Leukemia inhibitory factor as an anti-apoptotic mitogen for pluripotent mouse embryonic stem cells in a serum-free medium without feeder cells

In Vitro Cell. Dev. Biol.--Animal 41:19-28, January and February 2005 �9 2005 Society for In Vitro Biology 1071-2690/05 $18.00+0.00

LEUKEMIA INHIBITORY FACTOR AS AN ANTI-APOPTOTIC MITOGEN FOR PLURIPOTENT MOUSE EMBRYONIC STEM CELLS IN A SERUM-FREE MEDIUM

WITHOUT FEEDER CELLS

MIHO FURUE] TETSUJI OKAMOTO, YOHEI HAYASHI, HITOSHI OKOCHI, MANABU FUJIMOTO, YASUFUMI MYOISHI, TAKANORI ABE, KIYOSHI OHNUMA, GORDON H. SATO, MAKOTO ASASHIMA, AND J. DENRY SATO

Department of Biochemistry and Molecular Biology, Kanagawa Dental College, Yokosuka 238-8580, Japan (M. Furue), Department of Molecular Oral Medicine and Maxillofacial Surgery, Division of Frontier Medical Science, Graduate School of Biomedical Sciences,

Hiroshima University, Hiroshima 734-8553, Japan (T. 0., Y. M.), Department of Life Sciences (Biology), Graduate School of Arts and Sciences, The University of Tokyo, Tokyo 153-8902, Japan (g H., K. 0., M. A.), Department of Tissue Regeneration, Researclr Institute,

lnternatioaal Medical Center of Japan, Tokyo 162-8655, Japan (H. 0., M. Fujimoto), Department of Biological Science, Graduate School of Science, The University of Tokyo, Bunkyo-ku, Tokyo 113-0033, Japan (T. A., M. A.), Ministry of Fisheries, Massawa, Eritrea (G. H. S.),

International Cooperative Research Project (1CORP)-Japan Science and Technology Agency (JST), The University of Tokyo, Tokyo 153- 8902, Japan (M. A.), and Marine Cell Line and Stem Cell Program, Mount Desert Island Biological Laboratory, Salisbury Cove,

Maine 04672 (J. D. S.)

(Received 8 February 2005; accepted 18 February 2005)

SUMMARY

We have developed a serum-free medium, designated ESF7, in which leukemia inhibitory factor (LIF) clearly stimulated murine embryonic stem (ES) cell proliferation accompanied by increased expression of nanog and Rex-1 and decreased FGF-5 expression. These effects were dependent on the concentration of LIF. The ES cells maintained in ESF7 medium for more than 2 yr retained an undifferentiated phenotype, as manifested by the expression of the transcription factor Oct-3/4, the stem cell marker SSEA-1, and alkaline phosphatase. Withdrawal of LIF from ESF7 medium resulted in ES cell apoptosis. Addition of serum to ESF7 medium promoted ES cell differentiation. Addition of BMP4 promoted ES cell differentiation into simple epithelial-like cells. In contrast, FGF-2 promoted ES cell differentiation into neuronal and glial-like cells. Under serum-free culture conditions, LIF was sufficient to stimulate cell proliferation, it inhibited cell differentiation, and it maintained self-renewal of ES cells. Because this simple serum-free adherent monoculture system supports the long-term propagation of pluripotent ES cells in vitro, it will allow the elucidation of ES cell responses to growth factors under defined conditions.

Key words: ES; mouse embryonic stem cells; serum-free; LIF; nanog; 0ct-3/4.

INTRODUCTION

Pluripotent murine embryonic stem (ES) cells (Evans and Kauf- man, 1981; Martin, 1981) have the capacity to give rise to all the cell types in the developing mouse, and they have the potential to generate all differentiated cell types in vitro. Thus, ES cells have represented a model of mammalian development that is amenable to biochemical and molecular analyses (Gardner and Brook, 1997; Smith, 2001; Tanaka et al., 2002). Recently-, ES cell biology has become a focus for biotechnology. Major limitations of the use of stem cells in a variety of biotechnological applications are the cur- rent requirement for feeder ceils and the relative lack of knowledge regarding the long-term maintenance of pluripotency. An experi- mental system in which rules governing pluripotency and differentiation could be deciphered would encourage the development of new stem cell applications. Our previous studies in Xenopus have shown that the developmental fate of undifferentiated cells can be decided by concentration-sensitive inducing factors (Asashima et

1To whom con'espondenee should be addressed at Department of Bio- chemistry and Molecular Biology, Kanagawa Dental College, 82 Inaoka-cho, Yokosuka 238-8580, Japan. E-mail: [email protected]

al., 2000; Furue and Asashima, 2004) and most of those factors are present during both early mouse and Xetmpt~s development (Trem- blay et al., 2000; Tiedemann et al., 2001). Although the properties of pluripotent mammalian cells have been investigated for over 30 yr, the guidance of pluripotent cell differentiation during early mouse development is still poorly understood.

Murine ES cells are commonly maintained on primal3: etnb~3,onic fibroblast feeder cells in culture medium supplemented with serum and leukemia inhibitory factor (LIF). Previous studies demonstrated that self-renewal and pluripoteney could be maintained under these culture conditions (Smith et al., 1988; Williams et al., 1988). In the absence of LIF, ES cells differentiate spontaneously in serum- containing culture medium. Culture protocols have been developed recently that permit the derivation of some types of cells from undifferentiated ES cells. However, because these procedures require the cultivation of ES cell aggregates (Wiles and Johansson, 1999), conditioned medium containing undefined components (Lake et al., 2000), or feeder cells (Kawasaki et al., 2000), they produce variable and heterogeneous results. Differentiation protocols based on knowledge of ES cell responses to specific growth factors are not currently available for most types of mouse cells.

19

20 FURUE ET AL.

TABLE 1

COMPOSITION OF ESF NUTRIENT MEDIUM"

Amino acids L-Alanine 2.225 L-Arginine 50 L-Arginine HCI 94.75 L-Asparagine H20 16.2525 L-Asparagine 8.325 L-Cysteine HC1 H20 8.78 L-Cysteine 2HC1 47.5725 L-Glutamic acid 8.675 L-Glutamine 549.65 L-Hydroxyprotine 19.375 L-Histidine 23.165 L-Hyroxyproline 5 L-Isoleucine 65.935 L-Leucine 68.225 L-Lysine HCl 92.175 L-Methionine 19.87 L-Phenylalaine 37.99 L-Proline 13.625 L-Serine 31.125 L-Threonine 55.525 L-Tryptophan 9.76 L-Tyrosine 42.36 L-Valine 54.825

Vitamins Para-aminobenzoic acid 0.25 Biotin 0.05185 Calcium pantothenate 2.1825 Choline chroride 6.24 Folic acid 2.575 Inositol 16.85 Nicotinic acid amide 2.25925 Pyridoxal HC1 2 Pyridoxine HC1 0.2655 Riboflavin 0.2595 Thiamine HCI 2.335 Vitamin B~2 0.34125 Hypoxanthine 1.02

