tumour cell line are indicated. FPKM log2 fold change of each differentially expressed TF were marked by different sizes and color. C Bar plots showing the fractions of putative TP63 binding sites in all putative TFBSs located within the 3 groups of gained-in-tumour ATAC-seq peaks (Left bar in each panel: ATAC-seq peaks gained in tumour cells and located at contact-decreased loci; middle bar in each panel: ATAC-seq peaks gained in tumour cells and located at contact-unchanged loci; right bar in each panel: ATAC-seq peaks gained in tumour cells and located at contact-increased loci). Significance: **P < 0.01, ***P < 0.001, two-sided Fisher’s exact test. D Bar plots showing the fractions of ATAC-seq peaks colocalized with P63 ChIP-seq peaks in the 3 groups of loci. Loci were grouped by the number of other distal loci that showed concomitantly increased, unchanged or decreased contact frequencies in tumour cells. ChIP-seq data is from Sato et al. (Cancer Research, 79(24), 6084–6100.). Significance: *P < 0.05, **P < 0.01, ***P < 0.001, two-sided Fisher’s exact test.

Fig. 5. TP63 anchors multiple tumour-specific chromatin loops that instruct MYC oncogene activation in radiation-induced tumour cells.

A Hi-C contact maps of the MYC locus in BEP2D, BERP35T1 and BERP35T4 cells. Differential contact maps between each type of tumour cells and normal BEP2D cells are shown below the Hi-C contact maps of the corresponding tumour types. The red arrows mark the locus where a new TAD boundary in each radiation-induced tumor cells was established. The green arrows mark the anchors of the tumour-specific hub of chromatin loops. B Expression level of MYC in BEP2D, BERP35T1 and BERP35T4 cells. C Genome-browser like plot of the MYC locus showing RefSeq genes, ATAC-seq, TP63, CTCF, Cohesin CUT&Tag tracks of BEP2DNC, BEP2DOE and BERP35T1 cell lines.

Fig. 6. Effects of TP63 overexpression on gene expression, chromatin accessibility and chromatin structure.

A MYC expression difference before and after TP63 overexpression. pValue is calculated using DEseq2. B Western blot of MYC before and after TP63 overexpression. The experiment was repeated three times independently with similar results. Uncropped and unprocessed scans of all blots are provided as a Source Data file. C Venn diagram shows the overlap between BERP35T1 up-regulated genes and BEP2D OE up-regulated genes. D TP63 CUT&Tag, ATAC-seq, CTCF CUT&Tag and Cohesin CUT&Tag heatmap before and after TP63 overexpression. According to ATCA-seq signal foldchange, TP63 peaks were divided into 4 groups, including strongly increase (foldchange≥4), slightly increase (2≤foldchange<4), not significant (0.5<foldchange<2) and decrease (foldchange≤0.5). E APA of contact mediated by different group of TP63 peaks. F Gene expression foldchange mediated by different group of TP63 peaks. Box limits represent 25th percentile, median and 75th percentile. Whiskers represent 5th and 95th percentile. Points represent outliers beyond whiskers. Significance: **P < 0.01, two-sided Wilcox-test. Source data are provided as a Source Data file.