Non-transmissible MV Vector with Segmented RNA Genome Establishes Different Types of iPSCs from Hematopoietic Cells

Abstract

Recent advances in gene therapy technologies have enabled the treatment of congenital disorders and cancers and facilitated the development of innovative methods, including induced pluripotent stem cell (iPSC) production and genome editing. We recently developed a novel non-transmissible and non-integrating measles virus (MV) vector capable of transferring multiple genes simultaneously into a wide range of cells through the CD46 and CD150 receptors. The MV vector expresses four genes for iPSC generation and the GFP gene for a period of time sufficient to establish iPSCs from human fibroblasts as well as peripheral blood T cells. The transgenes were expressed differentially depending on their gene order in the vector. Human hematopoietic stem/progenitor cells were directly and efficiently reprogrammed to naive-like cells that could proliferate and differentiate into primed iPSCs by the same method used to establish primed iPSCs from other cell types. The novel MV vector has several advantages for establishing iPSCs and potential future applications in gene therapy.

This new non-transmissible and non-integrating measles virus vector, which can transfer multiple genes simultaneously into a wide range of cells through the CD46 and CD150 receptors and induce primed or naive-like pluripotent stem cells from hematopoietic cells in the same condition, will definitely contribute to the gene and cell therapy.

Introduction

The measles virus (MV), a member of the genus Morbillivirus in the family Paramyxoviridae, has a strong infectious capability in human epithelial cells, T cells, B cells, and dendritic cells, and it uses signaling lymphocytic activation molecule (SLAM, CD150)¹ or nectin4² as its receptor. MV is an enveloped virus with a non-segmented, negative-strand RNA genome. The genome contains six genes that encode the nucleocapsid (N), phospho- (P), matrix (M), fusion (F), hemagglutinin (H), and large (L) proteins,³ which are tandemly linked and separated by non-transcribed intergenic sequences. Each gene is transcribed sequentially but respectively toward the 5′ terminus by viral RNA-dependent RNA polymerase, which enters the single promoter at the 3′ terminus of the genome and recognizes the start and end sequences at each gene boundary. Since the initiation of transcription at each gene boundary is efficient but not perfect, the expression of downstream genes is less abundant than that of upstream genes. The H and F surface glycoproteins of MV are responsible for receptor binding and subsequent membrane fusion, respectively, and these proteins are used to retarget lentiviral vectors.⁴ Although MV causes a highly transmissible and severe acute systemic infection in humans, the infection can be controlled by effective vaccination; indeed, measles was eliminated using live attenuated vaccines in several countries.³

Previously, we generated MVs with two or three segmented RNA genomes.⁵ These MVs with segmented genomes could accommodate multiple extra genes or large genes. In previous work, we showed that certain substitutions in the H protein increase the binding efficiency to the CD46 receptor⁶ and that two substitutions in the M protein facilitate the assembly of infectious particles in infected cells via a strong interaction of the M protein with the cytoplasmic tail of the H protein.⁷ Based on these findings, we developed a novel viral gene transfer vector using the MV to achieve efficient and safe gene transfer in various types of human cells.

To evaluate the newly developed vector, we established induced pluripotent stem cells (iPSCs).⁸ Human iPSCs (hiPSCs)⁹ have properties similar to those of embryonic stem cells (ESCs), which grow rapidly and differentiate into three germline layers. These cells are valuable for improving our understanding of human disease and regenerative medicine, and they serve as a source for transplantation and pharmacology.¹⁰ The establishment of iPSCs from various tissues, including hematopoietic cells, was described previously.^11,12 Various methods have been described, including the use of the Sendai virus vector,¹³ which belongs to the same Paramyxoviridae family as the MV.¹⁴ Regardless of the involvement of viral or non-viral vectors in establishing iPSCs, ectopic expression of multiple genes, such as OCT4, SOX2, KLF4, MYC, and LIN28, for more than 2 weeks is necessary.¹⁵

Results

Construction of a Recombinant MV Vector

The MV-OKSLG vector, which expresses EGFP, human OCT3/4, SOX2, KLF4, and L-MYC, was constructed (Figure 1A). The MV-OKSLG vector was uniquely constructed to contain two RNA segments to enable the delivery of large or multiple transgenes.⁵ The genome-encoded H protein, termed H8, contains specific nucleotide substitutions (N390I, N416D, N481Y, and E492G)⁶ that allow it to use CD46, which is ubiquitously expressed in human cells, as a receptor. The M protein, termed M64/89, also contains specific substitutions (P64S and E89K) to reduce cytotoxicity associated with virus-induced cells to cell fusion and promote virus assembly.⁷ Removal of the F protein from the vector genome converted MV-OKSLG into a non-transmissible vector. MV-OKSLG transduced exogenous genes into >50% of human fibroblasts (BJ cells) or peripheral blood (PB)-derived stimulated T cells under our experimental conditions, and all these transgenes were detectable at 1 day after transduction and thereafter in both cell types (Figures 1B–1D). The expression levels of these genes decreased progressively depending on the order of the genes in the vector from the 3′ end to the 5′ end (Figure 1E). Transduction with increasing doses of MV-OKSLG also increased the GFP frequencies in both cell types, and there was no significant difference in transduction efficiency between the two cell types at each MOI (Figure 1F). Although transgene expression peaked at day 3 after transduction and decreased gradually thereafter, transgenes in more than half of the T cells were detectable after 1 week (Figure 1G), and a significant fraction of these cells expressed the transgenes for >2 weeks (data not shown).

Figure 1.

Multiple Genes Are Expressed from the MV Vector

(A and B) The MV-OKSLG vector (Le, leader sequence; Tr, Trailer sequence) (A) and GFP images 3 days (MOI = 3) after transduction into fibroblasts (right) and stimulated T cells (left) (B). Scale bars, 500 μm. (C and D) Expression of GFP (C) and four genes in GFP⁺ T cells (D) (red lines) compared with the non-transduced controls (black lines). (E) Relative expression (target/measles N gene). Data are presented as averages from three independent experiments. (F) GFP⁺ fibroblasts (white bars) and T cells (black bars) with MOI escalation (n = 4). (G) The persistence of GFP⁺ T cells (black line, MOI = 3; red line, MOI = 5) (n = 3). Data are presented as mean ± SEM.

