Pericellular matrix of decidua-derived mesenchymal cells: A potent human-derived substrate for the maintenance culture of human ES cells

Authors

  • Tomoko Nagase,

    1. Organogenesis and Neurogenesis Group, Center for Developmental Biology, RIKEN, Kobe, Japan
    2. Department of Pediatrics, Graduate School of Medicine, Gifu University, Gifu, Japan
    Current affiliation:
    1. Life Science Research Center, Gifu University, Gifu 501-1193, Japan
    Search for more papers by this author
  • Morio Ueno,

    1. Organogenesis and Neurogenesis Group, Center for Developmental Biology, RIKEN, Kobe, Japan
    2. Department of Ophthalmology, National Center for Geriatrics and Gerontology, Obu, Japan
    Search for more papers by this author
  • Michiru Matsumura,

    1. Organogenesis and Neurogenesis Group, Center for Developmental Biology, RIKEN, Kobe, Japan
    2. Division of Human Stem Cell Technology, Center for Developmental Biology, RIKEN, Kobe, Japan
    Search for more papers by this author
  • Keiko Muguruma,

    1. Organogenesis and Neurogenesis Group, Center for Developmental Biology, RIKEN, Kobe, Japan
    Search for more papers by this author
  • Masatoshi Ohgushi,

    1. Division of Human Stem Cell Technology, Center for Developmental Biology, RIKEN, Kobe, Japan
    Search for more papers by this author
  • Naomi Kondo,

    1. Department of Pediatrics, Graduate School of Medicine, Gifu University, Gifu, Japan
    Search for more papers by this author
  • Daisuke Kanematsu,

    1. Departments of Regenerative Medicine, Institute for Clinical Research, Osaka National Hospital, Osaka, Japan
    Search for more papers by this author
  • Yonehiro Kanemura,

    1. Departments of Regenerative Medicine, Institute for Clinical Research, Osaka National Hospital, Osaka, Japan
    Search for more papers by this author
  • Yoshiki Sasai

    Corresponding author
    1. Organogenesis and Neurogenesis Group, Center for Developmental Biology, RIKEN, Kobe, Japan
    2. Division of Human Stem Cell Technology, Center for Developmental Biology, RIKEN, Kobe, Japan
    • Organogenesis and Neurogenesis Group, Center for Developmental Biology, RIKEN 2-2-3 Minatojima-minamimachi, Chuo, Kobe 650-0047, Japan
    Search for more papers by this author

Abstract

In routine culture, human embryonic stem (hES) cells are maintained on either feeder cells or special culture substrates such as Matrigel. However, to expand hES cells for clinical applications, it is desirable to minimize animal-derived materials in the culture for safety reasons. In this report, we show that the pericellular matrix prepared from human decidua-derived mesenchymal cells (PCM-DM) is a potent substrate material that supports the growth and pluripotency of hES cells as efficiently as Matrigel does. This supporting activity of PCM-DM is stable and can be preserved for several months in the refrigerator. PCM-DM-based culture is compatible with non-conditioned commercial defined medium, and with the maintenance of dissociated hES cells in the presence of ROCK inhibitor. Since decidual mesenchymal cells can be prepared and expanded in a large quantity, PCM-DM is a practical human-derived substitute for the animal-derived substrates for use in clinical-grade culture of hES cells. Developmental Dynamics 238:1118–1130, 2009. © 2009 Wiley-Liss, Inc.

INTRODUCTION

Human pluripotent cells such as ES (hES) cells and iPS (hiPS) cells are strongly anticipated as excellent resources for cell therapies of intractable diseases (Wobus and Boheler,2005; Murry and Keller,2008). In addition, hES/iPS technology is important for other medical researches such as drug discovery as well as for basic studies bridging the gap between mouse embryology and human development. The derivation of hES cells and their maintenance culture are performed routinely in many laboratories, while hES cells are induced to differentiate into medically useful types of cell types such as dopaminergic neurons (Perrier et al.,2004; Ueno et al.,2006), cardiomyocytes (Laflamme et al.,2007), and islet cells (Kroon et al.,2008).

There are several technical difficulties that must be overcome before hES cells can be used for human patient therapies. With respect to safety issues, for instance, the possibility of tumorigenesis from grafted cells must be minimized by removing tumor-forming undifferentiated cells and purifying the population of desired cells (Kolossov et al.,1998; Choo et al.,2008). Another equally important safety issue is the contamination of xenogenic pathogens and immunogens (Martin et al.,2005). Xenogenic pathogens such as latent viruses may cause unexpected serious infections while xenogenic immunogens could promote acute rejection after transplantation (Petersen et al.,1998).

Unlike mouse ES (mES) cells, human pluripotent stem cells require feeder cells or special substrate matrix for their maintenance culture while the molecular and cellular bases for this difference remain largely elusive (discussed later in the Discussion). In many laboratories, hES cells and hiPS cells are maintained on animal-derived feeder cells such as mouse embryonic fibroblasts (MEFs) (Thomson et al.,1998; Reubinoff et al.,2000; Takahashi et al.,2007), which carry the risks mentioned above. Previous studies have provided two categories of alternatives to MEF. One is human feeder cells. For instance, a primary culture of foreskin fibroblasts has been successfully used in the derivation of hES cells and their maintenance cultures (Hovatta et al.,2003). Maintenance-supporting activity has been also reported for several other types of human cells including adult fallopian tube epithelial cells, fetal skin fibroblasts (Richards et al.,2002), bone-marrow mesenchymal cells (Cheng et al.,2003), and adult skin fibroblasts (Richards et al.,2003). However, the difficulty in the availability of human-derived tissues is a hindrance for their routine use. In addition to the requirement for proper informed consent from the patients, the amount of obtainable tissues is generally small and the cell passage number is often limited, which is problematic in general for primary culture of human cells. Moreover, the maintenance activity of human feeder cells often varies from batch to batch and it is difficult to achieve uniform and reproducibly high activity.

