Induction of integration-free human-induced pluripotent stem cells under serum- and feeder-free conditions

Human-induced pluripotent stem cells (hiPSCs) have shown great potential toward practical and scientific applications. We previously reported the generation of human dental pulp stem cells using non-integrating replication-defective Sendai virus (SeVdp) vector in feeder-free culture with serum-free medium hESF9. This study describes the generation of hiPSCs from peripheral blood mononuclear cells to increase the donor population, while reducing biopsy invasiveness. From 6-d-old primary culture of peripheral blood mononuclear cells (PBMCs) with IL-2, hiPSCs were established using SeVdp(KOSM)302L with recombinant Laminin-511 E8 fragments under serum-free condition. The established PBMC-derived hiPSCs showed pluripotency and differentiation ability both in vivo and in vitro. In addition, we evaluated microarray data from PBMC- and dental pulp–derived hiPSCs. These hiPSCs will be beneficial for characterizing the molecular mechanisms of cellular differentiation and may provide useful substrates for developing cellular therapeutics.


Introduction
Human-induced pluripotent stem cells (hiPSCs) (Takahashi and Yamanaka 2006) research holds tremendous potential for regenerative medicine, drug discovery, and disease modeling with the added advantage of circumventing the ethical concerns regarding the use of pluripotent cells derived from embryos. Ever since Yamanaka and colleagues established a cell bank of donor-derived hiPSCs, rather than making them for each patient (Andrews et al. 2014), the demand for highquality hiPSCs has risen significantly.
hiPSCs are typically induced and cultured from feeder cells in serum-containing medium in conventional culture systems. Murine-derived feeder cells are widely used to maintain the pluripotency of hiPSCs, and human-derived feeder cells are also used. However, these cells are unsuitable for stem cell maintenance (Price 2017) as the feeder cells are cultured in medium supplemented with fetal bovine serum or proprietary serum replacements. The use of culture medium containing undefined or unknown components has limited the understanding of development and cellular differentiation (Nims and Harbell 2017).
Atsuko Hamada, Eri Akagi and Sachiko Yamasaki contributed equally to this work DNA-integrative retroviral and lentiviral vectors, first described by Yamanaka (Takahashi and Yamanaka 2006), have been used widely in cell reprogramming because they stably express transgenes, resulting from the chromosomal insertion of the vector (Takahashi et al. 2007) (Yu et al. 2007) (Lowry et al. 2008). However, the therapeutic potential of hiPSCs is complicated by the potential risks posed by continuous expression of transgenes and by random genome integration of viral vectors. Recently, a number of procedures have been introduced to generate genetically non-integrative or unmodified hiPSCs. These approaches involve chemicals or plasmid, episomal, or viral vectors (Kaji et al. 2009) ) (Zhou et al. 2009) (Jia et al. 2010) (Warren et al. 2010) (Yu et al. 2011); however, they show extremely low efficiency in generating iPSCs.
To prevent the risk of contaminating hiPSCs with unknown viruses, unknown substances, and unpredictable genome insertions, as well as to standardize a culture method under defined conditions, we previously developed and reported culture systems capable of maintaining both the undifferentiated status and pluripotency of embryonic stem cells (ESCs) and iPSCs without using serum or feeder cells or retroviruses from dental pulp cells (DPCs) (Yamasaki et al. 2016), eliminating the risk of activating nearby oncogenes by gene insertions. To further simplify the generation of hiPSCs and to increase the donor population, we report here the generation of hiPSCs using peripheral blood mononuclear cells (PBMCs) from seven healthy adult donors of various ages as sources of hiPSCs. This procedure was carried out without invasive and painful biopsies.
