1.12 Transcriptome Sequencing Reveals Novel Mutations and Differential Gene Expression in Stereotyped Subsets of Chronic Lymphocytic Leukemia

Advances in next-generation RNA-sequencing have revealed the complexity of transcriptomes by allowing both coding and non-coding (nc)RNAs to be analyzed; however, limited data exist regarding the whole transcriptional landscape of chronic lymphocytic leukemia (CLL). In this pilot-study, we evaluated RNA-sequencing in CLL by comparing 2 subsets which carry stereotyped B-cell receptors and vary in their clinical course, i.e. subset #1 (n = 4) and subset #4 (n = 4). Subset #1 cases are uniformly unmutated, express IGHV1/5/7 clan genes together with IGKV1-39/IGKV1D-39 genes and a stereotyped heavy complementarity determining region 3 (VH CDR3) of 13-14 amino acids. These patients have a significantly worse outcome compared to patients expressing the same IGHV genes but without stereotypy. In contrast, mutated subset #4 patients (IGHV4-34/IGKV2-30 usage and VH CDR3 of 20 amino acids) have a low median age at diagnosis, express surface IgG and have an indolent disease. Overall, RNA-sequencing generated 76-140 × 106 reads per sample and evaluation of gene expression, alternative splicing, ncRNAs and mutations was performed. Our analysis revealed that 153 genes were differentially expressed between the subsets, with subset #1 displaying an overall higher gene expression profile compared to subset #4. Five genes, ITGA4, TLR7, TNFSF8, WNT9A and LPL, were subsequently selected for validation using real-time quantitative PCR in 12 subset #1 and 11 subset #4 cases. Interestingly, expression analysis also revealed that 76 ncRNAs were differentially expressed e.g. SNORD48 and SNORD115, however in these instances the significantly higher expression was observed in subset #4 cases. In addition, use of the SplitSeek algorithm enabled detection of numerous subset-specific splice variants (n = 410), including a novel splice-isoform of MSI2. Subsequent mutational analysis revealed 16 -30 missense mutations/sample and such mutations were found in genes with a strong potential for involvement in CLL pathogenesis e.g. ATM and NOTCH2. For validation we selected 8 exonic/non-synonymous mutations and all mutations were confirmed by Sanger sequencing of genomic DNA. In conclusion, our study demonstrated that RNA-sequencing not only leads to the rapid detection of mutations within coding regions of genes potentially relevant to CLL pathogenesis, but also enabled us to identify differential subset-specific expression of genes, splice-variants and ncRNA. Evidently, such high-throughput approaches are vital to fill the gaps in our knowledge of CLL pathogenesis and the mechanisms behind disease progression.