Even for cancer biologists rooted in fundamental research, having a view of how your work might one day benefit patients is often ingrained. Sometimes the path from question to therapy is clear early on; more often it takes time for the right scientific and technical pieces to align. For University of Edinburgh researcher Steve Pollard, more than a decade of fundamental work ultimately converged to inform an emerging therapeutic strategy.

Trained in developmental neurobiology, with a passion for stem cells and their transcription factors, Pollard began studying glioblastomas twenty years ago after seeing that genetic programs used to build brains during development are re-awakened in this devastating type of tumour. “You have a sort of foetal tissue state in the adult tissue,” he says, “which is part of the reason it doesn’t stop growing.”

Pursuing this idea, Pollard’s research thrived. He and his colleagues identified several transcription factors and gene switches that drive brain cells into an immature and aggressively cancerous neural stem cell-like state.

“But all the way through this, I’m not really thinking about therapeutics at all,” he says. “Because it’s not clear how you would do it. I’m working on these gene switches deep in the nucleus that are redundant; easy to compensate for; and almost impossible for drugging through classic approaches with small molecules.”

Then, in March 2017, Pollard attended a seminar by some colleagues.

He entered carrying some preconceived ideas about gene therapy. “The problem with gene therapy had always been delivery,” he says. The fear was that if you couldn’t infect all of a tumour, it would simply grow back – just as a glioblastoma does after a surgeon removes most of it. “That’s why I wouldn’t touch it with a barge pole,” Pollard says.

But then Pollard watched Edinburgh neuroscientists Sir Adrian Bird and Stuart Cobb describe their plans to treat the congenital neurological condition Rett Syndrome using an adeno-associated virus. This AAV, they said, could deliver DNA to correct the underlying genetic fault across the entire brain. A slide showing how many cells of a mouse’s brain it transformed blew Pollard away.

Hearing that AAVs were now a mature, clinic-ready technology – well, Pollard pictured using them to deliver therapeutic transgenes, which would instigate a cancer’s demise, into brain tumours. And he thought to himself, “Doing glioblastoma should be easy.”

Crucially, Pollard thought he also had a way to ensure these transgenes were only switched on in malignant cells and not in surrounding healthy brain cells that the AAV would also infect.

“Everyone – including me probably – thought that was a bit of a long shot,” Pollard says. “But it works really well. It works fantastically well.”

He christened the switches, Synthetic Super-Enhancers – or SSEs. They can be designed for any cell type, including diseased cell states, to strongly and selectively drive expression of any chosen gene.

Then in 2019, CRUK and its commercial arm (Cancer Research Horizons) contacted Pollard about potentially commercialising his research. Now, Trogenix – the Edinburgh-based biotech that Pollard co-founded – is planning a glioblastoma clinical trial.

They are testing a viral immunotherapy – which works brilliantly in mouse and human tissue models. It uses an AAV to deliver two transgenes. One kills infected cancer cells. The other awakens the immune system, helping train it to destroy remaining glioblastoma cells.

Based on the novel SSE platform, advances in AAVs and their delivery, and the maturing science of immunotherapy, Trogenix CEO, Ken Macnamara says, “This company couldn’t have existed five years ago.”

The name Trogenix, Macnamara explains, was inspired by the story of the Trojan horse. “There were only 30 soldiers in the Trojan horse,” he says. “It wasn’t that they killed everybody in the city. What they did was open the gates to the army. And we’re opening the gates to the immune system from within the cell.”

Researchers had always struggled to culture glioblastoma cells in vitro, creating a barrier to studying the disease. Pollard asked whether they might grow in conditions used to maintain neural stem cells. “That worked like a dream,” he says, “pretty much straight off the bat.”

These cultures offered Pollard new ways to study glioblastoma, and his mentor – stem cell biologist Austin Smith – supported him to pursue the disease, setting him up with collaborators and encouraging him to apply for funding.

“I was very, very lucky,” Pollard says. “I mean, I’ve been lucky my whole career – but very lucky at that time that Cancer Research UK supported me with a project grant, realising that these are good model systems, and Austin provided the freedom to explore this.”

Generating cell lines from dozens of glioblastoma patients allowed Pollard to figure out key transcription factors and gene switches that characterise these tumours.

Significantly, nearly all glioblastomas share this distinctive transcription factor combination regardless of the exact – often different between people – mutations that triggered the malignancy. This means that unlike a therapy predicated on a particular driver mutation – which would be applicable to only a minority of patients – an intervention utilising the transcription factor profile would be useful for essentially all glioblastoma patients.

Soon, Pollard had his own lab. This operated at UCL for three-and-half-years before he returned to Edinburgh in 2013, drawn partly by opportunities to work closely with local clinicians. There, CRUK again backed the work, awarding him a five-year senior research fellowship. “That was a big deal,” he says, “because with a five-year grant, you can really focus and double down on both the basic mechanisms, the models, and maybe even think about therapeutics.”

