Human Induced Pluripotent Stem Cells on Autologous Feeders

Background For therapeutic usage of induced Pluripotent Stem (iPS) cells, to accomplish xeno-free culture is critical. Previous reports have shown that human embryonic stem (ES) cells can be maintained in feeder-free condition. However, absence of feeder cells can be a hostile environment for pluripotent cells and often results in karyotype abnormalities. Instead of animal feeders, human fibroblasts can be used as feeder cells of human ES cells. However, one still has to be concerned about the existence of unidentified pathogens, such as viruses and prions in these non-autologous feeders. Methodology/Principal Findings This report demonstrates that human induced Pluripotent Stem (iPS) cells can be established and maintained on isogenic parental feeder cells. We tested four independent human skin fibroblasts for the potential to maintain self-renewal of iPS cells. All the fibroblasts tested, as well as their conditioned medium, were capable of maintaining the undifferentiated state and normal karyotypes of iPS cells. Furthermore, human iPS cells can be generated on isogenic parental fibroblasts as feeders. These iPS cells carried on proliferation over 19 passages with undifferentiated morphologies. They expressed undifferentiated pluripotent cell markers, and could differentiate into all three germ layers via embryoid body and teratoma formation. Conclusions/Significance These results suggest that autologous fibroblasts can be not only a source for iPS cells but also be feeder layers. Our results provide a possibility to solve the dilemma by using isogenic fibroblasts as feeder layers of iPS cells. This is an important step toward the establishment of clinical grade iPS cells.


Introduction
Human pluripotent stem cells, both embryonic stem (ES) cells and induced Pluripotent Stem (iPS) cells, are generally maintained on mouse embryonic fibroblasts (MEF), which are mitotically inactivated by treatment with mitomycin C or c-ray irradiation [1][2][3]. However, usage of mouse feeder cells may transfer exogenous antigens, unknown viruses, or zoonotic pathogens to iPS cells. In fact, non-human sialic acid N-glycolylneuraminic acid (Neu5Gc), which is potentially immunogenic, was detected on the surface of human ES cells maintained on MEF feeder [4]. Although feeder-free culture of human ES cells has been reported, it may lead to chromosomal instabilities of human ES cells [5,6]. To avoid these issues, human fibroblasts from neonatal foreskin or ES cell-derived fibroblast-like were used to support self-renewal of human ES cells [7][8][9][10][11]. However, one still have to concern about existence of unidentified pathogens, such as viruses and prions in these non-autologous feeders. Since iPS cells are generated from fibroblasts, it would be ideal if the same fibroblasts can be used for the generation and maintenance of iPS cells.

