Structural Variations Identified By Whole Transcriptome Sequencing (RNA-Seq) in Childhood ALL with Different Response to Therapy

Introduction. Acute Lymphoblastic Leukemia (ALL) is the most frequent type of childhood leukemia. It is a multi-step process, characterized by the expansion of a pre-leukemic clone, accumulating cooperative genetic events required for the full transformation and clinical manifestation. Recently, the technological advances in genome-wide profiling techniques have allowed a better understanding of its molecular basis and heterogeneity. However, incidence and cure rates greatly differ among children, reflecting diverse responses to drug treatment and distinguishing risk groups. This defines the need for molecular investigations to better understand leukemia biology and improve risk prediction.

Aim. We applied a whole-transcriptome sequencing approach (RNA-seq) to characterize low- (LR) versus high-risk (HR) patients, to identify new genetic lesions associated to different early response to therapy.

Methods. Total RNA was extracted from primary leukemic blast samples of 10 pediatric ALL patients (4LR and 6 HR, according to minimal residual disease monitoring), included in the Italian AIEOP-BFM ALL2000 protocol. Genome-wide DNA profiling was performed by Affymetrix Cyto2.7M Arrays, RNA-seq was carried out by Illumina GAIIx platform, and validations were performed using independent approaches, such as RT-PCR and FISH. Fusion events were detected using FusionMap software, followed by a custom computational pipeline for the reduction of false positives and the identification of the most likely fusion candidates. Potential interest for leukemia was explored by testing the occurrence of these candidate fusions and con-joined genes in other RNA-seq datasets from different tumors and normal blood samples (i.e.: 15 melanomas, 2 melanocytes, 45 CEU individuals from 1K Genomes Project, plus 25 AML and 12 ALL).

Results. We sequenced the transcriptome of 10 childhood ALL cases, not carrying other clinical or genetic risk factors. We performed a comprehensive whole-transcriptome analysis, comprising identification of fusion transcripts, alternative splicing and SNPs. Priority was given to fusion transcripts, which could originate from intra- or inter-chromosomal rearrangements, since they might represent potential prognostic markers or therapeutic targets for personalized treatments. We identified 127 fusion candidates. Strikingly, 123 out of 127 events were identified as intra-chromosomal, 119 of which were involving two contiguous genes or with overlapping loci (the so-called “con-joined genes”). Among the four intra-chromosomal events, the NUP214-ABL1 fusion, previously found in T-ALL and responsive to kinase inhibitors, was here identified and validated in one HR B-ALL patient, thus opening new perspectives for targeted treatment options. Finally, among the four inter-chromosomal events, the novel PAX5-POM121C fusion was identified and validated in one LR patient. Both intra- and inter-chromosomal fusions resulted private or low-frequent events, not recurrent in other tumor types, nor in normal blood samples. Among the con-joined genes, we identified a subset of 22 events not present in melanoma nor in normal blood samples, but common to the external AML and ALL datasets.

Conclusion. RNA-seq represents one of the most comprehensive approaches to identify genetic alterations carried by leukemia clones. Our analyses identified novel fusion genes, originated by either inter- or intra-chromosomal rearrangements, as well as a considerable number of con-joined genes. Further evaluations will address SNPs, mutations, gene expression changes and splice variants that could be related to a different risk of relapse, and the feasibility of the screening of these candidates on a larger population of consecutive clinical cases.