Inorganic salts Sodium chloride 6599,75 Potassium chloride 355,9 Calcium chloride (anhydrous) 108,305 Calcium nitrate 4H20 25 Magnesium chloride (anhydrous) 14.305 Magnesium sulfate (anhydrous) 61.055 Sodium phophate monobasic (anhydrous) 54.35 Sodium phosphate dibasic (anhydrous) 235.51 Ferric nitrate 9H~O 0.05 Cupric sulfate 5H~O 0.000625 Ferrous sulfate 7H20 0.2085 Zinc sulfate 7H20 0.216

Others Glutathione 0.25 Linoleic acid 0.021 Thioctic acid 0.0525 Putrescine 2HC1 0.04025 Thymidine 0.1825 Glucose (anhydrous) 2325.5 Sodium pyruvate 110 Phenol red 6.56 HEPES ~ 3574.5 NaHCQ 2000

Values are given in mg/L. b HEPES, N-2-hydroxyethyl piperazine-N'-2-ethane-sulfonic acid.

Wiles and Johansson (1999) and Ying et al. (2003a) reported that

populat ions with a non-ES cell morphology arose when ES cells

were mainta ined for more than three passages in the absence of

feeder ceils in chemical ly defined serum-free media containing LIF

(10 ng/ml), In this study, we descr ibed a serum-free culture medi-

um, designated ESF7, in which a novel mitogenic activity of LIF

was uncovered and ES cell self-renewal and pluripotency were

main ta ined after serial cultivation in the absence of feeder cells.

Using this culture system, it will be possible to elucidate the de-

velopmental responses of ES cells to specific environmenta l stimuli.

MATERIALS AND METHODS

The ES-D3 cell culture. Murine ES-D3 cells (Doetschman et al., 1985) (ATCC, CRL-1934) were euhured on 0.1% gelatin-coated dishes (Cell & Molecular Technologies, Phillipsburg, N J) in Complete ES medium (CEM; Dulbecco modified Eagle medium supplemented with 15% fetal bovine serum [FBS], L-glutamine, 2-mercaptoethanol [2-ME], nucleosides, nonessential amino acids, and 10 ng/ml LIF [Cell & Molecular Technologies]). Later, ES- D3 ceils were maintained on 0.15 mg/ml type I collagen-coated dishes (Nitta Gelatin Co., Osaka, Japan) in ESF7 medium, which consisted of a nutrient medium, designated ESF (Table 1) containing six factors (6F) (Sato et al., 1987, 2002) and 3 ng/ml LIF (ESGRO @, Chemicon International, Temecula,

CA). The 6F consisted of 10 Ixg/ml bovine insulin, 5 txg/ml human transferrin, 10 fxM 2-aminoethanol, 10 l.zM 2-ME, 20 rrM sodium selenite, and 9.4 l~g/ml oleic acid complexed with fatty acid-free bovine serum albmnin (FAF-BSA) (all from Sigma Chemical Co., St. Louis, MO). Each facto*" was made as a sterile lO0X concentrate and stored at 4 ~ C. The cells were incubated at 37 ~ C in a humid atmosphere of 5% CO 2. The ES-D3 cells were harvested in 0.001% trypsin-O.01% ethylenediaminetetraacetic acid (EDTA) in phosphate-buffered saline (PBS), and the trypsin was inactivated with 0.1% soy- bean trypsin inhibitor (Sigma). The cells were reseeded at 2 • 105 cells in collagen-coated 75-cm 2 flask (Corning, Coming, NY). Culture medium was changed every 2 d, and the cells were subcultured evecy 4 d.

The ES-D3 cell proliferation. To test LIF for mitogenic activity on ES-D3 cells, the cells were seeded in a type I collagen-coated 24-well plate (Falcon, Oxhard, CA) at a cell density of 5 • 10 a cells/well in ESF7 without LIF or in a gelatin-coated 24-well plate in CEM without LIF (CEM[-]). Increasing concentrations of LIF were added to each well, and the cells were counted after 5 d in culture with a Coulter particle counter (Beckman Coulter, Hia- leah, FL). To generate growth curves ES-D3 cells were seeded at 5 • 10 "~ cells/well in ESF7 medium in a collagen-coated 24-well plate or in CEM in a gelatin-coated 24-well plate. The plates were incubated at 37 ~ C in a humid atmosphere of 5% COs, and cells in triplicate wells were counted daily.

To detect DNA profile in ES-D3 cells cultured in the presence or absence of LIE ES-D3 cells were inoculated on collagen-coated dishes in ESF7 medium and cultured for 2 d. The cells were washed and given ESF7 medium without LIF for an additional 4 d. The cells were harvested, washed with

SERUM-FREE CULTURE FOR ES CELLS

TABLE 2

PRIMERS USED FOR REAL-TIME RT-PCR ~

21

Gene Primer sequence

nanog Rex- 1 0ct3/4 F ~ GAPDH

5'-aagcctttccatgtggggca-3' 5'-ttcaacttgcgcacccacat-3' 5'-ttctgcggagggatggcata-3' 5'-gctgtgtctcaggggattgt-3' 5'-acccagaagactgtggatgg-3'

5'-atggagcggagcagcattcc-3' 5'-cctgcctttgcgtgggttag-3' 5'-tttccactcgtgctcctgcc-3' 5'-cactctcggcctgtcttttc-3' 5'-cacattgggggtaggaacac-3'

RT-PCR, reverse transeriptase-polymerase chain reaction; GAPDH, reduced fom~ of guanosine adenine dinueleotide phosphate.

PBS, and fixed with ice-cold ethanol. Then, the cells were collected and incubated in extraction buffer (45 mM Na~HPQ, 2.5 mM citric acid, 0.I% Triton X-IO0) for 20 rain at 37 ~ C. The cells were stained with 50 pxg/ml propidium iodide for 30 rain at the room temperature and analyzed with an Epics Ahra flow cytonmter (Beckman Coulter) (Hots et al., 1994).

The 129/SV ES cell culture. Passage 10 ES cells derived from 129/SV mouse (Cell & Molecular Technologies) were thawed and cultured on fibroblast feeder cells (Asahi Technoglass Co., Tokyo, Japan) in CEM. The cells were harvested in 0.1% type IA-S collagenase (Sigma) and then dispersed with 0.01% EDTA. The cells were l~seeded in ESF7 medium (20 ng LIF/ ml) on flasks precoated with 0.5 mg/ml type I collagen (Nitta Gelatin Co.). In proliferation experiments, 129/SV ES cells were seeded at 4 • 104 cells/ well in ESF7 medium in a collagen-coated 24-well plate and cells in triplicate wells were counted daily.