Results and Discussion
To examine whether human fibroblasts support self-renewal of human iPS cells, we treated four independent human fibroblast lines (1388, 1392, 1503 and NHDF; see Table S1) and SNL cells [12] with mitomycin C, and seeded them on culture plates (Fig. S1). Then, we plated 201B7 iPS cell line [2] derived from 1388 fibroblasts onto these feeder cells with standard density (1:5 dilutions). The passage number of iPS cells was 20 at this point. All the five cell lines of feeder cells were supportive for undifferentiated growth of iPS cells at least 19 additional passages (Fig. 1A). The percentage of TRA-1-60 (a marker for undifferentiated ES cells and iPS cells) positive colonies was similar among different human fibroblasts and SNL cells (Fig. 1B). No significant differences were observed in the plating efficiencies (Fig. 1C). In iPS cells at passage 2 after switching onto various HDF feeders, no significant re-activation of transgenes was observed (Fig. S2). In addition, reverse transcription polymerase chain reaction (RT-PCR) showed that the expression of ES cell marker genes such as OCT3/4, SOX2 and NANOG were equally to those of H9 ES cells at passage 19 [1] (Fig. 1D).
Conditioned medium (CM) of MEF or SNL allows feeder-free culture of iPS cells. To test whether CM of fibroblasts could maintain self-renewal of iPS cells without feeder cells, we seeded 201B7 iPS cells onto Matrigel-coated plates in CM from each human fibroblast line or SNL. As a control, we used nonconditioned medium supplemented with bFGF. Cells in nonconditioned medium failed to form tightly packed colonies, whereas those in each CM grew healthily with typical undifferentiated ES-like morphologies (Fig. S3A). RT-PCR revealed that iPS cells maintained in each CM expressed undifferentiated ES cell marker genes such as OCT3/4, SOX2, NANOG and TERT at similar levels to those in iPS cells or human ES cells cultured on SNL feeder layers (Fig. S3B). Quantitative PCR (qPCR) confirmed that no significant alternations in the expression levels of OCT3/4, SOX2 and NANOG transcripts among CM from different feeders (Fig. S3C). These data demonstrated that human neonatal and adult fibroblasts could be utilized as feeder cells of human iPS cells.
Next, we examined whether human iPS cells could be established without non-autologous feeder cells. We introduced the four reprogramming factors into the four human fibroblast lines by retroviral transduction. Six days after infection, we plated the transduced cells at 5610 5 cells on 100-mm dishes either with SNL feeders, with isogenic human fibroblast feeder, or without feeder cells. Next day, we started cultivation using human ES cell culture medium. In the plates without feeders, the plated cells became confluent within a several days and showed an appearance resembling feeder cells. Around three weeks after transduction, ES-like colonies began to emerge on the feeder cell-like layer ( Fig. 2A). We observed no significant differences in the numbers of ES-like colonies among on SNL feeders, on isogenic fibroblast feeders or feeder-free condition (Table S2). On day 25 after transduction, we picked up ES-like colonies from the plates without feeders and transferred them onto new plates with mitomycin C-treated each parental fibroblasts as feeders. Human iPS cells derived from each of the four fibroblast lines used in this study grew normally and maintained the undifferentiated morphologies on corresponding autologous feeders for at least 18 passages (Fig. 2B).
RT-PCR showed that established clones at passage number 5 expressed endogenous OCT3/4, SOX2, NANOG and TERT transcripts at similar levels to those in 201B7 iPS cells, which were established on SNL feeder cells, and H9 ES cells (Fig. S4A). The retroviruses of the four factors were effectively silenced, which is a hallmark of complete reprogramming (Fig. S4B). Even after additional 15 passages, the expression of OCT3/4, SOX2 and NANOG in these iPS cells were comparable to those of ES cells and iPS cells maintained on SNL feeder cells (Fig. 2C). Immunoprecipitation assay with anti-methylated cytosine antibody revealed that the promoter regions of pluripotent-associated genes such as OCT3/4 and NANOG locus were almost completely unmethylated in iPS cells established and maintained on the autologous feeders at passage number 5, like in H9 ES cells (Fig.  S5). In addition, iPS cells generated with the autologous feeders showed normal karyotypes at least after 26 times passages (Fig.  S6).
To evaluate pluripotency of iPS cells generated and maintained on autologous feeders, we performed in vitro differentiation assay. These iPS clones formed embryoid bodies using the floating culture condition. After 16-day differentiation, we detected SOX17 (endoderm), a-smooth muscle actin (a-SMA, mesoderm) and NESTIN (ectoderm) positive cells in the culture (Fig. 2D, Fig.  S7). We also confirmed that undifferentiated markers such as OCT3/4, SOX2 and NANOG decreased and other differentiated markers such as AFP, PDGFRa and PAX6 increased (Fig. S8). In addition, we injected iPS cells at passage number 9 into testes of immune-deficient mice for teratoma formation. After 8 to 12 weeks, all clones we tested developed teratoma containing various tissues including gut-like epitheliums (endoderm), cartilages (mesoderm) and neural rosettes (ectoderm) (Fig. 2E). These data confirmed pluripotency of iPS cells, which were established and maintained on the autologous feeder cells.
To examine the compatibility of iPS cells and feeder cells, we plated iPS cells derived from the four HDF onto mitomycin Ctreated parental fibroblasts or SNL cells with all the possible combinations. After six days, we stained the cells with TRA-1-60 antibody, and counted the number of positive colonies. More than 80% of colonies in all the combinations showed morphology of undifferentiated ES-like cells and were positive for TRA-1-60 (Fig.  S9).
Our results demonstrated that human iPS cells can be generated and maintained on autologous fibroblasts as feeder layers. Furthermore, human iPS cell can be generated even without any additional feeders, since non-reprogrammed fibroblasts can serve as feeders. The maintenance of iPS cells can also be achieved without feeders, by using conditioned medium of human fibroblasts. All of the tests performed in this study revealed that iPS cells derived from four independent HDF maintained autologous feeders kept pluripotency during at least 100 day culture (Table S3). The most of reprogrammed cell colonies on isogenic feeders are uniformly undifferentiated (Fig. S9).
In general, we can obtain more than ten millions of fibroblasts from 5-mm square skin biopsy at the passage number 3 with our standard protocol [13]. For iPS cell generation with retroviruses, we need less than 1610 5 fibroblasts. An alternative method using an episomal vector system requires one million of fibroblasts [14]. Thus, we have enough amounts of surplus fibroblasts to be used as autologous feeder cells. We did notice that iPS cells cultured on SNL feeders are easier to passage than those on MEF and human fibroblast feeders.
Our data that all four HDF lines tested in this study could support both generation and maintenance of iPS cells, does not guarantee that every fibroblasts can be used as feeders cells for human pluripotent cells [15]. We also tested that whether 14 HDF lines were supportive for maintenance of ES cells and iPS cells. Both KhES3 ES cell line and 201B7 could grow normally on eleven out of 14 HDF lines, two MEF lines (ICR and C57BL6) and SNL [16] (Table S1, Fig. S1). In co-culture with three lines (1554, 1616 and TIG107), at least either KhES3 or 201B7 could not stay at undifferentiated state even at passage number 2. At least, among on 11 supportive fibroblasts and MEF and SNL, no significant differences were observed in growth rate of ES cells and iPS cells. Unsupportive lines are indistinguishable from HDF lines by their morphologies or growth speed. Probably, support activity for self-renewal of ES cells and iPS cells do not depend on at least passage number and donor's sex or race. Unsupportive lines tested in this study are derived from donors at 68, 77 and 81 years old. On the other hand, we found that HDF derived from donors at 69 and 73 years old could support maintenance of undifferentiated state of both ES cells and iPS cells. Further detailed analyses will be required for decision whether donor's ages of feeder cells are important or not.
Recent study by Unger and colleagues demonstrated that iPS cells derived from human fetal fibroblasts could be established and maintained on isogenic feeder cells [17]. On the other hand, we assessed that neonate and adult fibroblast-derived iPS cells can be generated and maintained on autologous feeders. Inspection of established iPS cells is always necessary before clinical trials even in autologous cell transplantation therapy. However, the culture system established in this study demonstrated that fibroblasts from an individual could play dual roles as source of iPS cells and feeder cells, probably contributing to efficient processing for clinical grade-pluripotent stem cells. Actually, the system still includes animal components such as albumin, insulin and trypsin. Xenofree culture is basically required for therapeutic usage of iPS cells. However, even if it is for removing the xenogenic components, the culture condition should not be oppressive for pluripotent cells. Isogenic culture from the start is one of true worth of iPS cells because autologous fibroblasts can not normally inhabit when ES cells are established from blastocysts. Our result is an important step toward the generation of clinical-grade human iPS cells suitable for future medical applications.

