"V体育2025版" Respecifying human iPSC-derived blood cells into highly engraftable hematopoietic stem and progenitor cells with a single factor

In Vitro Induction of MLL-AF4 Can Impart Self-Renewal and Lymphoid Potential to iPSC-Derived Blood Cells. (V体育2025版)

We optimized our previously established reprogramming method (18) to a feeder-free condition and derived iPSCs from peripheral blood (PB)-mobilized HSPCs (CD34-iPSC) and mononuclear cells (MN-iPSC). Pluripotency and normal karyotyping were verified on both iPSCs (Fig. S1 A–E). For in vitro differentiation of iPSCs toward hematopoietic cells, we also modified our prior strategy (19) to a feeder-free and xeno-free culture condition (Fig. 1A). The round-shaped hematopoietic cells could be detected from day 8, when they began to detach from the plate and reached the maximal number at day 12 (Fig. 1B). We then collected the detached cells at day 12 and performed colony-forming assay and flow cytometry analysis. Various colony-forming units could be observed by 2 wk of plating (Fig. 1C). Flow cytometry analysis showed that CD34-iPSC–derived blood cells had a relatively higher population of CD45⁺CD34⁺ cells (Fig. 1D and Fig. S1F), which might be ascribed to the epigenetic memory of iPSC related to its somatic cell source (20). Noticeably, the lineage distribution was invariably and predominantly skewed to CD33⁺ myeloid lineage, corroborating the myeloid-biased property of iPSC-derived blood cells under the regularly cytokine-directed differentiation condition.

Therefore, we constructed an MLL-AF4-GFP plasmid whose expression was doxycycline (Dox) dependent (Fig. S1G). MLL-AF4-GFP and reverse tetracycline response transactivator (rtTA) were cotransfected into iPSC-derived blood cells followed by a 3-d Dox induction (Fig. S1H). After sorting out GFP⁺ and GFP⁻ cell populations, we extended the induction for another week with cell number counting at each day. In striking contrast with GFP⁻ population, GFP⁺ cells significantly increased until reaching the plateau after 4 d of expansion (Fig. S1I). We next examined their self-renewal ability by the serial-replating assay. Within the 7 d of Dox induction after cell sorting, GFP⁺ and GFP⁻ cells were collected at day 1, day 4, and day 7 for the first round plating and continued with the secondary replating after 2 wk of colony forming in the methylcellulose medium. The total number of colonies formed by GFP⁺ cells obviously exceeded GFP⁻ cells (Fig. 1E). Specifically, the later collected GFP⁺ cells were able to form more colonies at the first round of plating, and these colonies could also give rise to more secondary colonies, demonstrating that the induction of MLL-AF4 could confer self-renewal potential to the targeted cells. Flow cytometry analysis was performed on GFP⁺ cells collected at day 4 of Dox induction. The increased CD34⁺ and CD43⁺ populations again revealed their enhanced stemness (Fig. 1F and Fig. S1J). More importantly, the CD10⁺ lymphoid population was remarkably increased and reached more than 10% among CD45⁺ cells for both CD34-iPSC and MN-iPSC. These revealed that in vitro induction of MLL-AF4 could impart self-renewal and lymphoid potential to iPSC-derived blood cells.