The ES-D3 differentiation in vitro. The ES-D3 cells were inoculated on 2 ~g/em 2 laminin (Biomedical Technologies Inc., Stoughton, MA), Stoughton, Mass, coated plates in ESF7 medium minus LIF and BSA-oleic acid and supplemented with 10 ng/nfl fibroblast growth-factor-2 (FGF-2) (Upstate Bin- technology Inc., Lake Placid, NY) or 10 ng/ml Bone molphogenetic protein- 4 (BMP4) (R&D Systems Inc., Minneapolis, MN), or the cells were inoculated on 0.1% gelatin-coated dishes in ESF basal medimn supplemented with N2 (Invitrogen, Carlsbad, CA) and B27 (Invitrogen) supplements without LIF. Ceils cultured with FGF-2 and 100 ng/ml heparin (Sigma) for 2 d were then cuhured without FGF-2 and heparin for 4 d or nmre. A total of 0.5 mg/ml FAF-BSA was added with BMP4. The FGF-2 and BMP4 were added daily, and the medium was replaced every 2 d. Cultures were scored for cells with differentiated morphologies on day 6.

Alkaline phosphatase activity. Cells plated on collagen-coated dishes in ESF7 medium or on gelatin-coated dishes in CEM wele euhured for 6 d with medium renewal every 2 d. Alkaline phosphatase (AP) activity in the cells was detected using a Fast Red substrate kit (Nichirei, Tokyo, Japan), as described by Toumadje et al. (2003).

Im,umohistochemical analysis. To detect pluripotent stem cell marker antigens, cells were euhured under the conditions described above and fixed in 4% (w/v) parafo~Tnaldehyde. Fixed cells were stained with a mouse anti- SSEA-1 antibody (Kyowa Medex Co., Tokyo, Japan). To detect the intracel- lular antigen Oct-3/4, fixed cells were permeabilized with 1% (w/v) saponin (Sigma) and then stained with a mouse anti-Oct-3/4 antibody (Transduetion Laboratories, Lexington, KY). Primary antibodies were observed with per- oxidase-conjugated Simple Stain MAX PO | goat anti-nmuse immunoglobulin G (IgG) (Nichit~i) and 3-amiuo-9-ethylcarbazol (Niehirei). For the cell surface differentiation antigen A2B5, unfixed cells were incubated with mouse anti- A2B5 antibody (Chemieon) for 15 min at 37 ~ C and then incubated with fluorescein isothioeyanate (FITC)-eonjugated goat anti-mouse IgG (Sigma) for 15 min at 37 ~ C. For intraeellular differentiation antigens, fixed cells were permeabilized with 0.2% Triton X-100 in PBS containing 0.2% BSA and 5% goat serum for 15 min and then immunostained with antinestin antibody (BD Bioseiences Immunoeytometry Systems, Franklin Lakes, NJ), mouse anti-13- tubulin III (Chemienn), rabbit anti-glial fibrillary acidie protein (GFAP) (San- ta Cruz Binteehnology Inc., Santa Cruz, CA), rabbit antikeratin (Niehirei), mouse antieytokeratin 14 (a gift from Prof. E. B. Lane) (Purkis et al., 1990), or mouse anticytokeratin 18 (Progen Bioteehnik GmbH, Heidelberg, Ger- many). Primary antibodies were detected with FITC-conjugated goat anti- mouse IgG (Sigma) or FITC-conjugated goat anti-rabbit IgG (Sigma). 4,6- Dianfidino-2-phenylindole (Sigma) was used as a counterstain.

Flow cytometric analysis of 0ct3/4, A2B5, and keratin expression. To detect Oct-3/4 antigen, cells were plated on collagen-coated dishes in ESF7 medium or in ESF medium supplemented with FBS and 2-ME or on gelatin-coated

dishes in CEM. The cells were cultured for 6 d with fresh medium added every 2 d. Analysis of Oct-3/4 expression was done according to Toumadje et ah (2003) using an Epics Ahra flow cytometer (Beckman Coulter). To analyze the expression of A2B5 antigen and keratin, cells cultured under the condition described above were dissociated with trypsin-EDTA and then immunostained with anti-A2B5 antibody or antikeratin antibody, as described above.

Real-time reverse transcriptase-polymerase chain reaction. Cells were plated on collagen-coated dishes in ESF7 medium containing increasing concentrations of LIF and cultured for 4 d. Total RNA extracted from the cells was treated with ribonuclease-free deoxyribonuclease I (GIBCO/BRL, Carland, CA). Real-time reverse transcriptase-polymerase chain reaction (RT-PCR) was then performed on an ABI PRISM 7700 (Applied Biosystems, Foster City, CA) using SYBR Green PCR Master Mix (QIAGEN, Hilden, Germany) according to the QuantiTect SYBR Green Kit instructions. Primer pairs used are listed in Table 2. The efficiency of complementary DNA synthesis from RNA was assessed on the basis of real-time RT-PCR products for reduced form of guanosine adenine dinucleotide phosphate. The expression of individual messenger RNAs was calculated relative to expression in cells cub tured in ESF7 medium containing 3 ng/ml LIE

RESULTS

Effects of cuhure conditions on ES-D3 cell morphology. The ES-

D3 cells were main ta ined cont inuously for more than 2 yr on type

I col lagen-coated flasks in the absence of a feeder layer in ESF7

medium, which consis ted of ESF nutr ient med ium supplemented

with 6F and 3 ng/ml LIE After three initial passages in ESF7 me-

d ium following propagation in se rum-supp lemen ted medium, most

of the colonies of densely packed ES-D3 cells exhibi ted a smooth outline consis tent with the cells retaining an undifferentiated mor-

phology (Fig. 1A). W h e n ES-D3 cells were cul tured for 6 d on

gelat in-coated flasks in the absence of a feeder layer in CEM, the

cul tures contained a mixture of cells with an undifferentiated mor-

phology and cells with the morphologies of fibroblasts, epithelial

cells, and neurons (Fig. 1B).

Expression of pluripotent stem cell markers in ES-D3 cells. After

being serially cult ivated in ESF7 med ium for 1 ran, most of the ES-

D3 cells expressed AP activity (Fig. 1C). Immunohis tochemica l ex-

aminat ion showed that most of the ES-D3 cells cul tured in ESF7

med ium expressed Oct-3/4 protein (Fig. 1E) and the s tem cell an-

t igen SSEA-1 (Fig. 1G). W h e n the cells were cul tured in CEM for

6 d, fewer cells expressed AP activity (Fig. 1D) or immunoreact ivi ty

for Oct-3/4 (Fig. 1F) and SSEA-1 (Fig. 1H). After serial cultivation

for more than 2 yr in ESF7 medium, the morphology of these ES

cells did not change dur ing subsequen t serial cultivation in ESF7

med ium (Fig. 1/) and the cells cont inued to express Oct-3/4 ant igen

(see below) and to exhibit AP activity and SSEA-1 ant igen (data not

shown).

