Direct reprogramming of murine fibroblasts to hematopoietic progenitor cells

Abstract

Recent reports have shown that somatic cells, under appropriate culture conditions, could be directly reprogrammed to cardiac, hepatic, or neuronal phenotype by lineage-specific transcription factors. In this study, we demonstrate that both embryonic and adult somatic fibroblasts can be efficiently reprogrammed to clonal multilineage hematopoietic progenitors by the ectopic expression of the transcription factors ERG, GATA2, LMO2, RUNX1c, and SCL. These reprogrammed cells were stably expanded on stromal cells and possessed short-term reconstitution ability in vivo. Loss of p53 function facilitated reprogramming to blood, and p53(-/-) reprogrammed cells efficiently generated erythroid, megakaryocytic, myeloid, and lymphoid lineages. Genome-wide analyses revealed that generation of hematopoietic progenitors was preceded by the appearance of hemogenic endothelial cells expressing endothelial and hematopoietic genes. Altogether, our findings suggest that direct reprogramming could represent a valid alternative approach to the differentiation of embryonic stem cells (ESCs) or induced pluripotent stem cells (iPSCs) for disease modeling and autologous blood cell therapies.

Graphical abstract

Figure 1

Screen for Hematopoiesis-Inducing TFs (A) Schematic representation of experimental strategy. Murine embryonic fibroblasts (MEFs) were prepared from day 14.5 embryos, and murine adult fibroblasts (MAFs) were prepared from adult ear skin. Cells expressing the surface markers CD31, CD41, c-KIT, and CD45 were excluded from the starting populations. Sorted cells were transduced with a cocktail of all TFs. After 21 days of culture, cells were analyzed for hematopoietic cell-surface markers, clonogenic capacity, and cellular and nuclear morphology. (B) Bright-field images of untransduced and all TFs transduced MEFs and MAFs at day 12. (C) Relative gene expression levels of indicated genes with respect to β-actin in untransduced (Un), all-TF-transduced MEFs, control bone marrow (BM), and control E10.5 aorta-gonad-mesonephros (AGM) region cells. Data presented are representative of one out of three independent experiments performed in triplicate (n = 3; mean ± SD). (D) Number of hematopoietic colonies generated by 50,000 all-TF-transduced day 21 harvested MEFs/MAFs (left). Mean ± SEM from two independent experiments performed in triplicates is shown (n = 2). Representative bright-field images of the different types of colonies observed (right). (E) Cellular morphology of day 21 transduced MEFs/MAFs and ESC-derived hematopoietic cells. Arrowheads depict indicated morphologies. Gr, granulocyte; M, macrophage; P, progenitor; E, erythrocyte; Mk, megakaryocyte, Scale bars represent 50 μm. Asterisks indicate significant differences (Student’s t test; ^∗∗∗p < 0.0005).

Figure 2

Five TFs Induce Reprogramming to Blood (A) Bright-field images of five TFs transduced MEFs/MAFs at day 12 (left). Number of round cell colonies observed per 15,000 transduced MEFs/MAFs (right, n = 5, mean ± SEM). (B) FACS analysis of untransduced and five-TF-transduced MEFs and MAFs at day 21. (C) Number of hematopoietic colonies generated by 50,000 five-TF-transduced day 21 harvested MEFs/MAFs (left, n = 4, mean ± SEM). Representative bright-field images of the different types of colonies observed (right). (D) Cellular morphology analyses of day 21 transduced MEFs/MAFs and control bone marrow (BM) derived cells. Arrowheads depict indicated morphology. Gr, granulocyte; M, macrophage; P, progenitor; E, erythrocyte; Mk, megakaryocyte. Scale bars represent 50 μm. (E) Phagocytic capacity of CD45/CD11b double-positive cells. Asterisks indicate significant differences (Student’s t test; ^∗∗p < 0.01).

