Repeated human deciduous tooth-derived dental pulp cell reprogramming factor transfection yields multipotent intermediate cells with enhanced iPS cell formation capability

Human tissue-specific stem cells (hTSCs), found throughout the body, can differentiate into several lineages under appropriate conditions in vitro and in vivo. By transfecting terminally differentiated cells with reprogramming factors, we previously produced induced TSCs from the pancreas and hepatocytes that exhibit additional properties than iPSCs, as exemplified by very low tumour formation after xenogenic transplantation. We hypothesised that hTSCs, being partially reprogrammed in a state just prior to iPSC transition, could be isolated from any terminally differentiated cell type through transient reprogramming factor overexpression. Cytochemical staining of human deciduous tooth-derived dental pulp cells (HDDPCs) and human skin-derived fibroblasts following transfection with Yamanaka’s factors demonstrated increased ALP activity, a stem cell marker, three weeks after transfection albeit in a small percentage of clones. Repeated transfections (≤3) led to more efficient iPSC generation, with HDDPCs exhibiting greater multipotentiality at two weeks post-transfection than the parental intact HDDPCs. These results indicated the utility of iPSC technology to isolate TSCs from HDDPCs and fibroblasts. Generally, a step-wise loss of pluripotential phenotypes in ESCs/iPSCs occurs during their differentiation process. Our present findings suggest that the reverse phenomenon can also occur upon repeated introduction of reprogramming factors into differentiated cells such as HDDPCs and fibroblasts.


Results
Generation of HDDPC-derived iPSCs. We successfully obtained 6 HDDPC lines (termed 'P01-P06') from the patients, which were all found to exhibit low staining when assessed by histochemical analysis for ALP activity, except P01 (Table 1). We first examined whether HDDPCs could be reprogrammed to form iPSCs after single transfection with Yamanaka's four reprogramming factors. The time-line for iPSC generation is shown in Fig. 1a. After single transfection of HDDPCs by electroporation, they were maintained in α-modified minimum essential medium (MEMα)/20% foetal bovine serum (FBS) for 15 days and thereafter in iPS medium for more 15 days. Of 6 HDDPC lines tested, only one (P01) exhibited distinct iPSC colonies at about 20 days after transfection ( Fig. 1b-i,ii; Table 1). Immunohistochemical staining of these established iPSC colonies with antibodies demonstrated that they expressed stem cell markers such as OCT3/4, SOX2, TRA-1-60, and SSEA-4 ( Fig. 1c). Furthermore, when outgrowth was allowed in the tissue culture dish from P01-derived EBs, the spread cells from the EBs exhibited various types of differentiated cells comprised of three germ layers (endoderm, mesoderm, and ectoderm) (Fig. 1d). However, the other 5 lines (P02-P06) retained a similar morphology as their parental cells even after 30 days following the first transfection.
We considered that the failure in acquiring iPSCs from the remaining 5 HDDPCs was because they had not yet reached the condition at which they could be readily reprogrammed to form iPSCs after transfection with reprogramming factors. Notably, Yan et al. demonstrated that repeated transfection with the reprogramming factors could increase the efficiency of iPSC generation 4 . We therefore tested this possibility by assessing whether the 5 lines (P02-P06) that were refractory to iPSC generation after single transfection could be successfully converted to form iPSCs upon repeated transfection with the reprogramming factors. For this purpose, cells were first transfected with Yamanaka's four reprogramming factors, then at 5 days after transfection were harvested and subjected to the 2nd round of transfection using the same factors. These doubly-treated cells were maintained in MEMα/20% FBS for 5 days and then in iPS medium for another 15 days (Fig. 1a). Of the five HDDPC lines tested, three (P02, P04, and P06) exhibited formation of iPSC-like colonies approximately 20 days after the 2nd-round transfection ( Fig. 1b-iii; Table 1). The other lines (P03 and P05) retained a form of fibroblastic morphology even after 30 days after the 2nd transfection. For exploring the possible conversion of these cells to iPSCs, a 3rd round of transfection was performed using the P05 line, as shown in the time-line shown in Fig. 1a. Approximately 20 days after the 3rd transfection, P05 successfully exhibited formation of iPSC-like colonies (Fig. 1b-iv; Table 1).