Effects o f LIF and serum on ES-D3 cell proliferation. W h e n seed-

ed in ESF7 minus LIK ES-D3 cells did not proliferate over a 5-d

22 FURUE ET AL.

FIG. 1. The ES-D3 cells in culture. Phase-contrast photomicrograph of ES-D3 cells cultured in ESF7 medium on a collagen-coated plate after three passages in ESF7 after culture in complete embryonic stem medium (CEM) (A) and ES-D3 cells cultured in CEM on a gelatin-coated plate (B). Stem cell marker expression: alkaline phosphatase activity in ES-D3 cells cultured in ESF7 medium (C) and in CEM (D), Oct-3/4 protein expression in ES-D3 cells cultured in ESF7 medium (E) and in CEM (F), and SSEA-1 protein expression in ES-D3 cells cultured in ESF7 medium (G) and in CEM (H). Nuclei were stained with hematoxylin (C-H). Phase-contrast photomicrograph of ES-D3 cells after 2 yr serial cultivation in ESF7 medium (/). Bar, 100 Dxm.

period, whereas the addition of LIF to the medium clearly stimulated ES-D3 proliferation (Fig. 2A). Maximal growth stimulation oc- curred at 1 ng/ml LIE In addition, the number of morphologically undifferentiated cell colonies increased in a LIF concentration-dependent manner (Fig. 2B). In CEM(-), which contained 15% FBS, the cells proliferated maximally in the absence of LIF and LIF concentrations up to 10 ng/ml had no influence on ES-D3 proliferation (Fig. 2A). As in ESF7 medium, LIF added to CEM(-) increased the number of undifferentiated ES-D3 colonies in a concentration-dependent manner (Fig. 2C). These results indicated that ES-D3 ceils absolutely requited LIF for proliferation in the absence of serum, and they suggested that serum components other than LIK insulin, transferrin, and oleie acid em'npensated for the mito- genie activity of LIK However, in both culture media, LIF promoted the formation of undifferentiated colonies. The ES-D3 cells exhib- ited comparable growth properties in ESF7 medium and in CEM (Fig. 2D). In CEM, ES-D3 cells proliferated with a population doubling time of 9 h, whereas in ESF7 medium, they proliferated with a doubling time of 11.8 h. The maximmn cell densities achieved

were the same in both media. Flow cytometric analysis of the DNA content of ES-D3 cells before and after LIF withdrawal from ESF7 medium suggested that the cells died by apoptosis (Fig. 3).

Analysis by flow cytometry of 0ct-3/4 expression in ES-D3 cells. After being serially passaged for 6 mo in ESF7 medium, more than 95% of ES-D3 cells expressed Oct-3/4 protein as measured by flow cytometry (Fig. 4, ESF7-A). After serial cultivation in ESF7 medimn for more than 2 yr, more than 95% of the ceils still expressed Oct- 3/4 protein (Fig. 4, ESF7-B). In contrast, less than 85% of ES-D3 cells expressed Oct-3/4 in CEM after 6 d in culture (Fig. 4, CEM). In ESF medium supplemented with 15% FBS, less than 60% of cells were positive for Oct-3/4 after 6 d in culture (Fig. 4, +FBS). Thus, serum rapidly promoted the differentiation of ES-D3 cells as measured by decreased levels of Oct-3/4 protein.

Marker gene expression in ES-D3 cells. We determined the relative expression of the 0ct-3/4, Rex-l, nanog, and Fgf-5 genes in ES-D3 ceils using real-time RT-PCR (Fig. 5). The 0ct-3/4 gene expression fluctuated only slightly in ES-D3 cells cultured in ESF7 medium. However, LIF increased the expression of both the Rex-1 and nanog genes in a concentration-dependent manner. In contrast, Fgf-5 gene expression decreased rapidly in response to increasing concentrations of LIF such that Fgf-5 transcripts were not detected in ES-D3 cells treated with LIF at more than 1 ng/ml. These results indicated that LIF inhibited cell differentiation and maintained self- renewal of ES cells in this serum-free medium.

Culture of 129/SV ES cells in ESF7 medium. The ES cells (passage 10) derived from 129/SV mouse, which were cuhured on feeder cells in CEM, were inoculated on type I collagen-coated flasks in ESF7 medium containing 20 ng/ml LIF without feeder cells (Fig. 6A). These ceils expressed AP activity (Fig. 6B) and reacted with antibodies to Oct-3/4 (Fig. 6C), and they proliferated with a population doubling time of approximately 14 h (Fig. 6D). Thus, this independently derived mouse ES cell line proliferated in ESF7 medium and expressed stem cell markers. These results suggest that ESF7 medium can be used generally to propagate undifferentiated mouse ES cells.

Induced differentiation of ES-D3 cells. When ES-D3 cells were inoculated on gelatin-coated dishes in FBS-supplemented ESF basal medium or in FBS-supplemented ESF7 medium minus LIK the cells differentiated into fibroblast-like cells, epithelial-like cells, and neuron-like cells (Fig. 7). Thus, ES-D3 cells propagated in serum-free ESF7 medium remained pluripotent. When we washed away LIF from ES-D3 cells precultured in serum-containing CEM, the cells also differentiated. When we washed away LIF from ES- D3 cells precultured in ESF with N2 and B27 supplements, the ceils differentiated into nemonal and glial cells (Fig. 8) after 6 d in culture, as previously described (Ying et al., 2003b). In contrast, when ES-D3 cells were inoculated on collagen-coated dishes in ESF7 medium in the absence of LIE the cells did not proliferate, differentiate, or survive. As estimated by trypan blue exclusion, LIF withdrawal resulted in the death of 78% of the ceils 'after 4 d (data not shown), and no spontaneous differentiation was observed in these cultures (Fig. 9A).

The ES-D3 cells were stimulated to differentiate when inoculated on laminin-eoated dishes in ESF7 medium minus LIF and oleie acid that was supplemented with growth factors or cytokines. The FGF-2 withdrawal stinmlated cells grown in the presence of FGF- 2 and heparin to differentiate into nem'onal and glial-like cells (Fig. 9B). These induced populations of cells were immunoreaetive with

SERUM-FREE CULTURE FOR ES CELLS 2 3

Fro. 2. Phenotypic characterization of ES-D3 cells. (A) The effect of leukemia inhibitory factor (LIF) on ES-D3 cell proliferation. The ES-D3 cells were seeded in a 24-well dish at 5 • 10 z cells/well in ESF7 (open circles) or complete embryonic stem medium (CEM) without LIF (CEM[-]) (closed circles) supplemented with increasing concentrations of LIE Cell numbers were counted after 5 d in culture. Values are the mean -+ SD for three measurements. (B) Phase-contrast photomicrograph of ES-D3 cells cultured in ESF7 containing the indicated concentrations of LIE In the complete absence of LIF, the cells proliferated poorly and underwent apoptotic cell death. In ESF7 medium, undifferentiated cell colonies increased in number in response to increasing LIF concentrations. (C) Phase-contrast photomicrograph of ES-D3 cells cultured in CEM(-) supplemented with the indicated concentration of LIF. The number of undifferentiated cell colonies increased in a concentration-dependent manner in response to LIE Bar, 100 txm. (D) The ES-D3 cell proliferation. The ES-D3 ceils were seeded in a 24-well dish at 5 • 103 cells/well in CEM (closed circles.) and ESF7 medium (open circles). Cells in triplicate wells were counted for 6 d. Values are the mean -+ SD.

