Discovery Stories of RET Fusions in Lung Cancer: A Mini-Review

In 2004, a chemical inhibitor of the kinase activity of EGFR was reported to be effective in a subset of lung cancer patients with activating somatic mutations of EGFR. It remained unclear, however, whether kinase fusion genes also play a major role in the pathogenesis of lung cancers. The discovery of the EML4-ALK fusion kinase in 2007 was a breakthrough for this situation, and kinase fusion genes now form a group of relevant targetable oncogenes in lung cancer. In this mini-review article, the discovery of REarrangement during Transfection fusions, the third kinase fusion gene in lung cancer, is briefly described.


INTRODUCTION
Somatic mutations cause cancer via multiple mechanisms, including point mutations, insertions, deletions, and gene rearrangements. In non-small cell lung cancer (NSCLC), one of the most common causes of cancer-related deaths, these oncogenic mutations are usually mutually exclusive, and generally only a single major driver mutation is found in each case. In addition, such a cancer usually depends on the signal pathway stimulated by the principal oncogene for its survival (oncogene addiction) (Weinstein, 2002). In 2004, it was reported that a chemical inhibitor of the kinase activity of EGFR was effective in a subset of lung cancer patients with activating somatic mutations of EGFR (Lynch et al., 2004;Paez et al., 2004). NSCLC in which EGFR inhibitors are effective preferentially develop in Asian and non-smoker populations, generally lacking other targetable driver mutations (Paez et al., 2004;Pao et al., 2004;Shigematsu et al., 2005). In addition to EGFR mutations, kinase fusion genes have become a group of relevant oncogenes in NSCLC, because targeted inhibition of oncogenic kinase fusion proteins also leads to growth inhibition of the cancer cells and regression of the patient's tumor.
Gene fusion was known to be a major mechanism of oncogenesis in hematopoietic neoplasms and sarcomas (Mitelman, 2000). Various types of fusion oncogenes were reported (Mitelman et al., 2007) after the identification of the BCR-ABL1 fusion kinase in chronic myelogenous leukemia (Bartram et al., 1983). In contrast, it remained unclear for a long time whether such fusion oncogenes also play a major role in the pathogenesis of epithelial tumors. The discovery of the EML4-ALK fusion kinase in NSCLC via inv(2)(p21p23) was a breakthrough in this scenario (Soda et al., 2007). Moreover, several small molecules, such as crizotinib (Kwak et al., 2010;Shaw et al., 2013) and alectinib Takeuchi et al., 2016;Hida et al., 2017), showed improved survival outcomes in ALK fusion-positive NSCLC patients. These clinical successes suggested that targeting specific fusion kinases was a promising strategy also for treating carcinomas (epithelial cancers). Representative fusions in epithelial tumors are listed in Table 1.
Receptor tyrosine kinases including ALK usually comprise an extracellular receptor domain, a transmembrane domain, and an intracytoplasmic tyrosine kinase domain. The receptor domain binds to ligands, resulting in dimerization of the kinase protein. Then, the dimerized proteins are autophosphorylated and stimulate the RAS-MAPK-ERK and PI3K-AKT pathways to promote cell proliferation, migration, and differentiation. A receptor tyrosine kinase gene rearrangement gives rise to the expression of the fusion kinase protein if the 5 -partner gene fuses with the 3 -kinase gene in an in-frame fashion. These fusion kinases can be oncogenic when they retain the kinase domain and are dimerized through the 5 partner, because this dimerization mimics that of the wild-type receptor tyrosine kinases through ligand binding. Consequently, a fusion kinase is constitutively expressed, dimerized, and autoactivated, and its downstream signaling promotes cell proliferation and survival.

RET FUSION
REarrangement during Transfection (RET) was identified by Takahashi et al. in 1985 as a proto-oncogene that underwent rearrangement during the transfection of DNA extracted from human T-cell lymphoma into NIH-3T3 cells (Takahashi et al., 1985). RET is a receptor tyrosine kinase encoded by a gene located on 10q11.22 (Ishizaka et al., 1989), and physiologically plays an important role in the development of neurons and kidneys. The first RET fusion in human cancer samples, CCDC6-RET, was identified in papillary thyroid carcinoma by Grieco et al. (1990). RET fusions are detected in 13-43% of papillary thyroid carcinomas (Kondo et al., 2006), and at least 12 RET fusions have been reported so far ( Table 2).

