Identification of RET gene fusion by exon array analyses in “pan-negative” lung cancer from never smokers

Dear Editor,

The incidence of lung cancer from never smokers has increased dramatically in China nowadays. Strikingly, approximately 30% of the lung cancer patients in East Asian population are never smokers^{1, 2}. The majority of these patients are females with lung adenocarcinomas². Identification of oncogenic drivers, which the tumors are “addicted to" and rely on for survival, has significantly reformed the current strategies for lung cancer treatment in clinic and initiated the era of personalized therapy³. Therapeutics specifically targeting EGFR mutations, frequently observed in never smoker patients with lung cancer, have been very helpful in improving the clinical symptoms as well as the progression-free survival^{4, 5, 6}. Similarly, patients with lung tumors positive for ALK fusions also benefit from ALK-targeted therapy^{7, 8}.

Our previous efforts have constructed a quite comprehensive map of those essential oncogenic drivers in 52 lung adenocarcinomas from never smokers⁹. We have uncovered the oncognic drivers in about 90% of these lung tumors including mutations of EGFR, HER2, KRAS, as well as EML4-ALK fusion⁹, thus providing a strong clinical guidance for molecular-targeted therapy for this subset of disease. However, there is still about 10% (5/52) of these never smoker patients were “pan-negative” for all known oncogenic driver mutations and could not benefit from the effective targeted therapy in clinic.

Similar to oncogenic gene mutations, gene fusions such as EML4-ALK or CD74-ROS1 are also essential for lung cancer development^{8, 10}, and serve as effective therapeutic targets. Since great efforts have been paid in searching gene mutations, we instead focus on our efforts in identification of novel oncogenic gene fusions. Previous studies have demonstrated that exon array analyses are capable of detecting gene fusions based on the differential expression of the exons located at either side of the breakpoint, which is frequently resulted from genomic translocation^{7, 11, 12}. For example, in the case of ALK fusion, the expression levels of ALK exon 1-20 and exon 21-29 flanking the breakpoint are significantly different and can be readily detected by exon array analyses^{7, 11, 12}. Therefore, we performed exon array (Affymetrix Exon 1.0) using all the five “pan-negative” samples plus another 12 samples with known oncogenic drivers to search novel oncogenic gene fusions. We initially identified about 1 000 potential gene fusions from exon array analyses. Since most of known oncogenic drivers are kinases, we manually went through all the heatmaps of those potential kinase fusions. Interestingly, we identified one potential RET fusion, with an obvious expressional change between exon 11 and exon 12, in the “pan-negative” lung cancer sample 181LC (Figure 1A). We then preformed the 5′ RACE assay to detect the partner of this potential RET fusion. We found that the DNA band obtained from 5′ RACE (about 1.4 kb) is actually the fusion of RET exon 12 to CCDC6 exon 1 (Figure 1B-1C), which has been previously reported in human thyroid carcinomas¹³. This RET fusion is derived from somatic genetic alteration since it is undetectable in paired normal lung tissue 181NL (data not shown). To clone the genomic breakpoint, we further designed a series of primers (22 forward primers at CCDC6 intron 1 with 1-3 kb intervals, and a reverse primer at RET exon 12) and performed long-range PCR using genomic DNA from the “pan-negative” lung cancer sample 181LC. Interestingly, we found that the intron 1 of CCDC6 is fused to a part of RET exon 11 at genomic DNA level (Figure 1D-1E), which results in the expression of CCDC6-RET fusion (CCDC6 exon 1 fused to RET exon 12) after RNA splicing. Previously studies have shown that RET signaling pathway promotes cell survival and cell proliferation through RAS-ERK pathway and PI3K-AKT pathway¹⁴. RET fusions, mainly found in papillary thyroid carcinomas¹⁵, are oncogenic drivers and capable of transforming thyroid epithelial cells in vitro as well as inducing papillary thyroid carcinoma in transgenic mice^{16, 17, 18}. We found that the CCDC6-RET fusion from lung cancer sample 181LC is undetectable in the rest of 4 “pan-negative” samples as well as those with known oncogenic driver mutations from never smokers (data not shown), consistent with the mutual exclusive pattern of oncogenic drivers. Together with our recent identification of the CD74-ROS1 fusion in another “pan-negative” sample^{9, 19}, we have further improved our original work and uncovered the oncogenic drivers in about 94% (49/52) of these lung adenocarcinoma from never smokers¹⁰: EGFR mutations (78.8%), HER2 mutations (3.8%), KRAS mutations (1.9%), EML4-ALK fusions (5.8%), CD74-ROS1 fusion (1.9%) and CCDC6-RET fusion (1.9%) (Figure 1F).

Figure 1.

Identification of CCDC6-RET fusion in the “pan-negative” lung adenocarcinomas from never smokers and the construction of a more comprehensive spectrum of oncogenic drivers in this subset of lung cancer. (A) Exon array analyses of 5 samples “pan-negative” for oncogenic drivers and 12 samples with known oncogenic drivers have identified a potential RET fusion in lung cancer sample 181LC. The potential fusion point is indicated by yellow arrow. (B) 5′ RACE analyses of the “pan-negative” lung cancer sample 181LC showed a PCR band about 1.4 kb, which is obviously different from the wild-type RET band (about 960 bp). (C) Sequencing result confirmed the CCDC6-RET fusion in lung cancer sample 181LC. (D) Long-range PCR detection of the genomic breakpoint of CCDC6-RET fusion in lung cancer sample 181LC. (E) Sequencing result showed the detailed genomic breakpoint of CCDC6-RET fusion in lung cancer sample 181LC. (F) A more comprehensive spectrum of oncogenic drivers is constructed for the cohort containing 52 lung adenocarcinomas from never smokers. (G) Detection of CCDC6-RET fusion in 24 “pan-negative” samples identified from the 202 lung adenocarcinomas from never smokers. Except for lung cancer sample 181LC, 261LC also harbors the CCDC6-RET fusion. LC: lung cancer. NL: normal lung.

Recently we have expanded the study of oncogenic mutation spectrum from the original 52 sample set to a large cohort with additional 150 samples and identified a total of 24 “pan-negative” lung adenocarcinomas from never smokers¹⁹. In an effort to detect the RET fusion in these “pan-negative” lung cancer samples, we found another lung cancer sample 261LC positive for CCDC6-RET fusion (Figure 1G). No other RET fusions including the recently reported KIF5B-RET fusion were found in these “pan-negative” samples²⁰ (data not shown). Interestingly, both patients with RET fusion are females; one is 46 years old with lung adenocarcinoma and the other is 61 years old with bronchioloalveolar carcinoma. Collectively, about 1% (2/202) of lung adenocarcinomas from never smokers harbors the CCDC6-RET fusion. Our data have not only provided a method for detection of novel oncogenic gene fusions but also constructed a more comprehensive map for oncogenic drivers in this subset of disease, which potentially helps develop the strategies for molecular-targeted therapies in clinic.

Experimental materials and methods are depicted in the Supplementary information, Data S1.

Supplementary Material

Supplementary information, Data S1

Materials and methods