The second alternative to MEF feeder cells is the use of extracellular matrices as the culture substrate. If its maintenance-supporting activity is strong enough, a matrix substrate is superior to feeder cells in both handling and quality control. Matrix substrate is cell-free and circumvents the problem of the contamination of allogenic feeder cells in hES cells (or their derivatives). In addition, a matrix substrate is generally stable and, if a large-scale preparation is feasible, the same batch of pericellular matrix (pre-tested for activity and safety) could be repeatedly used.

Matrigel, produced from a mouse tumor cell line (EHS sarcoma), is a most commonly used extracellular matrix and has a potent maintenance-supporting activity in human ES cell culture when used with MEF-conditioned medium (Xu et al.,2001). However, this matrix is animal-derived and the culture cannot be xeno-free. Defined matrices such as fibronectin are also maintenance-supporting (Xu et al.,2001; Lu et al.,2006), but are usually not as potent as Matrigel. While recombinant vitronectin was recently reported to support hES cell maintenance at least in a short-term culture, this matrix needs to be produced in a mouse myeloma cell line (Braam et al.,2008).

Here, we introduce the PeriCellular Matrix of Decidua-derived Mesenchymal cells (PCM-DM) as an excellent human-derived material for culture substrate. These cells are isolated from human fetal membranes, which can be routinely obtained in a large quantity from obstetric practices with a relatively simple informed consent form. The human fetal membrane consists of three main layers, the amniotic (epithelium and mesenchyme; inner), chorionic (middle), and decidual (outer) membranes (see Supp. Fig. S1, which is available online) (Malak and Bell,1994). We initially examined three different types of cells derived from human fetal membranes: amniotic epithelial cells (AECs), amniotic (+chorionic) mesenchymal cells (AMCs), and decidual mesenchymal cells (DMCs). Of these, the DMCs grew efficiently and could be reproducibly prepared in the largest number from a human fetal membrane specimen, and showed potent maintenance-supporting activity, which was similar to that of MEFs (described in Supp. Figs. S2 and S3).

To further improve the versatility and controllability, we have investigated the maintenance-supporting activity of the pericellular matrix of DMCs (PCM-DM). We demonstrate that PCM-DM is a versatile and practical human-derived substitute for animal-derived matrix substrates such as Matrigel and the pericellular matrix of MEFs (Klimanskaya et al.,2005).

RESULTS

Efficient and Reproducible Preparation of Decidual Mesenchymal Cells (DMCs) From Human Fetal Membranes

Human fetal membranes were obtained from normal deliveries and processed under the donors' informed consent in accordance with the Helsinki Declaration and institutional guidelines. The decidual portions (maternally derived) and the chorio-amniotic membranes (zygotically derived) were separated manually and thoroughly rinsed (Supp. Fig. S1). The decidual tissues were cut into small pieces with scissors and subjected to enzymatic digestion (collagenase, dispase, and DNase I). The dispersed cells were collected, cultured, and expanded as described in Materials and Methods. These cells grew well and exhibited a uniform mesenchymal morphology on a culture dish (Fig. 1A). Their mesenchymal nature was also confirmed by their stress fiber-pattern of phalloidin staining and expression of vimentin, while they were negative for the epithelial marker CK19 (Chen et al.,2007) and for the zygotically derived tissue marker HLA-G (Le Bouteiller et al.,1999) (Fig. 1B–H), showing that at least the majority of these cells were decidua-derived mesenchymal cells.

Figure 1.

Morphology and character of decidua-derived mesenchymal cells (DMCs). A: Typical phase-contrast image of the human DMCs used in this study. Scale bar = 200 μm. B: Fluorescent image of human DMCs stained with phalloidin. Scale bar = 50 μm. CH: Immunofluorescent images of human DMCs stained with an anti-vimentin Ab (green; C,E,F,H) and an anti-cytokeratin 19 (red; D) or anti-HLA-G Ab (red; G; HLA-G is a fetal derivative-specific marker). Merged image shows nuclei stained by TO-PRO-3 iodide. Scale bar = 20 μm. I: Representative growth curves of human DMCs. Graphs show growth curves after the first passage. Three independent preparations of human amniotic epithelial cells (AECs), amniotic mesenchymal cells (AMCs), and DMCs were examined. The cells were seeded at 2×105 cells/100-mm-diameter dish at every passage and cultured until the cells reached confluency. At passaging, cells were resuspended, counted, and reseeded into a new dish at the same density. The total cell number was calculated from the harvested cell number.

These DMCs grew efficiently and reproducibly (reproducible growths of three independent preparations of DMC are shown in Fig. 1I). In addition, the cells were continuously expanded for 100 days, and multiplied approximately 1010 times during this period. The cells at log phase could be easily stored by the conventional freezing method with DMSO, and were recovered efficiently by thawing and replating. Importantly, when used as feeder cells, the DMCs had a reasonable maintenance-supporting activity for hES cell culture similar to that seen with MEF feeder cells (Supp. Fig. S2; on the DMC feeder, hES cells expanded 2,000–3,000-fold in 6 weeks and maintained the expression of undifferentiated-state markers).