antibodies to nestin (Fig. 9D), ~-tubulin IIl (Fig. 9E), A2B5 antigen (Fig. 9F), and GFAP (Fig. 9G), but they did not react with antikeratin antibody (Fig. 9/-/). These results indicated that ES-D3 cells maintained in ESF7 could differentiate into neuronal and glial cells. In contrast, BMP4 supplemented with FAF-BSA induced the cells to differentiate into epithelial-like cells (Fig. 9C). These induced cells did not react with antibody to A2B5 antigen (Fig. 9/), but they were immunoreactive with antibodies to keratin (Fig. 9,/) and cytokeratin 18 (Fig. 9K). The cells did not react with anti-cytokeratin 14 (Fig. 9L), indicating that BMP4 induced ES-D3 cells to differentiate into simple epithelium. These results demonstrated that mouse ES cells maintained in ESF7 medium without feeder cells for more than 2 yr could be induced by the addition or withdrawal of soluble growth factors and cytokines to differentiate along multiple cell lineages.

Quantitation by flow cytometry of A2B5 and keratin expression in differentiated ES-D3 cells. When ES-D3 cells propagated in the presence of FGF-2 and heparin in ESF7 medium in the absence of LIF and oleic acid were cultured for 4 d without FGF-2 and heparin, more than 70% of the cells reacted with anti-A2B5 antibody-, as measured by flow cytometry (Fig. 9M). More than 80% of the ES-D3 cells reacted with an anti-keratin antibody when the cells were cultured in the presence of BMP4 supplemented with FAF- BSA (Fig. 9N). Thus, under defined culture conditions, a large majority of pluripotent mouse ES cells in a population could be di- rected to differentiate along specific pathways.

DISCUSSION

We have described previously a variety of serum-free media that can be used to propagate and analyze differentiated cells (Hayashi

24 FURUE ET AL.

FIG. 3. Flow cytometric analysis of the deoxyribonucleic acid (DNA) content of ES-D3 cells. The ES-D3 cells cultured for 2 d on collagen-coated dishes in ESF7 medimn were washed and given ESF7 medium without leukemia inhibitory factor (LIF) for an additional 4 d. The cells were stained with propidium iodide and analyzed with a flow cytometer. Withdrawal of LIF resulted in nuclear fragmentation consistent with apoptotic cell death. The DNA profiles of cells before (left panel) and after (right panel) LIF withdrawal are shown.

and Sato, 1976; Barnes and Sato, 1980; Furue et al., 1994, 2001; Sato et al., 2002). In this study, we described a serum-free culture medium that supported the serial cultivation of undifferentiated pluripotent mouse ES-D3 cells in the absence of feeder cells for more than 2 yr. The mouse ES-D3 cell line (Doetschman et al., 1985) has been cultured on gelatin-coated flasks in the absence of a feeder

layer in CEM. Under those culture conditions, ES-D3 cells tended to differentiate spontaneously and thus the cells remained self-re-

newing and pluripotent for a limited number of passages. In a direct comparison of the phenotypes of ES-D3 cells in ESF7 medium on

8

~ o i

,?,,, , ESFT-B,

I I

[.v/ \

OCT314 m,

FIG. 4. The 0ct-3/4 expression profile of ES-D3 cells. Flow cytometric profiles of Oct-3/4 expression in ES-D3 cells serially cultured in ESF7 medium for 6 mo (ESF7-A) or for more than 2 yr (ESF7-B); ES-D3 in complete embryonic stem medium (CEM), and ES-D3 cells cultured in ESF + fetal bovine serum (FBS) and 2-mercaptoethanol (+FBS) for 6 d.

i

mmmN II II II II

~o 1o

i = 0 5 . . t . o l l 0.1 0

2.01 narlOg

tO

1.5 tO

0.5

0 0.1 tO 3.0 10

LIF (n~ml)

F|G. 5. Effect of leukemia inhibitory factor (LIF) on gene expression in ES-Dg cells in ESF7 medium. Expression of 0ct-3/4, Rex-I, nanog, and Fgf- 5 in ES-D3 cells cultured in ESF7 medium was analyzed by real-time reverse transcriptase-polymerase chain reaction. The expression of individual genes was calculated relative to that in eells cultured in ESF7 medium supplemented with 3 ng/ml LIE

collagen-coated flasks and in CEM on gelatin-coated flasks, we found that a larger proportion of cells grown in ESF7 medium re-

tained an undifferentiated morphology. The undifferentiated nature of the ES ceils was confirmed by determining the proportion of cells

expressing AP activity and the stem cell marker Oct-3/4 (Niwa et al., 2000). These results suggest that serum components stimulate

ES cell differentiation in vitro rather than simply allowing the man- ifestation of a default differentiation program. We examined whether

ESF7 medium was suitable for other mouse ES cells, and we found that low-passage ES cells from 129/SV and C57/BL6J (data not shown) embryos proliferated in the absence of feeder cells without differentiating in ESF7 medium. To determine whether ESF7 medium is generally appropriate for the propagation of mouse ES cells

will require further testing. In addition, we do not know whether ESF7 medium will support the derivation of ES cell lines.

Doetschman et al. (1985) reported that the majority of the original clonal ES-D3 cells were karyotypically abnornlal with a minority (45%) of cells having a diploid complement. We found that more than 87% of ES-D3 cells were hyperdiploid before serial cultivation in ESF7 medium, with more than 20% of the cells being tetraploid (data not shown). These findings are consistent with reports that

mouse embi"yo ceils acquired karyotypic abnormalities when serially passaged in serum-supplemented medium (Todaro and Green, 1963; Loo et al., 1987). After propagation in ESF7 medium for more than 2 yr, no tetraploid cells were observed.

SERUM-FREE CULTURE FOR ES CELLS 25

FIG. 6. Phenotypic characterization of 129/SV embryonic stem (ES) cells in ESF7 medium. (A) 129/SV ES cells (passage 12) cultured in ESF7 medium for 3 d. Stem cell marker expression: alkaline phosphatase activity (/3) and Oct-3/4 protein (C) expressed in 129/SV ES cells (passage 14) cultured in ESF7 medium for 3 d. (D) The 129/SV ES cell proliferation. The ES cells were seeded in a 24-well dish at 4 X 104 cells/well in ESF7 medium. Cells in triplicate wells were counted for 5 d. Values are the mean -+ SD. Bar, 100 Ixm.