Based on our preliminary experimental results, we performed hiPSC generation from PBMCs using Laminin-E8 in completely serum-free culture conditions after 6 d of primary culture. The system can maintain the pluripotency of the cells and retain their potential to differentiate into all three embryonic germ layers, as well as DPC-derived hiPSCs (Yamasaki et al. 2016). We also compared microarray data from PBMCand DPC-derived hiPSCs. This culture system will be beneficial for evaluating the molecular mechanisms under defined conditions. Furthermore, the combination of this system and the generation of patient-derived iPSCs will contribute to attaining new knowledge to overcome various disorders.
Isolation and culture of PBMCs in serum-free medium We obtained human blood samples from healthy volunteers at Hiroshima University Hospital for the use of the blood to generate hiPSCs in accordance with approved guidelines. PBMCs were prepared by density gradient centrifugation in a Histopaque 1077 (Sigma-Aldrich) and cultured in RD6F serum-free medium supplemented with IL-2 (CELEUK, Takeda Pharmaceutical Co., Osaka, Japan) (Sato et al. 1987;Okamoto et al. 1996) for 0-6 d at 37°C in a humidified atmosphere of 95% air/5% CO 2 .
Induction of PBMC-hiPSCs with SeVdp(KOSM)302L An outline of experimental procedure for inducing PBMC-hiPSCs with SeVdp(KOSM)302L (Nishimura et al. 2017) is shown in Fig. 1. WT-PBMCs were transfected at a density of 1 × 10 5 cells with SeVdp(KOSM)302L, which does not integrate into the host genome, at an MOI of 6 for 2 h at 32°C in RD6F medium in a 48-well plate (BD Biosciences, Falcon®, Franklin Lakes, NJ) in a humid atmosphere of 95% air/5% CO 2 . The transfected cells were collected by centrifugation at 200×g for 5 min and then seeded onto a Laminin-E8 (0.5 μg/cm 2 ) (Nippi, Tokyo, Japan)-coated 6-well plate (BD Biosciences, Falcon®), in hESF9 medium at 38°C as described above. The medium was exchanged every other day. At approximately 14 d after transduction, detected colonies were mechanically picked with a 200-μL pipette and further cultured in a Laminin-E8-coated 4-well plate (Thermo Fisher Scientific, Waltham, MA) in hEFS9 with TGF-b1 or activin A. The PBMC-hiPSCs were passaged every 5-7 d by mechanical procedure as described above. The reprograming efficiency was determined as the alkaline phosphatase (ALP) positive colonies per total number of infected cells. For the control culture, transfected PBMCs were seeded onto mitomycin-C-treated mouse embryonic fibroblasts (MEFs) (Millipore) as feeder layer cells under KnockOut™ Serum Replacement (KSR) (Thermo Fisher Scientific)-based culture conditions.
ALP staining ALP staining was performed using a Fast Red substrate kit (Nichirei Biosciences, Inc., Tokyo, Japan) according to the manufacturer's protocol. Images of the dish were acquired using LUMIX (Panasonic, Osaka, Japan) and positive areas were detected using ImageJ software (National Institutes of Health, Bethesda, MD).
In vitro and in vivo differentiation of PBMC-hiPSCs The in vitro and in vivo differentiation of PBMC-hiPSCs was performed as described previously (Yamasaki et al. 2014).
Briefly, to confirm the in vitro differentiation capacity of PBMC-hiPSCs, an embryoid body assay was performed. Undifferentiated PBMC-hiPSCs were cultured in hESF6 without FGF2, heparin, and TGF-b1 or activin A in low-attachment 96-well plates (Sumitomo Bakelite Co., Ltd., Tokyo, Japan) for 4-5 d. Then, 3-5 embryoid bodies (EBs) were transferred to gelatin-coated 35-mm dishes and further cultured for another 21 d in hESF6. The medium was changed every 3-5 d. The cells were fixed and stained with the antibodies shown in Table 1. For in vivo studies, PBMC-hiPSCs were injected into the dorsal flank of SCID (CB17/Icr-Prkdcscid/CrlCrlj) mice (1 × 10 6 cells/100 μL of the cell suspension). Approximately 10 wk after injection, the teratoma were surgically dissected, fixed using phosphate-buffered saline solution containing 4% formaldehyde, and embedded in paraffin. Then each section was stained with hematoxylin/ eosin and Alcian blue/PAS. The histological findings were evaluated using a Nikon ECLIPSE E800 microscope (Tokyo, Japan) and photographed using a Leica DC500 camera (Leica Microsystems AG, Wetzlar, Germany).