Transduction Efficiency into Hematopoietic Lineages

Various types of human cells were transduced with MV-OKSLG, and a clear relationship between MV receptor expression and gene transduction efficiency was demonstrated (Table S1). Because high transduction efficiency into hematopoietic cells was observed, we compared the transduction efficiency of two vectors, the MV-OKSLG vector and the Sendai virus (SeV-GFP) vector, in primary human blood cells including T and B cells. Both MV-OKSLG and SeV-GFP showed high transduction efficiencies into CD14⁺ monocytes. In non-cultured CD3⁺ T cells, CD19⁺ B cells, and CD15⁺ granulocytes, transduction efficiency was higher for MV-OKSLG than for SeV-GFP, whereas CD56⁺ natural killer (NK) cells were resistant to MV-OKSLG transduction (Figure 2A; Figure S1A). However, non-cultured T cells, B cells, and NK cells expressed CD46 but not CD150 (Table S1). We focused on the high affinity of MV-OKSLG for T cell subsets.¹⁶ MV-OKSLG showed remarkable transduction efficiency into cord-blood (CB)-derived naive T cells compared with SeV-GFP (Figures S1B and S1C), and this higher transduction efficiency was also observed with peripheral naive and memory T cells (Figure 2B). CB-derived CD34⁺ cells expressing both CD150 and CD46 were also transduced with MV-OKSLG (Table S1). GFP expression was observed in more than 90% of CD34⁺ cells on day 1 and persisted for >4 weeks (Figure 2C); however, the number of GFP⁺ cells increased to peak level on day 14 and decreased rapidly thereafter (Figure 2D). During this period, the percentage of CD34⁺ cells decreased gradually due to their differentiation, dropping to 2% on day 14 and 0.2% on day 42 (data not shown).

Figure 2.

Transduction Efficiency of MV in Hematopoietic Cells

Fresh peripheral-blood-derived mononuclear cells were transduced with MV-OKSLG (MOI = 5) and SeV-GFP (MOI = 5) without cytokine stimulation. (A) Frequency of GFP⁺ cells in various hematopoietic lineages. *p < 0.01 (two-tailed t test); n.s., not significant. Data are presented as mean ± SEM. Each symbol represents a distinct experiment. (B) Frequency of GFP⁺ cells in the population of naive and stem cell memory (N) (CD45RA^high CD197^high), central memory (CM) (CD45RA^low CD197^high), effector memory (EM) (CD45RA^low CD197^low), and effector (E) (CD45RA^high CD197^low) cells. All data were obtained 2 days after transduction in cells from three healthy donors. *p < 0.01 (two-tailed t test). Data are presented as mean ± SEM. Each symbol represents a distinct donor. (C) GFP⁺ cell frequency of MV-infected CD34⁺ cells or MV-free controls over time. Each line represents an independent culture condition and a distinct experiment. (D) Time course of GFP⁺ cell number in total cultured cells from 10⁴ CD34⁺ cells in the presence of cytokines. Each line represents a distinct experiment.

Generation of Primed iPSCs from Differentiated Cells

BJ cells and stimulated T cells were transduced with the MV-OKSLG vector and cultured in human ESC maintenance medium. GFP-positive cell aggregates (Figures S2A and S2B) were formed from both cell types on day 20, and GFP-negative human ESC-like colonies (primed iPSCs) were generated from the GFP-positive cell aggregates on day 27 (Figures 3A, 3B, S3A, and S3B). Among these colonies, we analyzed three clones (-iPSC1, -2, and -3) from the transduced BJ cells and three clones (TMP-iPSC1, -2, and -3) from the transduced T cells. These iPSCs were cultured for more than 20 passages and maintained their GFP-negative ESC-like morphology (Figures 3C, 3D, and S3C); they expressed pluripotency markers but not MV genes (Figures 3E–3H, S2C, S2D, S2H, and S3D–S3I). The gene expression patterns of these iPSCs were similar to those of human ESCs (Figure 3L). Demethylation occurred in the original fibroblasts (Figure S2I), and no remarkable gene rearrangements or chromosomal aberrations were observed in these cells (Figures 3M and S2J). Moreover, these iPSCs underwent tridermal differentiation in vitro (Figures S2E–S2G and S3J–S3L) and in vivo (Figures 3I–3K). Each single T cell receptor repertoire was present in each TMP-iPSC, confirming that TMP-iPSCs were derived from individual terminally differentiated T cells (Figure S3M).

Figure 3.

Characterization of Fibroblast-Derived Primed iPSCs

(A and B) Morphology (A) and GFP images (B) of a primary colony (scale bars, 200 μm). (C and D) Morphology (C) and GFP images (D) of a colony after five passages (scale bars, 500 μm). (E–H) Expression of the following pluripotency markers in FMP-iPSC1: AP (E), NANOG (F), OCT3/4 (G), and SSEA-4 (H) (scale bars, 200 μm). (I–K) H&E staining of FMP-iPSC1-derived tissues, intestinal type columnar epithelium with goblet cells (endoderm, ×200) (I), cartilage (mesoderm, ×200) (J), and hair follicles with sebaceous glands (ectoderm, ×200) (K). (L) Comparison of the global gene expression patterns between the FMP-iPSC1 cells and ESCs (khES-3). (M) Representative karyotype of FMP-iPSC1.

Development of Naive-like Pluripotent Stem Cells

CB-derived CD34⁺ cells were also transduced with MV-OKSLG and cultured under the same conditions. In these cells, unlike T- or BJ-derived iPSCs, GFP-positive cell aggregates were observed at 2 weeks after transduction, and the aggregates grew and retained GFP expression for >4 weeks (Figures 4A and 4B). Colonies were harvested on day 24 and cultured on mouse embryonic fibroblasts (MEFs) in human ESC maintenance medium with human fibroblast growth factor (FGF)2, human leukemia inhibitory factor (LIF), a glycogen synthase kinase (GSK) inhibitor, and a MAPK/ERK kinase (MEK) inhibitor (2iL). Dome-shaped GFP-expressing colonies (MV-HPC-iPSCs) were observed on the MEFs. Completely dissociated single cells could be passaged over 20 times without Rock inhibitor¹⁷ (Figures 4C and 4D). The GFP-positive naive-like colonies (Figure 4E) induced the formation of GFP-negative human ESC-like colonies (Figures 4F and 4G) in response to a change to culture medium without 2iL. Furthermore, these lines were established from four independent healthy donors, including frozen or fresh CB cells. One representative line, naive-like iPSC (NL-iPSC)1, was analyzed. NL-iPSC1 cells expressed TFE3 in the nucleus, similar to mouse ESCs¹⁸ (Figure 4H), and pluripotency markers including NANOG, OCT3/4, and SSEA-4, similar to human ESCs (Figures 4I and 4K). In addition, some colonies expressed SSEA-1, similar to mouse ESCs (Figure 4L). The NL-iPSC1 cells also differentiated into the three GFP-negative germ layer tissues (Figures 4M–4O).

Figure 4.