DISCOVERY OF RET FUSIONS IN LUNG CANCER
In 2012, the first RET fusion in lung cancer, KIF5B-RET, was reported independently by 4 groups from Korea (Ju et al., 2012), Japan (2 groups) (Kohno et al., 2012;Takeuchi et al., 2012), and the United States (Lipson et al., 2012). Ju et al. (2012) examined tissue and peripheral blood samples from a 33-yearold Korean never-smoking male with lung adenocarcinoma. The patient was negative for EGFR and KRAS mutations, and the EML4-ALK fusion gene, which were the three well-known driver mutations in lung adenocarcinoma at that time. Fiftytwo fusion transcripts were called by transcriptome analysis in the patient's adenocarcinoma. Out of 52 fusions, they could detect a corresponding genomic rearrangement only for KIF5B-RET fusion (KIF5B exon 16;RET exon 12 fusion variant. K16;R12) by whole genome sequencing. Additionally, they performed transcriptome analysis in 5 lung adenocarcinomas that were negative for EGFR and KRAS mutations and EML4-ALK, and identified one more case with KIF5B-RET fusion transcript (K15;R12). Furthermore, they found another KIF5B-RET-positive case (K23;R12) in 15 "double-negative (negative for EGFR mutation and EML4-ALK but KRAS status unknown)" lung adenocarcinomas by RT-PCR. Based on their detection rate, they estimated that the fusion might exist in approximately 6% of lung adenocarcinomas.
The following three studies were published in the same issue of the same journal, reflecting the "fusion kinase discovery race in major carcinomas" in those days. In the three studies, the frequency and oncogenicity of KIF5B-RET were more specifically evidenced, and growth inhibition analyses using cell lines and RET inhibitors were performed. Kohno et al. (2012) at the National Cancer Center researchers in Japan performed whole-transcriptome sequencing of 30 lung adenocarcinomas to identify new fusion genes that could be targeted for therapy. As a result, they discovered a KIF5B-RET fusion transcript in 1 out of 30 cases. In addition, 289 Japanese lung adenocarcinomas were screened by RT-PCR and Sanger sequence analyses, and the KIF5B-RET fusion gene was identified in 5 cases. In total, they identified 6 KIF5B-RET-positive cases out of 319 lung adenocarcinomas (1.9%), and 4 fusion variants in these 6 tumors. They also examined lung adenocarcinomas in the United States and Norway, and detected a KIF5B-RET transcript in one of the 80 (1.3%) subjects from the United States, but not in the 34 from Norway. They exogenously expressed a KIF5B-RET transcript (KIF5B exon 15;RET exon 12 variant. K15;R12) in the H1299 human lung cancer cell line and showed that Tyr905 was phosphorylated in the absence of serum stimulation. This phosphorylation was suppressed by vandetanib, a tyrosine kinase inhibitor to several receptor tyrosine kinases, including RET. They also showed that expression of exogenous KIF5B-RET induced morphological transformation and anchorage-independent growth of NIH-3T3 cells, which was suppressed by vandetanib. Lipson et al. (2012) analyzed genomic DNA extracted from 24 formalin-fixed paraffin-embedded (FFPE) specimens of NSCLC by capture sequencing targeting 2,574 coding exons of 145 cancer-relevant genes and 37 introns of 14 frequently rearranged genes in cancer. They identified a KIF5B-RET transcript (K15;R12), generated via an 11,294,741-bp pericentric inversion on chromosome 10 in a lung adenocarcinoma from a 44-year-old never-smoking man of European ancestry. They detected KIF5B-RET fusions by RT-PCR in 1 of 121 (0.8%) European-ancestry and 9 of 405 (2%) Asian subjects, all of whom were never or limited former smokers. They estimated an overall occurrence rate of 2.0% (95% CI 0.8-3.1%). Four transcript variants were reported by them: K15;R12, K16;R12, K22;R12, and K15;R11. Ba/F3 cells, which are dependent on interleukin-3 (IL-3) for growth, that expressed KIF5B-RET were transformed and lived without IL-3. The cells were sensitive to sunitinib, sorafenib, and vandetanib, which are multi-target kinase inhibitors that inhibit RET.   