A Novel, Immortal, and Multipotent Human ... - Wiley Online Library

TISSUE-SPECIFIC STEM CELLS A Novel, Immortal, and Multipotent Human Neural Stem Cell Line Generating Functional Neurons and Oligodendrocytes LIDIA DE FILIPPIS,a GIUSEPPE LAMORTE,a EVAN Y. SNYDER,b ANTONIO MALGAROLI,a ANGELO L. VESCOVIa,c a Department of Biotechnologies, Fondazione Centro San Raffaele del Monte Tabor, Milan, Italy; bStem Cell and Regeneration Program, Burnham Institute for Medical Research, La Jolla, California, USA; cDepartment of Biotechnologies and Biosciences, Universita` degli Studi di Milano, Bicocca, Italy Key Words. Neural differentiation • Neural stem cell • Proliferation • Stem cell culture

ABSTRACT The discovery and study of neural stem cells have revolutionized our understanding of the neurogenetic process, and their inherent ability to adopt expansive growth behavior in vitro is of paramount importance for the development of novel therapeutics based on neural cell replacement. Recent advances in high-throughput assays for drug development and gene discovery dictate the need for rapid, reproducible, long-term expansion of human neural stem cells (hNSCs). In this view, the complement of wild-type cell lines currently available is insufficient. Here we report the establishment of a stable human neural stem cell line (immortalized human NSCs [IhNSCs]) by v-myc-mediated immortalization of previously derived wild-type hNSCs. These cells demonstrate three- to fourfold faster proliferation than wild-type cells in response to growth factors but retain rather similar properties, including multipotentiality. By molecular biology, biochemistry, immunocytochemistry, fluorescence micros-

copy, and electrophysiology, we show that upon growth factor removal, IhNSCs completely downregulate v-myc expression, cease proliferation, and differentiate terminally into three major neural lineages: astrocytes, oligodendrocytes, and neurons. The latter are functional, mature cells displaying clear-cut morphological and physiological features of terminally differentiated neurons, encompassing mostly the GABAergic, glutamatergic, and cholinergic phenotypes. Finally, IhNSCs produce bona fide oligodendrocytes in fractions up to 20% of total cell number. This is in contrast to the negligible propensity of hNSCs to generate oligodendroglia reported so far. Thus, we describe an immortalized hNSC line endowed with the properties of normal hNSCs and suitable for developing the novel, reliable assays and reproducible high-throughput gene and drug screening that are essential in both diagnostics and cell therapy studies. STEM CELLS 2007;25:2312–2321

Disclosure of potential conflicts of interest is found at the end of this article.

INTRODUCTION NSCs are highly undifferentiated, mitotically competent, multipotent brain precursors that play a key role both in neurogenesis and tissue cellular homeostasis throughout life [1–7]. Their existence has been described in both developing and adult nervous system of all mammals, including humans. One principal characteristic of NSCs is their inherent ability for extensive self-renewal that, under appropriate conditions, permits their long-term proliferation and expansive growth in culture. By means of chemically defined medium and mitogenic stimulation with epidermal growth factor (EGF) [8], either alone or combined with fibroblast growth factor 2 (FGF2) [9 –14], NSCs have been isolated and propagated extensively from the embryonic, neonatal, and adult rodent brain [15–17] and from the fetal and adult human central nervous system (CNS) [12, 17–22]. These human NSCs (hNSCs) have been proposed as an abundant and renewable source of brain cells for cell-based therapeutic approaches. However, their relatively slow growth kinetics somewhat limits their use in biotechnology-oriented applications, which require both rapid and reproducible production of large amounts of human brain cells. Nontransforming, oncogene-mediated immortalization of human neural precursors has been proposed to overcome these problems, allowing for the

