Fusions in solid tumours: diagnostic strategies, targeted therapy, and acquired resistance

Abstract

Structural gene rearrangements resulting in gene fusions are frequent events in solid tumours. The identification of certain activating fusions can aid in the diagnosis and effective treatment of patients with tumours harbouring these alterations. Advances in the techniques used to identify fusions have enabled physicians to detect these alterations in the clinic. Targeted therapies directed at constitutively activated oncogenic tyrosine kinases have proven remarkably effective against cancers with fusions involving ALK, ROS1, or PDGFB, and the efficacy of this approach continues to be explored in malignancies with RET, NTRK1/2/3, FGFR1/2/3, and BRAF/CRAF fusions. Nevertheless, prolonged treatment with such tyrosine-kinase inhibitors (TKIs) leads to the development of acquired resistance to therapy. This resistance can be mediated by mutations that alter drug binding, or by the activation of bypass pathways. Second-generation and third-generation TKIs have been developed to overcome resistance, and have variable levels of activity against tumours harbouring individual mutations that confer resistance to first-generation TKIs. The rational sequential administration of different inhibitors is emerging as a new treatment paradigm for patients with tumours that retain continued dependency on the downstream kinase of interest.

Figure 1 |. Discovery of oncogenic fusions.

Timeline depicting the identification of selected oncogenic fusions in various malignancies and the methods used to detect them. Structural rearrangements and later, gene fusions, were first identified in patients with chronic myeloid leukaemia, from whom tumour-cell-containing blood harbouring the t(9;22)(q34;q11) translocation was easily obtainable for cytogenetic analysis. Interchromosomal gene exchanges, resulting in translocations, were subsequently discovered using similar techniques in sarcomas (such as t(11;22)(q24;q12) and t(17;22)(q22;q13)) and carcinomas. Technologies for guided fusion detection evolved over time to include a variety of assays, including fluorescence in situ hybridization (FISH) and reverse-transcriptase PCR (RT-PCR) analysis, enabling the additional detection of intrachromosomal events that produce fusions (for example, those resulting in the EML4–ALK and KIF5B–RET fusions). Currently, the adoption of unbiased fusion-detection technologies, such as next-generation sequencing and anchored-multiplex PCR, have enabled the detection of increasing numbers of fusion events, both in the research setting and in the clinic. DFSP, dermatofibrosarcoma protuberans.

Figure 2 |. Structural modelling of putative targeted therapy resistance mutations.

a | In silico structural modelling of the tyrosine-kinase domains of ALK, ROS1, RET, TRKA, TRKB, and TRKC indicates a wide range of resistance mutations that cluster within regions including the ATP-binding pocket and the solvent front. The amino acids that replace wild-type residues following mutation are shown in red. b | Homology alignment demonstrates that several of the identified resistance mutations arise in paralogous residues across genes, suggesting that similar mechanisms of drug resistance can develop across different fusions. c | Most resistance mutations seen in ROS1, RET, TRKA, TRKB, and TRKC have paralogous resistance mutations identified in ALK. The specific paralogous amino acids shown in part b are listed here.

Figure 3 |. Activity profile of kinase inhibitors targeting specific resistance mutations.

The in vitro sensitivity of specific mutations to selected kinase inhibitors is indicated in the coloured boxes (green indicates an IC₅₀ <50 nM, yellow indicates an IC₅₀ of 50–200 nM, red indicates an IC₅₀ of >200 nM). Coloured dots represent reports of clinical activity (green indicates tumour shrinkage, red indicates disease progression) that either conflict with or have been reported in the absence of available in vitro data. The wide variations in activity against mutations that emerge in the setting of resistance highlights the clinical utility of molecular profiling of tumours at the time of progression as a potential means of directing sequential kinase-inhibitor therapy, assuming that results are received within a reasonable timeframe. Note that the cell lines used for in vitro assays varied, and that the listed resistance mutations in RET were identified in cancers harbouring RET-activating mutations, but have yet to be observed in tumours from patients, to our knowledge.

Figure 4 |. Distribution of kinase fusions across primary tumour sites.

Gene rearrangement is a common event in cancer; however, rearrangements that result in a pathogenic fusion gene for which targeted therapy is potentially available are considered the most clinically significant events. The distribution of selected clinically-relevant fusions, specifically those involving ALK, ROS1, RET, NTRK1/2/3, FGFR1/2/3, and BRAF/CRAF are shown here. The frequency of these fusions varies substantially between different tumours, ranging from isolated case reports to fusion events that are found in most tumours of a particular histology. The presence of clinically-relevant gene fusions across a wide range of malignancies supports the use of basket-type approaches to the design of clinical trials involving targeted therapies.