TRKing the path to a tumour agnostic drug target

In early 1998, my lab at BC’s Children’s Hospital set out to solve a practical problem in diagnostic pediatric pathology—how could we differentiate between a benign lesion called fibromatosis and a malignant tumour called congenital or infantile fibrosarcoma (CFS or IFS), both of which occur in very young children under 2 years of age, but which are often indistinguishable under the microscope. While fibromatosis can be treated with surgery alone, or even non-interventional follow-up, CFS/IFS requires surgery plus chemotherapy, and can metastasize and potentially kill the patient. Doing surgery and administering chemotherapy in young infants can cause many potential side effects, and so avoiding these aggressive therapies is of paramount importance in pediatric oncology. We therefore set out to find a genetic marker to distinguish fibromatosis from CFS/IFS. To make a long story short, heroic work by Stevan Knezevich, a PALM PhD student in the lab, along with others, identified a novel t(12;15) translocation that generated an oncogenic gene fusion between the ETV6 transcription factor, and TRKC, a receptor tyrosine kinase (later renamed NTRK3) that is normally expressed mainly in the CNS. This resulted in a seminal article (Knezevich et al, Nature Genetics, 1998), reporting expression of ETV6-NTRK3 (EN) fusions in CFS/IFS (PMID:9462753), and the first report of NTRK protein fusions as a recurrent alteration in human cancers. We went on in a series of articles to show that detecting EN fusions in tumour tissues effectively differentiates CFS/IFS from fibromatosis. So, we had achieved our initial goal, and discovered some interesting new cancer biology. Genetic analysis of EN fusions is now used worldwide as a molecular diagnostic tool for CFS/IFS.

Then, however, came a bit of a surprise. I was contacted by my colleague and friend, Dr. Doug Horsman at BCCA, who said he had noticed a similar t(12;15) translocation in a case of secretory breast carcinoma (SBC), a rare variant of ductal breast carcinoma. We went on to show that indeed, SBC harbours similar ETV6-NTRK fusions, which was surprising as oncogenic fusions are conventionally thought to be specific for one tumour type. This work was performed by both Stevan and Cristina Tognon, who had joined the lab as new post-doc, and resulted in us publishing an article demonstrating ETV6-NTRK3 as the first and only known translocation-associated etiologic event in breast cancer, showing direct relevance of our pediatric cancer work to adult malignancies (Tognon et al, Cancer Cell, 2002). We then collaborated with Dr. Stu Orkin at Harvard University, who called me up one day and said, “let’s prove that ETV6-NTRK3 is an oncogenic driver”. Working with Stu, a mouse-modeling expert, we published with Stu that 100% of mice expressing ETV6-NTRK3 developed breast cancer (Li et al, Cancer Cell, 2007). This would prove to be important later for therapeutic development of drugs to target NTRK fusion proteins.

Remarkably, ETV6-NTRK3 fusions were subsequently reported by others in virtually 100% of a subgroup of salivary gland tumors (leading to a major reclassification of these tumours to the designation of mammary analogue secretory carcinoma, or MASC), and in rare cases of infantile brain tumors, colorectal carcinomas, lung cancers, pancreatic carcinomas, leukemias, skin tumours, melanomas, thyroid carcinomas, and neuroendocrine tumours, highlighting the broad relevance of ETV6-NTRK3 in human cancers. Indeed, it is estimated that ~12% of thyroid carcinomas express ETV6-NTRK3, and it was reported that survivors of the Chernobyl nuclear accident that developed thyroid carcinoma demonstrated an extremely high incidence of ETV6-TRK3 fusions, which has never been explained. Moreover, with the advent of next-generation sequencing, many labs began to report gene fusions involving the other members of the TRK/NTRK family, namely NTRK1 and NTRK2. It is now estimated that NTRK1, 2, and 3 fusions occur in ~1% of human cancers, leading to a new classification of “NTRK fusion cancers”. Notably, each NTRK fusion is thought to activate both proliferation (RAS-ERK) and survival pathways (PI3K-AKT), as shown in the figure, much of which was gleaned from our biochemical studies of ETV6-NTRK3. This, plus the mouse studies proving that ETV6-NTRK3 is a bona fide oncogenic driver, directly prompted the pharmaceutical industry to begin thinking about whether NTRK inhibitors might be a tractable therapeutic strategy in clinical oncology.

Since NTRK proteins are tyrosine kinases, Loxo Oncology began to work on NTRK tyrosine kinase inhibitors. They successfully achieved this by developing Larotrectinib, which as it turns out is a pan-NTRK inhibitor and inhibits each type of NTRK fusion. Larotrectinib was then licensed by Bayer Pharmaceuticals worldwide. After a number of very promising international clinical trials across diverse tumour types, including pediatric patients with CFS/IFS, this drug was approved by the FDA on Nov. 26, 2018, and by Health Canada in the Spring of 2019. Given the broad number of potential indications, this agent can justifiably be hailed as a tumour type agnostic drug.

In summary, this has been something of a crazy and fascinating journey. What started out as a fairly straightforward molecular diagnostic question has turned into a much broader bench-to-bedside deliverable, and one that I can say gives me a great deal of satisfaction to be a part of. The most gratifying aspect of the story is that the work ultimately has helped patients, and this has to be the major goal of any translational research voyage that we embark on as biomedical scientists.