Integrated Molecular Analysis of Myelodysplastic Syndromes Using Whole Genome Sequencing

Background

Myelodysplastic syndromes (MDS) are a heterogeneous group of myeloid malignancies characterized by refractory cytopenias with marrow dysplasia, which frequently progress to acute myeloid leukemia (AML). Although poorly understood in the previous era, the molecular events that underlie the pathogenesis of MDS have been intensively studied using advanced genomics in the past decade and are now fully catalogued into an array of well-defined functional pathways. However, mostly obtained through exome/targeted-capture sequencing, our knowledge about these molecular events is largely confined to those of single nucleotide variations (SNVs) and short indels, as well as arm-level copy number lesions, mostly within the coding sequences. Alterations in the non-coding regions, particularly a diversity of structural variations, in MDS genomes remain to be investigated in most part, even though the relevance of such lesions has recently been unequivocally demonstrated for other cancer types through large-scale whole genome sequencing (WGS) studies. Unfortunately, however, only a small number of MDS samples have been fully analyzes and inspected for genetic alterations using WGS.

Patients and Methods

In the present study, we performed an integrated, unbiased molecular study of 60 MDS cases, using whole genome sequencing (WGS) in combination with exome and transcriptome sequencing as well as methylome analysis. Paired tumor/germline DNA were obtained from patients' bone marrow and buccal smear samples. Sequencing data were analyzed using novel in-house pipelines, which were tuned to optimize detection of complex structural variations (SVs) and abnormalities in non-coding sequences. For some patients, multiple longitudinal materials were obtained along with their clinical course.

Results

WGS identified SNVs across the entire genome with a mean of 5.7/Mb/genome with a clear predominance of age-related C to T transitions, followed by other signatures. The spectrum of major targets of somatic mutations successfully recapitulated the previously reported one in MDS, including those involving splicing factors (SRSF2, SF3B1, U2AF1, and ZRSR2), epigenetic regulators (DNMT3A, ASXL1, TET2, BCOR, and EZH2), transcription factors (RUNX1, ETV6, and CUX1), signal transducing molecules (NRAS, KRAS, FLT3, PTPN11, CBL), and other critical molecules (TP53, NPM1, and STAG2). Moreover, other somatic variants within the coding regions were also identified that had already been reported in other human cancers but not in MDS, such as NCOR2X, MUC6, and TIAM2. The analysis of SVs unexpectedly revealed the complexity of MDS genomes. Most of the MDS genomes analyzed had a heavy burden of SVs including tandem duplications, deletions, translocations, and inversions, with a mean of 7.2/genome, which was far more than expected from conventional cytogenetics and array-based karyotyping. Complex rearrangements were common, frequently converging into particular chromosomes, suggesting multiple genetic events at a single genetic insult. Known targets of SNVs and indels were often affected by SVs, which largely escaped from conventional exome and targeted-capture sequencing, including RUNX1, TET2, FHITand other genes, suggesting that conventional platforms may substantially underestimate the frequency of alterations for some genes. Concomitant transcriptome analysis allowed to correlated abnormal splicing with somatic intronic events otherwise undetectable. Furthermore, comprehensive analysis of genomic aberrations in longitudinal samples enabled us to delineate the clonal architecture of the cellular population in MDS and their dynamics during the AML progression or clonal changes caused by AZA treatment.

Conclusions

Integrated molecular analysis using WGS and other platforms revealed the complexity of MDS genomes previously unexpected and reveal novel genetic alterations. Our results should help to extend our knowledge about the genomic landscape of MDS and provide novel insights into the molecular pathogenesis and clonal dynamics of MDS.