RNA isolation and reverse transcription gene expression
Total RNA was isolated from PBMC-hiPSCs using TRIzol RNA Isolation Reagent (Thermo Fisher Scientific) according to the manufacturer's protocol. The cDNAs were synthesized from 1 μg of total RNA using high-capacity RNA to cDNA master mix (Applied Biosystems, Foster City, CA). RT-PCR was performed with KOD-FX neo (Toyobo, Osaka, Japan) using primers ( Table 2). The PCR products were sizefractionated by 1.5% agarose gel electrophoresis and imaged with the ChemiDoc Touch Imaging System (Bio-Rad, Hercules, CA). RT-qPCR was carried out on an AiraMx Real-time PCR system (Agilent Technologies, Santa Clara, CA) using FastStart Universal Probe Master (ROX) (Roche Diagnostic K.K., Basel, Switzerland). Each 10-μL reaction contained 5.0 μL of FastStart Universal Probe Master  Table 3), 0.2 μL of each primer (25 μM) (Roche); 0.5 μL of cDNA template (~25 ng/μL); and 4.0 μL of RNase-free dH 2 O. The cycle program for product amplification was 1 cycle of 95°C for 10 min (hot-start activation), followed by 40 cycles of 95°C for 30 s (denaturation), 55°C for 1 min (annealing), and 72°C for 1 min (extension).
Immunocytochemistry To examine the pluripotency of the PBMC-hiPSCs, immunocytochemistry was performed as described previously (Yamasaki et al. 2014

Results
Reprogramming efficiencies The obtained PBMCs on Days 0, 3, and 6 with interleukin-2 (IL-2) supplemented RD6F were u s e d f o r h i P S C c e l l r e p r o g r a m m i n g w i t h t h e SeVdp(KOSM)302L vector at an MOI of 6, using Laminin-E8 under completely feeder-free and serum-free culture conditions. After 25 d, we stained each well with an ALP staining kit (Nichirei) (Fig. 2). The reprograming efficiency was calculated as ALP-positive colonies/total number of transfected cells × 100 (%). Under serum-free culture conditions, the efficiency values were recorded as 0.024% in 0-d culture, 0.051% in 3-d culture, and 0.0033% in 6-d culture ( Fig. 2A). Under serum-supplemented condition, the values were 0% in 0-d culture, 0.0035% in 3-d culture, and 0.0025% in 6-d culture (Fig. 2B). Under serum-free conditions, the reprograming efficiency was 0.25-0.31% (DPC) and 0.008-0.1% (PBMC); under serum-supplemented conditions, these values were 0.30-0.44% (DPC) and 0.01-0.19% (PBMC), respectively. To ensure future applications and technical ease of handling, subsequent induction of PBMC-hiPSCs was performed on 6-d primary culture. Generation of integration-free PBMC-hiPSCs The schedule for PBMC-hiPSC induction is summarized in Fig. 1. At 14 d post-transfection of SeVdp(KOSM)302L into PBMC, small ESC-like colonies emerged (Fig. 1b, c), which grew to 2-3 mm, at which point several colonies were selected and mechanically passaged (Fig. 3). Each colony has a smooth edge and numerous tightly packed cells characterized by large nuclei and little cytoplasm.

Microarray analysis Cluster analysis was performed on
PBMCs cultured for 0, 3, and 6 d with IL-2-supplemented RD6F serum-free medium: PBMC-hiPSCs induced under serum-free condition (hESF9), PBMC-hiPSCs induced under serum-supplemented conditions (hKSR), and DPC-hiPSCs induced under serum-free conditions (hESF9). The gene expression patterns in PBMC group and PBMC-hiPSCs group were similar; PBMC-hiPSCs exhibited a very different expression pattern from that of PBMCs before induction, and PBMC-hiPSCs showed a similar expression pattern as DPC-hiPSCs (Fig. 9). Next, the gene expressions of PBMC-hiPSC isolated under different culture conditions were analyzed. WT4-PBMC iPSC induced in hESF9 displayed expression pattern more similar to that of WT3-PBMC hiPSC induced in hESF9 than that of WT4-PBMC hiPSC induced in serum-supplemented condition (Fig. 9). In other words, the gene expression pattern of PBMC-hiPSCs established under serum-free conditions is more similar to that of PBMC-hiPSCs established under serum-free conditions from another individual than that of those established under serum-supplemented conditions from the same individual. In summary, differences in culture conditions are concluded to have a greater effect on gene expression differences than differences in individuals from whom the cells were sourced.