Figure 3

Multilineage Potential of Five-TF-Reprogrammed Cells (A) Acquisition of hematopoietic cell surface markers during the course of reprogramming. Average percentages of cells expressing CD41, c-KIT, and CD45 are represented at indicated intervals after five-TF transduction of MEFs (n = 2, performed in duplicate; mean ± SD). (B) Number of hematopoietic colonies generated by 50,000 reprogrammed MEFs from day 7 to day 21 (n = 2, performed in duplicate; mean ± SEM). (C and D) Multilineage potential of five TFs reprogrammed day 12 sorted c-KIT⁺ cells. (C) Sorted c-KIT⁺ cells were cultured under erythroid and myeloid culture conditions for 1 week and FACS analyzed for cell-surface markers. (D) Morphology of colonies and cells generated by sorted c-KIT⁺ cells. (E) FACS and cellular morphology of cells derived from day 12 sorted single c-KIT⁺ cell expanded on OP9 for 2 weeks (n = 3). (F) Percentage reconstitution in peripheral blood determined by detection of donor-derived GFP⁺ cells after 2 weeks of transplantation in two individual mice per group (1 and 2) either with GFP⁺ or c-KIT⁺ reprogrammed cells.

Figure 4

Reprogramming p53-Null MEFs (A) Number of round cell colonies observed at day 6 per 15,000 transduced WT, p53^−/−, and p16/p19^−/− MEFs (n = 4; mean ± SEM). (B) Number and types of hematopoietic colonies generated by 50,000 five-TF-reprogrammed p53^−/− MEFs harvested from day 6 to 15. Data presented are mean ± SEM of triplicates in a representative experiment (n = 2). (C) Morphology of colonies and cells obtained from five-TF-transduced day 12 sorted c-KIT⁺ p53^−/− MEFs. (D) Average percentage of TER119 and CD11b/GR1 double-positive cells generated by day 12 sorted c-KIT⁺ cells from WT, p53^−/−, and p16/p19^−/− transduced MEFs (n = 3; mean ± SEM). (E) Acetylcholinesterase staining of megakaryocytes derived from bone marrow or generated by day 12 sorted c-KIT⁺ reprogrammed p53^−/− MEFs. (Fi) FACS analysis of c-KIT⁺ sorted reprogrammed p53^−/− MEFs after expansion on OP9 in lymphoid medium. (Fii) BCR rearrangements detection in B220⁺/CD19⁺ sorted reprogrammed p53^−/− MEFs, control untransduced MEFs (Un) and spleen cells. (Gi) FACS analysis of c-KIT⁺ sorted reprogrammed p53^−/− MEFs after expansion on OP9-DL1 in lymphoid medium. (Gii) TCR rearrangement detection in sorted CD25⁺ reprogrammed p53^−/− MEFs, control untransduced MEFs (Un) and thymus cells. Asterisk(s) represents statistical significance (Student’s t test; ^∗p < 0.05, ^∗∗p < 0.01).

Figure 5

Fibroblasts Are Reprogrammed to Blood via an Intermediate Hemogenic Endothelial Stage (A–D) Relative gene expression levels of pluripotent (A), endothelial (B), endothelial/ hematopoietic (C), and hematopoietic (D) markers with respect to β-actin at indicated days after five-TF transduction and in control bone marrow (BM) and aorta-gonad-mesonephros (AGM) region cells. Data presented are mean ± SD from one representative experiment (n = 3). (E) Immunostaining of day 5 five-TF-transduced MEFs for CDH5 (red) and DAPI (blue). (F) Schematic strategy to determine hemogenic potential of CDH5⁻cKIT⁻ and CDH5⁺cKIT⁻ five-TF-transduced cells (top). Number, type, and morphology of hematopoietic colonies generated by sorted cells reaggregated with irradiated OP9 stromal cells. Data presented are mean ± SEM of triplicates of a representative experiment (n = 2). Scale bars represent 50 μm.

Figure 6

Global Transcriptome Analyses of Reprogrammed Cells Day 8 CDH5⁺ (CDH5-positive) and day 12 c-KIT⁺ (c-KIT-positive) flow-sorted cells were compared to untreated (Un) MEFs in duplicate by mouse Affymetrix exon arrays. (A) Hierarchical clustering of differentially expressed genes (DEGs) among untreated MEFs, day 8 sorted CDH5⁺, and day 12 sorted c-KIT⁺ cells. Genes specific to specified cell types are shown on the right. (B) Gene set enrichment analysis (GSEA) for endothelial gene signature in untreated, CDH5⁺, and c-KIT⁺ transcriptomes. (C) Normalized enrichment score (NES) values obtained after performing GSEA for indicated gene sets comparing CDH5⁺ and c-KIT⁺ transcriptome data sets. (D) Unsupervised hierarchical clustering of DEGs in our transcriptome data sets along with expression in HSCs from different hematopoietic organs (McKinney-Freeman et al., 2012).