establishment of fast-growing hNSC lines [23]. Among the genes bearing potential ability for immortalization, v-myc has emerged as a most suitable candidate [23], whereas the Tag, bcl-XL, and c-myc genes have proven ineffective. Thus, the immortalized hNSC lines currently available have mostly been obtained through v-myc-mediated immortalization [24 –26]. Despite these encouraging results, the possibility of using these cells for reliable high-throughput screening in drug discovery and CNS therapeutics hinges on the expected ability to generate terminally differentiated, functional, mature neuronal cells. Such an ability has not been described to date, and the type of neuronal phenotype generated by these cells remains undefined. In addition, currently available cell lines rarely if ever give rise to oligodendrocytes. To provide an answer to these issues, we resorted to immortalizing wild-type hNSCs using a v-myc vector [23, 25] and compared their properties with those of their normal counterparts. The immortalized human NSCs (IhNSCs) that we have established are multipotent, remain growth factordependent, expand three- to fourfold faster than their parental cells, and differentiate terminally upon growth factor removal. The IhNSCs generate neurons that belong to the GABAergic, glutamatergic, cholinergic, and, to a much lesser extent, catecholaminergic phenotypes. It is noteworthy that these neurons are mature, are endowed with full synaptic machinery, and elicit both evoked and spontaneous action potentials. In addition,

Correspondence: Angelo L. Vescovi, Ph.D., Department of Biotechnology and Biosciences, University of Milano-Bicocca, Piazza della Scienza 2, I-20126 Milano, Italy. Telephone: ⫹39-02-6448-3351; Fax: ⫹39-02-7004-31033; e-mail: [email protected] Received January 17, 2007; accepted for publication May 30, 2007; first published online in STEM CELLS EXPRESS June 7, 2007. ©AlphaMed Press 1066-5099/2007/$30.00/0 doi: 10.1634/stemcells.2007-0040

STEM CELLS 2007;25:2312–2321 www.StemCells.com

De Filippis, Lamorte, Snyder et al. IhNSCs also demonstrate an inherent propensity to generate large numbers (up to 20% of total cell counts) of genuine oligodendrocytes upon differentiation. These findings identify a novel, invaluable tool for studies in the human neuropharmacology and neurogenesis areas and reduce the need to resort to the systematic use of primary human fetal tissue.

MATERIALS

AND

METHODS

The hNSC cultures (parental cells) used in this study were isolated and propagated from the diencephalic and telencephalic brain regions of a white human fetus at 10.5 weeks gestational age and were previously described by Vescovi et al. [19]. To induce hNSC differentiation, individual spheres were mechanically dissociated and transferred at a density of 2.5 ⫻ 104 cells per cm2 onto Matrigel (BD Biosciences, San Diego, http:// www.bdbiosciences.com)-coated chamber-slides in the presence of 20 ng/ml FGF2. After 72 hours, cultures were shifted to NS-A control medium (CM) containing 2% fetal calf serum and grown for 2 weeks.

Propagating Gene v-myc and Infection PK-VM-2 is a replication-defective, infective retroviral vector described previously [23, 25], coding for both avian myc (p110 gag-myc or v-myc) driven by the MoLV-long terminal repeat (LTR) promoter and aminoglycoside transferase (conferring resistance to neomycin; neor) driven by the SV40 promoter. The amphitropic retroviral particles were packaged in GP⫹envAM cells cultured in hNSC medium [23]. Retroviral transduction with v-myc was carried as described in Villa et al. [26] on parental hNSCs that had undergone 22 passages in vitro. Following G418 selection, aliquots of these bulk cell lines were cryopreserved in complete medium containing 10% dimethyl sulfoxide. To induce IhNSC differentiation, individual spheres were mechanically dissociated and transferred onto laminin (Roche, Basel, Switzerland, http://www.roche-applied-science.com)-coated glass coverslips at a density of 1 ⫻ 104 cells per cm2 in the presence of FGF2 (20 ng/ml). Cultures were shifted after 72 hours to NS-A CM and further cultured up to 5 weeks to obtain a mixture of neural cells containing astrocytes, neurons, and oligodendrocytes [12]. To induce the dopaminergic phenotype in the neuronal progeny, cultures were shifted after 72 hours from plating to NS-A control medium containing 75% of B49 cell-conditioned medium (B49C.M.) ⫹ leukemia inhibitory factor (20 ng/ml) and cultured for 2 weeks.