To gain more insights into their properties, we analyzed the transcriptomic signatures of MLL-AF4 transfected blood cells derived from iPSCs after 4-d induction of Dox. We also compared the RNA sequencing (RNA-Seq) data with mobilized HSPCs and the publicly available gene expression data for human cord blood (CB) HSCs and other progenitors (21, 22). Principal component analysis (PCA) placed MLL-AF4 transfected and nontransfected blood cells derived from iPSCs separately from each other, where the former group was closer to the lymphoid-primed multipotent progenitor and HSPC, while the latter group, to the common myeloid progenitor (Fig. 1G). This revealed the phenotypic transition from myeloid-biased progenitor state to multipotent stem and progenitor state after the induction of MLL-AF4. In addition, MLL-AF4 transfected blood cells derived from CD34-iPSC clustered closer to the bona fide HSCs than that of MN-iPSC, implying a more complete conversion of CD34-iPSC–derived HSPCs to the stem cell state. Gene set enrichment analysis (GSEA) indicated that MLL-AF4 could impart self-renewal and definitive hematopoiesis properties to the iPSC-derived blood cells (Fig. 1H). Correlation matrix further showed that these MLL-AF4 transfected blood cells clustered closest to bona fide HSCs than to multilymphoid progenitors (Fig. S1K). Genes that represented the key signatures of HSCs (1–4, 23) or directly regulated by MLL-AF4 (24) were significantly up-regulated in transfected blood cells (Fig. 1I and Fig. S1L), confirming the direct and potent influence exerted by MLL-AF4. These observations thus warranted our attempt to examine the in vivo reconstitution ability of MLL-AF4 transfected blood cells derived from both iPSCs.

MLL-AF4 Alone Is Sufficient for Enabling the Potent and Multilineage Engraftment of iPSC-Derived Blood Cells.

We next examined the engraftability of iPSC-derived blood cells with or without MLL-AF4 transfection. Given that MLL-AF4 was an oncogene, to avoid any potential risk of tumorigenicity, we enforced the transient expression of MLL-AF4 without integration by using either plasmid or mRNA transfection. We also introduced the TFs (EARSM, namely, ERG, HOXA9, RORA, SOX4, and MYB) that could facilitate the short-term engraftment of iPSC-derived myeloid progenitors (1) to assist MLL-AF4 and examine whether MLL-AF4 alone was sufficient for imparting engraftability to iPSC-derived blood cells. Newborn NOD-Scid-Il2rg^null (NSG) mice were used for xenotransplantation, given that they were more supportive for hematopoietic reconstitution and lymphopoiesis than adult mice (25). To induce the expression of EARSM and/or MLL-AF4 plasmid, both of which functioned in a Dox-dependent way, their transduced cells were induced by Dox for 48–72 h before transplant, and continued the induction by adding 2 mg/mL Dox to the drinking water of maternal mice for 2 wk so that the transplanted pups could acquire Dox through feeding (Fig. 2A).

Eight weeks after transplant, we measured the human cell chimerism in the bone marrow (BM) of recipient mice. Surprisingly, only the MLL-AF4 plasmid group (6/10) and EARSM plus MLL-AF4 plasmid group (6/10) showed a significant chimerism (>20%) of donor CD45⁺ cells (Fig. 2B). Nontransfected iPSC-derived blood cells failed to achieve the same engraftment on their own (0.18% of chimerism), even under the assistance of EARSM (0.38% of chimerism) or MLL-AF4 mRNA (0.22% of chimerism). Also, introducing EARSM did not much influence the effect of MLL-AF4 plasmid (both around 20% of chimerism), but adding MLL-AF4 plasmid to EARSM greatly influenced the outcome (from 0.38% to 20% of chimerism). In addition to the potent engraftment, the MLL-AF4 plasmid group also reestablished the multilineage hematopoiesis in the transplants (Fig. 2C). Myeloid bias was not detected as both B-cell (around 60%) and T-cell (around 1%) lineages were successfully reconstituted. Long-term HSC (CD34⁺CD38⁻CD90⁺CD49f⁺) compartment was also highly detectable within the CD34⁺CD38⁻ HSPC population. In addition to the BM, the multilineage hematopoiesis pattern could also be found in the spleen and PB of the recipient mice (Fig. 2D). Among recipients with engraftment (>1% of chimerism in the BM was considered engrafted), the CD34-iPSC group showed relatively higher chimerism than the MN-iPSC group (Fig. 2E). These observations suggest that only transient induction of MLL-AF4 without the assistance of other factors was sufficient to enable the potent engraftment of iPSC-derived blood cells. However, this effect was most likely dose dependent in terms of the different performance between the MLL-AF4 plasmid and mRNA. Since the decaying of in vitro synthesized mRNA in transfected cells was usually within several hours (26), its short lifetime appears to limit its effectiveness. Thus, we focused on the MLL-AF4 plasmid group for the following experiments.

iHSPCs but Not Primary HSPCs Are Prone to Leukemic Transformation During the Long-Term Posttransplant Period.