Discussion
We previously reported induction of hiPSCs from fibroblasts derived from gingiva or DPCs derived from extracted teeth (Yamasaki et al. 2016;Yamasaki et al. 2014). However, the extraction of skin, mucous, or dental pulp involves surgical procedure, and, in particular, target patients may not have any teeth for dental pulp extraction. Since our eventual goal is to establish a general method to induce hiPSCs from genetic disorder patients, the donor population must be bigger. Hence, it is important that we improve hiPSC induction procedure by reducing biopsy invasiveness. However, the induction efficiency of hiPSCs from PBMCs is extremely low compared to that of skin fibroblasts using conventional induction methods (Staerk et al. 2010). We induced PBMC-hiPSCs multiple times using conventional methods but failed to obtain positive results. The

SeVdp(KOSM)302L non-integrating viral vector used in this
report encodes all four reprograming factors (Klf4, Oct3/4, Sox2, c-Myc) to achieve efficient factor expression. By incorporating this expression vector into a reprogramming protocol that included serum-free and feeder-free culture conditions, we induced and successfully recovered hiPSCs from PBMCs. The efficiency of hiPSC induction from PBMCs was approximately one half to one third of that from DPCs, but was sufficiently high for routine use. Induction of PBMC-hiPSCs has been reported using other viral expression vectors in combination with feeder cells in serumsupplemented medium. Brown et al. (2014), using serum-free conditions, also reported an induction efficiency of 0.005%. In each of these cases, the efficiency of inducing PBMC-hiPSCs was extremely low as compared to the results of the present study in which induction efficiency was 0.008-0.1% using SeVdp(KOSM)302L carrying four genes in one vector.
Previously, we examined a variety of extracellular matrix molecules (ECMs) including fibronectin, type I collagen, and gelatin for their suitability for DPC-hiPSC induction and confirmed that fibronectin was the most suitable (Yamasaki et al. 2014). Prior to this study, we re-examined PBMC-hiPSC induction by adding Laminin-E8 to previously characterized ECMs under serum-free condition. We successfully generated hiPSCs on all ECMs; in particular, colonies appeared more rapidly and the induction efficiency was higher in the presence of Laminin-E8 (supplement data). While Matrigel is also widely used in hiPSC induction protocols as ECM under serum-free medium, its use is potentially problematic because Matrigel contains many undefined factors that may affect cell growth and differentiation. Laminin is the main component of  Matrigel and Laminin-E8 is the minimum fragment conferring integrin-binding activity. Miyazaki have reported that Laminin-E8 has the potential to promote the survival of hiPSCs by robust adhesion via integrins and phosphorylation of AKT, ERK1/2, and FAK, which are highly phosphorylated in human pluripotent stem cells (Miyazaki et al. 2012). Thus, Laminin-E8 was adopted as ECM for our system.
Prior to inducing hiPSCs from blood cells, granulocyte macrophage colony-stimulating factor (GCSF) was added to stimulate the proliferation of specific T cells, B cells, and granulocytes using cytokines, such as stem cell factor and IL-3 (Loh, et al., 2009;Seki, et al., 2010). Based on a culture method similar to that used to activate lymphocytes for cell therapy performed on patients with oral cancer in our department, natural killer (NK) cells and cytotoxic T cells with strong cytotoxic activity were stimulated to proliferate with IL-2 in this study. By inducing hiPSCs from these cells, we also aimed to differentiate hiPSCs into cells with high cytotoxic activity for cell therapy. When inducing hiPSCs from PBMCs in RD6F medium containing IL-2, considerable individual differences were observed in induction efficiencies. This tendency was marked in reprogrammed cells cultured on MEF feeder cells. This suggests that lymphocytes with high cytotoxic activity attacked non-autologous mouse feeder cells or self-lymphocytes transfected with Sendai virus, resulting in reduced induction efficiency. Supporting this hypothesis, differences in marker gene expression determined by DNA microarray between individuals with the highest and lowest induction efficiencies revealed that individuals with lower induction showed higher expression of cytotoxic activity-related genes. Although hiPSC induction was carried out using activated lymphocytes with high cytotoxic activity with the goal of future cell therapy, when aiming to induce hiPSCs from PBMCs with high efficiency, induction should be performed immediately after isolation without IL-2 stimulation. In addition, the potential problems associated with isolating hiPSC from PBMC on mouse feeder cells can be avoided by using our feeder-free and serum-free culture method. Finally, our reprogramming protocol that does not use an integrating viral vector is preferable for generating hiPSCs for clinical applications. Our defined culture system is advantageous for studying the regulation of cell differentiation and for generating and maintaining pluripotent cells for clinical applications. The combination of this system and the generation of patient-derived iPSCs will contribute to providing us with new knowledge to overcome various disorders. Furthermore, disease-relevant cells differentiated from patient-derived iPSCs will become a powerful tool for investigating pathogenesis in vitro and accelerate the development of effective therapies.