Naive-like iPSCs Derived from CD34⁺ Cells

(A and B) Morphology (A) and GFP images (B) of a primary colony (scale bars, 200 μm). (C and D) Morphology (C) and GFP images (D) of a colony after five passages (scale bars, 200 μm). (E–G) Images of NL-iPSC1 on day 0 without 2iL (scale bars, 500 μm) (E) and the morphology (F) and GFP images (G) on day 14. (H–O) Staining for TFE3 and nuclei (red and blue, respectively) (H), NANOG (I), OCT3/4 (J), SSEA-4 (K), SSEA-1 (L), AFP (M), vimentin (N), and nectin (O). Scale bars: (H)–(L), 50 μm; (M), 100 μm; (N) and (O), 200 μm.

Gene Expression Properties of MV-HPC-iPSCs

We performed RNA-sequencing (RNA-seq) analysis on cell clones of the newly established TMP-iPSC4 cells, MV-HPC-iPSCs (NL-iPSC1 and NL-iPSC2), and MV-HPC-iPSC-derived primed (NL-iPSC1-primed) iPSCs. We compared the results with published datasets of naive and primed human pluripotent stem cells (hPSCs), including iPSCs and ESCs,^19,20 as well as human inner cell mass (ICM) cells,²¹ by principal-component analysis. We sorted three populations from TMP-iPSC4 and NL-iPSC2 based on GFP intensity (negative, low expression, and high expression) and examined them separately. However, we could not identify any differences among these three populations in TMP-iPSC4 or NL-iPSC2 (Figure 5A). NL-iPSC1 in the naive condition (knockout serum replacement [KSR] + basic FGF [bFGF] + 2iLIF) and NL-iPSC1-primed cells cultured under the primed condition (KSR + bFGF) had transcriptomes close to that of hiPSC-rRset²⁰ (iPSCs cultured in commercial RSeT media); however, NL-iPSC-primed cells were closer to primed hPSCs, whereas the three populations of NL-iPSC2 were clustered with hiPSC-rNHSM²⁰ (iPSCs cultured in NHSM media²²) (Figure 5A), suggesting that our NL-iPSC1 and NL-iPSC2 cells are similar to previously reported naive cells. Meanwhile, TMP-iPSC4 was very similar to primed hPSCs. Next, we addressed whether our iPSCs shared gene expression features of the naive or primed hPSCs. We identified 22 common genes that were the most differentially expressed in naive versus primed hPSCs by comparing their expression levels among three different datasets from ESCs and iPSCs19, 20, 21 (Figure 5B). Again, we found that our TMP-iPSC4 cells had profiles very similar to those of the primed hPSCs. Interestingly, gene expression in NL-iPSC2-GFP^negative and NL-iPSC2-GFP^low cells was shared with the primed hPSC group, whereas gene expression in NL-iPSC1-primed cells was close to that in the naive hPSCs, suggesting that these cells might be partially, but not completely, primed. NL-iPSC2-GFP^high and NL-iPSC1 cells were clustered with the naive hPSCs. These data suggested that GFP expression was an indicator of the naive-like state and that our GFP^high cells were very close to the naive state.

Figure 5.

Comparative Expression Analysis of Naive-like iPSCs

(A) Principal-component analysis of RNA-seq data from our primed (TMP-iPSC4-GFP^neg, TMP-iPSC4-GFP^low, TMP-iPSC4-GFP^high, and NL-iPSC1-primed) or naive-like (NL-iPSC1, NL-iPSC2-GFP^neg, NL-iPSC2-GFP^low, and NL-iPSC2-GFP^high) iPSCs, indicated in red, along with registered RNA-seq data from late blastocyst,²¹ morula,²¹ H9,¹⁹ H9-Rset,¹⁹ hiPSC-Primed,²⁰ hiPSC-rNHSM,²⁰ hiPSC-rRset,²⁰ hiPSC-rt2iLGoY,²⁰ and hiPSC-r5iLAF²⁰ cells. Data were performed on datasets normalized against primed cells of hESCs, H9 cells, or primed hiPSCs from each dataset. PC1 and PC2, the first two components. (B) Heatmap representation for 22 genes differentially expressed in primed and naive cells. The 22 genes were chosen from a database as reported previously. The color indicates relative expression level (log₂ fold change versus the expression level in primed cells). The primed or naive-like cell lines that we established are indicated in red.

Properties of the MV-OKSLG Virus Genome and Reprogramming Genes

To clarify the derivation of GFP expression in MV-HPC-iPSCs, we examined both RNA and DNA harvested from NL-iPSC1 by PCR using primers specific for the measles N and H genes (Table S2). Both the N and H genes were present as RNA (Figure 6A) but were not detected in DNA (Figure 6B), indicating that infecting MV-OKSLG existed as RNA but not as DNA (including chromosomally integrated genome DNA) in host cells. We also examined the relationship between expression of MV-OKSLG and that of transgenes. We quantified the expression level of transcripts in MV-HPC-iPSCs (NL-iPSC1 and newly established NL-iPSC3) and MV-HPC-iPSC-derived primed (NL-iPSC1-primed and NL-iPSC3-primed) cells using specific primers for the MV H gene or the GFP gene (Table S2). MV-HPC-iPSCs expressed MV-OKSLG vector genomes, whereas MV-HPC-iPSC-derived primed cells did not express any of these genes (Figures 6C and 6D). We also examined the expression of MV vector-derived exogenous reprogramming genes, as well as endogenous reprogramming genes. We detected expression of the three exogenous reprogramming genes in MV-HSC-iPSCs (Figures 6E–6G). Endogenous expression of these reprogramming genes was detected in both MV-HPC-iPSCs and MV-HPC-iPSC-derived primed cells (Figures 6H–6J). Interestingly, endogenous KLF4 was more highly expressed in these naive-like MV-HPC-iPSCs than in MV-HPC-iPSC-derived primed cells, whereas endogenous OCT3/4 and SOX2 were expressed at lower levels in the former than in the latter. This could reflect an important role of KLF4 in regulating naive pluripotency.²⁰

Figure 6.

Characteristics of Vector Genome with Reprogramming Genes in MV-HPC-iPSCs

(A) PCR products of measles N and H genes, with GAPDH gene amplified from cDNA of NL-iPSC1 or non-transduced peripheral blood mononuclear cells (PBMCs). An MV-OKSLG plasmid was used as a positive control. (B) Measles N and H genes in DNA. The HBB gene was used as an internal control. (C–J) Relative expression (target/GAPDH gene) of measles H (C), GFP (D), exogenous OCT3/4 (E), exogenous SOX2 (F), exogenous KLF4 (G), endogenous OCT3/4 (H), endogenous SOX2 (I), or endogenous KLF4 (J) genes in NL-iPSC1, NL-iPSC1-primed, NL-iPSC3, and NL-iPSC3-primed cells. (C–J) Data represent mean ± SD.

