When broken strands of DNA get repaired improperly, they can stitch together in new “Frankengene” fusions that can cause cancer.
Researchers from two labs in the Human Biology division at Fred Hutch Cancer Center are working together to better understand how one of these notorious fusions — ZFTA-RELA — drives rare brain tumors in children called ependymomas.
The work, recently published in the journals PNAS and Neuro-Oncology, maps the deep biology of ependymoma tumors, mirrors that biology in a mouse model and reveals a new molecular vulnerability in ZFTA-RELA fusions that could be targeted with drugs.
The discovery is the result of a longstanding collaboration between the labs of rare cancer expert Taran Gujral, PhD, and brain cancer researcher Eric Holland, MD, PhD, who directs the Human Biology Division and holds the Pigott Family Endowed Chair.
Their collaboration showcases innovative methods developed at Fred Hutch that are well suited to overcoming the logistical challenges of studying rare cancers, which typically don’t attract investment from big pharmaceutical companies.
Ependymomas comprise about 10% of intracranial malignant tumors in children with 30% of cases diagnosed before the age of 3.
Though there are many subtypes of this tumor, they’re all usually treated with surgery and radiotherapy with minimal benefit from chemotherapy.
There’s no specific treatment for a rare, but particularly lethal variety of these tumors that harbor a Frankengene called ZFTA-RELA, which sounds even more intimidating spelled out: Zinc Finger Translocation Associated – RELA Proto-Oncogene.
That’s a common problem for research in rare cancers because there are fewer patient tumor samples to study and fewer genetically engineered preclinical models to study disease progression and response to treatment in living organisms.
Gujral, Holland and their colleagues overcame those hurdles in three steps, starting with an approach pioneered in the Holland Lab that classifies tumors based on their underlying biology rather than their appearance under a microscope.
Holland and Sonali Arora, MS, who runs the computational biology section of his lab, integrated gene expression data from more than 1,200 tumor samples gathered from North America and Europe of two pediatric brain cancers — medulloblastoma and ependymoma.
Using computational tools invented at Fred Hutch, Holland’s team simplified that information, which comprises millions of data points, and represented it graphically on a digital reference map.
But such a map is only useful if it can make fine-grained distinctions between molecular subtypes of tumors. To draw significant contrasts Holland suggested adding data from medulloblastomas so that they were comparing two diseases instead of one.
“That was [Holland’s] brilliant idea because then we could have that contrasting factor,” Arora said. “You want to have stark differences across our map.”
She and Holland and their colleagues analyzed gene expression data for the tumor samples collected in publicly available datasets. That data represented average gene expression across large groups of cells using a method called bulk RNA sequencing.
Another method — single-cell RNA sequencing — analyzes gene expression from individual cells that each receive a unique barcode to track their activity. That method reveals a finer grain of molecular detail than bulk RNA sequencing, but it takes a lot more time, money and complex analysis.
Holland and Arora’s approach showed that bulk RNA sequencing can produce quicker, cheaper, more clinically relevant results on a larger scale without sacrificing the detail needed to trace signaling pathways, biomarkers of disease and potential drug targets.
They performed single-cell RNA sequencing data from 25 individual tumor samples to confirm that they were capturing the deep tumor biology they needed to make a meaningful reference map using the bulk RNA method.
Federal Funding Helps Us Save Lives
At Fred Hutch, 70% of our research funding comes from federal grants, which are awarded based on scientific merit. This level of federal funding is a direct reflection of our proven ability to make landmark breakthroughs and our trusted ability to lead large scientific collaborations.
Their analysis of 370 ependymoma tumor samples across five independent studies revealed a subtype of tumor loaded with the ZFTA-RELA fusion called EPN-E1, which arises from ependymal cells that line the fluid-filled spaces of the brain and spinal cord.
More than 70% of samples harbor ZFTA-RELA fusions, making the subtype an attractive hunting ground for either new drugs or repurposed drugs.