Unlike the above-mentioned three studies, Takeuchi et al. (2012) identified KIF5B-RET fusions without nextgeneration sequencing analyses, but with traditional methods. They established an integrated platform of conventional histopathology and molecular pathology to identify fusion genes in various types of cancer. They performed fluorescence in situ hybridization (FISH) with their laboratory-made probes on tissue microarrays of various types of cancers. Using lung cancer tissue microarrays containing 1,528 samples, rearrangement of KIF5B was examined by a split FISH assay to discover new fusions, because they previously identified KIF5B-ALK fusions in lung cancer (Takeuchi et al., 2009) and thus hypothesized that KIF5B might fuse to other kinases in lung cancer. Twentyfour KIF5B split FISH-positive tumors were identified; among them, a KIF5B-RET transcript (K23;R12) was identified by 3 rapid amplification of cDNA ends (RACE). Then, 22 RET rearrangement-positive tumors were identified in 1,528 lung cancers by RET split FISH. Among the 22 cases, 12 KIF5B-RETpositive tumors were identified through a multiplex RT-PCR system that captures all possible KIF5B-RET fusions: 8 cases with K15;R12, and one case each with the K16;R12, K22;R12, K23;R12, and K24;R11. The presence of inv(10)(p11.22q11.2) was supported by a KIF5B-RET fusion FISH assay in all 12 of these tumors. In lung cancer, they also identified CCDC6-RET, which is the first RET fusion identified in thyroid cancer (Grieco et al., 1990). In a routine pathology diagnosis during the study period, a pathologist in the group encountered an adenocarcinoma with a mucinous cribriform pattern that is a histopathological marker for the presence of EML4-ALK (Inamura et al., 2008). The case was, however, negative for ALK fusion and was positive for CCDC6-RET, as determined by FISH and inverse RT-PCR. In the remaining 10 tumors, another CCDC6-RET-positive tumor was identified by RT-PCR. In total, 14 RET fusion-positive tumors (13 out of the 1,528 tumors tested, and one additional tumor found through a routine pathology diagnostic service) were identified. RET fusions existed in 0.9% (13 out of 1,482) of the NSCLCs and 1.2% (13 out of 1,119) of the adenocarcinomas. The researchers demonstrated the oncogenicity of all the 5 KIF5B-RET fusion variants they identified through a focus formation assay and a mouse subcutaneous transplantation assay using NIH-3T3 cells expressing each KIF5B-RET variant. KIF5B-RET (K15;R12) transfected Ba/F3 cells grew in the absence of IL-3. Vandetanib inhibited the proliferation of cells expressing K15;R12 but not the proliferation of cells expressing EML4-ALK.

CONCLUDING REMARKS
REarrangement during Transfection-positive lung cancers constitute a small subset of lung adenocarcinomas showing clinicopathological features similar to those of other fusion kinase-positive lung cancers. Since their discovery, several trials for RET-positive lung cancer have been conducted using kinase inhibitors including vandetanib, cabozantinib, sorafenib, sunitinib, lenvatinib, ponatinib, and dovitinib. Although some clinical benefits were observed, efficacy was limited compared with that shown by EGFR and ALK inhibitors. The abovementioned agents used in earlier trials are multi-kinase inhibitors, and are notably more effective to VEGFR, EGFR, and KIT than RET. Therefore, off-target dose limiting toxicity caused frequent dose reduction and discontinuation. RET inhibitors with more specificity and hence less off-target toxicity are currently undergoing clinical and preclinical development.

AUTHOR CONTRIBUTIONS
The author confirms being the sole contributor of this work and has approved it for publication.