Takayuki Takahata

An MDS-derived cell line and a series of its sublines serve as an in vitro model for the leukemic evolution of MDS

Myelodysplastic syndromes (MDS) are clonal hematopoietic stem cell disorders characterized by ineffective hematopoiesis and increased risk of progression to acute myeloid leukemia (AML). Occurrence of somatic mutations leads to formation of an abnormal clone with impaired differentiation, and additional driver mutations lead finally to AML [1]. We previously established a myelodysplastic cell line MDS92 from the bone marrow of an MDS patient with deletion of 5q chromosome[del(5q)] [2, 3]. MDS92 cells proliferated in the presence of interleukin(IL)-3 with a tendency for gradual maturation and represented karyotypic abnormalities including del(5q) (actually der(5)(5;19)), monosomy7, and a point mutation at codon12 of NRAS gene [2]. These characteristics are exclusively compatible with the property of MDS. Later, a blastic subline MDS-L was established from MDS92 [4]. MDS-L cell line has contributed to the molecular study of MDS with del(5q) [5, 6] and therapeutic mechanisms of lenalidomide [4, 7]. MDS-L has also been utilized as an in vivo model by transplanting experiments into immunodeficient mice [8] and for investigation of various drugs expected for treatment of MDS [9, 10]. In addition to MDS-L, we isolated several blastic sublines independently of one another from the parental MDS92, and further obtained MDS-L-2007 and MDS-LGF from MDS-L in the presence and absence of IL3, respectively (Fig. 1a). The morphology and surface markers of the cell lines are shown in Supplementary Figure 1 and Supplementary Table 1, respectively. In Fig. 1a, MDS-Gen implied the mid-stage cultured cells from bone marrow of the original patient (Pt-BM) for 3 months in the presence of IL-3 for aiming at establishment of MDS cell lines. We confirmed that the series of cell lines were all originated from Pt-BM by the short-tandem repeat profiling (Supplementary Table 2). We performed comprehensive, comparative exome analyses on the series of samples by the next-generation sequencing (NGS) method. Total effective reads per sample ranged from 35,548,000 to 58,711,000 and the average sequencing depth ranged from 46 to 77. Candidate somatic mutations, excluding dbSNP, were detected in 406 genes, and 178 mutations were found in common in all ten samples. Main results are summarized in Supplementary Table 3. As notable mutations, a homozygous, splice donor site mutation of TP53(c.672+1G>A; COSM6906 in COSMIC) was present in all samples, including Pt-BM, together with homozygous TP53(c.74+14T>C) mutation. MDS-L lost normal TP53 protein probably due to these mutations (Supplementary Figure 2). CEBPA(Q311stop) mutation (heterozygous; COSM29221) and NRAS(G12A) mutation (heterozygous; COSM565) were not detected in Pt-BM, but they emerged clonally at MDS-Gen stage, and both mutations were inherited by all subsequent cell lines. Whether CEBPA–mutant and NRAS–mutant clones were originally latent in Pt-BM or whether these mutations newly emerged during in vitro culture is an important question. To find the answer, we performed ultra-deep-targeted sequencing of CEBPA and NRAS in Pt-BM, and the results were as