Accelerating the search for new glioblastoma treatments

As a young boy, Nader Sanai watched on as his aunt died from a glioblastoma brain tumour. She was in her late twenties. It left his family devastated.

“It’s a terrifying disease,” says Sanai, now a neurosurgeon based in Phoenix, Arizona, where he runs the Ivy Brain Tumour Institute. “It’s not just about losing the ability to do certain things like walk or move, or communicate. You lose the very elements of yourself – your own self-identity. It’s your whole thought process that gets degraded.”

His aunt’s ordeal and death left a lifelong impression on him, motivating him to go to medical school and ultimately specialise in neurosurgery, to try to make a difference with people facing the same fate. But progress in the field has been limited.

“What’s just completely absurd to me is that my aunt’s experience 40 years ago is almost identical to what my patients go through today,” he says. “I feel embarrassed, quite honestly, to acknowledge that because I’m part of this field, and I feel a sense of responsibility to change that.”

There have, Sanai acknowledges, been improvements in surgery and radiotherapy techniques. “These fields have become really quite well developed. But there have been few to any bright spots in terms of new medical therapies.”

This lack of progress in drug development stands in stark contrast to other cancer types. For example, between 2000 and 2023, the US FDA approved 26 new breast cancer treatments. For glioblastoma, just two new drugs – temozolomide and bevacizumab – have been licensed over the same period.

Overcoming nihilism

Underpinning this slow progress, argues Richard Mair, a consultant neurosurgeon at Addenbrookes Hospital in Cambridge who heads up a research group at the Cancer Research UK Cambridge Institute, is a series of overlapping ‘nihilisms’.

Then, Mair says, there’s ‘pharmaceutical’ nihilism – the reluctance of drug manufacturers to trial existing drugs for brain cancer patients, on the grounds that they’re exceptionally fragile.

“Of course, brain tumour patients are very, very delicate – if you have a bleed from the tumour, or swelling, all of that can cause incredibly serious problems for the patient –more so than for other tumour types,” says Mair. “And if they get lots of complications, that can cause problems with taking the drug forward commercially. So again, these are things that have been historically used to exclude brain cancer patients from trials.

“And then there’s financial nihilism, the idea that there aren’t enough patients with brain cancer to make a business model to get a drug to market,” he says.

Nader Sanai agrees with this assessment. “When your sample size is small, you have limited bites of the apple. In a larger field like breast cancer, for example, a 1,000-person trial is something that can be completed within months or years. And if it’s negative, you can quickly go on to the next one.

“In brain cancer, a 600-patient trial may take half a decade from start to finish,” he says. “And if it’s negative, the setback can mean there’s much less appetite to go for the next one.

Ultimately, says Sanai, for pharma companies, there’s limited return on investment in going for less common, hard-to-treat cancers like glioblastoma, versus a lot of risk. “So, as a field, we have to do better in trying to de-risk that process for industry.”

Challenging biology

But ‘de-risking’ drug development first means overcoming some even more fundamental challenges – namely the incredible complexity of the brain, and of tumours that grow within it.

“But while I wouldn’t want to understate the challenge, there’s really been an incredible amount of effort to understand the biology of glioblastoma. I don’t think it’s an exaggeration to say that it is actually one of the best analysed cancers in humans.

“So, I wouldn’t say that the lack of clinical advance is a consequence of the inadequacy of the basic science around the disease,” says Sanai. “What we have is a translation problem.”

At the root of this, is a lack of clinically relevant models to generate data that advances drug discovery, says Juanita Lopez, a Consultant Medical Oncologist at The Royal Marsden NHS Foundation Trust and Group Leader in Early Phase Drug Development at The Institute of Cancer Research, and for whom developing new glioblastoma therapeutics is a current focus.

“The models that we have had, up until maybe five or so years ago, didn’t accurately clinically recapitulate the biology of human brain tumours,” says Lopez.

“We are starting to make progress. We now have a few syngenic, clinically relevant mouse models, but we don’t still have the richness of, say, breast or lung cancer, where there are hundreds of relevant models.”

This, Lopez says, allows researchers in these fields to rapidly cross-validate pre-clinical findings and find reliably targetable biology – something that the brain tumour community has struggled to do.