Generation of intermediate stem-like cells from primary HDDPCs.
From the results of the previous experiment, it was shown that the HDDPCs that failed to convert to iPSCs after the single transfection with Yamanaka's four reprogramming factors could be successfully reprogrammed to form iPSCs when subjected to Scientific RepoRts | (2019) 9:1490 | https://doi.org/10.1038/s41598-018-37291-2 repeated transfections. These findings suggest that HDDPCs (which are refractory to iPSC generation after single transfection) come to reach the condition at which cells are easily reprogrammed to iPSCs after repeated transfection. In other words, they are in an intermediate state between fully differentiated cells such as parental HDDPCs and iPSCs. Notably, ALP expression is closely associated with undifferentiated immature cells including somatic stem cells and ESCs/iPSCs 4,10 . Furthermore, after transfection with the reprogramming factors, ALP-positive cells first appear prior to the formation of OCT3/4 positive iPSC-like colonies 10 . This suggests the potential of ALP as a useful marker for defining cells at an intermediate state.
Therefore, we next examined the possible appearance of ALP-positive HDDPCs after transfection with Yamanaka's four reprogramming factors. In this case, we employed the P02 line, which is negative for ALP activity and found to be refractory to iPSC formation after the 1st-round transfection with the reprogramming factors (Table 1). On 3, 5, 7, and 9 days after transfection, cells were fixed and subjected to cytochemical staining for ALP activity (Fig. 2a). Gradual increase in ALP activity was clearly discernible as days in culture proceeded (upper columns in Fig. 2b). For the 2nd-round transfection, cells at 5 days after the 1st transfection were harvested and immediately subjected to retransfection. On 3, 5, 7, and 9 days after the 2nd-round transfection, cells were fixed for the detection of ALP activity. As shown in the case of the 1st transfection, a gradual increase in ALP activity was also seen as days in culture proceeded (middle columns in Fig. 2b). This pattern was observed as well when triple transfections were performed using the same line (lower columns in Fig. 2b). The number of ALP-positive cells/well in the double and triple transfections groups was significantly higher than that in the single transfection group when cytochemical staining was performed on 7 and 9 days after transfection (Fig. 2c). Notably, the proliferation rate in the single transfection group was also higher than those in the double and triple transfection groups (Fig. 2d). These findings suggest a correlation between repeated transfections with the reprogramming factors and elevated numbers of ALP-positive cells.
To test whether cells growing after repeated transfections with the reprogramming factors might also express other pluripotency-rerated markers such as OCT3/4, SOX2, NANOG, KLF4, and TRANSP, reverse transcription-polymerase chain reaction (RT-PCR) analysis was carried out using the samples (P02) harvested on 9 days after single, double, or triple transfection. Increased expression of these pluripotency-rerated markers was already discernible in HDDPCs 9 days after the single transfection (Fig. 2e). The levels of these mRNAs appeared to be almost equivalent to that found in the established iPSCs 4,11 . Notably, OCT3/4 and TRANSP tended to show an increase in mRNA expression as the transfection was repeated. However, regardless of pluripotency-related gene expression, HDDPC morphology remained fibroblastic, similar to its parental cells, at 9 days after the 1st transfection. Therefore, we considered that these cells, termed as "hiTSC-D", were in an intermediate state between HDDPCs and iPSCs.

