The power of partnership – the unique alliance taking on brain cancer

The course of Richard Gilbertson’s career was set in motion roughly thirty years ago. During his second week as a medical student, Gilbertson witnessed a young girl die of a brain tumour. And ever since, he has strived to understand the disease.

At first, Gilbertson split his time between clinical practice and basic research. But in 2000 – moving from the UK to St. Jude Children’s Research Hospital in Memphis – he dedicated himself entirely to research. This, he says, was how he thought he could have the greatest impact on the lives of children with cancer.

Gilbertson wanted to better define the genetics of brain tumours. Initially, his work focused on the cancer that had killed that girl – medulloblastoma, the second most common paediatric brain tumour. But later, he started to also explore the third most common type: ependymomas. He has been instrumental in uncovering the multiplicity of existing medulloblastoma and ependymoma subtypes.

Professor Richard Gilbertson is a paediatric oncology clinician scientist and Director of the Cancer Research UK Cambridge Centre.

“We recognised that ependymoma were born in stem cells that line the ventricles of the brain,” he says. “That was the first time that that had been shown. And so, we got interested in ependymoma – what causes it, where it comes from.”

That quest took a major step forward in 2014, when Gilbertson’s team at St. Jude described a mutation that was present in – and likely caused – around two-thirds of ependymomas occurring in the front most part of the brain.

The mutation resulted from chromosome 11 splintering, then re-annealing in such a way that two genes were thrust together to generate a unique, carcinogenic fusion protein.

Now – having moved back to the UK to become director of the CRUK Cancer Centre in Cambridge – Gilbertson hopes he is on the verge of introducing a brand-new diagnostic tool: an antibody that binds this fusion protein.

If it passes through its final stages of development, the antibody could transform the clinical care of children and adults diagnosed with ependymomas. In addition, being able to easily distinguish fusion protein-negative and -positive ependymomas will facilitate clinical trials in patients with precisely defined subtypes of this cancer.

“This project happened, because we had the opportunity to set up the CRUK Children’s Brain Tumour Centre of Excellence,” Gilbertson says, “where we were always thinking about the diagnostic side.”

But when it came to the concrete process of actually making the antibody, it was another CRUK venture that took centre stage: The Antibody Alliance Laboratory.

The AAL – which opened in September 2015, shortly after Gilbertson’s move to Cambridge – was born of a unique collaboration between CRUK and AstraZeneca. Staffed jointly by CRUK- and AstraZeneca-scientists, the AAL was founded to run collaborative projects between academics and bioengineers who had at their disposal the full might of AstraZeneca’s antibody discovery platform.

“The academic community hold a vast knowledge around oncology biology and have novel insights into cancer targets,” says Maria Groves, the lab’s director. “As part of the AAL collaboration, they propose ideas for novel cancer targets and if that project is accepted into the Alliance, we work with them very closely.”

Groves recalls that when Gilbertson proposed targeting the ependymoma fusion protein, it wasn’t just the robust scientific rationale for doing so that stood out. “One of the things that was really noted,” she says, “was he was incredibly passionate about delivering this putative diagnostic.”

An in-fusion of data

Ependymomas are first classified according to where in the central nervous system they occur – the spinal cord, the rear of the brain (posterior fossa) or the frontmost part of the brain (supratentorial). While ependymomas affect both children and adults, the most aggressive ones develop most frequently in children under the age of ten.

Despite being the third most common paediatric brain cancer, they remain relatively rare – about 50 cases are diagnosed annually in the UK and roughly 185 per year in the US. Such numbers make it difficult to run clinical trials deter pharma from investing heavily in the disease; and make gaining sufficient materials for doing research challenging.

“Gilbertson and his colleagues had found a gene that distinguished one class of supratentorial ependymomas from another”

No effective chemotherapy regimen has been developed for ependymoma. Instead, the current standard of care is maximal resection followed by radiotherapy. The survival rate following such treatment is around 60%. This number does vary according to where the tumour is growing – and successful treatment typically comes at a high cost.

The course of radiation used is so intensive, it’s referred to as ‘devastating radiotherapy’.  “Our conventional radiotherapy is very damaging for a child,” says Gilbertson, “they don’t form long-term relationships; they don’t graduate from high school; they don’t have long-term jobs; and they often remain dependent.”

Clearly, there is a need for new treatment options.

Gilbertson’s 2014 study built on work showing that, despite their shared histology, spinal, posterior fossa and supratentorial ependymomas had different gene expression profiles and different mutational loads.

To dig deeper into this heterogeneity, his group examined dozens of supratentorial and posterior fossa tumours using whole genome sequencing and RNA sequencing. The standout finding was that in all nine supratentorial tumours that underwent whole genome sequencing, a region of chromosome 11 had undergone a catastrophic rearrangement. And in 8 of 9, this had made a little-known gene now termed ZFTA abut the well-characterised gene RELA.

RNA-seq revealed a transcript – formed by splicing – that encoded a putative fusion protein. And in a larger sample, this transcript was absent from posterior fossa tumours and present in around two-thirds of supratentorial tumours.

When Gilbertson’s group expressed the ZFTA-RELA fusion protein in neural stem cells, they became malignant. “You can make tumours using the fusion with nothing else, which is pretty remarkable,” says, Gilbertson. “Typically, the dogma is you need three or four mutations to make a cancer. In fact, you don’t.”