But first the researchers needed to mirror that subtype in mice to discover exactly how ZFTA-RELA fusions drive tumor growth and whether they could find any weak links in those signaling chains that could be targeted with a drug.
The researchers used a system that delivered the ZFTA-RELA fusion into specific cells in genetically engineered mice from Holland’s lab to study how the tumors grew.
They used a harmless engineered virus that travels only to brain cells that haven’t specialized yet, delivering the ZFTA-RELA fusion. The resulting tumors in the mouse model accurately reflected the human disease.
The next step was to uncover the key molecular signaling pathways driving growth in ZFTA-RELA tumors, which are orchestrated by enzymes called kinases that regulate many cellular processes, including growth and proliferation.
Drugs that block kinases and interrupt those pathways fueling tumor growth have transformed cancer therapy in the last quarter century.
The first kinase inhibitor — imatinib — was approved in 2001 to attack another Frankengene, the BCR-ABL fusion that plays a role in chronic myeloid leukemia. The 100th kinase inhibitor was approved late last year.
That rapid expansion of FDA-approved kinase inhibitors has generated a mountain of molecular data about the drugs and the kinases they block, which Gujral’s lab leverages to not only discover new uses for approved drugs but also to discover new cancer biology, which is especially important in rare cancers.
Kinase inhibitors typically target one kinase, but they often block additional kinases, creating overlapping patterns of inhibition that can be untangled with computer models.
To find the kinases that matter in ZFTA-RELA tumors, Gujral’s team tested 37 kinase inhibitors on the tumor cells and observed how the cells responded.
They used a machine-learning technique developed in Gujral’s lab to train a computer model with the results of those tests on the molecular data of 298 kinases to predict which inhibitors would likely work and then confirmed the model’s accuracy on individual samples.
They zeroed in on a dozen kinases that appeared to be especially important for tumor cell growth and one of them stood out as particularly promising as a potential drug target.
MERTK is a kinase known to regulate signaling pathways that enhance cell survival and prevent cells’ normal “self-destruct” procedures when they’re damaged or no longer needed, but its role in tumor cells isn’t well known.
Subsequent experiments in the mouse model confirmed MERTK is uniquely required for the survival and proliferation of ZFTA-RELA fusion ependymoma cells.
Once the team established that MERTK signaling is active and contributes to pro-survival pathways, they tested a clinical-grade MERTK inhibitor capable of crossing the blood-brain barrier, a semi-permeable membrane that protects the brain from toxins and pathogens carried by circulating blood.
The compound, which is not yet FDA-approved but has passed safety trials, reduced expression of both MERTK and the ZFTA-RELA fusion, significantly reducing tumor growth in mice.
The researchers speculate that a drug targeting that vulnerability in humans could not only kill cancer cells, but activity in the microenvironment surrounding the tumor that also promotes growth.
Gujral said the collaboration with Holland, which unites deep tumor biology mapping with machine-learning amplified kinase testing, advances rare cancer research by making the most of scarce resources.
“Over the past decade, we have co-authored several papers together and collaborated on research grants,” Gujral said. “We have students and postdoctoral researchers working across both groups, which further reinforces the collaborative spirit.”
This work was supported by philanthropy, the National Institutes of Health and a grant from The Ben and Catherine Ivy Foundation.
John Higgins, a staff writer at Fred Hutch Cancer Center, was an education reporter at The Seattle Times and the Akron Beacon Journal. He was a Knight Science Journalism Fellow at MIT, where he studied the emerging science of teaching. Reach him at jhiggin2@fredhutch.org or @jhigginswriter.bsky.social.
Are you interested in reprinting or republishing this story? Be our guest! We want to help connect people with the information they need. We just ask that you link back to the original article, preserve the author’s byline and refrain from making edits that alter the original context. Questions? Email us at communications@fredhutch.org
Every dollar counts. Please support lifesaving research today.
For the Media