We continued to examine the long-term engraftment in the recipients of iHSPCs over 4 mo and observed that CD45⁺ donor cells dramatically increased in the BM of recipient mice, and the discrepancy of chimerism between CD34-iPSC and MN-iPSC disappeared as the time of posttransplant prolonged (Fig. 3A and Fig. S2A). By analyzing the long-term engrafted donor cells in the BM of the recipient mice at 16 wk posttransplant, we found a significant decrease on the myeloid lineage, and that the stem/progenitor and B-cell compartments markedly enlarged (Fig. 3B). From the shot-term to long-term engraftment, the increase of the B-cell compartment was consistently observable in the BM, spleen, and PB of the recipient mice (Fig. S2B), and splenomegaly was a common feature for most of the recipients during the long-term engraftment period (Fig. S2C). Moreover, when examining the cellular morphology of the BM cells, hardly any myeloid or erythroid elements could be observed; instead the marrows were overwhelmed with lymphoid cells with an abnormally high nuclear/cytoplasmic ratio (Fig. S2D). These abnormal lineage reconstitution patterns suggested a B-cell leukemia transformation might occur during the long-term engraftment period. To further investigate their properties, we also conducted secondary transplantation on newborn mice. Eight weeks after transplant, 3 of 10 recipients of the CD34-iPSC group and 2 of 10 recipients of the MN-iPSC group were dead. Five of 7 and 6 of 8 of the remaining recipients were successfully engrafted. In the BM, all of the engrafted recipients showed the same high chimerism of CD45⁺ cells (>60%), most of which coexpressed CD10 and CD19 with little or no expression of CD33 or CD3 (Fig. S2E). Splenomegaly was a common feature among the recipients, some of which also developed tumors in the intestine (Fig. S2F). Wright-Giemsa staining of the harvested BM cells showed a severe shortage of erythrocytes and myeloid cells, but presented uniform-looking lymphocytes with abnormally high nuclear/cytoplasmic ratio, as we observed in the primary recipients (Fig. S2G). Together, these observations revealed that the iHSPCs were prone to leukemic transformation during the long-term posttransplant period.

Given that persistent expression of MLL-AF4 in human CB HSCs was unable to trigger leukemia (15), it was intriguing that our transient induction of MLL-AF4 in iHSPCs could possibly lead to the leukemia transformation. It should be pointed out that we did not detect the existence of MLL-AF4 either in the in vivo-derived BM cells as early as 5 wk posttransplant or the in vitro transfected cells after a 2-wk induction (Fig. 2H), indicating that MLL-AF4 was only transiently introduced without being integrated into the genome of the targeted cells.

To further investigate whether our induction and/or transplant conditions were undesirably favorable for leukemogenesis, we conducted the same transplant experiments on primary HSPCs (PB mobilized CD34⁺ HSPCs). Nontransfected or empty vector transfected primary HSPCs were transplanted as the control group. As previously documented, the engraftment of mobilized HSPCs in immunodeficient mice was less efficient than that of CB or BM HSPCs (27, 28); only ∼5% of chimerism was detected in the BM of the control group (8/10) after 12–16 wk of transplantation, and much less in the PB (around 0.1%) and spleen (around 1%). In contrast, the MLL-AF4 transfected group (4/5) showed remarkably enhanced chimerism in the BM (33.1%), PB (7.66%), and spleen (14.2%) (Fig. 3C). To further interrogate whether MLL-AF4 transfected primary HSPCs underwent the leukemic transformation, we compared the hematopoietic reconstitution pattern between MLL-AF4 transfected and control groups. Among the engrafted CD45⁺ cells, the MLL-AF4 group had a fourfold increase in CD34⁺ stem cell compartment, in which the more primitive HSC population of CD34⁺CD38⁻ cells significantly expanded (Fig. 3D). However, induction of MLL-AF4 did not influence the lineage distribution, and the T-cell and myeloid cell compartments were maintained at the same levels as in the control group. Moreover, diverse cell types with normal morphology were detected in the harvested BM cells of MLL-AF4–treated group (Fig. S3A) with an absence of splenomegaly (Fig. S3B). To further confirm their “none-transformed” identity, we also performed secondary transplantation under the same condition as what we did for iHSPC recipients. Despite the HSPC expansion, lineage distribution was normal (Fig. S3C) and the harvested BM cells also showed diverse and normal morphology (Fig. S3D). Again, no splenomegaly was observed (Fig. S3E). Therefore, these results demonstrated that the current scheme of transfecting MLL-AF4 in vitro and inducing its expression in vivo did not inherently favor the leukemogenesis on the targeted cells.