Pollard invested quickly and heavily in the then new technology of CRISPR. This was great for genetically manipulating cell lines. But it also enabled his lab to address something else that had long held the glioblastoma field back: the lack of clinically relevant mouse models.

Introducing a combination of mutations found in human glioblastomas into murine brain stem cells yielded cells that aggressively formed tumours when transplanted into mice’s brains.

One of the things that makes the human disease so deadly – half of people diagnosed with glioblastoma die within 12-18 months, and over 90% within five years – is that the cancers very powerfully evade immune surveillance.

The mouse model replicated this, offering new ways to dissect the escape mechanisms. A glioblastoma, Pollard summarises, becomes full of suppressive macrophages, which are essentially tricked into responding as if it were a healing wound – something that needs to be fixed, not destroyed. Consequently, T cells that ordinarily attack and destroy cancer cells don’t enter the tumour. Or, if they do, the local environment tells them not to kill.

This all suggested that glioblastomas should respond to immunotherapies that mobilise a person’s immune system against their tumour.

Gene therapy

Also, through the 2010s, Pollard was honing SSEs – the technology on which Trogenix stands. He speaks gleefully of this project – of going across town to work with DNA assembling robots and other gizmos offered by the synthetic biology labs. “I love that,” he says, “I’m a molecular geneticist, a science geek.”

The idea was that if you knew the transcription factors that specified the cell type that you wanted your SSE to be active in, then you stitched together fragments of gene switches known to bind those transcription factors.

“They become a very, very strong switch,” Pollard says, “because they’re like a landing pad for all the transcription factors that define that cell type and drive gene expression.”

The possibilities offered by SSEs were reframed by seeing how effective AAV delivery vectors could be. “Rather than just thinking could we get an SSE reporter for marking stem cells,” Pollard says, “We asked could we use it in a gene therapy and drive a payload?”

As this worked progressed, SV Health Investors came knocking in 2019 after CRUK had earmarked the approach as having potential clinical and commercial promise.

“The core technology was then – and remains – the SSE,” Pollard says. That is, a platform – a way of designing and making gene switches that can provide specific gene expression in any disease-relevant cell type.

At that time, he says, there was a lot of money around and high interest in gene therapies. And very quickly, Cellinta was born.

It just happened that the company launched in February 2020, just as a novel coronavirus was dragging the world to a bewildered halt.

Lock-down was a great opportunity to read the literature and consider what exactly an AAV should deliver to glioblastomas, Pollard says. But for just about everything else, it was a terrible time, and Cellinta folded in autumn 2021.

Pollard says that in hindsight, despite the SSE approach being fairly mature, that company likely formed too early. “There were too many moving parts, too many unknowns.”  Most notably, it formed before the therapeutic payloads had been decided, let alone validated. None-the-less he pauses and wryly says: “It might not have been too early if Covid hadn’t happened. Things you’ll never know….”

Payload decision

As the world re-opened and the SSE IP reverted to the University of Edinburgh, it was the payload and its validation that occupied Pollard. After much deliberation they settled on two: Herpes Simplex Virus Thymidine Kinase (HSV-TK) and Interleukin-12 (IL-12).

HSV-TK phosphorylates the orally given prodrug, ganciclovir, to yield a compound that is lethally incorporated into a cell’s DNA. This means any cell making HSK-TK is selectively killed when a mouse or person is given ganciclovir.

The cytokine IL-12 potently activates the immune system, signalling that it should go into attack mode and learn to remember what it is attacking. That is, IL-12 should instruct immune cells to react to a glioblastoma, not as if it were a regenerating wound but as a foreign body that needs to be obliterated.

Trogenix CEO Macnamara says that helpfully both payloads have been used clinically – albeit separately – giving important insights into their safety and efficacy.

In fact, an HSV-TK vector got to a phase 3 trial for glioblastoma. “The tumour shrank but then just bounced back again six months later,” Macnamara says, “because if you just kill by itself,you can’t kill every single cell. It has to be in combination with the immune training.”

IL-12 made it to a phase 2. “It had some super responders,” Macnamara says, “but it’s got systemic toxicity.”

Here, though, the cytokine is made only in the glioblastoma, which should preclude systemic issues. Moreover, the only cells making IL-12 are fated to die as they also make HSV-TK. “The upshot,” says Pollard, “is you have a self-limiting and self-regulating, self-dosing situation.”

To test the strategy, Pollard’s lab used the genetically engineered mice they’d made to model glioblastoma and its complex immunobiology.

It was spectacularly successful, yielding the sort of survival curves that draw gasps at conferences. “The data were just compelling,” Pollard says. “It’s much, much better than I ever would have thought.”