Serum components can influence the effects of exogenously added factors on cell growth and differentiation. Serum contains variable amounts of soluble growth and differentiation factors including activins and fibroblast growth factors, and we found that in the absence of serum, aetivin A promoted ES-D3 cell proliferation and the expression of the genes Lefiy-1, brachury, and goosecoid (data not shown). A previous study by Niwa (2001) found that LIF maintained ES cell self-renewal and pluripotency but did not affect ES cell proliferation in serum-containing medimn. We confirmed that LIF had little influence on cell proliferation in FBS-supplemented culture medium. However, in ESF7 medium, LIF was essential for ES-D3 survival and it clearly stimulated cell proliferation in a concentration-dependent manner. The expression of Oct-3/4, a member of the POU transcription factor family, has been reported to be tightly restricted to pluripotent cells in the mouse life cycle (Niwa et al., 2000). The Oct-3/4 is believed to cooperate with LIF in maintaining ES cell pluripotency, but it is still not clear how these

FIG. 7. Differentiation of ES-D3 cells cultured in ESF nutrient medium supplemented with fetal bovines serum and 2-mercaptoethanol. (A) Fibro- blast-like cells. (B) Epithelial-like cells. (C) Neuron-like cells. Bar, 100 Ixm.

FIG. 8. The ES-D3 cell differentiation in ESF nutrient medium with N2 and B27 supplements. The ES-D3 cells were seeded in 0.1% gelatin---coated dishes in ESF medium supplemented with N2 and B27 supplements without LIE (A) Phase-contrast photograph of the cells in ESF medium + N2 B27 in the absence of LIE (B) Nestin-positive cells. (C) 13-Tubulin III-positive cells. (D) A2B5-positive cells. (E) Glial fibrillary acidic protein-positive cells. (F) Keratin-positive cells. Bar, 100 txm.

components work together (Niwa, 2001). Niwa (2001) suggested that LIF does not support ES cell renewal by maintaining Oct-3/4 expression. Our demonstration that 0ct-3/4 gene expression was not affected by LIF over a 100-fold range in concentration also suggests that 0ct-3/4 expression is not directly regulated by a LIF signaling pathway.

Recently, Ying et al. (2003a) reported that LIF was insufficient to block differentiation and maintain pluripotency in serum-free medium. The majority of ES cells cultured in a conventional serum- fi'ee nutrient medium with N2 (Bottenstein and Sato, 1979) and B27 (Brewer et al., 1993) supplements converted to Soxl-positive neural precursor cells in 4 d (Ying et al., 2003b). The addition of LIF reduced but did not prevent neural differentiation. However, the addition of LIF and BMP4 to N2B27 medium suppressed the neural determination of ES ceils and promoted self-renewal for multiple passages (Yiug et al., 2003a). The B27 supplements enhanced ES cell viability but were not required for self-renewah The N2 supplements consist of insulin, transferrin, progesterone, selenium, and putrescine. The B27 supplements contain corticosterone, putrescine, trido-l-thyronine, DL-a-tocopherol (vitamin E), catalase, and other components in addition to insulin, transferrin, progesterone, selenium, putrescine, ethanolamine, and BSA. Given that N2 and

26 FURUE ET AL.

FIG. 9. Differentiation of ES-D3 cells induced by FGF-2 and BMP4. (A) Phase-contrast photomicrograph of ES-D3 cells in ESF7 medium 2 d after leukemia inhibitov factor (LIF) withdrawal. (B) Phase-contrast photomicrograph of ES-D3 cells differentiating into neuron-like cells after the withdrawal of FGF-2 from ESF7 lnedium (minus LIF and bovine serum albumin [BSA]- oleic acid) supplemented with FGF-2 and heparin. (C) Phase-contrast photomicrograph of ES-D3 cells differentiating into epithelial-like cells after 6 d in ESF7 medimn (minus LIF and BSA--oleic acid) supplemented with BMP4 and fatty acid-free BSA. (/9) Nestin-positive cells cultured as in panel B. (E) 13-Tubulin Ill-positive cells cultured as in panel B. (F) The A2B5-positive cells cultured as in panel B. (6) Glial fibrillary acidic protein-positive cells cultured as in panel B. (H) Anti-keratin antibody staining of cells cultured as in panel B. (1) Anti-A2B5 antibody staining of cells cultured as in palvel C. (J) Keratin-positive cells cultured as in panel C. (K) Cytokeratin 18- positive cells cultured as in paTwl C. (L) Anti-cytokeratin 14 antibody staining of cells cultured as in panel C. (M) Flow cytometric analysis of A2B5 expression by ES-D3 cells cultured as in panel B (blue) or panel C (red). (N) Flow cytometric analysis of keratin expression by ES-D3 cells cultured as in panel B (blue) or panel C (red). Bar, 100 Ixm.

B27 supplements were developed originally to support the maintenance and growth of neuronal ceils (Bottenstein and Sato, 1979; Brewer et al., 1993), it is plausible that some of the supplements in N2B27 medium direct ES cells toward a neural fate, which is

suppressed by BMPs. We found that LIF withdrawal promoted the differentiation of ES-D3 ceils cultured in ESF nutrient medium with N2 and B27 supplements into neuronal and glial cells alter 6 d in culture, as described previously (Ying et al., 2003b).

In ESF7 medium, LIF was sufficient to inhibit cell differentiation and to maintain self-renewal of murine ES-D3 cells. The inhibitory effect of LIF on cell differentiation was manifested in its effects on gene expression. Real-time RT-PCR analysis demonstrated that LIF increased the expression of Rex-1 in a concentration-dependent manner. Rex-l, a zinc finger transcription factor, can be detected in the inner cell mass but not in primitive ectoderm (Pehon et al., 1998), and it is believed to be a marker of undifferentiated ES cells. The Fgf-5 gene expression, which occurs in the primitive ectoderm (Pelton et al., 1998), rapidly decreased in ES-D3 cells in response to LIF in a concentration-dependent manner. Furthermore, by real- time RT-PCR analysis, the expression of the homeobox transcription factor nanog, which was identified recently as a pluripotency sustaining factor (Chambers et al., 2003; Mitsui et al., 2003), increased in response to LIF in a concentration-dependent manner. This latter result suggests that nanog is located downstream of LIF in a signaling pathway in murine ES cells. Cavaleri and Scholer (2003) have hypothesized that nanog, Oct-3/4, and LIF define three dif- ferent transcriptional pathways operating in ES cells. Our culture system has allowed the discovery of a novel mitogenic activity of LIF for mouse ES cells, and it has provided evidence linking LIF to the regulation of nanog in the maintenance of pluripotency. This system should permit further elucidation of signaling pathways in- volved in maintaining ES cell pluripotency.

