Genome-wide recruitment to Polycomb-modified chromatin and activity regulation of the synovial sarcoma oncogene SYT-SSX2

Abstract

Background: SYT-SSX is the oncogene associated with synovial sarcoma (SS), a stem cell disease. SYT-SSX is thought to be responsible for sarcoma initiation and development. It interacts with components of Polycomb and SWI/SNF complexes, the two epigenetic controllers that maintain the heritable status of differentiation-specific genes in the stem/progenitor cell. Through these associations SYT-SSX is thought to alter gene expression programs by epigenetic mechanisms. Recently, we reported that SYT-SSX2 reprograms mesenchymal stem cells and myoblasts by dictating their commitment to the neural lineage while disrupting their normal differentiation. This reprogramming was due to the direct occupancy of proneural genes by the SYT-SSX2 nuclear complex. To gain a clear understanding of SYT-SSX2 control of gene expression networks, we conducted a thorough genome-wide analysis to determine the mechanism of its recruitment and identify signature sets of epigenetic markers that would predict its targeting and transcriptional activity.

Results: SYT-SSX2 was recruited to distinct loci across all chromosomes, and an overwhelming number of Polycomb-modified sites enriched with the trimethylated histone H3 on lysine 27 (H3K27me3) formed the main recruiting module for SYT-SSX2. Not all SYT-SSX2/H3K27me3-occupied genes had altered expression, denoting the requirement for additional signals upon oncogene binding. Differential binding and epigenetic patterns distinguished upregulated and downregulated genes. Most activated genes had SYT-SSX2 sites enriched with H3K27me3 within their body or near their transcription start site (TSS) whereas a majority of downregulated genes were characterized by SYT-SSX2/H3K27me3-rich regions at long-range, or by modifications associated with transcription activation within the gene body or near the TSS. Hierarchical and functional clustering identified H3K27me3 as the dominant epigenetic marker associated with SYT-SSX2 binding and gene expression. Notably, this analysis revealed a cluster of upregulated neuronal genes densely covered by H3K27me3, consistent with programming toward the neural lineage by SYT-SSX2 observed previously.

Conclusions: The data analysis revealed that Polycomb complexes or their modified chromatin and their stably silenced differentiation programs seem to be the main target for SYT-SSX2, suggesting that their perturbation is at the center of tumorigenesis driven by the oncogene. Further research into this mechanism is crucial to the full understanding of SS biology.

Figure 1

Chromosomal distribution of SYT-SSX2 peaks. Diagrams of chromosomes 2, 4, and 15. Chromosomes were subdivided into 500kb bins, and the number of SYT-SSX2 ChIPSeq peaks was tabulated for each bin. Black bars represent individual bins, and the height of each bar is proportional to the number of SYT-SSX2 peaks contained within that window. Scale bar shown represents a distance of 10MB. Red and green stars, respectively, depict the location of upregulated and downregulated genes determined by microarray analysis that could be annotated to SYT-SSX2 occupied regions.

Figure 2

Overlap of SYT-SSX2 peaks with epigenetic markers. Datasets that identified regions of enrichment for histone modifications and PolII binding [28] were compared with the locations of the SYT-SSX2 peaks using the Galaxy analysis suite to determine sites of overlap. The bar graph shows the percent of SYT-SSX2 peaks that overlap with 10 epigenetic markers. Ubiquitylated histone H2 lysine 120 (H2UB) is a marker for transcription elongation. Mono, di, and trimethylated histone H3 lysine 4 (H3K4me1, H3K4me2, H3K4me3), as well as acetylated histone H3 lysine 9 (H3K9Ac), acetylated histone H3 lysine 18 (H3K18Ac), acetylated histone H4 lysine 12 (H4K12Ac), and RNA Polymerase II (PolII) all mark gene promoters and enhancers. Trimethylated histone H3 lysine 36 (H3K36me3) is a marker for active transcription in gene bodies. Trimethylated histone H3 lysine 27 (H3K27me3) is the Polycomb-specific modification associated with chromatin silencing.

Figure 3

Differential pattern of binding between upregulated and downregulated genes targeted by SYT-SSX2. 425 upregulated and 223 downregulated genes were bound by SYT-SSX2 ranging from within the gene body up to 200Kb upstream of the TSS. Top panel shows the percentage of genes that are occupied by SYT-SSX2 within given distance windows. Bottom panel depicts a bar graph showing the percentage of genes for which the associated SYT-SSX2 occupied regions overlap with the given epigenetic markers. Red bars represent upregulated genes, and yellow bars signify downregulated genes.

Figure 4

Hierarchical clustering of differentially regulated genes. Signatures for each differentially regulated gene were derived from the extent of overlap between SYT-SSX2 peaks and the epigenetic markers within the gene body or in 5Kb bins upstream of the TSS up to -50Kb (defined as coverage ratio; each bin has a coverage ratio for each epigenetic marker). Signatures were used as the input for hierarchical clustering using Cluster 3.0. Heatmaps were generated using Java Treeview. Each row represents a gene, and each column represents an epigenetic marker for a given bin (for example H3K27me3 from 0-5Kb or PolII in gene). Coverage ratios range from 0-1 with black squares representing bins with a ratio of 0 and red squares representing a ratio of 1. Top panel: Clustering of upregulated genes. Inset is enlarged portion of the heat map, and the epigenetic markers included in this region correspond to H3K27me3 at all distances and within the gene. Gene names are listed to the right. Highlighted in red are genes that are involved in neural development and function. Bottom panel: Clustering of downregulated genes. Yellow and blue ovals isolate 2 distinct clusters corresponding to genes with SYT-SSX2/H3K27me3 overlapping regions and SYT-SSX2 peaks overlapping with modifications associated with transcriptional activation, respectively. The position and identity of the epigenetic marker(s) comprising a particular cluster are listed to the right.

Figure 5

Models of SYT-SSX2 recruitment and activity. A)Model of recruitment and activity at an upregulated gene. SYT-SSX2 is recruited primarily to silent genes involved in the specification of other lineages by interactions with Polycomb Repressive Complexes (PRC1, dark blue oval; PRC2, light blue oval). In the presence of certain lineage-specific transcription factors (LSTF, red oval) and/or by interaction with lineage-specific enhancers (Enhancer, orange rounded rectangle), SYT-SSX2 association with the gene results in transcriptional activation. B) Model of recruitment and activity at a downregulated gene. Top) SYT-SSX2 may be recruited by interactions with Polycomb Repressive Complexes at a distance from the gene TSS. The gene itself is marked by activating histone modifications (Active, pink circle), and due to the presence of other regulators (?, black rounded rectangle), SYT-SSX2 carries out transcriptional silencing. Bottom) SYT-SSX2 may be directly recruited to the proximal regulatory region through association with activating histone modifications or other unidentified factors. These other factors also provide a signal to SYT-SSX2 such that it mediates repression.