Clonal Analysis IhNSC neurospheres were mechanically dissociated (⬎200 extrusion cycles through a p200, Gilson yellow-tip pipette) to yield a single-cell suspension. Cells were plated by serial, limiting dilution steps [27] to obtain a statistical distribution of ⬍1 cell per well in 96-well plates (BD Biosciences) in complete growth medium containing EGF and FGF-2. Single wells were scored by direct visualization under a phase-contrast microscope, and only those wells containing a single cell were taken into account. Slightly less than 40% of the wells containing a single cell gave rise to a clonal cell line. Three of these clones were chosen as representative and characterized by growth curve analysis and differentiation assay, yielding comparable results.

Immunocytochemistry Cultures were fixed in freshly buffered 4% paraformaldehyde. After blocking with 10% normal goat serum, cultures were incubated overnight at 4°C in the following antibodies: ␤-tubulin III (monoclonal antibody [mAb] MMS-435P; 1:400; Babco), glial fibrillary acidic protein (GFAP) (polyclonal antibody [pAb]; 1:400; DAKO, Glostrup, Denmark, http://www.dako.com), galactocerebroside C (mAb MAB345; 1:100; Chemicon, Temecula, CA, http://www. chemicon.com), O4 (mAb MAB342; 1:100; Chemicon), tyrosine hydroxylase (TH) (mAb; 1:100; NCL-TH; Novocastra Ltd.,

www.StemCells.com

2313

Newcastle upon Tyne, U.K., http://www.novocastra.co.uk), GABA (pAb; 1:1,000; Sigma-Aldrich), glutamatergic acid (glutamate; pAb; 1:3,000; Sigma-Aldrich), choline acetyltransferase (ChAT) (pAb; 1:100; Chemicon), microtubule-associated protein 2 (MAP2) (mAb; 1:400; Chemicon), and KI67 nuclear antigen (pAb; NCL-KI67p; 1:1,000; Novocastra). After rinsing in phosphate-buffered saline (PBS), cultures were incubated for 45 minutes at room temperature in the following secondary antibodies: Alexa 488 (A11008; anti-rabbit; 1:1,000; Molecular Probes Inc., Eugene, OR, http://probes.invitrogen.com) and/or Alexa 546 (A11030; antimouse; 1:1,000; Molecular Probes). Cultures were then stained with Hoechst (0.03 ␮g/ml; Sigma-Aldrich) for nuclear staining. Microphotographs were taken using a Zeiss Axiovert 200 direct epifluorescence microscope (Carl Zeiss, Jena, Germany, http://www.zeiss. com). Data are reported as percentages of labeled cells over the total number of nuclei ⫾ SE. Each value represents the average of three independent experiments.

Electrophysiological Recordings IhNSCs grown on glass coverslips, were continuously perfused with a Tyrode’s solution (adjusted to 305 mOsm and pH 7.4) containing (in mM): NaCl, 119; KCl, 5; CaCl2, 2; MgCl2, 2; Hepes, 25; glucose, 30. Patch electrodes (2–5 MW) contained (in mM): Kgluconate 110; MgCl2 5; NaCl 10; EGTA or BAPTA 0.6 –10; ATP 2; GTP 0.2; HEPES 49; adjusted to pH 7.2 and 290 mOsm. Cell currents and action potentials were recorded in the cell-attach and whole cell modalities using an Axopatch 1D amplifier (Axon Instruments/Molecular Devices Corp., Foster City, CA, http://www. moleculardevices.com). Tetrodotoxin (0.5 mM; Latoxan, F) was used to suppress action potentials. Reagents were obtained from Sigma-Aldrich, except where noted. Traces were filtered at 2–5 KHz and stored using a digital tape recorder. Data were digitized off-line at 10 –70 KHz after using a low-pass filter at 3–5 KHz.