Comparisons with Frequencies of iPSC Generation between MV Vector and SeV Vector

Using MV-OKSLG or the commercially available SeV virus reprogramming vector (CytoTune-iPS2.0), we compared the efficiencies of iPSC generation between MV-OKSLG and SeV vector from the same cell sources, including stimulated or unstimulated T cells or CB-derived CD34⁺ cells (Table 1). Efficiencies were calculated by dividing the number of colonies by the number of cells seeded on MEF-coated dishes. We analyzed three colonies from each condition and confirmed that they were iPSCs by staining for NANOG. The efficiencies of establishing primed iPSCs from stimulated T cells, and establishing MV-HPC-iPSCs from CB-CD34⁺ cells using MV-OKSLG, were comparable (0.002%–0.012% in primed cells; 0.003%–0.005% in MV-HPC-iPSCs). The iPSC-establishing efficiency of MV-OKSLG was 3- to 10-fold lower than that of the SeV vector with stimulated T cells, and 30–70 times lower than when using CB-CD34⁺ cells. However, we obtained a few colonies from non-stimulated T cells using MV-OKSLG, but no colonies using the SeV vector, due to the low transduction efficiency of the SeV vector in these cells. Moreover, we obtained no naive-like colony from CB-CD34⁺ cells using the SeV vector. These results suggested that MV-OKSLG had some advantages for iPSC establishment from hematopoietic cells, e.g., non-stimulated T cells or CB-CD34⁺ cells, although iPSC generation by MV-OKSLG from stimulated T cells, which are the most common cell source for iPSCs, was less efficient than iPSC generation using the SeV vector.

Table 1.

Efficiencies of iPSC Establishment from the Same Cell Source in the Same Condition Using MV-OKSLG and SeV

Host Cells	Vector	Number of Seeded Cells	Number of Colonies	Efficiency (%)	Shape
IL2-Stimulated T Cells
Donor #1	MV-OKSLG	100,000	12	0.012	primed-like
	SeV	10,000	15	0.150	primed-like
Donor #2	MV-OKSLG	100,000	6	0.006	primed-like
	SeV	50,000	8	0.016	primed-like
Donor #3	MV-OKSLG	200,000	4	0.002	primed-like
	SeV	50,000	12	0.024	primed-like
No Simulated T Cells
Donor #1	MV-OKSLG	100,000	2	0.002	primed-like
	SeV	100,000	0	0.000	primed-like
Cord Blood Cells
Donor #1	MV-OKSLG	100,000	5	0.005	naive-like
	SeV	10,000	23	0.230	primed-like
Donor #2	MV-OKSLG	100,000	3	0.003	naive-like
	SeV	20,000	18	0.090	primed-like
Donor #3	MV-OKSLG	100,000	4	0.004	naive-like
	SeV	5,000	14	0.280	primed-like

For all entries, MOI = 5.

Discussion

We showed that the novel MV vector could transduce multiple genes simultaneously and express all transgenes for a period long enough to establish iPSCs. The new MV vector transduced genes into various human cell types using the CD46 ubiquitous receptor without damaging transduced cells. In particular, without any preliminary culture, the MV vector could transduce genes into various hematopoietic cells, including hematopoietic stem/progenitor cells (HPCs) and both naive T cells and stem cell memory T cells, which are more effective and suitable than other memory cells for gene therapies against cancer.^23,24 The removal of the F protein gene and substitution of the M protein gene from wild-type MV conferred the property of non-transmissibility on the MV vector and enforced and prolonged expression of transgenes. The MV vector required only 1 h to transduce genes into various cell types. The expression levels of transgenes decreased progressively according to the order of the genes in the vector. Although wild-type MV is highly pathogenic in humans, the non-transmissible MV vector can be used safely, because homologous recombination events are extremely rare and, indeed, may never occur among MV strains.25, 26, 27, 28 Our data (Figure 6B) also demonstrated that the chromosomal integration events did not occur in our MV-OKSLG-transduced cells, including MV-HPC-iPSCs expressing GFP.

Several methods for generating iPSCs from human hematopoietic cells have been developed. These systems include granulocyte-colony-stimulating-factor-mobilized human CD34⁺ cells using retrovirus vector,¹¹ human PB mononuclear cells (PBMCs) using lentivirus vector²⁹ or Dox-inducible lentivirus vector,¹² and human CD133⁺ cord blood (CB) cells using retrovirus vector.³⁰ Although the iPSC-establishing efficiencies of these methods were the same as that of MV-OKSLG (0.0002% to 0.02%), the transgenes from these retrovirus or lentivirus vectors are inserted into the host genome and may run the risk of tumorigenesis³¹ or changing the gene expression levels of some endogenous genes. Integration-free methods, such as transduction with plasmids,³² episomal vectors,³³ RNA,³⁴ or proteins,³⁵ require frequent transduction and might have low transduction efficiencies in hematopoietic lineages, resulting in low iPSC-generation efficiencies. The combination of stimulated T cells with SeV vector has proven to be the most efficient method for iPSC generation from hematopoietic cells.³⁶ Although MV and SeV belong to the same Paramyxoviridae family, the establishment of iPSCs with the MV vector occurred through a different process from that of the Sendai virus vector.¹³ In the case of the MV vector, the rank order of reprogramming gene expression levels was important for the generation of iPSCs.³² We tested another MV vector (MV-KOSLG) containing a different order of reprogramming genes (Figure S4A) and confirmed that all transgenes were expressed (Figure S4B). However, we were unable to generate iPSCs from BJ cells, stimulated T cells, or CD34⁺ cells. Cells transduced with MV-KOSLG failed to express SSEA-4 and Tra-1-60 on day 10 after transduction, suggesting that the reprogramming event in these cells did not occur.

The efficiency of ESC-like colony generation from fibroblasts or stimulated T cells using MV-OKSLG was lower, and the process required more time than that using the SeV vector³⁶ (the efficiencies of establishing both primed and NL-iPSCs with MV-OKSLG ranged from 0.002% to 0.024%). However, MV-OKSLG could establish primed iPSC without host genome integration similar to SeV. Because the receptors for infection differ between SeV and MV, iPSC establishment from certain cells was possible with MV-OKSLG but not with the SeV vector. For example, despite a very low efficiency, we successfully established primed iPSCs from unstimulated T cells, which was not possible using the SeV vector, as shown in Table 1.

One surprising feature was that MV-OKSLG could generate NL-iPSCs under the same conditions used for establishing primed iPSCs. The RNA-seq data in Figure 5 indicated that our NL-iPSCs were close to previously reported naive cells, although the qPCR data in Figures 6C–6J showed that exogenous genes were still expressed. Interestingly, we could not establish NL-iPSCs with the SeV vector under the same experimental conditions (Table 1). MV-HPC-iPSCs retained MV-OKSLG vector genomes even after iPSC establishment (Figure 6C) and expressed both exogenous and endogenous transgenes, although we could not identify any relationship between the exogenous gene transduction and the endogenous gene expression (Figures 6C–6J). This development did not require additional cytokines such as human LIF; inhibitors such as ERK1/2 inhibitors, GSK inhibitors, transforming growth factor (TGF)-beta inhibitors, or p38 inhibitors;^22,37 or the forced gene expression of NANOG, LIN28, KLF2, RAR-gamma, or LRH1,^38,39 and these genes were not expressed in cord-blood-derived CD34⁺ cells (data not shown). The proposed methods for establishing primed and NL-iPSCs from hematopoietic cells may improve our understanding of the events involved in the reprogramming of somatic cells to naive or primed pluripotent stem cells.

