Detection of novel fusion-transcripts by RNA-Seq in T-cell lymphoblastic lymphoma

Abstract

Fusions transcripts have been proven to be strong drivers for neoplasia-associated mutations, although their incidence in T-cell lymphoblastic lymphoma needs to be determined yet. Using RNA-Seq we have selected 55 fusion transcripts identified by at least two of three detection methods in the same tumour. We confirmed the existence of 24 predicted novel fusions that had not been described in cancer or normal tissues yet, indicating the accuracy of the prediction. Of note, one of them involves the proto oncogene TAL1. Other confirmed fusions could explain the overexpression of driver genes such as COMMD3-BMI1, LMO1 or JAK3. Five fusions found exclusively in tumour samples could be considered pathogenic (NFYG-TAL1, RIC3-TCRBC2, SLC35A3-HIAT1, PICALM MLLT10 and MLLT10-PICALM). However, other fusions detected simultaneously in normal and tumour samples (JAK3-INSL3, KANSL1-ARL17A/B and TFG-ADGRG7) could be germ-line fusions genes involved in tumour-maintaining tasks. Notably, some fusions were confirmed in more tumour samples than predicted, indicating that the detection methods underestimated the real number of existing fusions. Our results highlight the potential of RNA-Seq to identify new cryptic fusions, which could be drivers or tumour-maintaining passenger genes. Such novel findings shed light on the searching for new T-LBL biomarkers in these haematological disorders.

Figure 1

Venn diagrams showing the number of fusion transcripts identified by the three detection methods (TopHat-fusion, Chimerascan and EricScript) in nine T-LBL samples and two control thymuses.

Figure 2

Validation of the fusion junction sequences of selected fusion transcripts by Sanger sequencing. Vertical black-bars indicate the fusion junctions. All validations were performed at the transcript level, with the exception of KANSL1-ARL17A that was validated both at transcript and genomic levels (d and e, respectively).

Figure 3

Levels of expression of 16 genes around the breakpoint of the RIC3-TCRBC2 fusion. LMO1 gene showed significant over-expression in comparison with its level of expression in control thymuses.