Generation of intermediate stem-like cells from human fibroblasts. Similar to the protocol for
HDDPCs, human skin-derived fibroblasts were fixed and subjected to cytochemical staining for ALP activity on days 3, 5, 7, and 9 after transfection (Fig. 3a). Gradual increase in ALP activity was also clearly discernible as days in culture proceeded (Fig. 3b). As shown in the case of the 1st transfection, a gradual increase in the frequency of ALP activity was also seen as days in culture proceeded (Fig. 3c). The number of ALP-positive cells/ well, examined on days 7 and 9 after transfection, in the group subjected to double transfections was significantly  higher than that in the group subjected to single transfection (Fig. 3c). Similar to hiTSC-D, we termed these ALP-positive human fibroblast-derived intermediate state cells as "hiTSC-F".  and Supplemental Dataset. These data indicate a close similarity in gene expression profile between HDDPCs and hiTSC-D (Fig. 4a), which is in contrast to that between skin-derived fibroblasts and hiTSC-F ( Fig. 4b) or between HDDPCs and skin-derived fibroblasts (Fig. 4c). Unsupervised hierarchical clustering of the gene expression profiles of these cells also confirmed these notions: namely, hiTSC-D clustered more closely with HDDPCs than with fibroblasts ( Fig. 4d).

Multi-differentiation potency of intermediate cells from HDDPCs.
To test whether these intermediate cells possessed multi-differentiation potential, cells (P02) harvested 13 days after the 1st transfection were subjected to differentiation induction to osteoblastic or neurogenic lineage, as shown in Fig. 5a. Von Kossa staining revealed the presence of mineralisation (stained brown-black) in the cells treated with STK3 ™ (Fig. 5b- whereas no deposition was observed in cells cultured in the normal medium ( Fig. 5b-iv). Furthermore, the presence of calcific deposition by cells of an osteogenic lineage was detected when the cells were treated with osteogenic differentiation medium STK3 ™ for 7 days and then stained with Alizarin red S (Fig. 5c-ii). In contrast, however, in the control culture, no such alteration was observed ( Fig. 5b-i). We also detected the presence of neuronal cells, as revealed by Nissl staining (Fig. 5b-iii; arrows in Fig. 5b-iv) in the P02 cells cultured for 7 days in STK3 ™ but not in the control culture ( Fig. 5b-i,ii). We next assessed the in vivo growth capacity of the P02-derived intermediate cells by xenogenic intrapancreatic parenchymal transplantation into nude mice, which allows the growth and differentiation of a small number of iPSCs as well as proliferative tumours such as the pancreatic cancer cell line SUIT-2 and murine teratocarcinoma F9 cells in vivo 12 . When the P02 cells at 13 days after reprogramming factor transfection were transplanted into the pancreas of four nude mice (approximately 10 6 cells/site; a total of three sites injected), no visibly discernible tumours could be detected in the pancreas for four months (data not shown).

Discussion
The development of iPSCs has shed light on the possibility of obtaining autologous pluripotent embryonic-like stem cells, circumventing the need for somatic cell nuclear transfer-treated embryos 2,13 . However, the reprogramming process appears highly inefficient and likely depends on many factors including the age, type, passage number, and origin of the primary cells used 14 . For example, our previous attempt to reprogram a total of six primary HDDPC lines by transfection with plasmids carrying Yamanaka's four reprogramming factors, only two were successfully reprogrammed to form iPSCs whereas the other four lines were refractory to this treatment 4 . Moreover, we found a close association between a higher ALP activity of cells and the susceptibility to reprogramming to form iPSCs 4 . As ALP activity is closely associated with immature undifferentiated cells such as ESCs/iPSCs and stem cells 15 , we considered that HDDPCs showing strong ALP activity may be abundantly enriched with stem cells, which may in turn be more susceptible to form iPSCs after transfection with the reprogramming factors. In accordance with this notion, in the present study we found that only an ALP-positive HDDPC line (P01) that was newly established was successfully induced to form iPSCs upon initial transfection (Table 1). However, the other five lines (P02-06), which were judged as those with low or absent ALP activity, failed