Introduction. High hyperdiploidy (51-67 chromosomes) and ETV6/RUNX1 -fusion are the two most common primary genetic events in pediatric B-cell precursor acute lymphoblastic leukemia (BCP-ALL), together found in more than half of the cases. In this study, we performed a proteogenomic analysis of these two subtypes, with the aim of understanding the interplay between somatic genomic aberrations, in particular copy number variants (CNVs), and gene and protein expression in leukemogenesis.
Materials and methods. Eighteen high hyperdiploid and nine ETV6/RUNX1 -postive childhood BCP-ALL cases were subjected to whole exome or whole genome sequencing, SNP-array analysis, RNA-sequencing, and proteome analyses using high-resolution isoelectric focusing liquid chromatography mass spectrometry. Furthermore, SNP array analysis and RNA-sequencing data for an additional 29 high hyperdiploid and 31 ETV6/RUNX1 -positive cases were analyzed.
Results. Expression data for 12,314 genes and 8,480 proteins were obtained, with 8,224 genes/proteins overlapping between the datasets. There was a clear linear relationship between the fold changes on the transcript and protein levels for high hyperdiploid cases over ETV6/RUNX1 -positive cases (Spearman correlation coefficient=0.62). For individual gene/protein pairs, genes and proteins involved in specialized functions such as hematopoiesis and amino acid metabolism showed high correlations, indicating a general regulation on the RNA level, whereas genes and proteins involved in basic cell functions such as the spliceosome, the ribosome, and oxidative phosphorylation showed lower correlations. Unsupervised principal component analyses showed separate clustering of high hyperdiploid and ETV6/RUNX1 -positive cases on both the RNA and protein levels. Dosage effects related to the chromosomal gains in high hyperdiploid ALL were clearly apparent, affecting genes both in cis (i.e., loci in the CNV) as well as in trans (i.e., loci outside the CNV). These effects were lower on the protein level than on the RNA level, showing additional layers of control for protein expression. A total of 2,627 genes and 1,098 proteins were differentially expressed between high hyperdiploid and ETV6/RUNX1 -positive BCP-ALL. Alternative splicing events were also investigated and 787 differentially expressed isoforms were identified. Allele-specific expression (ASE) analysis was performed to investigate the cis-effect of gene expression and 18 ASE events were observed in high hyperdiploid cases. Enrichment analyses on the whole gene/protein dataset showed upregulation of translation and proteasome-associated genes and proteins in high hyperdiploid cases, likely associated with an increased protein turn-over and possibly associated with an aneuploidy stress response. Overexpression of the cytokine receptors FLT3 and CRLF2 was also seen and the cytokine receptor interaction pathway was enriched. RAG1/RAG2-associated somatic recombination was enriched in ETV6/RUNX1 -positive BCP-ALL, in line with previous studies showing frequent activation of this pathway. Furthermore, pathways related to chromatin modifications and chromosome structure were markedly downregulated in high hyperdiploid BCP-ALL, suggesting a general deregulation of transcriptional control. In particular, low levels of CTCF and the cohesin complex, which are important regulators of chromatin structure, were seen in both gene and protein expression data.
Conclusions. In this study, we have characterized the proteogenomic landscape of the two most common forms of childhood ALL. Our data provides new insight into pathogenesis of high hyperdiploid and ETV6/RUNX1 -postive BCP-ALL and may be used to identify novel targets for treatment.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.