These insights will inform novel strategies for tackling tumours that contain the ZFTA fusion protein. But back in 2014 what mattered was that Gilbertson and his colleagues had found a gene that distinguished one class of supratentorial ependymomas from another.

Gilbertson’s data suggested that these two types are caused by different mechanisms. But soon data from another group – with access to a larger number of patient samples – upped the ante by showing that the people whose cancers contained the fusion protein fared considerably worse than those whose cancers did not. People with fusion protein-negative tumours have about an 80% chance of surviving – people with positive cancers survive at a rate of around 20%.

Finding a reliable way to distinguish the two tumour types immediately felt urgent. “First of all,” says Gilbertson, “you could tell patients who are fusion negative that they have maybe an 80% chance of being cured. At the moment, we can’t do that, because we don’t know which patient they are until we’ve done the genomic testing.”

“This is a story of discovery and collaboration, with two different groups coming together that had completely different skills and knowledge to enable the discovery of something that has the potential to be very valuable.”

But as well as a diagnostic, Gilbertson says, distinguishing patients by tumour type will allow the respective groups to partake in targeted clinical trials.

In particular, those with the fusion protein should be prioritised for tests of novel medicines. “We’ve done a very large, high-throughput drug screen of about 1.2 million compounds – beginning at St. Jude, and now working with the Institute of Cancer Research through the CRUK Children’s Brain Tumour Centre of Excellence – with the aim of making completely new therapeutics that target this disease,” says Gilbertson. “These are the patients who we should be testing this in.”

Towards a diagnostic

For a diagnostic to have real impact, thinks Gilbertson, it must be reliable, straightforward to use and easily accessible.

This, he says, ruled out relying on either whole genome sequencing or RNA sequencing, because “that automatically restricts that diagnostic capability to labs that currently have the ability to do that – which are a minority of labs around the world.” The same applies to more recent, somewhat simpler, methods for staining DNA methylation.

Therefore, Gilbertson decided the best tactic was to use the standard pathology technique of immunohistochemistry. That required developing an effective antibody – and thankfully as Gilbertson was deciding on this, CRUK and AstraZeneca were planning to open the AAL.

Gilbertson learnt about this in 2015 when he met representatives of Cancer Research Horizons Together, they agreed that he should apply to work with the AAL. In January 2016, his proposal was greenlit, becoming one the lab’s first projects.

The cornerstone of the AAL’s technical capabilities is AstraZeneca’s proprietary phage display library, which contains around 100 billion fragments of human antibodies that can be screened for binding to a recombinant target protein.

Dr Maria Grove, Head of CRUK-AstraZeneca Antibody Alliance Laboratory.

A fusion protein was, however, a special challenge. “This was a unique fusion protein,” says Groves. “But the un-fused proteins are also present within the cells, within the nucleus, so we wouldn’t want our antibodies to cross react with the two proteins involved in the fusion.”

This meant that the team had to target the junction between RELA and ZFTA, the only exclusive sequence in the whole fusion protein. They managed, and having done just that, a panel of tens of antibodies was given to Gilbertson’s lab for the next stage of testing.

“They found some antibodies that looked quite promising. But the staining was quite weak,” says Groves. Uncertain whether it was best to look for more candidates or to work on improving the performance of these initial hits, the team decided on the latter, taking the two best and working to better them.

“That’s when they got this fantastic data,” says Groves. “When Richard describes first seeing the results, he says he nearly fell off his chair! There was really strong nuclear staining with one particular antibody.”

“The AAL had refined our initial ‘blush’ of staining on tumour sections to absolutely striking and clear binding in sections of tumours that expressed the fusion protein,” Gilbertson says. “We knew then we were in with a chance for this to work.”

Groves says: “This is a story of discovery and collaboration, with two different groups coming together that had completely different skills and knowledge to enable the discovery of something that has the potential to be very valuable.”

The next phase

That confidence led to two patents being filed on the antibody, in 2019 and 2020. Now, the final hurdle is to confirm that the antibody works as well in human brain sections as it does in mouse.

This relatively straightforward experiment is complicated by the rarity of supratentorial ependymomas and Gilbertson has had to seek tissue samples from international collaborators to complete the work. “Although ependymomas are the third commonest brain tumour in children, they are still rare,” says Gilbertson. “Also, the fusion is only seen in tumours that occur in the front of the brain. Most of the ones in children arise at the back. So, its rarest of rare. Finding enough tissue from those cases that we can test antibody on very precious samples is very hard.”

It is exactly this – rarity – which makes the development of this diagnostic antibody an excellent match for CRUK’s mission to meet unmet clinical needs even when they only affect a small patient population.

“CRUK is a very special organisation,” says Gilbertson. “You can pick up the phone and talk to the CEO of CRUK. It gives you this sense that you’re partnering – and they’re receptive and they’re responsive. It makes them much more agile in the research space.”

Nothing is certain for ependymoma patients – all this will depend on careful trials and clinical exploration. But Gilbertson points to the evolution of medulloblastoma treatment as a cause for optimism. “The thing that brings me to tears as a paediatric oncologist is that over the last ten years, St. Jude has backed off on radiotherapy, and all those patients are disease-free. They’re all cured, and they will have normal lives.”

Author:
Liam Drew is a writer and journalist covering biology and medicine. In 2020, he received the Association of British Science Writers’ Award for Best Engineering and Technology Reporting.