to form iPSCs (Table 1). We then addressed whether it was possible to convert these five lines (P02-06) to form iPSCs? Based on their ALP-negative status, we considered that application of conditions to elevate ALP activity might constitute a plausible option. Following the results of Yan et al. 4 , we expected that repeated transfection with the reprogramming factors would result in generation of ALP-positive cells and enhanced potential for iPSC generation. Consistent with this, three of five lines tested exhibited successful generation of iPSCs (Fig. 1b-iii; Table 1). Furthermore, P05, a line that failed to convert into iPSCs even after the 2nd-round transfection, was successfully converted to form iPSCs after the 3rd-round transfection (Fig. 1b-iv; Table 1). These results led us to suppose that repeated transfection of HDDPCs might constitute a useful tool for the enrichment of ALP-positive stem cells that are more conducive to iPSC generation. Accordingly, when ALP activity in the ALP-negative P02 line was assessed cytochemically during the process of reprogramming, a gradual increase in ALP activity was observed ( Fig. 2b; upper  columns; Fig. 2c). Notably, this pattern was also detected when the P02 line was subjected to 2nd-and 3rd-round transfections ( Fig. 2b; middle and lower columns; Fig. 2c). Furthermore, the frequency of ALP-positive cells increased as the number of transfections increased (Fig. 2c). Analysis of pluripotency-related gene expression by RT-PCR also demonstrated that the HDDPCs were in a stem-like cell state at 9 days after transfection with the reprogramming factors, which was very similar to that of iPSCs at the molecular level (Fig. 2e) although their morphology remained fibroblastic, which differed from that of iPSCs. Taken together, these findings indicate that repeated transfections are beneficial for converting ALP-negative cells to stem-like cells as well as for iPSC generation.
Notably, it was possible to obtain iPSCs after the 4 th -round transfection with four reprogramming factors. However, 7 out of 10 lines tested failed to form iPSC-like colonies on passage 7 and afterward (Supplemental Fig. 1). Furthermore, it was reported that repeated transfection with reprogramming factors can lead to generation of iPSCs in naïve state (Supplemental Fig. 1). Unfortunately, we failed to detect such naïve state iPSCs after the 4 th -round transfection.
Notably, ALP activity still appeared to be low in the cells at 3 days after the 3rd-round transfection (Fig. 2b  lower column; Fig. 2c). However, these cells corresponded to the cells at 12 days after the 1st transfection (Fig. 2a), which already showed relatively strong ALP activity, similar to Day 9 cells (Fig. 2b, upper columns). This implies that the ALP expression observed in the Day 9 cells may be labile, with rapid loss of pluripotency-related gene expression potentially occurring immediately after re-seeding. Subsequent analyses should therefore assess whether the Day 9 cells immediately after re-seeding can continue to express pluripotency-related genes.
Similar to the case of transfection of HDDPCs using reprogramming factors, the rate of ALP-positive human skin-derived fibroblasts after the 2 nd -round transfection was higher than that of cells obtained after the single transfection (Fig. 3c). The number of ALP-positive fibroblasts gradually increased after the first transfection. This tendency was also seen when the 2 nd -round transfection was performed: the number of ALP-positive cells appeared to peak at 7 and 9 days after transfection (Fig. 3c). These findings suggest that fibroblasts can also be reprogrammed partially, and similar to the case of HDDPCs, ALP activity was also a useful marker for cells at early stage of de-differentiation.
Cell proliferation after transfection with reprogramming factors appeared to differ between HDDPCs and fibroblasts. For example, HDDPCs exhibited faster cell proliferation after the 1 st transfection but not after the 2 nd and 3 rd -round transfection (Fig. 2d). On the other hand, fibroblasts exhibited faster cell proliferation than their untransfected parental cells on day 8 and afterward after the 2 nd -round transfection (Fig. 3d). It remains unknown why the mode of cell proliferation differs between these two cells, but as shown in the results of microarray analysis, it may reflect the difference in gene expression profile between them.
