Science Spotlight

The NTRK2 splice variant TrkB.T1 plays a vital role in human glioma

From the Holland lab, Human Biology Division

Alternative splicing of a gene transcript results in protein variants that can have similar, distinct, or even opposite functions. From an evolutionary perspective, alternative splicing is beneficial because it maximizes protein diversity to accommodate complex biological processes without increasing genome size. However, in cancer, the splicing mechanism is often highjacked to produce aberrant proteins that promote tumor formation and progression. Yet despite their oncogenic properties, the clinical relevance of individual splice variants and the mechanisms by which they can interfere with cellular pathways to promote cancer is poorly understood. The Holland lab at the Human Biology Division is interested in understanding the molecular basis of brain tumors. In a recent study, led by postdoctoral fellow Dr. Siobhan Pattwell, the group reported that a truncated splice variant of the neurotrophic tyrosine receptor kinase 2 (NTRK2) gene plays a role in the formation and progression of gliomas, a common type of human brain tumors. Their study was recently published in the journal Nature Communications

The full-length product of the NTRK2 gene which encodes the tropomyosin receptor kinase B protein (TrkB.FL) has well-known functions determining neural stem cell fate both in normal brain and gliomas. The gene also produces several splice variants, including TrkB.T1, a truncated isoform that is missing the kinase domain and contains a unique 11-amino-acid terminal sequence. The authors hypothesized that, since NTRK2 is involved in several subtypes of human gliomas, its distinct isoforms could also drive cancer formation. The authors first used principal component analysis to compare gene expression data sets from The Genotype-Tissue Expression Project and The Cancer Genome Atlas of normal human brain, low-grade gliomas (LGG), and glioblastomas (GBM). Their analysis showed that TrkB.FL’s expression is similar across all samples; however, TrkB.T1 is the predominant isoform in LGG and GBM compared to normal brain. These results showed that, contrary to previous hypotheses in the field, TrkB.FL is not the sole contributor to oncogenesis. Dr. Pattwell explains the implications of these findings: “This work highlights the realization that an NTRK2 splice variant, lacking a kinase domain, is the predominant neurotrophin receptor isoform in human brain tumors.”  

Schematic of TrkB.FL and TrkB.T1 isoforms. The unique c-terminal in TrkB.T1 is noted.
Left: Schematic of TrkB.FL and TrkB.T1. The unique C-terminal 11-amino acid stretch can be used as an epitope to differentiate between isoforms. The novel antibody was recently patented with the help of Hutch Business Development. It will be used for future studies in the lab and to explore potential diagnostics that use TrkB.T1 as a cancer biomarker. Right: TrkB.T1 immunohistochemistry. In normal brain, the isoform was distributed in distinct clusters forming a “dot” pattern. In brain tumors, the isoform was diffused across the tissue. Image provided by Dr. Siobhan Pattwell

To dissect the role of NTRK2 isoforms in tumor formation, the authors identified gene classes that are differentially expressed in LGG and GBM, relative to normal brain, and are positively correlated with NTRK2 expression using gene ontology (GO) analysis. Some of the gene classes identified were associated with endocytic and vesicular transport, suggesting that the subcellular localization of TrkB.T1 may be relevant to its function and might differ between tumor and normal brain. Historically, assigning functionality to NTRK2 isoforms had been challenging due to the lack of reagents to discriminate between isoforms accurately. To correctly detect TrkB.T1, the researchers developed a novel antibody that recognizes the unique terminal 11-amino-acid sequence in TrkB.T1. When the investigators used the antibody in normal and brain tumors, they found that, in the normal brain, the isoform was distributed in distinct clusters forming a “dot” pattern. In brain tumors, the isoform was diffused across the tissue. Dr. Pattwell adds, “Using a novel antibody developed at Fred Hutch, we show that this variant, TrkB.T1, has a unique distribution in normal brain vs. brain tumors in both mice and humans.” The distinct localization pattern suggested potentially different roles of TrkB.T1 in normal brain and gliomas. 

The researchers then sought to explore the potential oncogenic role of TrkB.T1 in vivo. To this end, they used a mouse model in which gliomas are induced by delivering a virus encoding the platelet-derived growth factor (PDGF), an oncogene known to drive glioma formation, to a specific cell type in the brain. Delivery of TrkB.T1, in addition to PDGF, enhanced the formation of gliomas and decreased mouse survival. In vitro studies showed that TrkB.T1 could increase the time signaling pathways downstream of PDGF are active, which potentially contributes to the increased aggressiveness of tumor formation observed in the mouse model. Dr. Pattwell explained the new avenues of research coming out of this study: “These findings raise important questions surrounding the role of splicing choices, and the TrkB.T1 variant specifically, in cancer and development. Does TrkB.T1 play a role in embryonic development or in other cancers besides glioma? Does the overexpression of TrkB.T1 cause other cancers? What are the binding partners and downstream interactors that are critical for these variants oncogenic effect, and can they be targeted? Current and future work will address these questions.” 

This study was supported by grants from the National Institutes of Health, a Jacobs Foundation Research Fellowship, a grant from the American Cancer Society, the National Research Agency, and the Seve Ballesteros Foundation

Dr. Eric Holland is a member of the Fred Hutch/UW Cancer Consortium.

Pattwell, S.S., Arora, S., Cimino, P.J. et al. 2020. A kinase-deficient NTRK2splice variant predominates in glioma and amplifies several oncogenic signaling pathways. Nature Communications doi: 10.1038/s41467-020-16786-5