Deconstructing myelodysplastic syndromes

In this issue of Blood, Hsu et al show that myelodysplastic syndrome (MDS) cells can be efficiently reprogrammed into induced pluripotent stem cells (iPSCs) to capture the clonal intermediates that appeared during MDS progression.¹ 

Myeloid malignancies such as MDS arise following the accumulation of multiple genetic (and potentially also epigenetic) changes in hematopoietic stem cells (HSCs) that confer a competitive advantage to mutant clones.²  Historically, human MDS has been challenging to study because of the reduced ability of MDS hematopoietic stem and progenitor cells (HSPCs) to grow in vitro, engraft immunodeficient mice, or give rise to cell lines. Although we and others have shown that MDS patient HSPCs can engraft in immunodeficient mice and recapitulate features of human disease, engraftment levels are generally low, and serial transplantation of disease, the gold standard assay for HSC function, remains challenging even with newer humanized immunodeficient mouse strains.³ 

Given these challenges, it is not surprising that there is an absence of studies in primary MDS patient samples that investigate the functional impact of the order of mutations or the consequences of additional mutations on HSC self-renewal and differentiation in the context of prior mutations. This represents an important gap in our knowledge because the order of mutations has been shown to have an impact on clinical outcomes in myeloproliferative neoplasms.⁴  Indeed, in MDS, it is generally thought that the initial mutations, frequently in epigenetic modifiers such as DNMT3a and TET2, may determine the ability of subsequent mutations to alter the self-renewal or differentiation capacity of HSCs and promote the expansion of specific clones. Next-generation sequencing methods often lack the sensitivity to detect early, rare subclones. Reprogrammed iPSCs enable functional analysis of premalignant clonal intermediates because iPSC clones generated from HSCs of MDS patients can be differentiated into various hematopoietic cell types. Although reprogramming of cancer and MDS cells has been difficult, in recent years, successful reprogramming of MDS and acute myeloid leukemia (AML) cells has been achieved, which offers investigators the opportunity to characterize such clones from individual patients with unique mutational profiles.^5,6 

In the Hsu et al study, the investigators demonstrated that reprogramming of CD34⁺ MDS HSPCs using nonintegrating episomes transiently expressing the Yamanaka reprogramming factors OCT4, SOX2, KLF4, and cMYC⁷  can more efficiently generate iPSCs derived from MDS premalignant clonal intermediates than those using a Sendai viral vector. Importantly, although the Sendai virus could more efficiently reprogram HSPCs from MDS patients, the episomal system was able to more efficiently reprogram MDS clones. This is a significant advance, given previous studies in cancer and MDS/AML demonstrating that reprogramming of mutant clones is significantly more difficult than reprogramming normal HSPCs.^5,8  Highlighting the utility of the reprogramming approach, the authors made the following novel observations with respect to MDS biology: (1) mutations in SF3B1 can occur as a second hit in patients with multiple mutations, and mutations in SF3B1 can cooperate with EZH2 mutations to impair mitochondrial function and induce apoptosis, resulting in ineffective erythropoiesis; and (2) del(5q) can occur as an early cytogenetic lesion in patients with complex karyotypes, and it likely compromises genome stability by inducing persistent DNA damage and cooperates with TP53 mutations to increase chromosomal instability.

Although the potential utility of this technique has been convincingly demonstrated in the article by Hsu et al, several important caveats must be considered when using iPSC models to study MDS biology. First, as exciting as these studies are, they were performed with a relatively limited number of primary patient samples, so it will be important to confirm the reprogramming efficiency advantages of the episomal system using more patient samples, including comparison with lentiviral vectors, which were not included in these studies. Second, because hematopoietic differentiation from iPSCs represents fetal liver as opposed to adult hematopoiesis, it is possible that somatic mutations captured in reprogrammed iPSCs (ie, in vitro differentiated hematopoiesis) may not have effects identical to those in MDS patients because of developmental stage–specific modifiers of MDS biology. Third, reprogramming induces changes in the epigenome that could cause unpredictable effects in differentiated cells, confounding the interpretation of experiments. However, this may not be important if enough of the disease-potentiating epigenome is recapitulated in hematopoietic cells induced to differentiate from iPSCs, as has been shown for AML patient reprogrammed cells.^5,6  Another important concern will be the inability of such systems to capture the effects of the bone marrow microenvironment in MDS biology. This problem certainly is not unique to studying hematopoiesis using reprogrammed MDS HSCs, but this is nonetheless a major concern because there is a growing body of literature implicating the microenvironment as a driver of both aging and MDS.⁹  Finally, given the inability to create engraftable HSCs from iPSCs generated from normal HSCs, it will remain difficult to study the effects of sequential mutations on HSC self-renewal. Of course, such a limitation may be overcome with the development of methods for generating engraftable HSPCs.

Overall, these studies confirm that reprogramming of clonal intermediates represents an important and robust approach to studying mechanisms of MDS progression and transformation. Given that the order of disease-initiating mutations can be resolved by using single-cell genetic profiling technologies,¹⁰  the future utility and widespread adoption of reprogramming approaches will depend on continued efforts to develop improved methods to reprogram and capture clones harboring MDS-associated genetic changes, such as the episomal system described by Hsu et al. Given the types of important disease biology insights revealed by their studies, we anticipate that reprogramming approaches will become more widely adopted in the field of MDS and other hematologic malignancies.