Malignant Progression of an MDS-Derived Cell Line Serves As an in Vitro Model for the Leukemic Evolution of MDS

Myelodysplastic syndromes (MDS) have a risk of progression to acute myeloid leukemia (AML), but the deterioration mechanisms of MDS and the alteration points still remain to be elucidated. We previously established a myelodysplastic cell line, MDS92 from the bone marrow of an MDS patient, and after a long-term interleukin(IL)-3-containing culture of MDS92, five blastic sublines including MDS-L were isolated. From MDS-L, we obtained two sublines, MDS-L-2007 and MDS-LGF after culture in the presence and absence of IL-3, respectively. To investigate the mechanism of leukemic evolution, we applied a next-generation sequencing (NGS) to the series of cell lines for comprehensive, comparative exome analyses, and searched for the origin of mutations by ultra-deep target sequencing of the original patient bone marrow.

Whole exome sequencing and ultra-deep target sequencing demonstrated: (1) TP53 mutation was found in the patient bone marrow and this mutation was inherited by all subsequent cell lines; (2) CEBPA mutation was originally present in a small fraction of the bone marrow; (3) NRAS mutation emerged by chance during IL-3-containing culture; (4) HIST1H3C(K27M) mutation (Histone-H3-K27M) was newly detected at the generation of MDS-L from MDS92. H3-K27M mutation was detected in MDS-L-2007 but not in MDS-LGF.

We focused on H3-K27M mutation because it is frequently found in pediatric brain stem tumors and recently found in a small population of AML cases (Lehnertz et al. Blood. 2017). MDS-L cells were a mixture of H3-K27M-mutant and wild-type clones. When MDS-L was cultured in the presence of IL-3, the proportion of H3-K27M-mutant fraction gradually increased. In contrast, when MDS-L was cultured without IL-3, the proportion of H3-K27M-mutant fraction gradually decreased. To investigate the implication of H3-K27M mutation, we tried single cell cloning from MDS-L and secured four wild-type clones and seven H3-K27M-mutant clones. In all H3-K27M-mutant clones, there was a marked reduction in H3-K27me3/2. Expression of a tumor-suppressor molecule p16 was reduced in six of the seven H3-K27M-mutant clones. H3-K27M-mutant clones showed rapid growth in the presence of IL-3, but cell proliferation was suppressed without IL-3. Competitive growth experiment by co-culture of H3-K27-wild-type and H3-K27M-mutant clones in the presence or absence of IL-3 showed that H3-K27M-mutant clones were predominant in the presence of IL-3, whereas wild-type clones were sustained comparatively in the absence of IL-3. Treatment with EPZ-6438, an inhibitor of H3-K27 methyltransferase EZH2, caused growth suppression of H3-K27M-mutant clones as well as wild-type clones and involved obvious recovery of p16 expression in H3-K27M-mutant clones, which provides a possibility that p16 might be a therapeutic target for H3-K27M mutant. Although GSK-J4, an inhibitor of H3-K27 demethylase JMJD3, was reported to inhibit H3-K27M-mutated pediatric brain stem tumors, GSK-J4 exerted only non-specific growth inhibitory effect on both H3-K27M-mutant and wild-type clones.

Whole exome analyses indicated that the accumulation of oncogenic mutations seemed to have led to establishment of MDS cell lines. The finding that growth advantage of H3-K27M mutant was influenced by the presence or absence of IL-3 raised a possibility that even if neoplastic clones emerge, their expansion might be influenced not only by genetic/epigenetic status but by surrounding environmental factors including cytokines. This series of cell lines will be a useful tool as an in vitro model for leukemic evolution of MDS.