In contrast, amniotic epithelial cells (AECs) and chorio-amniotic mesenchymal cells (AMCs) grew slowly and only for limited cell cycles (Fig. 1I). In particular, the AECs expanded very little. Therefore, although the AECs and AMCs also had some maintenance-supporting activity for hES cells (our unpublished observations; the maintenance-supporting activity of AECs for monkey ES cells was previously reported; Miyamoto et al.,2004), in terms of quantitative supply, these two types of cells were considered to be much less suitable for practical use than the DMCs. Therefore, we focused on the use of DMCs in the following studies.

Pericellular Matrix of Decidual Mesenchymal Cells Efficiently Supports the Maintenance Culture of hES Cells in MEF-Conditioned Medium

DMCs were cultured on gelatin-coated plastic culture dishes, and lysed by deoxycholate treatment to prepare the pericellular matrix (Klimanskaya et al.,2005; Supp. Fig. S3) 3 days after they reached confluency. The extracted pericellular matrix of DMCs (PCM-DM) was then used for human ES cell culture, and its maintenance-supporting activity was compared with the conventional matrices fibronectin and Matrigel (Xu et al.,2001). hES cell clumps, partially dissociated by enzyme digestion (Suemori et al.,2006), were cultured for 6 days on these matrices (1.7×104 cells/well, 24-well plate) in MEF-conditioned medium (MEF-CM) (Xu et al.,2001). The hES cell colonies grew on PCM-DM as efficiently as on Matrigel and more efficiently than on fibronectin (Fig. 2A–C). During multiple passages (Fig. 1D), PCM-DM supported the growth of human ES cells with an efficiency similar to (or slightly better than) that of Matrigel (the population doubling times on PCM-DM and Matrigel were 60 and 72 hr, respectively). These findings showed that human PCM-DM has a growth-supporting activity comparable to that of the animal-derived Matrigel.

Figure 2.

PCM-DM is comparable to Matrigel for the maintenance culture of hES cells with MEF-conditioned medium. A–C: Feeder cell-free culture of hES cells on various culture surfaces in MEF-conditioned medium. A: Staining of alkaline phosphatase in human ES cells cultured on each culture surface. B: Cell counts on day 7. Dunnett's method was used for the comparison against control. Values shown on the graphs represent the mean ± SD. Gelatin coating was used as a negative control. C: Growth curve of hES cells on each culture surface. Cell numbers were counted on days 2, 4, and 6 in 24-well plates (n=3). D: Propagation efficiency of hES cells cultured in MEF-conditioned medium. PCM-DM supported propagation better than Matrigel.

We next examined the expression of undifferentiated-state markers in PCM-DM-cultured hES cells. hES cells on PCM-DM were small in size with a high nuclear/cytoplasmic ratio, and formed flat and cell-dense colonies (Fig. 3A). The cells in the periphery of the colonies often had lamellipodia-like protrusions toward the outside. The cells in the colonies were positive for typical pluripotency-specific markers such as alkaline phosphatase (Fig. 3B), Oct3/4 (Fig. 3C), Nanog (Fig. 3D; Nanog staining showed both strong and weak signals, as previously reported in Singh et al.,2007), TRA-1-60 (Fig. 3E), and SSEA4 (Fig. 3F).

Figure 3.

PCM-DM supports the maintenance of hES cells in an undifferentiated state. KhES1 cells were cultured in MEF-conditioned medium on PCM-DM dishes and analyzed after more than six passages. A: Phase-contrast view of the hES cell colonies. Alkaline phosphatase activity (B) and immunostaining of hES cell colonies with anti-Oct3/4 (C), anti-nanog (D), anti-TRA-1-60 (E), and anti-SSEA4 antibodies (F). Scale bar = 100 μm.

These observations indicated that the human PCM-DM possesses a potent ability to support the maintenance of hES cells in MEF-CM.

PCM-DM Maintains the Growth and Pluripotency of hES Cells in Non-Conditioned Serum-Free Medium

Since MEF-CM itself contains mouse-derived materials, we next examined whether PCM-DM-based hES cell culture is compatible with unconditioned serum-free medium. hES cell clumps were seeded on PCM-DM in STEMPRO hESC SFM (StemPro, hereafter; commercially available from Invitrogen), which is known to support the human ES cell maintenance on Matrigel-type matrices. Under these conditions, the hES cells formed colonies (Fig. 4A; a high-magnification view is shown in Supp. Fig. S4A) expressing the undifferentiated-state markers alkaline phosphatase, Oct3/4, Nanog, TRA-1-60, SSEA4 (Fig. 4B–E and Supp. Fig. S4B and C; the early differentiation marker SSEA1 was not expressed; Supp. Fig. S4D), and grew efficiently (Fig. 4F; although the growth was slightly slower than in MEF-CM, hES cells on PCM-DM in StemPro still expanded 105 folds in 6 weeks). In addition, these ES cells did not express the early neural markers CD133 (progenitors; Pruszak et al.,2007), N-cadherin or Pax6, the epidermal marker CK14, the early mesodermal marker Brachyury, or the early endodermal marker HNF3β (Supp. Fig. S5). Consistent with these immunostaining data, FACS analysis showed high expression of the undifferentiated markers SSEA3, SSEA4, and TRA-1-60, and very low expression of the differentiated marker SSEA1 in human ES cells cultured in StemPro on the PCM-DM (Supp. Fig. S6). These observations indicated that the PCM-DM-based culture permits hES cell culture to be maintained with human-derived material in the unconditioned commercial medium, in which highly purified BSA is the sole xenogenic component (see Experimental Procedures). Notably, the maintenance-supporting activity of PCM-DM was stable and could be preserved for at least 8 months after its preparation by keeping the plate in the refrigerator under semi-dry conditions (Fig. 4G–J).