In general, the fates of MV-infected cells fall into three patterns. The first is MV-induced cell death by apoptosis or cell lysis; the second is the disappearance of viruses (abortive infection) due to the innate immune responses of infected cells; and the third is persistent infection by MV. Persistent infection is commonly observed with MV, and establishment of persistently MV-infected cells has been used as a standard method for the study of this virus.^40,41 We found that our primed iPSCs, FMP-iPSCs, TMP-iPSCs, and NL-HPC-iPSCs-primed lost GFP expression after establishment. However, no gene silencing events or mechanisms have been reported in paramyxoviruses, including MV. Moreover, recombinant MV exhibits very high stability of encoded transgenes, allowing transgenes to be stably expressed for more than 12 passages.⁴² In contrast to primed iPSCs, MV-HPC-iPSCs expressed GFP after several passages. Although further studies will be needed to clarify why GFP expression differs between primed and naive-like cells, all three types of infection were observed in MV-OKSLG-infected cells, and a certain portion of cells that were persistently infected with MV seemed to initiate reprogramming to become primed or NL-iPSCs.

Collectively, the new MV vector exhibited unique features, and it is a promising new tool for the development of novel gene therapies.

Materials and Methods

Establishment of MV Vector Plasmids

The genomes of MV-OKSLG consist of two RNA segments, MV-NPM-OSM and MV-HL-K-EGFP (Figures S5A and S5B). MV-NPM-OSM encodes the MV N, P/V/C, and M genes and OCT3/4, SOX2, and L-MYC genes. The vector’s M gene was engineered to include P64S and E89K substitutions to promote virus assembly (vector production) and reduce virus-induced cell-cell fusion. The M gene was termed M64/89.⁷ Compared with the IC323-EGFP genome,^43,44 the enhanced GFP (GFP) and F genes were replaced with the OCT3/4 and SOX2 genes, respectively. The H and L genes were replaced with the L-MYC gene alone. MV-HL-K-EGFP encodes the MV H and L genes and the GFP and KLF4 genes. The vector’s H gene was engineered to include N390I, N416D, N481Y, and E492G substitutions to enable the vector to use CD46 as a receptor.⁶ The H gene was termed H8.⁶ Compared with the IC323-EGFP genome, the MV N, P, M, and F genes were replaced with the KLF4 gene alone. The major restriction enzyme recognition sites used for cloning are indicated. The SOX2 gene was inserted using NruI and PshAI sites (blunt-ended sites), and these sites became unavailable in MV-NPM-OSM because of the SOX2 gene insertion.

Hematopoietic Cell Isolation

PB and CB samples from healthy donors were collected after obtaining informed consent, and experiments were approved by the ethical committee of Kyushu University (approval number: 22-82) and The Institute of Medical Science, The University of Tokyo (approval number: 27-32-0910). Granulocytes and monocytes were collected after the lysis of red blood cells using ACK buffer containing 8.29 g/L NH₄Cl (Nacalai Tesque, Kyoto, Japan), 1 g/L KHCO₃ (Nacalai Tesque), and 37 mg/L EDTA-2Na (Nacalai Tesque). MNCs from peripheral and CB were obtained using lymphocyte separation solution (Nacalai Tesque) as recommended by the manufacturer. HPCs were harvested from cord-blood-derived MNCs using a Dynabeads CD34 Positive Isolation Kit (Thermo Fisher Scientific, Waltham, MA, USA).

Cell Culture

Fibroblasts (BJ cells), 293T cells, SH-SY5Y cells, and BEAMS-2B cells were purchased from the American Type Culture Collection. MEF feeder cells were obtained from E13.5 C57BL/6 mice (Charles River Laboratories Japan, Kanagawa, Japan). The 409B2 iPSCs were obtained from the RIKEN BioResource Center. Fibroblasts, 293T cells, SH-SY5Y cells, and MEF cells were maintained in DMEM (Nacalai Tesque) containing 10% fetal bovine serum (FBS; Thermo Fisher Scientific). BEAMS-2B cells and BHK-T7/9 cells⁴⁵ (a gift from Dr. Ito and Dr. Sugiyama from Gifu University) were cultured in minimum essential medium-α (α-MEM; Thermo Fisher Scientific) with 10% FBS. Vero/CD150/F cells, which express human SLAMs and MV F proteins by transduction of pCAGGS-T7-F and pCXN2-SLAM into Vero cells, were maintained in DMEM with 7.5% FBS and 500 μg/mL G418 (Nacalai Tesque). These cells were cultured at 37°C in a 5% CO₂ incubator (Panasonic, Tokyo, Japan) and passaged every 4 days. All human ESCs and primed iPSCs were maintained on mitomycin C (Wako, Osaka, Japan)-treated MEF feeder cells in human ESC/iPSC maintenance medium (HEMM) (DMEM plus Ham’s nutrient mixture F-12 [DMEM-F12, 1:1; Nacalai Tesque] supplemented with 1 × 10⁻⁴ M 2-mercaptoethanol [2-ME; Nacalai Tesque], 2 mM L-glutamine [Nacalai Tesque], 1% non-essential amino acid solution [NEAA; Nacalai Tesque], 4 ng/mL human basic FGF2 (Peprotech, Rocky Hill, NJ, USA), and 20% KNOCKOUT Serum Replacement (Thermo Fisher Scientific). The primed iPSCs were passaged weekly. The culture conditions for NL-iPSCs were based on a previous report.⁴⁶ Briefly, NL-iPSCs were maintained on mitomycin C-treated MEF feeder cells in HEMM containing 3 μM CHIR99021 (Tocris Bioscience, Bristol, UK), 1 μM PD0325901 (Sigma-Aldrich, St. Louis, MO, USA), and 1,000 U/mL human LIF (Nacalai Tesque). The NL-iPSCs were cultured in an incubator at 37°C, 5% CO₂, and passaged every 4 days. Activated T cells were obtained from MNCs cultured with Dynabeads Human T-Activator CD3/CD28 (Thermo Fisher Scientific) in KBM 502 (Kohjin Bio, Saitama, Japan) or RPMI 1640 (Nacalai Tesque) containing 10% FBS and 175 U/mL Imunace (Shionogi, Osaka, Japan). Non-stimulated T cells and other hematopoietic cells were maintained in RPMI 1640 with 10% FBS. Cord-blood-derived CD34⁺ cells were cultured in StemSpan SFEM (Veritas, Tokyo, Japan) containing 100 ng/mL SCF (PeproTech), 10 ng/mL TPO (PeproTech), 100 ng/mL Flt-3L (Peprotech), SR-1 (StemCell Technologies, Vancouver, BC, Canada), and UM171 (STEMCELL Technologies). Mesenchymal stem cells were maintained in MF medium (Toyobo, Osaka, Japan).