Generation of iPS Cells
iPS cells were established from HDF as described previously with some slight modifications [2]. In brief, we firstly introduced mouse solute carrier family 7 (cationic amino acid transporter, y+ system), member 1 (Slc7a1) gene which encodes ecotropic retrovirus receptor by lentiviral transduction. Transfectants were plated at 2610 5 cells per 60 mm dish and incubated overnight. The next day, into the cells OCT3/4, SOX2, KLF4 and c-MYC were introduced by retroviral infection. Six days later, the cells were harvested by trypsinization, and plated at 5610 5 cells per 100 mm dish. The medium was replaced on the next day with hES cell medium, and cultured for another 20 days. At day 25 postinduction, ES-like colonies were mechanically dissociated and transferred on to 24-well plate on each isogenic feeder. We designated this point as passage 1.

Feeder Cells
We added phosphate buffered saline (PBS, Nacalai tesque) containing 12 mg/ml mitomycin C directly into fibroblast culture in subconfluent, and incubated at 37uC for 3 hours. After treatment, the cells were washed twice with PBS and harvested by trypsinization. The cells were plated at 1610 6 cells per 24-well plate, 6-well plate, 3 of 60 mm dishes or 100 mm dish.

Conditioned Medium
We plated fibroblasts at 3610 5 cells per 60 mm dish, and incubated overnight. Next day, the medium was replaced with 3 ml of hES medium, and incubated for 24 hours. After incubation, the supernatant of fibroblast culture was collected and filtered. We added 4 ng/ml bFGF before use.

Expression Analyses
We performed RT-PCR as described previously [2]. In brief, the cells were lysed with Trizol reagent (Invitrogen), and then total RNA was purified. RNA samples were treated with Turbo DNA free (Ambion) to remove genomic DNA contamination. One microgram of DNase treated RNA was used for first-strand complementary DNA (cDNA) synthesis with Rever tra ace -a-(Toyobo) and oligo dT 20 primer. qPCR was performed using SYBR Premix ExTaq II (Takara). Primer sequences were listed in Table S4 [2,20,21,22].

Methylation Assay
Four microgram of genomic DNA was mechanically shared by sonication, and boiled at 95uC for 10 minutes. Then shared genomic DNA was incubated with pan-mouse IgG magnetic beads (Invitrogen) -conjugated anti-5-methyl cytosine antibody (Eurogentec) supplemented with 5 mg/ml BSA and 25 mg/ml yeast tRNA (Ambion) overnight at 4uC. Beads were washed three times with PBS containing 0.05% TritonX-100. Beads were suspended in 0.15 ml of TE containing 1% SDS, and incubated at 65uC for 5 minutes. The elution was repeated with an additional 0.15 ml of 1% SDS/TE. The eluates were treated with Protease K at 50uC for 2 hours, and then extracted with phenol: chloroform: isoamyl alcohol, and purified by ethanol precipitation. Primer sequences are provided in Table S4.  Figure S8 iPS cells maintained on isogenic feeders (U) or differentiated by embryoid body formation (D) were lysed with Trizol reagent. Total RNA was purified and treated with DNase to remove genomic DNA contamination. One microgram of DNasetreated RNA sample was used for first-strand cDNA synthesis with oligo dT20 primer. PCR was performed with the primers listed in Supplemental Table 2.