We performed microarray analysis for possible differences in gene expression profile between HDDPCs and hiTSC-D, fibroblasts and hiTSC-F, or HDDPCs and fibroblasts. This cluster analysis demonstrated that HDDPCs and hiTS-D cells had the most similar molecular signatures among these 4 cells (see Fig. 4a,d). We have already indicated that HDDPCs are thought to contain a number of stem cells as they exhibited ALP activity and were enriched with transcripts derived from stemness factors such as OCT3/4 and SOX2 4 . This is in contrast to the case of fibroblasts, which are thought to have less amounts of stemness factors. Gene expression profile between fibroblasts and hiTSC-F appears to differ (see Fig. 4b,d), suggesting alteration in the mode of gene expression in fibroblasts after transfection with reprogramming factors. Furthermore, gene expression profile between HDDPCs and fibroblasts differed (see Fig. 4c,d), suggesting that both types of cells differ at the gene expression level.
Since different molecular transitions during reprogramming were first documented by the laboratories of Jaenisch and Hochedlinger in 2008 16,17 , the reprogramming process has been roughly grouped into three phases (i.e., initiation, maturation, and stabilisation) by Samavarchi-Tehrani et al. 18 . In the initiation phase, fibroblast-specific surface markers such as THY1 and CD44 are down-regulated, whereas the pluripotency markers ALP followed by SSEA1 are gained, along with reactivation of telomerase activity 16 . This process is also characterised by a loss of the somatic cell signature, such as decreased expression of transcription factors SNAIL1/2 and ZEB1/2 17,19,20 and the gain of an epithelial-associated miRNA-200 family 18,21 . Polo et al. also noted that ALP and FBXO15, early markers of pluripotent cells, gradually increased their expression, whereas endogenous transcripts of OCT3/4 and SOX2 were detectable only late during iPSC generation 10 . Furthermore, single-cell analysis of 26 genes as well as fluorescence-activated cell sorting analysis demonstrated that gene expression changes occur homogeneously at early (day 0-3) and late time points (day 9 onward), whereas they are heterogeneous at intermediate stages (day [6][7][8][9] 10 . In the present study, we showed an increased expression of ALP around 9 days after a single transfection (Fig. 2b, upper columns), which may correspond to the intermediate stages described by Polo et al. 10 . Taken together, it may be plausible to consider that the intermediate stage between HDDPCs and iPSCs should exist around 9 days after transfection, as depicted in Fig. 6. Cells at this stage are morphologically indistinguishable from their parental HDDPCs but express pluripotency-related markers such as ALP and Consistent with this, we observed that HDDPCs, at 13 days after singe transfection with the Yamanaka's four reprogramming factors, had the ability to differentiate into at least two types of cells, osteoblastic and neuronal cells (Fig. 5b). Moreover, intrapancreatic parenchymal inoculation of these cells produced no visible teratomas in five nude mice observed over a course of four months (data not shown). Thus, these HDDPC-derived intermediate stage cells were considered as iTSCs and are termed 'hiTSC-Ds (induced tissue-specific stem cells from deciduous tooth-derived dental pulp cells)' .
In conclusion, we successfully produced iPSCs from HDDPCs (judged as cells refractory to convert to iPSC formation by a single transfection) through repeated transfections with Yamanaka's four reprogramming factors. During the reprogramming process, we revealed the presence of intermediate cells, termed hiTSC-Ds, having the stemness properties of molecular profile, multipotentiality, and non-tumorigenicity. These cells will likely have potential application towards the study of mammalian dental tissue regeneration.

Animals. All animal experiments were performed in agreement with Niigata University Committee on
Recombinant DNA Security guidelines (permit no. SP00636 dated 1st Aug. 2016) and with Animal Care and Experimentation Committee of Niigata University approval (permit no. 28 No163-1 dated 24th Jun. 2016) according to the "Guide for the Care and Use of Laboratory Animals" of the National Academy of Sciences, USA. All surgeries were performed under three anaesthetics (medetomidine, midazolam, and butorphanol) 22 , and all efforts were made to minimise suffering. For intrapancreatic tumour cell inoculation, eight-to twenty-week-old immunodeficient female mice (Balb/c-nu/nu, CLEA Japan, Tokyo, Japan) were used.