MV Vector Production

The plasmids pCAGGS-T7-F, pCITE-IC-N, pCITE-IC-PΔC, pCITE-ko-9301B-L,⁴⁷ MV-NPM-OSM, and MV-HL-K-EGFP were transduced into BHK-T7/9 cells using Lipofectamine LTX (Thermo Fisher Scientific), and the medium was replaced with DMEM containing 10% FBS the following day. Two days later, the BHK-T7/9 cells were transferred onto Vero/CD150/F cells. Then, the Vero/CD150/F cells were cultured for approximately 20 days by passaging at 3-day intervals. After sufficient GFP expression was observed, the Vero/CD150/F cells were harvested and subjected to freezing and thawing. Then, the cell debris was removed by centrifugation at 1,500 rpm for 15 min. The vector-containing supernatants were obtained, frozen quickly, and stored at −80°C. The titers of MV vectors (transduction units per microliter) were determined by transduction into MRC-5 cells and counting of GFP-expressing cells at 2 days after transduction.

Gene Transduction

Sendai virus vectors (SeV-GFP: Plasm-Ex-AG; SeV-R: Cytotune-iPS 2.0; Medical & Biological Laboratories) were purchased and transduced as recommended by the manufacturer. All MV vectors were transduced as described in the following text. A total of 1 × 10⁵ cells were seeded onto a 0.1% gelatin (Sigma-Aldrich)-treated 12-well plate (Corning, Corning, NY, USA) and incubated at 37°C, 5% CO₂, for 4 h. Then, the culture medium was removed, and MV vector solution and fresh medium were added for a final volume of 500 μL. Then, the plates were centrifuged at 1,200 × g for 45 min to increase infection efficiency. After centrifugation, the supernatant was replaced with fresh medium and cultured at 37°C in a 5% CO₂ incubator.

Generation of iPSCs with MV-OKSLG Vectors

We generated TMP-iPSCs by transduction of CD3⁺ T cells with CMV-OKSLG vectors and NL-iPSCs from CB CD34⁺ cells. The procedures used to generate human primed iPSCs and NL-iPSCs using MV vectors were essentially as previously described,⁹ with minor modifications. Briefly, 1 × 10⁵ cells were transduced with reprogramming genes using the SeV-R or MV vector. Three days later, cells were harvested using 0.25% trypsin/EDTA (Nacalai Tesque) and seeded onto mitomycin C-treated MEF feeder cells in HEMM. The medium was replaced every other day until colonies appeared. Colonies were mechanically dissociated into small clumps and seeded onto mitomycin C-treated MEF feeder cells. To maintain the naive-like state, NL-iPSCs were cultured in HEMM supplemented with 2iL. To induce primed iPSCs (NL-iPSC primed), the cells were maintained in HEMM.

In Vitro EB Formation Assay

The in vitro embryonic body (EB) formation assay was performed as previously described with minor modifications.⁹ Briefly, human primed and NL-iPSCs were harvested using a cell scraper (Iwaki, Tokyo, Japan). The clumps were then transferred onto low-cell-binding 6-well plates (Thermo Fisher Scientific) and cultured in human ESC maintenance medium. The next day, the suspended clumps were harvested, and the medium was changed to human ESC maintenance medium without FGF2. The medium was changed every 3 days thereafter, and EBs were harvested at 14 days. The EBs were dissociated using 0.25% trypsin/EDTA and seeded onto Celltight-G 12-well plates (Sumitomo Bakelite, Osaka, Japan). Seven days later, the wells were washed two times with PBS followed by immunochemical staining.

Flow-Cytometric Analysis

To analyze MV vector-derived reprograming gene expression, MV vector-transduced T cells were permeabilized using a BD Cytofix/Cytoperm Fixation/Permeabilization Solution Kit (Thermo Fisher Scientific). The samples were then stained for 30 min at 4°C with anti-OCT3/4 (sc-5279, 1:50; Santa Cruz Biotechnology, Dallas, TX, USA), anti-SOX2 (sc-20088, 1:50; Santa Cruz Biotechnology), anti-KLF4 (ab72543, 1:50; Abcam, Cambridge, MA, USA), and anti-LMYC (AF4050, 10 μg/mL; R&D Systems, Minneapolis, MN, USA) and were sequentially stained with Alexa Fluor 546-conjugated anti-mouse immunoglobulin G (IgG) (1:500; Thermo Fisher Scientific), Alexa Fluor 546-conjugated anti-rabbit IgG (1:500; Thermo Fisher Scientific), and Alexa Fluor 546-conjugated anti-goat IgG (1:500; Thermo Fisher Scientific). The other samples were stained for 30 min at 4°C with phycoerythrin (PE)-conjugated anti-human PVRL4 (R&D Systems), allophycocyanin (APC)-conjugated anti-human CD3 (BioLegend, San Diego, CA, USA), APC-cyanin7 (Cy7)-conjugated anti-human CD4 (BioLegend), PE-Cy7-conjugated anti-human CD8 (BioLegend), PE-conjugated anti-human CD11b (BD Biosciences, Franklin Lakes, NJ, USA), APC-Cy7-conjugated anti-human CD14 (BioLegend), peridinin chlorophyll protein complex (PerCP)-Cyanin5.5-conjugated anti-human CD15 (BioLegend), APC-Cy7-conjugated anti-human CD19 (BioLegend), APC-conjugated anti-human CD45RA (BioLegend), PE-conjugated anti-human CD46 (Affymetrix-eBioscience, San Diego, CA, USA), PE-Cy7-conjugated anti-human CD56 (BioLegend), PE-conjugated anti-human CD150 (BioLegend), and PE-conjugated anti-human CD197 (BioLegend). PE-conjugated mouse IgG, PerCP-Cy5.5-conjugated mouse IgG, PE-Cy7-conjugated mouse IgG, APC-conjugated mouse IgG, and APC-Cy7-conjugated mouse IgG were used as isotype controls (all isotype antibodies were purchased from BioLegend). Cell populations were defined as follows: monocytes, CD14⁺ CD11b⁺; B cells, CD3⁻ CD19⁺ cells as B cells; T cells, CD3⁺ CD19⁻; neutrophils, CD15⁺; and NK cells, CD3⁻ CD19⁻ CD56⁺. Live cells were detected using 7-AAD (BD Biosciences). Labeled cells were analyzed using a BD FACSVerse system (BD Biosciences).