A number of previous studies demonstrated that LIF withdrawal immediately promoted ES cell differentiation (Doetschman et al., 1985; Wiles and Kelle~, 199i; Keller et al., 1993; Wiles, 1993; Smith, 2001). When we added FBS to the culture medium, withdrawal of LIF induced the ES-D3 cell differentiation into fibroblast- like cells, epithelial-like cells, and neuron-like cells. These results indicated the pluripotent nature of ES-D3 cells maintained in serum-free ESF7 medium. The ES-D3 ceils exposed to serum expressed molecular markers characteristic of cells derived from all three embryonic germ layers: the primitive ectodermal marker FGF- 5; the mesodermal marker brachury; and the endodermal marker GATA4 and GATA6 (data not shown). Ying et al. (2003b) reported that in the absence of serum, LIF withdrawal stimulated ES cell differentiation into neural precursors. However, withdrawal of LIF from ESF7 medium resulted in cell death, and no spontaneous differentiation of ES-D3 cells was observed. A majority of the cells died within 4 d of LIF withdrawal. Flow cytometric analysis of the DNA content of ES-D3 cells before and after LIF withdrawal from ESF7 medium suggested that the cells died by apoptosis. In contrast, the addition of growth factors and cytokines to ESF7 medium in the absence of LIF induced ES-D3 cells to differentiate. In response to BMP4, ES-D3 cells differentiated into epithelial-like cells, which were cytokeratin 18-positive and cytokeratin 14-negative. Because eytokeratin 18 is expressed in simple epithelium, and cytokeratin 14 is expressed in the basal layer of nmhilayered epithelia (Carette et al., 1991; Bagutti et al., 1996; Owens and Lane, 2003), these findings indicate that BMP4 induced ES cells to differentiate into simple epithelium. The FGF-2 induced ES cells to differentiate into the cells that expressed nestin, 13-tubulin III, A2B5, and GFAP. Nestin-positive cells have been reported to differentiate into neurons, astrocytes, and oligodendroeytes when mi-

SERUM-FREE CULTURE FOR ES CELLS 27

togens are withdrawn or when they are exposed to other factors (MeKay, 1997; Tropepe et al., 2001; Zhang et al., 2004). Progenitor cells restricted to a neuronal lineage have been reported to express [~-tubulin III (Zhang et al., 2004). Our findings confirm that FGF- 2 promotes ES cell differentiation into neuronal and glial precursor cells. These results demonstrate that mouse ES-D3 cells maintained

in ESF7 medium can be induced by soluble factors to differentiate along specific cell lineage pathways.

Taken together, our results show that by using ESF7 serum-free

medium, mouse ES-D3 cells can be propagated in vitro in an undifferentiated state in the absence of feeder cells for more than 2 yr and the cells can be induced to differentiate along predefiued paths by specific differentiation-inducing factors. We anticipate that this

culture system will be suitable for other ES cell lines. This study is a step forward and should stimulate further applications and im-

provements in the stem cell culture field. This culture system

should help advance our understanding of stem cell biology, facil- itate the development of reproducible differentiation protocols for ES ceils in culture, and promote practical applications in regener- ative medicine.

ACKNOWLEDGMENTS

We thank Prof. E. B. Lane (Cancer Resem'ch U.K. Cell Structure Research Group, School of Life Sciences, University of Dundee) for providing antibodies to cytokeratin 14. This study was supported by grants-in-aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Tech- nology of Japan to M. E, T. 0., and M. A., by a grant from the Ministry of Health, Labor and Welfare of Japan to H. O., M. E, and M. E, an Interna- tional Cooperative Research Project grant of the Japan Science and Tech- nology Agency to M. A., and by a grant from the Smoking Research Foun- dation to T. O. J. D. S. was supported by grants P20-RR16463, P20- RR016463, and P30-ES03828 from the National Institutes of Health and by a short-term fellowship from the Japan Society for the Promotion of Science.

REFERENCES

Asashima, M.; Ariizumi, T.; Matacinski, G. M. In vitro control of organogenesis and body patterning by aetivin during early amphibian development. Comp. Biochem. Physiol. B Biochem. Mol. Biol. 126:169-178; 2000.

Bagutti, C.; Wobus, A. M.; Fassler, R.; Watt, E M. Differentiation of embry- onal stem cells into keratinocytes: comparison of wild-type and beta 1 integrin-deficient cells. Dev. Biol. 179:184-196; 1996.

Barnes, D.; Sato, G. Serum-free cell culture: a unifying approach. Cell 22:649-655; 1980.

Bonenstein, J. E.; Sato, G. H. Growth of a rat neuroblastoma cell line in serum-free supplemented medium. Proc. Natl. Acad. Sci. USA 76:514-517; 1979.

Brewer, G. J.; Torricelli, J. R.; Evege, E. K.; Price, P. J. Optimized survival of hippocampal neurons in B27-supplemented Neurobasal, a new serum-free medium combination. J. Neurosci. Res. 35:567-576; 1993.

Carette, M. J.; Lane, E. B.; Ferguson, M. W. Differentiation of mouse embryonic palatal epithelium in culture: selective cytokeratin expression distinguishes between oral, medial edge and nasal epithelial cells. Differentiation 47:149-161; 1991.

Cavaleri, E; Scholer, H. R. Nanog: a new recruit to the embryonic stem cell orchestra. Cell 113:551-552; 2003.

Chambers, I.; Colby, D.; Robertson, M.; Nichols, J.; Lee, S.; Tweedie, S.; Smith, A. Functional expression cloning of Nanog, a pluripotency sustaining factor in embryonic stem cells. Cell 113:643-655; 2003.

Doetschman, T. C.; Eistetter, H.; Katz, M.; Schmidt, W.; Kemler, R. The in vitro development of blastocyst-derived embryonic stem cell lines: fornmtion of visceral yolk sac, blood islands and myocardium. J. Embryol. Exp. Morphol. 87:27-45; 1985.

Evans, M. J.; Kaufinan, M. H. Establishment in culture of pluripotential cells from mouse embtTcos. Nature 292:154-156; 1981.

Furue, M.; Asashima, M. Isolation of pluripotential stem cells from Xenopus embryos. In: Lanza, R., ed. Handbook of stem cells, vol. 1. San Diego, CA: Academic Press; 2004:483-492.

Fume, M.; Okamoto, T.; Ikeda, M.; Tanaka, Y.; Sasaki, Y.; Nishimura, K.; Sato, J. D. Primitive neuroectodemlal tumor cell lines derived from a metastatic pediatric tumor. In Vitro Cell. Dev. Biol. 30A:813~16; 1994.

Fume, M.; Zhang, 5(.; Okamoto, T.; Hata, R. I.; Asashima, M. Activin A induces expression of rat Sel-ll mRNA, a negative regulator of notch signaling, in rat salivary gland-derived epithelial cells. Biochem. Bio- phys. Res. Commun. 282:745-749; 2001.