Primary cell culture of HDPPCs and human fibroblasts. HDDPCs were collected from patients after obtaining written informed consent from their legal guardians; study protocols were conducted in accordance with the tenets of the Declaration of Helsinki and approved by the Ethical Committee for Use and Experimentation of the Niigata University Graduate School of Medical and Dental Sciences (permit no. 28-R21-6-20 dated 21st Nov. 2016). HDDPCs were collected from patients after obtaining informed consent from their legal guardians; study protocols were approved by the Ethical Committee for Use and Experimentation of the Niigata University Graduate School of Medical and Dental Sciences (permit no. 28-R21-6-20 dated 21 st Nov. 2016). HDDPCs were isolated as described previously 23 , with slight modifications 4 . Briefly, pulp tissue was removed from the deciduous teeth of four young patients and digested with a solution of 3 mg/mL collagenase type I (#17100-017; Invitrogen, Carlsbad, CA, USA) and 4 mg/mL dispase (#410810077, Roche Applied Science, Basel, Switzerland) in Dulbecco's phosphate-buffered saline (DPBS) (#D8537; Sigma-Aldrich Co., Dorset, UK) for 25 min at 37 °C. Isolated pulp cells were cultured in MEMα (#135-15175, Wako Pure Chemical Industries, Ltd., Osaka, Japan) with 20% foetal bovine serum (FBS), 100 μM L-ascorbic acid-2-phosphate (#323-44822; Wako), 50 U/mL penicillin, and 50 mg/mL streptomycin (herein referred to as 'MEMα/20% FBS') at 37 °C in 5% CO 2 . After 3-7 passages, HDDPCs were used for transfection experiments.

Generation of intermediate stem-like cells from primary HDDPCs.
HDDPCs (approximately 1 × 10 5 ) judged as having no ALP activity underwent the 1st transfection then were seeded onto a gelatin-coated 24-well plate (#4820-020, Iwaki Glass) at approximately 1 × 10 4 cells/well to check ALP and proliferative activity (Fig. 2a). Cells were cultured in MEMα/20% FBS throughout the experiment. Cells were fixed with 4% paraformaldehyde (PFA) (for checking ALP activity) or harvested at 3, 5, 7, and 9 days after seeding by trypsinisation (for checking proliferative activity). For repeated transfections (double transfection), at 5 days after the 1st transfection, cells were harvested by trypsinisation for the 2nd transfection with reprogramming factors. The treated cells were seeded onto a gelatin-coated 24-well plate for checking ALP and proliferative activity on the days indicated after seeding, as shown in Fig. 2a. For triple transfection, at 5 days after the 2nd transfection, cells were harvested for the 3rd transfection, as shown in Fig. 2a. After final seeding, cells were examined for ALP and cell proliferative activities as indicated. In some cases, cells were harvested at 19, 14, and 9 days after the final transfection for single, double, or triple transfection, respectively (Fig. 2a), for mRNA expression analysis using RT-PCR. HDDPCs 13 days post-1st transfection were also subjected to in vitro differentiation induction and in vivo teratoma formation, as shown below. Using the same procedure used for generation of HDDPC-derived hiTSCs, fibroblasts were transfected twice with the reprogramming factors and checked for ALP activity.

Proliferative rate of HDDPCs and fibroblasts.
To evaluate cell proliferation rates, cells (approximately 1 × 10 4 ) seeded onto a gelatin-coated 24-well plate were collected by trypsinisation after 1, 3, 5, 7, or 9 days and the cell number counted using a disposable haemocytometer. At least 3 wells per line were examined and the average cell number was plotted. The data were expressed as the mean ± standard error. Two groups were compared using the Student t-test. The differences in each group were considered significant if P < 0.05. All statistical analysis methods were performed in accordance with the relevant guidelines and regulations.