Figure 4.

PCM-DM supports the maintenance of hES cells in chemically defined medium. hES cells were cultured in StemPro on PCM-DM dishes and analyzed after more than six passages. A: Phase contrast view of hES cell colonies. B,C: Alkaline phosphatase activity of hES cells cultured in StemPro on a 4-well PCM-DM plate. D,E: Immunostaining of hES cell colonies, cultured in StemPro medium, with anti-Oct3/4 (D) and anti-nanog antibodies (E). F: Propagation efficiency of hES cells. StemPro supported the propagation of hES cells better than MEF-conditioned medium. G–J: The maintenance-supporting activity of PCM-DM was preserved for 8 months in the refrigerator under semi-dry conditions. G,H: Phase-contrast views. I,J: Alkaline phosphatase staining. G,I: Cells were cultured on PCM-DM that had been kept in the refrigerator for 3 weeks. H,J: Culture on 8-month-old PCM-DM. No significant reduction in the maintenance activity was observed. Scale bar = 100 μm.

We next tested whether hES cells maintained on PCM-DM in StemPro retained their pluripotency after multiple passages. After repeated passages (10 passages or more), hES cells were subjected to in vitro differentiation or teratoma formation assays. hES cells were differentiated in vitro into Brachyury+ mesodermal precursors (Fig. 5A; induced by serum; Watanabe et al.,2007), HNF3β+/E-cadherin+ endodermal precursors (Fig. 5B; induced by serum; Watanabe et al.,2007), and Nestin+/Pax6+ neuroectodermal progenitors (Fig. 5C; induced by SDIA culture; Kawasaki et al.,2000). In addition, the subcapsular injection of hES cells into SCID mouse testes caused the formation of teratomas (Fig. 5D) that contained mesodermal (e.g., cartilage; Fig. 5E), endodermal (e.g., mucous columnar epithelium; Fig. 5F), and ectodermal (e.g., brain neuroepithelium; Fig. 5G; pigment epithelium; arrows in Fig. 5D) derivatives. These findings demonstrated that human ES cells maintained with PCM-DM and StemPro remain pluripotent even after multiple passages. Furthermore, these cells showed a normal karyotype after multiple passages (Fig. 5H; 46 XX for KhES1).

Figure 5.

hES cells cultured on PCM-DM retain their pluripotency and normal karyotype after long-term culture in chemically defined medium. A,B: Immunostaining for mesodermal and endodermal markers of differentiating cells (day 6) on an 8-well chamber slide coated with collagen I and IV. Expression of the mesodermal marker Brachyury was seen in a number of differentiating cells (A). DAPI was used for nuclear staining (blue). Immunostaining of Hnf3β and E-cadherin in hES cell–derived endodermal tissues (B). C: Immunostaining for neural markers in differentiating cells on PA6 cells for 21 days. DG: hES cells were bilaterally injected into the SCID mouse testes. The teratomas contained a mixture of well-differentiated tissues including macroscopic pigment epithelium (D). Histological analysis showed the presence of (E) cartilage, (F) gut-like mucosa with columnar epithelium, and (G) neuroepithelium. H: G-banding analysis of hES cells cultured in StemPro on PCM-DM after 11 passages showed a normal karyotype (100%). Scale bar = (A,B,C,E,G) 100 μm, (D) 1 cm, (F) 50 μm.

Taken together, our results indicate that human PCM-DM can be used as an excellent substitute for animal-derived matrix substrates to maintain hES cells.

Dissociation Culture of hES Cells on PCM-DM in the Presence of a ROCK Inhibitor

One of the major differences between mouse and hES cells is that human ES cells frequently undergo apoptosis in dissociation culture. However, dissociation culture is necessary for several important culture procedures including cell cloning. We have recently shown that the ROCK (also called Rho kinase) inhibitor Y-27632 (Ishizaki et al.,2000) strongly suppresses the dissociation-induced apoptosis in hES cells and greatly increases their cloning efficiency (Watanabe et al.,2007). Therefore, we tested whether the PCM-DM-based culture was capable of supporting colony formation from dissociated hES cells in Y-27632-supplemented StemPro medium.

Dissociated hES cells cultured in StemPro efficiently formed colonies on PCM-DM when Y-27632 was present, but not in its absence (Fig. 6A,B; Y-27632 was added to the medium only for the first 2 days). At least in the case of dissociation culture with StemPro, the PCM-DM-based culture yielded an even better cloning efficiency than Matrigel-based or MEF-CM-based culture (Fig. 6C,D).

Figure 6.

Dissociated hES cells can be cultured on PCM-DM in the presence of ROCK inhibitor. A,B: Feeder cell-free hES cell culture on PCM-DM plates in StemPro medium. Colony formation from dissociated hES cells was clearly enhanced by Y-27632. Scale bar = 100 μm. Cell number (C) and colony formation (D) on Day 7 under each culture condition. Tukey's test was used for the repeated-measurement comparison between groups. Values shown on graphs represent the mean ± SD.

Thus, using PCM-DM as culture substrate, dissociated hES cells could be maintained with a simple treatment with ROCK inhibitor.