To better understand the molecular features of the transformed leukemic cells, we harvested BM cells from primary recipients engrafted with MLL-AF4–induced iHSPCs at different stages and sorted out CD45⁺ human cells for RNA sequencing. By measuring the mutations and comparing with in vitro differentiated blood cells transfected with MLL-AF4 and control groups (normal HSCs and mononuclear cells), we found leukemic mutations that were closely related to B-cell acute lymphoblastic leukemia (B-ALL) were distinctly enriched in in vivo-derived cells but absent from in vitro-derived cells or control groups (Fig. 3E). In the period of short-term (9 wk posttransplant) and midterm (12 wk posttransplant) engraftment, canonical B-cell–associated genes such as PAX5 and IKZF1 were randomly mutated over the coding region, splicing region, and 3′ UTR with variant frequencies, however, only resulting in single nucleotide variations (SNVs) or some minor indels, which were not considered leukemogenic on their own (29). However, by 16 wk after the transplant, various and randomly occurred fusion transcripts became highly detectable in different recipients (Fig. S4), suggesting a more unstable status of the genome at this time in contrast to the early engraftment period. As consistent with the PCR results (Fig. S2H), RNA sequencing did not detect any MLL-AF4 transcripts in the engrafted BM cells, indicating that the leukemic cells were not likely transformed by MLL-AF4.

To understand the entirely different behaviors of MLL-AF4 transfected primary HSPCs, we also examined their gene expression profiles. PCA analysis showed that wild-type and transfected HSPCs strongly correlated with CB HSCs [from publicly available data (22) as previously described], all of which had very divergent components from those of the iPSC group and B-ALL control (Fig. 3F). Also, gene sets involved in stem cell self-renewal, proliferation, and cell-cycle phase transition were highly enriched in MLL-AF4 transfected HSPCs (Fig. S5A). However, despite their facilitated cell-cycle transition, features of G0/G1 and G2/M checkpoints were also enriched while DNA damage was not (Fig. S5B), indicating a more stable genetic status of primary HSPCs in contrast to that of iHSPCs. Quantitative PCR also verified the up-regulated expression of self-renewal related marker genes (Fig. S5C).

By performing Gene Ontology (GO) analyses on gene signatures representing the top 1,000 most differentially expressed genes in in vivo-derived leukemic cells relative to the in vitro differentiated blood cells from iPSCs, we found that the most enriched molecular features were highly related to cancer, and particularly B-ALL (Fig. 3G). Especially, the feature of DNA damage was highly scored, indicating the genomic instability of the engrafted cells, which accounted for the accumulation of genetic lesions ranging from site mutations to fusion transcripts. We also analyzed the most enriched GO features of engrafted primary HSPCs. By comparing with the iHSPC group, we found that those features containing genes that were mostly down-regulated were highly correlated with B-ALL and cancer, such as the B-cell receptor (BCR) signaling pathway, DNA damage, and proliferation (Fig. 3H). On the other hand, features containing genes that were mostly up-regulated were strongly related to multilineage hematopoiesis. These molecular discrepancies implied that induction of MLL-AF4 could incur different biological events in primary HSPCs and iHSPCs, which underlined their inherently distinct cell identity.