Two decades ago, Poul Sorensen’s microscope was letting him down.
Like so many other pediatric pathologists around the world, Dr. Sorensen was finding it immensely difficult – and sometimes impossible – to distinguish congenital fibrosarcoma, a rare and deadly type of connective tissue tumour in very young children, from a similar but benign version called fibromatosis. Under the microscope, they looked pretty much the same. But patients with fibrosarcoma needed major surgery, chemotherapy or even both, while patients with fibromatosis could be left untreated.
“This was a real quandary for pathologists and pediatric oncologists, because they couldn’t tell the difference based on tools available at the time,” recalls Dr. Sorensen, a Professor of Pathology and Laboratory Medicine at UBC. “And if it’s one-year-old kid, the stakes are so high. Surgery or chemotherapy for a one-year-old is a far more delicate and risky undertaking than it would be for an adult.”
From that inquiry, Dr. Sorensen wound up making a discovery with long-term dividends – the development of a cancer drug approved today by the U.S. Food and Drug Administration, based on its demonstrated ability to shrink many different kinds of tumours.
“We scientists always dream that the stuff we do in the lab leads to more knowledge, but also has an impact on people’s health,” says Dr. Sorensen, the Johal Chair in Childhood Cancer Research at UBC and a Distinguished Scientist at BC Cancer. “In this case, it took 20 years, but it got done.”
An “umbrella” cancer drug
The drug, called larotrectinib (Vitrakvi), was developed by Connecticut-based Loxo Oncology. When Loxo and Bayer, which bought the marketing rights to larotrectinib and a follow-on drug, made a presentation to Health Canada last year, they asked Dr. Sorensen to provide expert testimony about the science underlying its effectiveness.
What makes larotrectinib’s approval so noteworthy is that it applies to a wide array of tumours, many of which would seem to have nothing in common except the mutation that spawned them – the very same mutation that Dr. Sorensen first revealed to the world in 1998.
It all began with Dr. Sorensen wondering if he could detect a differentiating feature between the two types of connective tissue tumours – a feature that wasn’t apparent at the microscopic level, but perhaps at the molecular level.
Dr. Sorensen examined the chromosomes of both types of growths and found something startling in the malignant version: a fusion mutation, the result of two chromosomes colliding, breaking and exchanging pieces with one another, creating a new gene in the process.
“It’s like splicing pieces of two different films together to form a new one,” Dr. Sorensen says.
He and his colleagues published their findings in Nature Genetics in 1998, providing pathologists around the world with a way of diagnosing fibrosarcoma.
But that was just the beginning.
An “always-on” enzyme
“We wanted to dig into what this new gene does,” he says. “It was a tumour driver, so we thought it might provide lessons about oncogenesis.”
Further investigations by Dr. Sorensen revealed that the new gene, called ETV6-NTRK3, leads to production of an enzyme that plays a role in driving cell growth and differentiation. This mutant enzyme is always on – it doesn’t need a biochemical stimulus. Since cancer by definition is uncontrolled cellular proliferation, that makes ETV6-NTRK3 an oncogene.
“It’s a ‘dangerous liaisons’ scenario, because the proximity of two chromosomes have created a new gene that gives cells a huge growth advantage,” Dr. Sorensen says.
He and his colleagues found several biochemical pathways that led to the “always on” state. Using existing drugs, they succeeded in blocking one pathway after another in vitro, but not all pathways at the same time, so the drugs failed to prevent tumour formation in mouse models. Snipping one wire wasn’t enough – all of them had to be cut.
But there was no single drug that could do that.
Not just one rare disease
Then BC Cancer scientist Doug Horsman came to him with intriguing news: Under the microscope, he observed a chromosomal change in a breast cancer biopsy that reminded him of what Dr. Sorensen’s team had found in childhood fibrosarcoma.
The discovery prompted them to go through BC Cancer’s biobank of breast cancer biopsies, which revealed more examples of the fusion mutation.
