Figure 3. SATB2 binds and regulates EMT and neural crest-associated genes.

(A) GO-term enrichment of SATB2-bound loci (GREAT analysis) from anti-SATB2 ChIp-seq on MCR:SATB2 tumors. Genes are defined as bound if SATB2 peaks are found within 3 kb of the transcriptional start (TSS) or end site (TES), and the gene body (GB). (B) H3K27Ac and H3K9me3 TSS +/- 2 Kb profile of H3K27Ac at SATB2 bound targets comparing MCR:SATB2 in red and MCR:EGFP in green. (C) ChIP-seq and significant differentially expressed genes RNA-seq (p<0.05, q < 0.05, FC>|1.5|) overlap was performed on primary zebrafish tumors to identify SATB2-bound genes that are misregulated. (D) qRT-PCR validation of transcriptionally altered SATB2-bound and associated targets associated to neural crest development, and the actin cytoskeleton and extracellular matrix. Symbols indicate single primary tumors, normalized to β-actin. *p<0.05, ***p<0.0001, ****p<0.0001. SATB2-bound genes are highlighted in purple. (E) Whole mount primary tumor immunohistochemistry for E-cadherin (red), scale bar is 100 μm, Vimentin (green), scale bar is 100 μm, Fibronectin 1 (red), scale bar is 50 μm, and Collagen 1 (red), scale bar is 50 μm, Dapi (blue). (F) Injection in MCR model as described in Figure 1A of MCR:EGFP, and MCR:SNAI2. (Log-rank (Mantel-Cox) test p<0.0001****). (G) Gross anatomy of zebrafish injected with MCR:SNAI2. Arrows show melanomas. (H) H and E representative histology of MCR:EGFP, MCR:SATB2, and MCR:SNAI2 tumors. Black bar is 100 µm.

Figure 3—figure supplement 1. Analysis of SATB2 expression and chromatin effect in human and zebrafish melanoma.

(A) Western blot analysis in MCR:EGFP, MCR:SATB2, and MCR:SATB1 zebrafish tumors of SATB2 expression and antibody cross reactivity with human SATB1 (in MCR:SATB1) and lack of detection of endogenous zebrafish satb1. The SC-81376 antibody was used for anti-SATB2 ChIP-seq in MCR:SATB2 tumors, since the other anti-SATB2 antibodies shown in Figure 2—figure supplement 3B did not work in immunoprecipitation. (B) HOMER motif prediction analysis of significant peaks (P<E-9) from anti-SATB2 ChIP-seq enrichment over input in MCR:SATB2 tumors showing the top motif to be AT-rich as expected. (C) HOMER known motifs enrichment analysis on anti-SATB2 ChIP-seq peaks with p<E-9 in MCR:SATB2 (D) Go-term enrichment analysis of genome-wide H3K27Ac peaks enriched in MCR:SATB2 over MCR:EGFP Around TSS Log2FC > 1(GREAT analysis). (E) RNA-seq analysis of satb2 expression in sorted crestin:EGFP⁺ developing neural crest cells (data source7). (F) Knock-down of satb2 in the transgenic neural crest reporter line Tg(sox10:mCherry) using a validated morpholino for satb2. While satb2 knock down did not affect melanocyte development (0/120), satb2 morphants displayed severe cranio-facial abnormalities as observed by overall reduced sox10 reporter expression, reduced pectoral fins, and head size (109/120) (arrows). (G) Example analysis of the chromatin state of a direct SATB2 target showing the pdgfab locus. ChIP-Seq track in zebrafish MCR:SATB2 for SATB2 (in black). ChIP-seq tracks for H3K27Ac and H3K9me3 in primary tumors MCR:SATB2 (red) and MCR:EGFP controls (green).