Until my son Michael was born, I had never heard of cystic fibrosis.
Two years later, I was on my way to London to join the Gene Therapy Consortium, helping to raise capital for a Phase 2b clinical trial involving Imperial College London and the universities of Oxford and Edinburgh.
I was a lawyer and finance executive, not a scientist. Raising money was something I understood. The chemistry behind gene therapy was another matter entirely.
So, on the way to the airport, I picked up a copy of Chemistry For Dummies.
It seemed a reasonable place to start.
I had no idea where that journey would eventually take me.
A problem worth solving
Cystic fibrosis (CF) is a genetic disease that particularly affects the lungs. The problem is a faulty gene which causes thick, sticky mucus to accumulate, leading to repeated infections and, ultimately, shortened lives.
The proposition behind gene therapy seems simple enough. Get a working copy of the gene into the lung cells and correct the problem.
Unfortunately, getting it there has proved extraordinarily difficult.
The gene responsible for CF was discovered in 1989. More than 35 years later, despite enormous investment and scientific effort, there is still no approved gene therapy for CF.
The usual approaches involve modified viruses or tiny fatty particles called lipid nanoparticles. Neither has yet delivered a successful treatment.
So I started Loxegen and went looking for something different.
I found a promising technology based on polymeric nanoparticles originally developed for cancer research. We adapted it to deliver genetic material into the lungs using a nebuliser.
We call them Non-Canonical Polymeric Particles, or NCPPs.
Millions of dollars, a $1.5 million Australian Government grant, support from charities and countless friends in the community later, we have particles that nebulise well and have successfully delivered different types of genetic material into a range of cells.
It’s only taken nearly a decade!
But having a promising technology and finding someone to pay for it are two rather different things.
The problem with fashion
Biotechnology investment follows trends. That’s understandable. Most drug development programmes fail, so investors tend to back technologies that have already demonstrated some success.
The COVID-19 vaccines made lipid mRNA nanoparticles fashionable, and much of the investment in CF gene therapy has followed them.
Which I find rather odd.
With a vaccine, you generally want to stimulate the immune system and produce a relatively short-lived response. With gene therapy, you usually want to avoid an immune response and achieve something lasting.
That’s not to say lipid nanoparticles can’t work in gene therapy, but success in vaccines doesn’t automatically make them the answer for gene therapy.
There have been setbacks in inhaled mRNA programmes, yet the investment trend continues.
We could sit around waiting for the pendulum to swing back towards alternative technologies, which it is starting to do rather rapidly.
Or we could find another way to make our technology commercially useful.
As it turns out, that may also improve our chances of eventually treating CF.
An unexpected opportunity
Artificial intelligence has become the latest big thing in biology, and with good reason.
AI can find patterns in data, suggest experiments and help scientists investigate questions that might otherwise take years to answer.
But there’s a catch.
AI needs data. Lots of it. And it needs to be reliable.
Biological experiments are notoriously difficult to compare. Scientists change ingredients, doses, cell types and procedures as they learn. Quite reasonably, they’re trying to answer scientific questions, not manufacture consistent datasets for computers.
I know. We’ve been doing it for years.
We started organising nearly a decade of Loxegen’s experimental results into a structured database and developed an AI model called Optilox Alpha to help us improve our formulations.
Then something occurred to us.
Perhaps the immediate opportunity wasn’t selling our nanoparticles as another gene delivery product.
Perhaps it was using them to generate the biological data that AI companies need.
From delivering genes to delivering answers
Our particles can carry different types of genetic material into cells, making it possible to investigate all sorts of biological questions.
And some of the things they can do are quite remarkable.
Fireflies glow because they produce a particular protein. We can take the genetic instructions for that protein, package them into our particles and have a mouse inhale them.
The mouse’s lungs then produce the same light-generating protein.
In other words, we can make a mouse’s lungs glow using a firefly gene.
They’re also remarkably robust.
We once shipped particles from Queensland in 43-degree heat to Chicago during a freezing winter. They then sat on a laboratory bench for seven months.
When the researchers finally used them, they worked just fine.
That’s useful if you’re trying to run lots of experiments without spending a fortune.
Instead of simply selling a delivery system, we could help researchers answer questions about how cells behave.
What happens when you introduce a particular gene? Does the cell survive? Does its behaviour change? How long does the effect last?
Each experiment generates data. Run enough carefully controlled experiments and you begin building something valuable for AI.
Our particles are relatively inexpensive, adaptable and straightforward to formulate. That makes them well suited to generating large numbers of biological experiments.
And the applications extend well beyond CF.
We’ve already signed collaborations and are talking to organisations involved in AI-driven biology, gene editing and research tools.
It’s early days, but the interest is encouraging.
AI is the tool. People make the difference.
There’s a lot of talk about AI replacing scientists.
I doubt it.
Even when I interrogate our own AI model, sometimes a result just doesn’t look right.
The mathematics might be perfectly sound, but experience tells you the model has misunderstood something, made a false assumption or drawn a conclusion from a questionable experiment.
That’s where people matter.
AI can process more information than any human being, but someone still has to understand the problem, challenge the assumptions and decide what to do next.
And sometimes the most important discovery is realising you’ve been asking the wrong question.
That’s certainly been our experience.
A different horizon
Michael is no longer the two-year-old who sent me off to London with a copy of Chemistry For Dummies.
The reason I started hasn’t changed. I’d still love to see our technology contribute to a treatment for CF.
But after nearly a decade, I’ve learnt that scientific progress and commercial success don’t necessarily follow the same path.
What began as an attempt to solve one problem may have given us a way to help researchers solve hundreds of others. And seeing as we don’t need to go through clinical trials we can do so for much less capital, we are a near term revenue business now.
And, rather neatly, generating better biological data may ultimately bring us closer to solving the original problem.
We still have plenty to prove. We need more customers, more experiments and some additional capital, although considerably less than we once did.
Not a bad outcome from buying Chemistry For Dummies at the airport.
About the author
Andrew Venables is Founder and CEO of Loxegen, an Australian biotechnology company developing non-viral, non-lipid gene delivery technology. A former partner at Minter Ellison, Andrew has worked in corporate law, finance and capital raising in Australia and internationally.
His involvement in gene therapy began after his son Michael was born with cystic fibrosis. Nearly a decade of research has produced Loxegen’s nanoparticle technology, an extensive biological dataset and Optilox Alpha, its AI-assisted research platform. Andrew is now developing partnerships that use Loxegen’s technology to generate biological data, while continuing to pursue its original goal of improving treatments for genetic disease.
www.loxegen.net