The dual payload constitutes an approach sometimes termed in situ vaccination. Because the intervention kills tumour cells and activates the immune system simultaneously, as the tumour cells die and spew out cancer-specific antigens, the immune system is primed to learn to recognise and attack any cell it subsequently encounters expressing those antigens. “It’s a training exercise,” Pollard says, “to say, ‘Look, immune system, this is what this tumour looks like’.”

This is critical given glioblastoma’s almost universal recurrence. And so, to test if mice formed effective immune memories, Pollard’s team took animals whose initial tumours had been previously destroyed by the AAV and injected their brains with a second bolus of cancerous cells. Remarkably, no second tumour took hold.

With these data in hand, Pollard, in 2023, again sought to start a company. The bedrock of this company would remain the SSE platform, but it would now also be based on this more mature preclinical foundation.

Despite having a much stronger proposal, Pollard says it was interesting to encounter a more cautious biotech sector than in 2019. But several backers – including Cancer Research Horizons – ensured that this exciting technology again began commercialisation.

Throughout the pitching process Pollard had gotten advice from Macnamara. Both Edinburgh residents, moving in similar circles, they met regularly, Macnamara dispensing insights into the start-up world. Then, when the gene therapy company Macnamara had worked for was acquired by a major pharmaceutical company, Macnamara’s appetite for start-ups and his belief in Pollard and SSEs saw him take the CEO role at Trogenix as the company launched in January 2024.

First trial

Trogenix has developed quickly over its short existence. Pollard praises Macnamara for this. “Eighteen months later, he’d essentially built a whole team – people, the manufacturers, CROs, the consultants…”

“We also had incredible support and expertise from our investors – particularly those at 4BIO Capital, who shaped both the scientific and strategic plans from the earliest stages.”

Pollard still runs his lab at the university. Officially, he works for Trogenix one-and-a-half days a week – though he confesses there are evenings and weekends to add to this.

Trogenix initially had a sponsored research agreement with Pollard’s lab, allowing it to add preclinical data. But this changed in October 2025 after Trogenix secured £70 million in Series A funding, allowing it to establish its own labs, with several people from Pollard’s lab becoming full-time employees.

Those staff won’t work solely on glioblastoma, as the company is already working up a larger pipeline. Nevertheless, most excitement is currently focussed on the upcoming glioblastoma trial.

Macnamara says four factors have accelerated their journey to this point. One, glioblastoma presents huge unmet clinical need. Two, there’ve been major advances in understanding the disease’s biology. Three, Trogenix has an approach that is technically feasible. “Steve’s proof of concept work is incredible,” he says.

Macnamara highlights a final preclinical experiment in which Pollard’s lab recently acquired resected glioblastoma tissue from the local hospital. Having devised ways to keep this tissue alive for over a week, they showed that the AAV worked superbly in these cells. “There’re very few experiments that I can do after that which could add meaning” says Pollard. “We just have to go into patients.”

“And four,” Macnamara says, “There’s a hugely supportive regulatory environment. Because there is no meaningful standard of care, the MHRA and FDA have been working with us very supportively.”

Consequently, this first phase 1 trial – to be conducted in Edinburgh and at Ohio State University – will serve both the UK and US regulators.

Slated to start early in 2026, the first people to receive the AAV will have recurrent glioblastoma, allowing the team to confirm that the treatment is safe at the selected doses, while also checking for biological activity.

Then, from probably around the fourth person onwards, trial participants will be people freshly diagnosed with the disease.

This is to most powerfully immunise people against their tumours, Pollard explains. Because today’s standard-of-care chemotherapy can damage the immune system, giving the viral immunotherapy first, means the immune activity and memory are activated before the system is compromised.

Another major advantage to this timeline is that roughly three weeks after a trial participant is given Trogenix’s immunotherapy, they will have their tumour removed. Interrogating the tissue that the neurosurgeon excises will give an immediate indication of whether the AAV is working. The team will be able to check for cell loss, for expression of HSV-TK and IL-12, and for other robust indicators that the immune system is responding as planned.

Trogenix should, therefore, have a good idea of where their therapy is heading within a year.

If initial results are disappointing, Pollard says, there will be various strategies to improve matters. The once sceptic on gene therapy says, “I like gene therapy. It’s very different to a small molecule, which either works or doesn’t usually. We can select the biology, we can control it, tune it, go back. We can iterate.”

Both Pollard and Macnamara stress just how desperate the need for new glioblastoma treatments is. After the company announced its recent funding and news of the trial, Macnamara says, “we were absolutely flooded with inquiries from patients, from families, from patient groups, asking about the trial, how they can know more – or actually just say thank you.”

With people sending in scans and asking how to get on the trial, Macnamara says, “It was a very short-lived celebration. It was very soon, ‘Right team, back to work.’”