ALp assay. Cytochemical ALP activity assay was performed using 4% PFA-fixed cells with an Alkaline Phosphatase Staining Kit II (#00-0055; STEMGENT, Cambridge, MA, USA), which utilised α-naphtholum-coupled diazonium salt to stain ALP per manufacturer protocol. Cells in the centre well portion were photographed and the stained and unstained cells were manually counted. A total of approximately 1 × 10 5 cells per well were counted. Three wells per transfection group were examined and the average cell number was plotted.
RT-PCR analysis. Total RNA from each sample was isolated using the RNA Mini Kit (#50204; Qiagen, Tokyo, Japan) as previously described in Murakami et al. 24 . A negative, no-template control (designated as -RT) was included for each reaction. Additionally, cDNA from MMiPS, an iPSC line established from HDDPCs in our laboratory 11 , was used as a positive control. PCR primers are listed in Supplementary Table 1  Gene microarray analysis. For the oligonucleotide-based microarray analysis, total RNA was extracted from cells using the RNA Mini Kit. Microarray analysis was performed with a 3D-Gene ® Human Oligo chip 25k (Toray Industries Inc., Tokyo, Japan). For efficient hybridisation, this microarray adopted columnar structure to stabilize spot morphology and to enable micro bead agitation. Total RNA was labelled with Cy5 using the Amino Allyl MessageAMP II aRNA Amplification Kit (Applied Biosystems, CA, U.S.A.). The Cy5-labelled aRNA pools were mixed with hybridisation buffer and hybridised for 16 h. The hybridisation was performed according to the supplier's protocols (www.3d-gene.com). The hybridisation signals were obtained using a 3D-Gene Scanner (Toray Industries Inc.) and processed by 3D-Gene Extraction software (Toray Industries Inc.). Detected signals for each gene were normalised using a global normalisation method described in the guide provided by the Toray Industries Inc. (the median of the detected signal intensity was adjusted to 25).
In vitro differentiation induction assay. Primary HDDPCs and HDDPCs 13 days after transfection were seeded onto a gelatin-coated 6-well plate containing MEMα/20% FBS. After culturing to 80-90% confluency, the medium was changed to differentiation-inducing medium. To test ability hiTSC ability to differentiate into neurogenic and osteogenic cells, the method described in our previous study was used 24 .
To induce EB formation, portions of HDDPC-iPSC colonies were dissected mechanically using a pipette tip under a stereomicroscope and then seeded onto an ultra-low attachment 96-well plate (#MS-9035X; Sumitomo Bakelite Co., Ltd., Tokyo, Japan) with DMEM/10% FBS. After 7 days of culture, emerging EBs were transferred onto gelatin-coated chamber slides (#154526JP, Iwaki Glass) and cultured for another 5 days in DMEM/10% FBS to allow enhanced differentiation into various cell types.
In vivo teratoma formation assay. The potential of HDDPCs to form teratoma in vivo at 13 days after the 1st transfection with reprogramming factors was assessed using a recently established method termed 'IPPCT' 12 , which allows growth and differentiation of a small number of iPSCs through grafting into the pancreatic parenchyma of nude mice by a glass injection needle under observation using a dissecting microscope.
Briefly, a solution (1-3 μL) containing HDDPCs (approximately 10 4 ) suspended in MEMα/20% FBS was injected into the pancreatic parenchyma of anesthetised Balb/c-nu/nu mice. The injections were performed at a total of 3 different sites in each pancreas. In the positive control, iPSCs (approximately 10 4 ) were also similarly injected. At 1.5 months after transplantation, the pancreas was inspected for the presence or absence of teratomas. When teratomas were detectable, they were dissected from the pancreas and fixed with 4% PFA at 4 °C for 4 days for preparing tissue sections for pathological analysis.