Alkaline Phosphatase Staining and Immunohistochemistry

Alkaline phosphatase staining was performed using an Alkaline Phosphatase Detection Kit (Merck Millipore, Billerica, MA, USA). For immunohistochemical staining, human primed iPSCs or EB cells were fixed with 4% paraformaldehyde in PBS (Nacalai Tesque) at 4°C for 30 min and then permeabilized and blocked in 0.1% Triton X-100 (Nacalai Tesque) containing 5% skim milk (Thermo Fisher Scientific) in PBS for 30 min on ice. The cells were first labeled with anti-human NANOG (AF1997, 10 μg/mL, R&D Systems), anti-human OCT3/4 (sc-5279, 1:50; Santa Cruz Biotechnology), anti-human SSEA-4 (sc-21704, 1:50; Santa Cruz Biotechnology), anti-human TRA-1-60 (sc-21705, 1:50; Santa Cruz Biotechnology) and anti-human TRA-1-81 (sc-21706, 1:50; Santa Cruz Biotechnology), anti-human α-Fetoprotein (MAB1368, 20 ng/mL; R&D Systems), anti-human vimentin (sc-6260, 1:50; Santa Cruz Biotechnology), anti-SSEA-1 (MAB4301, 1:50; Merck Millipore), anti-TFE3 (HPA023881, 1:50; Sigma-Aldrich), and anti-human nestin (MAB5326, 1:50; Merck Millipore) antibodies at 4°C overnight, followed by incubation with Alexa Fluor 546-conjugated anti-goat IgG (1:500; Thermo Fisher Scientific), Alexa Fluor 546-conjugated anti-mouse IgG (1:500; Thermo Fisher Scientific), or Alexa Fluor 546-conjugated anti-mouse IgM (1:500; Thermo Fisher Scientific) for 30 min at room temperature. The nuclei were then stained with DAPI (1:1,000; Dojindo, Kumamoto, Japan) and observed using a fluorescence microscope (BZ-9000; Keyence, Osaka, Japan).

Teratoma Formation and Histological Analysis

To produce teratomas, 1 × 10⁶ cells were inoculated into the testes of NOD.Cg-Prkdcscid Il2rgtm1Sug/Jic (NOG) mice (Central Institute for Experimental Animals, Kawasaki, Japan). After 9–13 weeks, the resected teratomas were fixed in 20% formalin and processed for paraffin sectioning. Teratomas were then stained with H&E.

PCR for Cellular DNA

Cellular DNA was extracted from MV vector-induced iPSCs and PBMCs using the DNeasy Blood & Tissue Kit (QIAGEN). MV-H, MV-N, and beta hemoglobin (HBB) were amplified by PCR and analyzed by 4% agarose gel electrophoresis. pUC19 DNA/MspI (HpaII) (Thermo Fisher) was used as a marker. Primers are listed in Table S2.

RT-PCR

Total RNA was extracted using the RNeasy Mini Kit (QIAGEN, Hilden, Germany), and cDNA was synthesized using a SuperScript III or IV First-Strand Synthesis System for RT-PCR (Thermo Fisher Scientific). Semiquantitative RT-PCR was performed with Takara Ex Taq (Takara Bio, Kusatsu, Japan) to determine the expression levels of genes of interest. All semiquantitative RT-PCR reactions were performed using a Mastercycler pro S (Eppendorf, Hamburg, Germany). qRT-PCR for Figure 1E was performed as described in the following text. Monolayers of Vero/CD150/F cells in 6-well cluster plates were infected with each MV vector and incubated for 2 days. Then, total RNA was extracted from virus-infected cells with the RNeasy Mini Kit and reverse-transcribed into cDNA using a PrimeScript RT Reagent Kit (Takara Bio). The cDNAs of the mRNAs were quantified using the LightCycler 480 SYBR Green I Master and a Light Cycler (Roche Diagnostics, Basel, Switzerland). qRT-PCR for Figures 6C–6K was performed in triplicate using Prime Time Master Mix (Integrated DNA Technologies, Coralville, IA, USA) and Prime Time qPCR assays (Integrated DNA Technologies) on an Applied Biosystem StepOnePlus Real-Time PCR System (Thermo Fisher Scientific). For each reaction, the ratio of each mRNA level relative to the level of GAPDH (Hs.PT. 39a.22214836; Integrated DNA Technologies) was calculated using the 2^−ΔΔCt method. All primer sets are shown in Table S2.

G-Band Analysis

The chromosomal G-band analyses were performed at Chromocenter (Tottori, Japan).

Bisulfite Sequencing

Genomic DNA was extracted using a DNeasy Blood & Tissue Kit (QIAGEN). The genomic DNA was treated with an EZ DNA Methylation-Gold Kit (Zymo Research, Irvine, CA, USA). The target regions of OCT3/4 and NANOG were amplified by PCR using ExTaq polymerase. The PCR products were cloned into the pGEM-Easy vector (Promega, Fitchburg, WI, USA) and sequenced. The primer sets were constructed as previously described,^9,48 as shown in Table S2.

TCR Rearrangement Analysis

The 5′ primers for TCR β gene amplification (BL-primer) were designed from leader signal sequences of human TCR β annotated by IMGT (http://www.imgt.org). The 3′ primers (BJ-1st or BJ-2nd primers) were designed from genomic DNA sequences of human TCR β locus sequences (http://www.ncbi.nlm.nih.gov/gene/6957). BL-primers bind to sequences encoding the leader peptide of TCR β. BJ-1st primers bind downstream of Jβ genes, and BJ-2nd primers bind upstream of BJ-1st primers. The BL-primer mix containing 39 BL-primers and one of 13 BJ-1st primers was used for the first PCR. In the second PCR, the ADP-primer, which binds to the adopter sequence at the 5′ end of all BL-primers, was used with one of 13 BJ-2nd primers. In the first PCR, 1 μL of 10 ng/μL genomic DNA from TMP-iPSCs was added to each PCR tube containing 19 μL of the first PCR mix (10 μL 2 × PrimeSTAR GC Buffer [Takara Bio], 6.5 μL nuclease-free water, 1.6 μL 2.5 mM dNTP [Takara Bio], 0.4 μL BL-primer mix [2 μM each], 0.4 μL 10 μM BJ-1st primers, and 0.1 μL 2.5 U/μL PrimeSTAR HS DNA Polymerase [Takara Bio]). The program for the first PCR was as follows: 98°C for 1 min and 35 cycles of 98°C for 10 s, 52°C for 5 s, and 72°C for 1 min. The first PCR products were diluted 10-fold in nuclease-free water and used for the second PCR. In the second PCR, 2 μL 0.1× first PCR products was added to each PCR tube containing 18 μL PCR mix (10 μL 2× PrimeSTAR GC Buffer, 5.5 μL nuclease-free water, 1.6 μL 2.5 mM dNTP, 0.4 μL 10 μM ADP primer, 0.4 μL 10 μM BJ-2nd primer, and 0.1 μL 2.5 U/μL PrimeSTAR HS DNA Polymerase). The program for the second PCR was as follows: 98°C for 1 min and 35 cycles of 98°C for 10 s, 52°C for 5 s, and 72°C for 30 s. The resultant PCR products were purified and sequenced. The BL primers are shown in Table S2.

