A healthy cell doesn’t turn cancerous for no reason. Somewhere deep within the recesses of the nucleus where the cell copies and transcribes DNA, something goes wrong. Whether from exposure to a carcinogen or a simple copying error, one nucleotide can get swapped for another. Sometimes whole chromosomes break into pieces and fuse together. The resulting DNA damage typically leads to cancer.

Ependymomas, a class of rare and aggressive childhood brain cancers, don’t seem to have a clear genetic cause. Some ependymoma subtypes such as supratentorial RELA-fusion associated group (RELA) ependymomas form because of a chromosomal break, but many others, including the rare posterior fossa ependymoma group A (PFA), have no known genetic cause.

“We have sequenced dozens and dozens of kids, and we have not found a single mutation that could potentially explain why these [PFA] tumors happen,” said Lukas Chavez, a pediatric brain cancer researcher at the University of California, San Diego.

Using a combination of deep sequencing approaches and three-dimensional genomics, Chavez and an international team of scientists identified new disease mechanisms and drug targets in RELA and PFA ependymomas (1). The findings, which the researchers reported in a preprint, will not only lead to new treatments for these aggressive cancers, but also underscore the value of understanding the physical arrangement of the genome in cancer.

Because Chavez and his team had not found any causative mutations of ependymomas, they investigated whether more global changes to the genome could cause these cancers. For example, many PFA tumors have lost the repressive epigenetic mark, H3 lysine 27 trimethylation, leading to improper gene expression, but this is not the case for all PFA tumors (2).

Besides epigenetic marks, how DNA folds in on itself inside the nucleus can also control gene expression. Two sections of DNA that may be located far apart on a linear piece of DNA, such as a gene promoter and the gene it regulates, will actually fold onto each other within the nucleus. This three-dimensional arrangement of the DNA is how the promoter regulates the expression of that gene. If a chromosome breaks and then fuses with a different piece of DNA, the gene and its promoter will no longer be in contact, breaking that regulation, which could lead to cancer.

“You can imagine that where they fuse together, those genes that live in that three-dimensional neighborhood are now exposed to regulatory elements from the other chromosome that fused,” explained Anthony Schmitt, an author of the preprint and a scientific leader at Arima Genomics. “It's like picking up your house and dropping it into a new neighborhood. Now, you're going to have tons of new neighbors that can influence your behavior.”

Chavez and his team hypothesized that these ependymoma subtypes result from three-dimensional changes to their genomes. Working with an international group of scientists from Germany and Canada, Chavez and his team collected 14 RELA and PFA ependymoma patient samples, a difficult task because these ependymoma subtypes are so rare. For example, RELA and PFA are just a subset of the ~200 ependymoma cases diagnosed in children in the United States every year.

To determine whether there were any structural variants in the genomes of RELA or PFA samples compared to healthy cells, the researchers performed deep sequencing and chromosome conformation capture (Hi-C) profiling, which allowed them to identify which sections of DNA were close together in the three-dimensional nucleus and compare it with changes in gene expression across the whole genome.

Chavez and his team identified known epigenetic changes in these ependymoma subtypes and more than twice as many genes with altered regulation than they knew about before performing Hi-C profiling.

“It was exciting to see that the 3D genomes of these ependymomas are clearly special,” said Chavez. “We were surprised to find quite a complexity of structural rearrangements that these tumors have that were not seen previously by standard DNA sequence analysis… It really gave us a much better global view on the complex rearrangements that we have seen in some of these tumors.”

In the RELA tumors, the researchers noticed that a chromosomal structural rearrangement put the gene REST Corepressor 2 (RCOR2) in a new genomic neighborhood. Their gene expression data revealed that RCOR2 is overexpressed in RELA tumors. When the researchers knocked down expression of RCOR2 in patient-derived RELA cells, the cells died. While there is no compound that targets the RCOR2 protein directly, RCOR2 exists in a complex with other proteins including histone deacetylase 1 and 2 (HDAC1/2). When the team treated RELA cancer cells with the HDAC inhibitor Entinostat, RELA cell growth slowed, highlighting a potential new therapeutic for RELA ependymomas.

When the researchers assessed the PFA tumors, they noticed a unique three-dimensional DNA cluster that linked genes and regulatory sequences in PFA cells. In this cluster, the gene integrin α6 (ITGA6), a cell surface receptor for the protein laminin, had increased expression in PFA cells compared to RELA cells or healthy brain tissue. When the team knocked out ITGA6 in PFA patient-derived cell lines, tumor cell proliferation decreased, leading the researchers to identify ITGA6 as a new drug target for PFA ependymoma.

The team also identified a chromosomal break in the laminin subunit γ1 (LAMC1) gene in PFA ependymomas, which put the gene in a new regulatory environment. Knocking out LAMC1 in PFA and RELA cells led to reduced growth in PFA cells but not RELA cells.

“They’ve discovered these new genes that will probably drive insight into how these tumors grow and how they behave, so I think that's really exciting,” said Sriram Venneti, a pediatric brain cancer and epigenetics researcher at the University of Michigan School of Medicine who was not involved in the study. “It strengthens the hypothesis that the 3D genome is a major driver of cancer in general.”

Chavez is excited to test whether targeting the genes they identified can slow tumor growth in animal models, and eventually they would like to see if the results replicate in humans. Like Venneti, he too would like to see three-dimensional genomics being used in the clinic to reveal the molecular mechanisms of tumors that have no clear genetic cause.

“What would be very promising is to apply this technology in a routine way to all clinical samples that we see in the hospital,” said Chavez. “That can actually have an impact on how patients might be treated in the future.”

References

1. Okonechnikov, K. et al. Oncogenic 3D genome conformations identify novel therapeutic targets in ependymoma. Preprint at: https://www.researchsquare.com/article/rs-88331/v1 
2. Panwalkar, P. et al. Immunohistochemical analysis of H3K27me3 demonstrates global reduction in group-A childhood posterior fossa ependymoma and is a powerful predictor of outcome. Acta Neuropathol  134, 705-714 (2017).