Epigenetic Memory Enables the Dominant Generation of Adult-Type Erythrocytes From Human Induced Pluripotent Stem Cells

It is well known that “globin switching” during erythropoiesis is associated with the pathophysiology of sickle cell anemia, as well as an approach to ameliorating some hemoglobinopathies. Furthermore, the process of globin switching represents an important paradigm for developmental gene regulation. Human embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs) are attractive tools for studying the ontogeny of human erythropoiesis because they exhibit in vitro differentiation toward various erythrocytes with embryonic (e), fetal (g) or adult (b) globin genes. However, earlier studies mainly showed that erythroid differentiation from human ESCs/iPSCs stopped at fetal erythrocytes and yielded few erythrocytes expressing b-globin (Chang, Blood 2006, 2010; Lapillonne, Haematologica 2010). In addition, successful differential reprogramming through somatic cell nuclear transfer and the use of iPSCs raised to the possibility that the poor yield of b-globin-expressing erythrocytes may result from incomplete genomic methylation or deregulated epigenetic modification. Recent studies using mouse iPSCs suggest the presence of “epigenetic memory” reflecting the tissue of origin. In that context, we attempted to establish an in vitro differentiation culture system that would preferentially yield adult-type erythrocytes derived from human iPSCs and would enable the study of b-globin gene switching during erythrocyte ontogeny. We initially established iPSCs from human dermal fibroblasts (HDFs) (f-iPSCs) or cord blood-derived CD34⁺/CD45⁺ cells (b-iPSCs) using retroviral vectors harboring human reprogramming factors (OCT3/4, SOX2, KLF4 and/or c-MYC). Gene expression analyses showed comparable patterns of gene clustering in f-iPSCs and b-iPSCs, and the b-globin gene was undetectable in either undifferentiated iPSC type. We then compared the hematopoietic colony forming capacities of 6 f-iPSC clones and 10 b-iPSC clones using a previously established culture system in which an “iPS-Sac” structure manifests an “in vitro hematopoietic niche” that generates multipotential hematopoietic progenitors. The b-iPSCs produced a higher number of iPS-sac structures and mixed-lineage or BFU-E colonies than f-iPSCs (Mixed colonies: 261±39.4 vs. 36±13.8, p<0.01; BFU-E: 192±35.4 vs. 11.8±6.7 per 1×10⁵ iPSCs, p<0.01). Moreover, as a result of terminal differentiation, erythrocytes were generated much more efficiently from b-iPSCs than f-iPSCs under erythrocyte-specific culture conditions. Finally, we selected two clones from each group to further analyze erythroid maturation and globin switching in erythrocytes generated from b-iPSCs and f-iPSCs. RT-PCR and immunochemical assays revealed limited differentiation by f-iPSC-derived erythrocytes (i.e., most were at the fetal stage), which is consistent with previous reports (Chang, Blood 2010), but b-iPSCs efficiently generated adult-type erythrocytes expressing b-globin (f-iPSCs, 21.0±4.7% vs. b-iPSCs, 54.7±4.2% b-globin⁺ p<0.01). In addition, the number of enucleated erythrocytes was higher from b-iPSCs than f-iPSCs. These data strongly suggest that genes regulating g- to b-globin switching are suppressed in f-iPSCs, possibly by epigenetic modification. Studies of the mechanisms underlying b-globin gene expression are clinically important, as they can provide the basis for potential gene therapeutic and reactivation strategies employing fetal globin genes to treat various hemoglobinopathies. Thus, b-iPSCs could be an abundant source of adult-type erythrocytes for use in clinical applications.

No relevant conflicts of interest to declare.