PCM-DM Also Supports the Maintenance of hiPS Cells in Chemically Defined Medium

Finally, we tested whether PCM-DM was applicable to the maintenance culture of human iPS cells. hiPS cells (a kind gift from Dr S. Yamanaka; 253G4, established without myc; Nakagawa et al.,2008) were cultured as cell clumps on PCM-DM or Matrigel in MEF-CM or StemPro. In both media, PCM-DM supported hiPS cell growth as efficiently as Matrigel did (Fig. 7A). hiPS cells cultured on PCM-DM expressed pluripotency-related markers such as alkaline phosphatase, Oct3/4, Nanog, TRA-1-60, and SSEA4 (Fig. 7B–F).

Figure 7.

Human iPS cells can be also maintained on PCM-DM in chemically-defined medium. A: hiPS cells (253G4) were cultured under various conditions. StemPro medium and a PCM-DM supported the efficient growth of hiPS cells. B: hiPS colonies plated on PCM-DM 4-well plate showed alkaline phosphatase activity. CF: hiPS colony on a 4-well plate (PCM-DM) in StemPro medium, Oct3/4 (C), Nanog (D), TRA-1-60 (E), and SSEA4 (F). Scale bar = 100 μm.

In conclusion, PCM-DM supports the maintenance of both types of human pluripotent cells (hES and hiPS cells) when used as the culture substrate. Thus, PCM-DM is a versatile human-derived material that can replace the animal-derived Matrigel substrate. On the other hand, we also found that the activity of human PCM-DM is not completely the same as that of Matrigel. Matrigel not only supports the maintenance of hES cells but is also applicable to several kinds of differentiation culture. For instance, whereas human ES cells differentiate into neural progenitors on Matrigel when cultured in low-growth factor medium, we have so far failed to observe efficient neural differentiation on PCM-DM (by day 23; our unpublished observations). Dot blot analysis showed that the PCM-DM contained substantial levels of fibronectin and collagen IV while laminin, heparan sulfate, and vitronectin were detected only at very low levels, if any (Supp. Fig. S7). This is in contrast to Matrigel, which is known to contain a high level of laminin (Supp. Fig. S7; >50% of the matrix components according to the manufacturer's document) in addition to the ECM components described above. The role of the very low laminin level in PCM-DM remains elusive, and no substantial differences in plating efficiency or colony growth were observed between PCM-DM and PCM-DM + laminin (coating at 1 μg/cm2) at least for short-term culture (Supp. Fig. S8). The similarities and differences in the activation of intracellular signals of hES cells cultured on Matrigel versus PCM-DM is an intriguing basic question for future study.

DISCUSSION

Human PCM-DM Can Substitute for Animal-Derived Matrix Substrates in hES and hiPS Cell Maintenance Culture

This study showed that human PCM-DM is a highly effective matrix substrate for the maintenance culture of human pluripotent stem cells. An important aspect of PCM-DM is that it supports hES cell maintenance in the animal component-minimized StemPro medium with an efficiency similar to that of the widely used matrix substrate Matrigel. This indicates the potential suitability of PCM-DM as a clinical-grade culture substrate for hES cells.

PCM-DM can be prepared in a large quantity from human DMCs, which are abundantly present in the human decidua (a typical yield is 2×103 cells/cm2 membrane; see the Experimental Procedures section). These cells grow efficiently in culture (doubling time 57 hr) and express typical mesenchymal markers such as phalloidin and vimentin (D.K. and Y.K., unpublished data). Starting with a 5 × 5 cm human decidua, 2 × 106 cells can be routinely obtained after a 12-day culture and >109 cells after 4–5 weeks; this number of cells is sufficient to prepare PCM-DM for > 3,000 3.5-cm plates. Owning to this availability, PCM-DM should be feasible for medical use in a quality-controlled manner (with pretests for its activity and safety), particularly given that the PCM-DM plates retain good activity for at least 8 months in the refrigerator.

Requirement of Proper Matrix Signaling in hES and hiPS Cell Culture

Human and mouse pluripotent stem cells differ in the requirements for their culture conditions. Unlike mouse stem cells, human pluripotent stem cells do not respond to LIF while Fgf signals play an important role in hES and hiPS cell culture (Wobus and Boheler,2005; Takahashi et al.,2007). mES cells form colonies efficiently from single cells whereas dissociated hES cells frequently undergo apoptosis and have low cloning efficiency (Watanabe et al.,2007).

Another large difference between human and mouse pluripotent cells in culture conditions is the requirement for matrix substrates. In the presence of LIF, mES cells grow well not only on the Matrigel substrate but also on dishes coated with poly-D-lysine, collagen, or even gelatin (Ogawa et al.,2004), which is not generally the case for hES cell culture. Whereas laminin and fibronectin have positive effects on hES cell maintenance (Xu et al.,2001; Lu et al.,2006), these ECMs promote differentiation in mES cells at least under some conditions (Hayashi et al.,2007). Then, what makes this difference?

Recent studies of mouse EpiStem cells (epiblast-derived pluripotent cells) have shed light on this issue. mES cells are derived from the inner cell mass (ICM) and keep many characteristics of their origin. For instance, mES cells do not have a clear apical-basal (A–B) polarity when cultured in vitro. The ICM lacks the basement membrane and does not exhibit A–B polarity before implantation. In contrast, the epiblast, a direct derivative of the ICM, has an evident A–B polarity with the apical side on the pre-amniotic cavity side. The basement membrane is formed between the epiblast and the visceral endoderm (reviewed in Hogan et al.,1994). Mouse EpiStem cells are derived from the epiblast of the post-implantation embryo and cultured under conditions similar to hES cell culture (Brons et al.,2007; Tesar et al.,2007). Interestingly, detailed analyses of mouse EpiStem cells in gene expression and cell behaviors (e.g., requirements in culture conditions) show a number of similarities not only to epiblast cells, but also to hES cells. This fact led to the persuasive proposal that hES cells represent human pluripotent cells related to the epiblast status rather than the ICM status (Brons et al.,2007; Tesar et al.,2007). This is also in agreement with the flat, simple columnar epithelial morphology of a hES/hiPS cell colony.