“This was surprising to us, because the dogma was that these fusions were very specific for a certain disease,” Dr. Sorensen says. “And here we had the same fusion mutation linked to cancer in two different types of tissue: connective tissue in fibrosarcoma, and epithelial tissue – which lines the outer surfaces of organs and blood vessels – in breast cancer.”
That news was published in Cancer Cell in 2002. And that led to a call from Stuart Orkin, a pediatric oncologist and hematologist at Harvard Medical School, the Dana Farber-Harvard Cancer Center and Boston Children’s Hospital.
“He said, ‘Let’s prove this is a driver of breast cancer,’” Dr. Sorensen recalls.
The scientists on opposite coasts teamed up to create a mouse model with the ETV6-NTRK3 fusion mutation, and every mouse – every single mouse – developed secretory breast cancer. Cancer Cell published that discovery in 2007.
ETV6-NTRK3’s relevance took on even more momentum when it turned up in about 40 per cent of the thyroid gland tumours of people exposed to radiation from the Cherynobyl nuclear disaster.
A commercially viable target
And like a trickle turning into a stream that grows into a river, the mutation was subsequently found in an increasing number of adult malignancies, including 100 per cent of certain types of salivary gland tumours, 100 per cent of a type of pediatric kidney cancer, and 10 per cent to 12 per cent of thyroid cancers.
Enter Loxo Oncology, though Dr. Sorensen had no idea that a company was targeting the mutation until he was approached by a Loxo executive at a 2014 conference.
“I was a bit surprised, because I hadn’t been following the literature,” Dr. Sorensen says. “But Michael told me they had been collecting data from next-generation sequencing projects, and realizing that this mutation was far more common than people had ever imagined. When you added it all up, this mutation was present in about 1 per cent of human tumours. That’s when it became a commercially viable target.”
Later, Loxo sent a sample of the drug to Dr. Sorensen, who tested it on cultures of cancerous cells.
“And it worked,” he says. “I wasn’t so surprised they found something. I was more surprised that they found something that was so effective.”
Hitting the mother ship
Loxo now says 22 different types of cancer harbour this fusion mutation, or variants of it, including colorectal cancers, lung cancer, brain tumours, and melanoma, leading to the creation of a new category of human tumours – NTRK fusion cancers. Larotrectinib, by inhibiting only the mutation and nothing else, maximizes “the potential for efficacy, while minimizing the risk of off-target toxicities.”
In clinical trial data presented to the American Society of Clinical Oncology in 2017, the drug shrunk the tumours in 76 per cent of patients with a different types of tumours. In 12 per cent, their tumours vanished entirely; 12 per cent saw their cancer stabilize; another 12 per cent experienced continued cancer growth.
At the Health Canada hearing earlier this year that took testimony about larotrectinib, the panel asked Dr. Sorensen whether it might be better to target the gene’s downstream pathways – in other words, blocking the always-on enzyme – rather than the gene itself.
“I could answer that immediately, because we had done those experiments, and that strategy had failed to stop oncogenesis,” Dr. Sorensen says. “Tumours cells might find ways to compensate for one blocked pathway by activating another one. But if you hit the mother ship, then it’s game over.”
Rarer, simpler – and more revealing
The approval of larotrectinib represents a new approach to cancer drugs and their regulation – one drug that is intended and approved for a wide variety of cancers, with the common link being their genetic driver.
Deciding whether a patient would be helped by such a drug would depend, of course, on the tumour being genetically analyzed – something that isn’t currently done with most cancer patients. But with the cost of sequencing falling with every passing year, such a strategy will make more and more sense.
Dr. Sorensen is philosophical that it took so long for something tangible to arise from his discoveries. After all, it began with a modest goal – trying to distinguish between a rare pediatric malignant tumour and its benign look-alike. Developing a new, wide-ranging treatment was far from his mind at first, if only because the tumour was so unusual.
“What’s really gratifying to me is that it validates the idea of studying rare tumours in pediatrics that other people don’t focus on,” he says. “These rarer tumours are often less genetically complex, so it’s easier to discern the driving pathways, revealing fundamental processes of cancer growth. It’s gratifying that this time around, we were able to see the bigger picture.”