“But the second issue is around toxicology – particularly in understanding the biodistribution and pharmacokinetics of a given drug. The problem is most toxicology is done in animals with intact blood-brain barriers” she says.

This has led to the conclusion that most drugs aren’t brain-penetrant. But that, Lopez says, may not be valid when a tumour is present.

“Brain tumour patients have a really deregulated blood-brain barrier, meaning all sorts of molecules can get across it,” and that includes complex molecules like antibodies, she says.

In fact, real-world data from patients with metastatic disease in their brain shows that many drugs can, and do, target these tumours.

“There’s now a long and growing list of drugs – chemotherapies like temozolomide, targeted agents like trametinib and dabrafenib, even large complex molecules like trastuzumab deruxtecan for Her2-driven cancers – that have shown efficacy against brain metastases in patients with a range of advanced cancers,” says Lopez.

‘Cancel culture’

But for patients with primary brain tumours, when it comes to early phase trials, there’s been what Lopez calls a form of ‘cancel culture’.

“Over the last 10 years, I’ve been involved in more than 300 first-in-human trials, many of which have enrolled patients with a fairly broad range of cancer types. And even when there’s strong preclinical data in, say, melanoma or bowel cancer, during a phase 1 clinical trial in which patients with other histologies are enrolled, we’d often see an exceptional response in a different tumour type. And in several cases, this prompted the manufacturer to alter the path of that therapeutic,” says Lopez.

“We need to find ways to give brain tumour patients the opportunity to take part, not least so that we can learn from them. If we’ve got all the information, in terms of their biopsies, cells and cultures, so let’s just allow them access to experimental therapy trials, and investigate and document their response,” she says. “Let’s let these patients be their own disease model.”

Turnaround times

Several years ago, at Addenbrookes Hospital in Cambridge, consultant neurosurgeon Richard Mair had been trying to answer a very different question: how best to deploy new technologies like next-generation genome and transcriptome sequencing to improve how patients with brain cancer are managed.

“People were asking, ‘what’s the point of doing this? It’s expensive. If we can set all this up, what’s the point?’,” he remembers.

“And we reasoned that there would be two main reasons: to help with diagnosis, and to help predict optimum treatment.”

The problem was, at the time, programmes aiming to answer this were taking around nine months to return results to clinicians, “by which point, many of the patients had died,” recalls Mair.

To try to speed things up, with funding from the Minderoo Foundation, in 2021 Mair set up the Precision Brain Tumour Programme, with the aim of assessing the feasibility and utility of whole genome and transcriptome sequencing for patients with brain cancer.

Initially building on existing biobank infrastructure and setting up a tissue processing pathway to enable the genome sequencing, Mair’s team followed by establishing an analysis pipeline with the genomic laboratory hub at the hospital. “And that meant we could get the genomes extracted, sequenced, returned, analysed and fed back to clinicians within twenty days,” says Mair – an advance that allowed them to begin to answer the question of clinical utility.

“From a diagnostic perspective, with over around 330 patients sequenced so far, we found the sequencing didn’t really help that much,” he says. “We had a couple of cases where it helped pin down the tumour type, and we had a couple of cases where it altered the initial diagnosis from low to high grade. But overall, sequencing didn’t seem to be useful for diagnosis.”

“That left therapy. But to answer that question, we hit another issue,” says Mair. By now, the Precision Brain Tumour Programme had enabled Mair to establish a UK-wide group of clinicians who’d become comfortable looking at genetic data on their patients and discussing its implications. “And it became very clear was that there wasn’t a good way of referring our patients into clinical trials,” he says.

But that began to change when Mair was introduced to Lopez by a mutual acquaintance, and the pair began discussing how to kick-start progress.

Getting Going against Glioblastoma

“I rang Juanita, and it turned out she was interested in seeing whether it was possible to trial existing targeted medicines in patients with known biomarkers in real time. And that’s what drew us together,” Mair recalls. “She was very interested in targeting drugs and I had a way of getting those targets in a clinically actionable time. It’s a great synergistic: me with my molecular hat, her with her clinical trials hat.”

With their complementary skillsets and “shared language” of understanding the patients and the biological challenges, Mair and Lopez were clear on the challenge they needed to solve.