This idea seems to explain, at least in part, the difference between hES and mES cells with respect to the substrate requirement: with respect to survival and growth, hES cells (epiblast-type) are more dependent on the presence of basement membrane-like matrices while mES cells (ICM-type) are not. In addition, a recent study has shown that the expression profiles of integrin subunits are different between hES and mES cells (Braam et al.,2008).

A remaining question is what is the essential signal(s) of the basement membrane matrix. Generally speaking, little is known still about the identities of the active components in pericellular or basement membrane matrices, mainly because of their complexity in molecular compositions. In hES cell culture, defined matrix substrates such as laminin and fibronectin facilitate the attachment and/or growth, but they are not necessarily as potent as Matrigel and PCM-DM, suggesting that the substrate-derived signals could be more complex than just Integrin signals. In addition to the macromolecular matrix components, growth factor molecules tethered to the matrices may also play crucial roles. In the same line, functional comparison between PCM-DM and visceral endoderm-derived matrices is a challenging research topic.

Future Research towards Medical Applications

One possible improvement of the PCM-DM method for medical application is to use a cell line of DMCs, instead of primary culture cells, because the quality control and mass production are even easier with an established line. Since human cell lines are generally more difficult to establish than mouse lines, the immortalization of human cells may require the transfer of promoting genes such as TERT (Bodnar et al.,1998; Nishiyama et al.,2007). Given that the PCM-DM method involves cell lysis, the use of genetically manipulated cells (especially the introduction of a human non-oncogene) would not substantially increase the a priori risk in human cell therapy.

Another important application of the PCM-DM method may be the derivation of hES cells and hiPS cells under feeder-free and serum-free conditions. We hope that this versatile human-derived matrix will contribute to the promotion of human pluripotent cell research for medical application in a variety of ways.

EXPERIMENTAL PROCEDURES

hES Cell Maintenance Culture

hES cells (KhES-1, KhES-3) (Suemori et al.,2006) were a gift from N. Nakatsuji and H. Suemori (Kyoto University) and were used in this study following the hES cell guidelines of the Japanese government. Undifferentiated hES cells were maintained on a feeder layer of MEFs (Invitrogen, Carlsbad, CA, http://www.invitrogen.com; inactivated with 10 μg/ml mitomycin C) in DMEM/F12 (Sigma Aldrich, Cambridge, U.K. http://www.sigmaaldrich/) supplemented with 20% KSR, 2 mM glutamine, 0.1 mM nonessential amino acids (Invitrogen), 5 ng/ml recombinant human bFGF (Wako, Osaka, Japan, http://www.wako-chem.co.jp/english/), and 0.1 mM 2-mercaptoethanol under 2% CO2.

For passaging, hES cell colonies were detached and recovered en bloc from the feeder layer by treating them with 0.25% trypsin and 0.1 mg/ml collagenase IV in PBS containing 20% Knockout Serum Replacement (KSR) (Invitrogen) and 1 mM CaCl2 at 37°C for 7 min, followed by tapping the cultures and flushing them with a pipette (Suemori et al.,2006). After two volumes of culture medium was added, the detached ES cell clumps were broken into smaller pieces (approximately 20 cells) by gently pipetting them several times. The passages were performed at a 1:4 split ratio.

For storage, the ES cell colonies were recovered en bloc (without further dissociation) from a 6-cm culture dish, suspended in 1 ml of ice-cold culture medium supplemented with 2 M DMSO, 1 M acetamide, and 3 M polypropylene glycol, and quickly frozen in a 2-ml cryogenic tube (BD Biosciences, San Jose, CA http://www.bdbiosciences.com/) by directly submerging the tube in liquid N2.

Preparation of Pericellular Matrix of Decidua-Derived Mesenchymal Cell (PCM-DM)

These procedures were carried out in accordance with the principles of the Helsinki Declaration and approvals to use human fetal membranes were obtained from the ethical committees of both the Osaka National Hospital and RIKEN CDB. Human full-term fetal membranes (FMs) were collected from normal healthy mothers at the Osaka National Hospital with written informed consent. The donor mothers' bloods were serologically tested for HBs, HCV, HIV, and syphilis. After extensive washing with phosphate-buffered saline (PBS), the decidual tissues were isolated by mechanically removing the chorio-amniotic membranes (AMs) from the FMs. The remaining FMs (decidual tissues) were dissected into small pieces and enzymatically dissociated in PBS containing collagenase (1 mg/ml; Invitrogen), dispase (1 mg/ml; Invitrogen), and DNase I (final 0.01%, Invitrogen) for 1 hr at 37°C with gentle shaking. The suspensions were filtered through a 40-mm nylon mesh (Cell Strainer, BD Biosciences) and single-cell suspensions (2×104 cells/ml) were propagated in 100-mm-diameter cell culture dishes using DMEM/F-12 (1:1)-based culture medium supplemented with 10% fetal bovine serum (FBS), 15 mM HEPES, and antibiotic-antimycotic (Invitrogen) at 37°C in 5% CO2. The culture medium was replaced twice a week and the cells were passaged using trypsin-EDTA (0.05%, Invitrogen) once a week. The human DMC used in this study had been cultured for 50 days or less in vitro (DIV).

