Research Article: Mutational landscape of canine B-cell lymphoma profiled at single nucleotide resolution by RNA-seq

Date Published: April 24, 2019

Publisher: Public Library of Science

Author(s): Diana Giannuzzi, Laura Marconato, Luciano Cascione, Stefano Comazzi, Ramy Elgendy, Sara Pegolo, Alessio Cecchinato, Francesco Bertoni, Luca Aresu, Serena Ferraresso, Noel F. C. C. de Miranda.


The genomic landscape in human B-cell lymphoma has revealed several somatic mutations and potentially relevant germline alterations affecting therapy and prognosis. Also, mutations originally described as somatic aberrations have been shown to confer cancer predisposition when occurring in the germline. The relevance of mutations in canine B-cell lymphoma is scarcely known and gene expression profiling has shown similar molecular signatures among different B-cell histotypes, suggesting other biological mechanisms underlining differences. Here, we present a highly accurate approach to identify single nucleotide variants (SNVs) in RNA-seq data obtained from 62 completely staged canine B-cell lymphomas and 11 normal B-cells used as controls. A customized variant discovery pipeline was applied and SNVs were found in tumors and differentiated for histotype. A number of known and not previously identified SNVs were significantly associated to MAPK signaling pathway, negative regulation of apoptotic process and cell death, B-cell activation, NF-kB and JAK-STAT signaling. Interestingly, no significant genetic fingerprints were found separating diffuse large B-cell lymphoma from indolent lymphomas suggesting that differences of genetic landscape are not the pivotal causative factor of indolent behavior. We also detected several variants in expressed regions of canine B-cell lymphoma and identified SNVs having a direct impact on genes. Using this brand-new approach the consequence of a gene variant is directly associated to expression. Further investigations are in progress to deeply elucidate the mechanisms by which altered genes pathways may drive lymphomagenesis and a higher number of cases is also demanded to confirm this evidence.

Partial Text

Lymphoma is the most frequent hematopoietic cancer in dog and comprises a heterogeneous group of malignancies of varying severity [1]. According to the World Health Organization (WHO) [2], the most frequent B-cell lymphoma (BCL) is diffuse large B-cell lymphoma (DLBCL) by far having the highest incidence, followed by marginal zone lymphoma (MZL), Burkitt lymphoma and follicular lymphoma (FL) [3,4]. Under a clinical perspective, BCLs are also categorized in aggressive (DLBCL and Burkitt Lymphoma) and indolent (FL and MZL) lymphomas. However, both classifications reflect partially the biological behavior. Considering these limitations, recent investigations have described the molecular bases underlying canine BCL. Frantz et al. [5] performed the first gene expression profiling (GEP) study highlighting molecular similarities between DLBCL and MZL. Next, for DLBCL, molecular signatures involving specific signaling pathways resembling the human counterpart (i.e. NF-kB, PI3K-AKT and JAK-STAT signatures) were identified [6,7,8]. Analysis of copy number variations (CNVs) via array comparative genomic hybridization (aCGH) has revealed recurrent gains in chromosome 13 and chromosome 31 [8,9,10] which appeared correlated to treatment response and survival [10]. Regarding epigenetic deregulation, a recent genome wide-DNA methylation study revealed three distinct DLBCL subgroups with different outcome [11].

Improvement of NGS technologies has increased understanding of somatic mutations in both cancer origin and evolution, and knowledge about germline variations as risk factors for cancer development. A large amount of RNA-seq data were generated in the last years and, in addition to gene expression profiling, several bioinformatics pipelines were developed to render transcriptomic data suitable for variants identification [25,26]. Although the highest sensitivity and specificity in identifying genomic variants are ensured by merging genomic and transcriptomic data [25,27], the relative high coverage of RNA-seq enables to identify both germline and somatic SNVs in expressed genes at lower costs [25,26] and a recent study has also proved the high overlap and reliability of RNA-seq variant calling compared to whole exome sequencing (WES) data [28].