Gardner, R. L.; Brook, E A. Reflections on the biology of embryonic stem (ES) cells. Int. J. Dev. Biol. 41:235-243; 1997.

Hayashi, I.; Sato, G. H. Replacement of serum by hormones permits growth of ceils in a defined medium. Nature 259:132-134; 1976.

Hots, M.; Gong, J.; Traganos, E; Drazynkiewics, Z. Flow cytometric detection of apoptosis: comparison of the assays of in situ DNA degradation and chromatin changes. Cytometry 15:237-244; 1994.

Kawasaki, H.; Mizuseki, K.; Nishikawa, S.; Kaneko, S.; Kuwana, Y.; Nak- anishi, S.; Nishikawa, S. I.; Sasai, Y. Induction of midbrain dopami- nergic neurons from ES cells by stromal cell-derived inducing activity. Neuron 28:31-40; 2000.

Keller, G.; Kennedy, M.; Papayannopoulou, T.; Wiles, M. V~ Hematopoietic commitment (luring embryonic stem cell differentiation in culture. Mol. Cell Biol. 13:473-486; 1993.

Lake, J.; Rathjen, J.; Remiszewski, J.; Rathjen, P. D. Reversible programming of pluripotent cell differentiation. J. Cell Sci. 113 (Pt. 3):555-566; 2000.

Loo, D. T.; Fnquay, J. 1.; Rawson, C. L.; Barnes, D. W. Extended cuhure of mouse embryo cells without senescence: inhibition by serum. Science 236:200-202; 1987.

Martin, G. R. Isolation of a pluripotent cell line from early mouse embryos cultured in medium conditioned by teratocarcinoma stem ceils. Proc. Natl. Acad. Sci. USA 78:7634-7638; 1981.

McKay, R. Stem ceils in the central nervous system. Science 276:66-71; 1997.

Mitsui, K.; Tokuzawa, Y.; Itoh, H., et al. The homeopmtein Nanog is required for maintenance of pluripotency in mouse epiblast and ES cells. Cell 113:631-642; 2003.

Niwa, H. Molecular mechanism to maintain stem cell renewal of ES cells. Cell Struct. Funct. 26:137-148; 2001.

Niwa, H.; Miyazaki, J.; Smith, A. G. Quantitative expression of 0ct-3/4 de- fines differentiation, dedifferentiation or self-renewal of ES cells. Nat. Genet. 24:372-376; 2000.

Owens, D. W.; Lane, E. B. The quest for the function of simple epithelial keratins. Bioessays 25:748-758; 2003.

Pehon, T. A.; Bettess, M. D.; Lake, J.; Rathjen, J.; Rathjen, P. D. Develop- mental complexity of early mammMian pluripotent cell populations in vivo and in vitro. Reprod. Fertil. Dev. 10:53,5-549; 1998.

Purkis, P. E.; Steel, J. B.; Mackenzie, I. C.; Nathrath, W. B.; Leigh, I. M.; Lane, E. B. Antibody markers of basal cells in complex epithelia. J. Cell Sci. 97 (Pt. 1):39-50; 1990.

Sato, J. D.; Barnes, D.; Hayashi, I., et at. Specific cells and their require- ments. In: Davis, J. M., ed. Basic cell culture: a practical approach. 2rid ed. Oxford, England: Oxford University Press; 2002:227-274.

Sato, J. D.; Kawamoto, T.; Okamoto, T. Cholesterol requirement of P3-X63- Ag8 and X63-Ag8.653 mouse myeloma cells for growth in vitro. J. Exp. Med. 165:1761-1766; 1987.

Smith, A. Embryonic stem cells. In: Marshak D. R.; Gardner, R. L.; Gottlieb, D., ed. Stem ceil biology. New York: Cold Spring Harbor Laboratory Press; 2001:205-230.

Smith, A. G.; Heath, J. K.; Donaldson, D. D.; Wong, G. G.; Moreau, J.; Stahl, M.; Rogers, D. Inhibition of pluripotential embryonic stem cell differentiation by purified potypeptides. Nature 336:688-690; 1988.

Tanaka, T. S.; Kunath, T.; Kimber, W. L., et al. Gene expression profiling of emblTco-derived stem cells reveals candidate genes associated with pluripotency and lineage specificity. Genome Res. 12:1921-t928; 2002.

Tiedenmnn, H.; Asashima, M.; Gmnz, H.; Knochel, W. Pluripotent cells (stem cells) and their determination and differentiation in early vertebrate embryogenesis. Dev. Growth Differ. 43:469-502; 2001.

2 8 FURUE ET AL.

Todaro, G. J.; Green, H. Quantitative studies of the growth of mouse embryo cells in culture and their development into established lines. J. Cell Biol. 17:299-313; 1963.

Toumadje, A.; Kusumoto, K. I.; Parton, A., et al. Pluripotent differentiation in vitro of murine ES-D3 embryonic stem cells, in Vitro Cell. Dev. Biol. 39A:449M~53; 2003.

Tremblay, K. D.; Hoodless, E A.; Bikoff, E. K.; Robertson, E. J. Formation of the definitive endoderm in mouse is a Smad2-dependent process. Development 127:3079-3090; 2000.

Tropepe, V.; Hitoshi, S.; Sirard, C.; Mak, T. W.; Rossant, J.; van der Kooy, D. Direct neural fate specification from embryonic stern cells: a primitive mammalian neural stem cell stage acquired through a default mechanism. Nem'on 30:65-78; 2001.

Wiles, M. V. Embryonic stem cell differentiation in vitro. Methods Enzymol. 225:900-918; 1993.

Wiles, M. V.; Johansson, B. M. Embryonic stem cell development in a chem- ically defined medium. Exp. Cell Res. 247:241-248; 1999.

Wiles, M. V.; Keller, G. Multiple henmtopoietic lineages develop from embryonic stem (ES) cells in culture. Development 111:259-267; 1991.

Williams, R. L.; Hilton, D. J.; Pease, S., et al. Myeloid leukaemia inhibitory factor maintains the developmental potential of embryonic stem cells. Nature 336:684--687; 1988.

Ying, Q. L.; Nichols, J.; Chambers, I.; Smith, A. BMP induction of Id proteins suppresses differentiation and sustains embryonic stem cell self-renewal in collaboration with STAT3. Cell 115:281-292; 2003a.

Ying, Q. L.; Sravridis, M.; Griffiths, D.; Li, M.; Smith, A. Conversion of embryonic stem cells into neuroeetodennal precursol~ in adherent monocuhure. Nat. Biotechnol. 21:183-186; 2003b.

Zhang, X.; Klueber, K. M.; Guo, Z.; Lu, C.; Roisen, E J. Adult human ol- factory neural progenitors cultured in defined medium. Exp. Neurol. 186:112-123; 2004.

Documents

Leukemia inhibitory factor as an anti-apoptotic mitogen for pluripotent mouse embryonic stem cells in a serum-free medium without feeder cells