The pericellular matrix (PCM-DM in this study) was prepared by a modification of the method described by Klimanskaya and others (Klimanskaya et al.,2005). Briefly, DMC were plated at a density of 3.5×104 cells/cm2 on culture plates precoated with 0.1% gelatin and cultured in DMEM/F12 supplemented with 10% FBS and 15 mM HEPES for 3 days. After being rinsed with PBS, the cells were treated at 4°C for 30 min with deoxycholate solution (0.5% sodium deoxycholate in 10 mM Tris-HCl, pH 8.0). After cell lysis, the dishes were washed six times with PBS by pipetting to flush off the cell debris. The treated dishes were stored under semi-dry conditions at 4°C. A schematic of the preparation procedure is shown in Supplemental Figure 3.

Feeder-Free Culture on Matrices in Conditioned and Non-Conditioned Media

MEF-conditioned medium (MEF-CM) was prepared by incubating the maintenance culture medium with confluent MEFs for 24 hr. Although the conditioned medium of DMCs also had a similar growth-supporting activity, MEF-CM was used in this study because it is widely used in this field. STEMPRO hESC SFM (Invitrogen) is a chemically defined medium whose exact components are not disclosed, and was prepared according to the manufacturer's instructions: DMEM/F12 supplemented with GlutaMAX (1×), STEMPRO hESC SFM Growth Supplement, 1.8% purified bovine serum albumin (we used the one purified to >99% by crystallization; Sigma), 8 ng/ml bFGF, and 0.1 mM 2-mercaptoethanol.

For the preparation of matrix-coated plates, plates were incubated with Matrigel (Matrigel Growth Factor Reduced, BD Biosciences, 1:30), fibronectin (human plasma fibronectin, Invitrogen, 5 μg/cm2), or gelatin (Sigma, 0.1%) at room temperature for 1 hr. Before seeding hES cells onto the plates, contaminating MEF cells from the previous culture were removed by incubating the hES cell suspension on a gelatin-coated plate at 37°C for 2 hr in the maintenance culture medium so that MEF cells would adhere to the plate. For the feeder-free cultures, 1.8 × 104 cells were plated in each well of a 24-well plate. Colonies were stained for alkaline phosphatase activity on day 6. Both hES cell lines (KhES-1, KhES-3) grew equally well on PCM-DM.

Dissociated Cell Culture of hES Cells

hES cell clumps were recovered by centrifugation, washed with PBS, incubated in TrypLE Select (Invitrogen) at 37°C for 5 min, and dissociated into single cells by pipetting. The dissociated cells were seeded onto 24-well plates at a low density (2,000 cells/well, 2 cm2), and cultured in 5% CO2 according to the manufacturer's instructions. Y-27632 was added to the culture medium at 10 μM for the first 2 days after dissociation.

Histochemical and Chromosomal Analyses of hES Cells

hES cell cultures were fixed in 4% paraformaldehyde for 15 min, permeabilized for 30 min in 0.1% Triton X-100, and blocked in 2% skim milk in phosphate-buffered saline. The cells were incubated with primary antibodies overnight at 4°C and visualized by incubation with fluorescein isothiocyanate (FITC)- or Cy3-conjugated secondary antibodies for 60 min at room temperature. Cell nuclei were counterstained with 4,6-diamidino-2-phenylindole. The alkaline phosphatase activity was determined using an alkaline phosphatase activity detection kit (Sigma) according to the manufacturer's instructions.

Chromosomal G-band analysis (300–400 band level) was performed by pretreating hES cells with 0.06 μg/ml colcemid for 3–4 hr and dissociating them in TrypLE Select, followed by incubation with 0.075 M KCl (hypotonic) for 10 min and fixation in Carnoy's fixative.

Antibodies for Immunostaining

Commercial antibodies were purchased from Developmental Studies Hybridoma Bank (Iowa City, IA, http://dshb.biology.uiowa.edu/) (HNF3β/mouse monoclonal/4C7/1:100), BD Biosciences Pharmingen (San Jose, CA) (Oct-3/mouse monoclonal/611202/1:200, N-cadherin/mouse monoclonal/610920/1:500), Millipore (Billerica, MA, http://www.millipore.com/) (TRA-1-60/mouse monoclonal/MAB4360/1:200, SSEA4/mouse monoclonal/MAB4304/1:200, human vimentin/goat polyclonal/AB1620/1:100), Covance (Berkeley, CA, http://www.covance.com/) (human Nestin/rabbit polyclonal/PRB-570C/1:1,000), R & D Systems (Minneapolis, MN, http://www.rndsystems.com/) (Pax6/mouse monoclonal/MAB1260/1:500, Brachyury/goat polyclonal/AF2085/1:50, human cytokeratin 14/mouse monoclonal/MAB3164/1:100), Reprocell (Tokyo, Japan, http://www.reprocell.com/en/) (human Nanog/mouse monoclonal/RCAB0003P/1:1,000), Takara (Shiga, Japan, http://www.takara-bio.com/) (E-cadherin/rat monoclonal/M108/1:50), SantaCruz, (SantaCruz, CA, http://www.scbt.com/) (human cytokeratin 19/mouse monoclonal/1:200), EXBIO (Vestec, Czech Republic, http://www.exbio.cz/) (HLA-G/mouse monoclonal/11-431/1:100), Miltenyi Biotec (Auburn, CA, http://www.miltenyibiotec.com) (PE human CD133/mouse monoclonal/130-080-801/1:100), and Invitrogen (Carlsbad, CA, http://www.invitrogen.com/) (AlexaFluor-568 phalloidin/A12380/1:40).

