Utilization of Peripheral Blood Cell-Free DNA in Myelodysplastic Syndromes: Clinical and Molecular Characteristics and Utilization for Genetic Analyses Using Conventional and Next-Generation Strategies

Background: Genetic mutations are detected in over 90% of patients with myelodysplastic syndromes (MDS). Some mutations may be critical not only for the pathogenesis of MDS, but also for disease progression to acute myelocytic leukemia and the acquirement of drug resistance; however, the detailed functions of these mutations remain unclear. To determine the roles of these mutations, genetic analyses using serially harvested tumor DNA are required. To date, genomic DNA from bone marrow (BM) cells has been utilized for genetic analyses of MDS patients, but bone marrow aspiration cannot be performed repeatedly because it causes physical pain in patients. Recently, peripheral blood cell-free DNA (PB-cfDNA) obtained from plasma and/or serum has received much attention as an alternative tumor DNA source, especially for solid tumors. Here, we demonstrate that PB-cfDNA may be used as an alternative DNA source instead of BM cells, that it reflects MDS disease status, and that the mutations in MDS can be successfully detected from PB-cfDNA by Sanger sequencing and next-generation targeted sequencing analyses.

Aims: 1) Confirmation of the molecular and clinical basis of PB-cfDNA in MDS; 2) detection of genetic mutations using PB-cfDNA with conventional and next-generation sequencing strategies; and 3) optimization of sequencing conditions using PB-cfDNA.

Methods: PB-cfDNA from patients with MDS and other related diseases (N = 33) and normal volunteer donors (N = 19) was analyzed. The concentration of PB-cfDNA was measured and correlations between the PB-cfDNAconcentration and clinical data were analyzed. DNA sequencing was performed using Sanger sequencing and targeted sequencing was performed using a TruSight Myeloid Sequencing Panel (Ilumina).

Results: The plasma PB-cfDNA concentration was significantly higher in MDS patients than in normal donors (p = 0.0405), and was significantly higher in those with a higher International Prognostic Scoring System (IPSS) score than in those with a lower score (p = 0.0339). The concentration of plasma and serum PB-cfDNA was significantly correlated with the serum lactate dehydrogenase (LDH)level (both p < 0.0001) and the blast cell count in PB (plasma, p = 0.0373; serum, p = 0.0274). Since PB-cfDNA showed a fragmented pattern reflecting its oligonucleosomal structure, amplification using primers for PCR-based sequencing analyses were optimized when the size of the PCR products were approximately 160 bp. For Sanger and targeted sequencing, 1 and 50 ng of PB-cfDNA, respectively, were required for each assay. Almost all genetic mutations detected by Sanger and targeted sequencing using BM cells from MDS patients (TET2, IDH2, SETBP1, U2AF1, SRSF2, NRAS, TP53) were also detected using the PB-cfDNA, but they were not detected in the germline controls. Serially harvested PB-cfDNA samples from a patient with refractory anemia with excess blasts-1 (RAEB-1) who underwent cord blood transplantation (CBT) after 5-azacytidine treatment were utilized for genetic analyses; they showed a correlation between PB-cfDNA concentration and the clinical course, and that the U2AF1 and SETBP1 mutations detected before CBT were no longer detectable 150 days after CBT.

Discussion: These data suggest that PB-cfDNA likely originates from MDS clones, which reflect their disease status, and that they can be utilized as a biomarker and an alternative promising source of tumor DNA instead of BM cells for genetic analyses; including not only conventional Sanger sequencing, but also next-generation targeted sequencing. However, since the PB-cfDNA concentration of some MDS patients was too low (less than 10 ng/1 mL plasma) to be used for targeted sequencing and/or whole exome sequencing, optimization of the sensitivity of sequencing analyses is still required.