Antibody Uptake Experiments To evaluate the spontaneous vesicular turnover occurring in putative axons, IhNSCs were differentiated on glass coverslips for 24 days in vitro (div) and incubated overnight in regular culture medium with the addition of antibodies specific for the intravesicular N-terminal portion of synaptotagmin-I (affinity-purified rabbit polyclonal serum; [28]) (1:500 dilution). In parallel, in vivo analysis of active presynaptic varicosities and surface GluR1 glutamate receptors was performed by incubating viable neurons for 2 hours in standard culture medium containing anti-synaptotagmin-1 antibodies (1:500 dilution) and anti-GluR1 antibodies for the extracellular epitope of the GluR1 molecule. Neurons were then fixed and processed for triple immunofluorescence using an anti-MAP2 antibody as counterstain (Fig. 7M) and specific fluorescent secondary antibodies. Cultures were then rinsed in Tyrode’s solution (37°C) followed by PBS at 4°C, fixed in 4% paraformaldehyde, and permeabilized for 1–2 hours at room temperature with 1% bovine serum albumin and 0.4% saponin in PBS. Retrograde staining of synapses and dendrites was done using antibodies specific for synaptotagmin-I (anti-goat), MAP2, and Tau (anti-mouse). Speciesspecific, fluorochrome (fluorescein isothiocyanate, CY5, or Texas Red) conjugated secondary antibodies were used at 1:100 dilutions (Chemicon and Jackson Immunoresearch Laboratories, West Grove, PA, http://www.jacksonimmuno.com). Microscopy and confocal imaging were done using an upright Zeiss Axioskop microscope (⫻60 oil-immersion objective) equipped with a Zeiss confocal system (Carl Zeiss).

RESULTS v-myc-Transduced hNSCs Display Enhanced yet Normal Growth Properties Human NSCs, previously isolated from human fetus forebrain at 10.5 post-conception weeks, heretofore referred to as “parental cells” [19], were transduced with a retroviral construct expressing both v-myc and neomycin-resistant genes; transgene expres-

2314

A Novel Immortalized Human Neural Stem Cell Line

Figure 1. Analysis of long-term proliferation in immortalized human neural stem cells (IhNSCs). (A): Growth curve of IhNSCs and parental human neural stem cell (hNSC) cultures in the presence of EGF and FGF2 in the culture medium. At every passage, 200,000 cells were plated and counted when they originated neurospheres. (B): The exposure of cell line IhNSCs to EGF and FGF2, alone and in combination. (C, D): Phase-bright microphotographs showing IhNSCs (C) and parental hNSC-derived neurospheres (D) 4 days after dissociation. Objective magnification, ⫻10. (E): Karyotype analysis of IhNSCs at 79 passages after infection. Abbreviations: Control T⫹D, hNSC; EGF, epidermal growth factor; FGF2, fibroblast growth factor 2; tot n., total number; v-myc T⫹D, InNSC.

sion was confirmed by reverse transcription-polymerase chain reaction (supplemental online Fig. 1). After selection in neomycin-containing medium, IhNSCs proliferated as neurospheres in the presence of EGF and FGF2, similar to parental cells [19]. However, their growth rate increased considerably, with a doubling time of 2–3 days as compared with the 8 –10 days observed with parental hNSCs (Fig. 1A) [19]. Cytofluorimetric analysis demonstrated an average IhNSC cell cycle time of approximately 40 hours, with well over 25% of the cells actively engaged in the S-phase (supplemental online Fig. 2). It is important to note that following immortalization and extensive subculturing, IhNSCs still retained a strict growth factor dependence. In fact, growth factor removal promptly triggered a progressive, terminal loss of proliferation capacity (Fig. 1B). Furthermore, karyotype analysis confirmed the preservation of a normal karyotype over passaging (Fig. 1E). Intriguingly, analysis by Western blot (Fig. 2I) or immunofluorescence (Fig. 2K, 2L), using two distinct anti-v-Myc antibodies, showed a progressive decline of the v-myc protein levels upon removal of growth factors and the consequent differentiation that this manipulation brought about in IhNSCs (Fig. 2I, 2L), in a fashion very similar to their parental counterpart. We ought to emphasize, however, that our conclusions regarding v-myc downregulation are restricted to a phenomenological observation of this event upon cessation of mitogenic stimulation by EGF and FGF2. In our study, retrovirus-mediated v-myc transfer and expression was driven by the murine leukemia virus (MLV) LTR. Hence, further analysis will be necessary to understand the mechanisms that might determine downregulation of the LTR or other promoters upon differentiation in vitro and in vivo [25, 26]. Taken together, these findings show that v-myc-mediated immortalization of cultured hNSCs leads to the establishment of IhNSCs endowed with an extensive and rapid proliferation capacity, which, in turn, do not display the characteristics of transformed cells, such as growth factor independence or aberrant karyotypic properties.