Microarray Analysis

Microarray analysis was performed at the Cell Innovator (Fukuoka, Japan). Briefly, total RNA was isolated using the RNeasy Mini Kit. The RNA samples were quantified with an ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA), and the quality was confirmed with an Experion System (Bio-Rad Laboratories, Hercules, CA, USA). Aliquots of total RNA were amplified and labeled using a Low-Input QuickAmp Labeling Kit (Agilent Technologies, Santa Clara, CA, USA). Carbocyanin 3-labeled RNA was hybridized with SurePrint G3 Human Gene Expression Microarray 8×60K v2 (Agilent Technologies), and the signals were detected using a DNA microarray scanner (Agilent Technologies). Raw signal intensities and flags for each probe were calculated from hybridization intensities (gProcessedSignal) and spot information (gIsSaturated, etc.), according to the procedures recommended by Agilent. Flag criteria on GeneSpring Software are as follows: Absent (A): “Feature is not positive and significant” and “Feature is not above background”; Marginal (M): “Feature is not uniform,” “Feature is saturated,” and “Feature is a population outlier”; Present (P): others.

The raw signal intensities of two samples were log₂-transformed and normalized by a quantile algorithm with the “preprocessCore” library package⁴⁹ in Bioconductor software.⁵⁰ We selected probes that showed a “P” flag in both samples. To identify up- or downregulated genes, the Z scores⁵¹ and ratios (non-log-scaled fold change) were calculated from the normalized signal intensities of each probe for comparison between control and experimental samples.

Then, the criteria for regulated genes were established as follows: for upregulated genes, Z score ≥ 2.0 and ratio ≥ 1.5-fold; for downregulated genes, Z score ≤ −2.0 and ratio ≤ 0.66.

CGH Array

Comparative genomic hybridization (CGH) arrays were performed at the Cell Innovator. Briefly, genomic DNA was extracted from BJM-iPS1 and BJ cells using a DNeasy Blood & Tissue Kit. A total of 1 μg of each genomic DNA sample was labeled using an Agilent SureTag Complete DNA Labeling Kit (Agilent Technologies) and hybridized onto an Agilent SurePrint G3 Human CGH Microarray 1×1M (Agilent Technologies). The Human CGH Agilent array contains 974,016 coding and noncoding sequences (design ID, 035606). Slides were scanned using an Agilent Autofocus Dynamic Scanner (Agilent Technologies). Array data from the SurePrint G3 Human CGH Microarray 1×1M were analyzed with Agilent Genomic Workbench Lite Edition v7 (Agilent Technologies) using the built-in Aberration Detection Method 2 (ADM-2). Scanned image files were loaded into Agilent Genomic Workbench Lite Edition v6.5 (Agilent Technologies) and analyzed for aberrant calls by selection as follows: algorithm ADM-2 (Threshold 6 or 3), Fuzzy Zero: ON, Feature Filters: DefaultFeatureFilterON, Design Filters: DefaultDesignFilterv1 ON, GC Correction (Window size 2 kb): ON, Centralization (Threshold 6, Bin size 10): ON, Combine: Intra array replicate ON, Aberration Filter DefaultAberrationFilterv1: ON.

RNA-Seq and Data Analysis

For RNA-seq, iPSCs were sorted on a FACSAria into three populations: GFP-negative, GFP^low, and GFP^high; alternatively, the cells were directly collected from the culture dish. Total RNA was extracted using the RNeasy Plus Mini Kit (QIAGEN). RNA-seq libraries were prepared from total RNA using the NEBNext Ultra RNA Library Prep Kit for Illumina (New England Biolabs, Ipswich, MA, USA). Purification of reaction products at each step was performed using AMPure XP paramagnetic beads (Beckman Coulter, Pasadena, CA, USA). Sequencing was performed on an Illumina HiSeq 2500 sequencer.

Published sequencing datasets SRP011546,²¹ ERP006823,¹⁹ and SRP115256²⁰ were downloaded from the Sequence Read Archive (SRA) database. All read data were checked with FastQC (v0.11.7) (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and trimmed with Trimmomatic (v0.38).⁵² The trimmed data were processed through RSEM (v1.3.1)⁵³ and aligned to the hg19 reference with bowtie2 (v2.3.4);⁵⁴ then, raw read counts of each gene were obtained. The raw read counts of each dataset were processed through the edgeR package (v3.26.3)⁵⁵ of Bioconductor, and normalized counts per million (cpm) were generated. To compare all study datasets, cpm were normalized against primed samples of each dataset (compute log₂ fold change from the average of primed expression value). The resultant values were used in principal-component analysis and to generate the heatmap image.

Statistical Analysis

All data are presented as the mean ± SEM, and p < 0.05 was considered significant. Statistical analysis was performed using Prism software (GraphPad Software, San Diego, CA, USA).

GEO Accession Numbers

The GEO accession number for the FMP-iPS1 versus khES3 microarray data is GEO: GSE79604. The GEO accession number for the CGH array data is GEO: GSE79762. The GEO accession number for the RNA-seq data is GEO: GSE134717.

Author Contributions

T.H., M. Tahara, R.K., M. Takeda, H. Kohara, J.L., Y.S., and K.T. designed the study. Y.Y., Y.H., and K.T. obtained informed consent from peripheral blood and cord blood donors and provided the peripheral and cord blood cells. M. Tahara and M. Takeda established all MV vectors. A.R. provided critical cDNA clones of human genes and antibodies. T.H. and J.L. were involved in most of the iPSC experiments. T.H., Y.M., J.L., S.M., and M. Tahara produced the MV and determined the virus titers. T.H. and J.L. generated iPSCs using the MV vector. H.H., H. Kishi, and A.M. analyzed TCR rearrangement. H.T., K.I., and H.S. performed epigenetic experiments. S.O. and Y.O. analyzed teratomas from iPSCs. J.L. and Y.S. performed transcriptome analysis. T.H., M. Tahara, J.L., Y.S., M. Takeda, and K.T. wrote the paper.

Conflicts of Interest

K.T. obtained research grants from Neopharma Japan Co. Ltd. and Takara Bio Inc.