In Vitro Differentiation Assay

For their in vitro differentiation into mesodermal or endodermal progenitors, hES cells were plated on a collagen-I-coated culture slide (BD Biosciences) coated with collagen IV, and cultured in DMEM/F12 supplemented with 10% FCS, 2 mM glutamine, 0.1 mM nonessential amino acids, and 0.1 mM 2-mercaptoethanol under 5% CO2 for 6 days. On day 6, about 5% of the cells were positive for Brachyury staining, and 2–5% of the cells were positive for Hnf3β staining.

For the neural differentiation medium in this study, we used the stromal cell–derived inducing activity (SDIA) method with PA6 feeder cells (Kawasaki et al.,2000,2002) in G-MEM medium, supplemented with 10% KSR, 2 mM glutamine, 1 mM pyruvate, 0.1 mM nonessential amino acids, 0.1 mM 2-mercaptoethanol, 100 U/ml penicillin, and 100 μg/ml streptomycin.

For in vitro differentiation into epidermal differentiation, hES cells were plated on a collagen-I-coated culture slide (BD Biosciences) coated with gelatin, and cultured in DMEM/F12 supplemented with 20% KSR, 2 mM glutamine, 0.1 mM nonessential amino acids, 0.1 mM 2-mercaptoethanol, and 0.5 nM BMP4.

Teratoma Formation Assay

Teratoma formation was examined by injecting approximately 1×105 hES cells subcapsularly into the testes of 8-week-old severe combined immunodeficient mice (SCID/CrlCrlj) using a Hamilton syringe. The mice were sacrificed after 12 weeks, and the tumors were excised and fixed in 4% paraformaldehyde. Paraffin sections (5-μm-thick) of the tumors (formed in 9 out of 10 testes) were evaluated histologically by HE staining for the presence of the three germ-layer derivatives.

FACS Analysis

Cells were dissociated to single cells by TrypLE select treatment, and stained with primary antibodies for 1 hr at 4°C. The cells were washed with medium, and then stained again with secondary antibodies labeled with FITC or Cy3 for 1 hr at 4°C. After then the cells were suspended in HBSS containing 1% BSA. Cells were counted by FACSAria (BD Biosciences). Dead cells were excluded by gating on forward and side scatter. Commercial antibodies were purchased from Millipore (SSEA1/mouse monoclonal/MAB4301/1:100, SSEA3/rat monoclonal/MAB4303/1:100, SSEA4/mouse monoclonal/MAB4304/1:100, Tra-1-60/mouse monoclonal/MAB4360/1:100). The cells stained with secondary antibody only were used as a negative staining control.

Dot Blot Analysis

PCM-DM was solubilized from the 15-cm culture dish bottom with 1.0 ml of 0.3M Tris (pH 7.2) containing 3% SDS at room temperature (finally, 36 μg protein/ml) (Parker et al.,1998). An aliquot of 400 μl of each solution was transferred to nitrocellulose membrane with a blotter system (Whatman Minifold I) (16 or 120 μg/cm2, and serial dilution by 2-fold). The membrane was incubated with a blocking solution consisting of 2% skim milk in TBST (TBS containing 0.05% Tween-20) for 30 min, then with each antibody in blocking solution at room temperature (RT) for 3 hr. Commercial antibodies were purchased from Millipore (fibronectin/mouse monoclonal/MAB88916/1:1,000, human laminin/mouse monoclonal/MAB88918/1:1,000, mouse laminin/rat monoclonal/MAB1905/1:1,000), American Research Products (Belmont, MA, http://arpl.com) (collagen Type IV/rabbit polycronal/03-10760/1:1,000), Thermo Fisher Scientific (Fremont, CA, http://www.thermo.com/) (heparan sulfate proteoglycan/rat monoclonal/RT-794-P0/1:1,000), and Epitomics (Burlingame, CA, http://www.epitomics.com/) (vitronectin/rabbit monoclonal/1783-1/1:1,000). After washing twice with TBST for 5 min each time, the membrane was incubated with HRP-conjugated antibodies (1:5,000 in blocking solution) at RT for 1 hr. After washing twice with TBST for 5 min each time, the bound antibody was developed to show the HRP product using ECL Western Blotting Detection Reagents (GE Healthcare Amersham). Total PCM-DM protein was determined by the BCA assay (Pierce). Images were acquired on a Fujifilm luminoimage analyzer (LAS-3000). The following proteins were used as positive controls in this assay: fibronectin (Sigma/F2006); collagen IV (BD/354245); human laminin (Sigma/L6274); mouse laminin (BD/354239); heparan sulfate proteoglycan (Sigma/H4777); and vitronectin (Sigma/V8379). BSA was used as a negative control. Matrigel contained 12.4 μg protein/μl.

Acknowledgements

We are grateful to Dr. Satoshi Ando for invaluable comments, to Dr. Chiaki Ban and Dr. Mami Yamasaki at Osaka National Hospital for providing the human fetal membranes, to Dr. Nobuyuki Shimozawa for encouragement, and to members of the Sasai laboratory for discussion and advice. This work was supported by grants-in-aid from MEXT, The Kobe Cluster Project and the Leading Project (Y.S. and Y.K.).

Ancillary