Ultimately, this led them to bid, successfully, for £3m funding from Minderoo and Cancer Research UK, in order to launch a first-of-its-kind platform trial for UK patients with glioblastoma: the next-Generation aGile Genomically Guided Glioma platform trial, or ‘5G’ for short.

“It’s a precision adaptive platform trial,” says Mair. “We’re using specific drug/biomarker pairings for each arm. And this is something that hasn’t been done traditionally in brain cancer. Instead, we’ve tended to give precision drugs to an untargeted population and hope that the stats would mean we’d see something.”

“It’s ‘adaptive’ because if we see signal early on, we can adjust the subsequent enrolment to recruit patients more likely to benefit to make sure we get a positive or negative signal that’s appropriate,” he says.

What’s particularly exciting, says Lopez, are three unique innovations baked into the trial design. “The first is that it allows for iteration. So, if, for example, we see patients on a given arm, who have both mutation A and B, never respond, but patients with A, but without B, do, it means we can recruit another cohort specifying A minus B. And we don’t have to go back to the regulators for approval for that new arm, it’s all part of the protocol. Nobody’s ever done that before,” she says.

The second allows the trial much greater flexibility with drug combinations. “For example, let’s say we see that the non-responders all have an additional aberration, and that’s targetable – say for example EGFR and NF1 – the trial allows us to try and hit that aberration too, just by opening a new sub-protocol.”

The third adaptation allows 5G to move promising-looking drug strategies in the frontline setting – right after a patient has had initial surgery and radiotherapy – rather than waiting for people to relapse. “We know that these tumours evolve over time – so why not try targeting them when they’re hopefully less heterogeneous and less evolved, and see if we can improve survival,” says Lopez.

To date, 5G has opened three separate arms. RUBY will test two drugs in combination that have recently shown efficacy in ovarian cancer: dual Raf/MEK inhibitor avutometinib alongside dafectinib, which targets FAK.

EMERALD will investigate amivantamab a bispecific antibody that targets both EGFR and MET which has already shown promise in lung cancer.

A third arm is planned which will look at a PI3 kinase inhibitor.

This arm highlights another strength of 5G – its design should allow insights into why a given treatment does or doesn’t work, an area where other trials have frequently fallen short. In the case of a PI3 kinase inhibitor, late last year a large US platform trial, GBM AGILE, tested one such drug – paxalisib – in patients with glioblastoma, but failed to show an overall survival benefit.

But crucially, points out Mair, GBM AGILE didn’t stratify patients by biomarker status, nor carry out the sort of deep biological analysis of patient samples planned on 5G. “The drug seemed to work initially for some, but then stopped working. But we don’t know why. As a community, we’re very good at running trials that say, this works or it doesn’t, but we’re not very good at finding out the ‘why’.”

Lopez points out another feature of the 5G trial: rather than being ‘first-in-human’, it’s ‘first-in-brain’ – these are existing therapies that have already been proven broadly safe, and to hit their targets in other cancer types. 5G sees their first use against primary brain tumours in patients with appropriate mutations for the first time.

The first tranche of preliminary results are expected later in 2025, says Mair. “The unfortunate thing with brain cancer is, you know pretty quickly whether your drug’s going to work,” he says.

A new hope

“I think the message I want to get out there, to the clinical community in the UK, is – please consider referring your patients for trials like 5G,” says Mair. “We need clinicians to empower their patients to be part of clinical trials in brain cancer.”

Ultimately, he says, this is all about hope – something that’s in short supply among people with brain tumours. “There’s a lot of charlatans and snake oil salesmen out there,” he says. “And we need to remove these vulnerable patients from their clutches and give them a viable alternative that doesn’t cost their life savings. They deserve a better chance of finding an alternative to standard of care, which we know will ultimately fail them.”

“One of the things the late Tessa Jowell said is that we need to enable and empower patients to make decisions on their care”, says Mair. “And that’s a real key driving force for what we’re trying to do.”

Over in Arizona, where his own brain tumour drug development work focuses on pharmacokinetic- and pharmacodynamic-driven early-phase clinical trials, Nader Sanai is adamant that it’s imperative that trials like 5G, and others around the world trying to bring forward new brain cancer medicines, are properly resourced and prioritised.

“Every day I see what my patients are going through, and it’s a reminder to me of what I saw when I was a young boy,” he says. “I think about it a lot. It’s why I’m here”.