IhNSCs Give Rise to Mature Neurons and Oligodendrocytes An important property of hNSCs is their multipotentiality, whereby the progeny of a neural stem cell are able to undergo terminal differentiation into neurons and both types of macroglia [19]. This property has yet to be allocated to the immortalized hNSCs currently available. Hence, we investigated the developmental potential and candidate multipotentiality of IhNSCs. The cells analyzed in this work were from a mixed, bulk population. Nonetheless, by means of limiting dilution standard techniques [27], we derived clonal cell lines from the bulk population. Three clones were extensively expanded and were then differentiated for up to 38 days in vitro, displaying proliferation and differentiation properties similar to those of the original bulk population. Downregulation of v-myc upon mitogen removal occurred in a similar fashion in both clonal and mixed cultures. The period of residual proliferation, which normally precedes terminal differentiation upon mitogen starvation in hNSCs, was extended in IhNSCs. Most likely related to the gradual v-myc downregulation observed in these cells upon mitogen removal, this residual proliferation in IhNSCs did cease entirely by day 40. By this time, only a small number of mitotic glial or neuronal precursors (maximum, 1.76% ⫾ 0.44% [SE] of total cell counts) could be detected (Fig. 5A), as shown by labeling with either anti-GalCO4 or ␤-tubulin III antibody (Fig. 5B). IhNSC differentiation following growth factor removal was initially investigated by immunofluorescence assays using antibodies directed against ␤-tubulin III and microtubule-associated protein 2 (MAP2) to identify neurons, GFAP to label astrocytes, and GalC or O4 to detect oligodendrocytes. Both bulk cultures and clonally expanded IhNSCs gave rise to the three major neural lineages (Fig. 3) and, notably, to a quite high number of neurons and oligodendrocytes (Fig. 4B, 4C). We observed the generation of unipolar, ␤-tubulin III- and MAP2-positive neurons by day 10. By day 17, MAP2-stained neurons displayed long, branched processes with numerous var-

De Filippis, Lamorte, Snyder et al.

2315

Figure 2. Analysis of spontaneous differentiation of human neural stem cells (hNSCs) and immortalized human neural stem cells (IhNSCs) in the absence of mitogenic factors. (A–H): Phase-bright microphotographs of hNSCs (E–H) and IhNSCs (A–D). Freshly dissociated NS (A, E) were cultured for 3 days in the presence of FGF2 (B, F) and terminally differentiated in the absence of mitogenic factors (C–D, G–H). (I): Western blot of cell lysates from IhNSCs for the analysis of v-myc expression in NS, D (corresponding to dissociated NS at 24 hours from dissociation), progenitors (third day), and differentiated cultures at 10, 17, and 24 days in vitro (div). ␣-Vinculin expression is also shown (J). (K, L): Downregulation of v-myc expression during differentiation of IhNSCs by immunofluorescence analysis with N88 v-myc antibody from stem cell stage (K) to 17 div differentiated progeny (L). Some aspecific signal comes from detached dead cells (C, D). Magnification, ⫻40. Abbreviations: D, doublets; NS, neurospheres.

icosities (Fig. 3D–3F). At day 17, the number of ␤-tubulin III-positive cells peaked at 20% of the total number (Fig. 4B) and remained stable up to day 24, after which a progressive increase in the dendritic MAP2 levels was observed, reflecting neuronal maturation (Fig. 4A, 4C). IhNSCs generated typical GFAP-positive astrocytes (Fig. 3A–3C). These colabeled with

neither ␤-tubulin III (Fig. 5C) nor MAP2 (supplemental online Fig. 3), as reported for transformed NSC-like cells [29]. The functional differences between IhNSCs and their parental hNSCs regarded a higher overall neuronal differentiation capacity in IhNSCs (␤-tubulin III⫹ cells, 18.22% ⫾ 1.93% vs. 8.95% ⫾ 0.95% [SE]) [26] together with a unique capacity to

Figure 3. The differentiation potential of immortalized human neural stem cells (IhNSCs). Immunofluorescence analysis of glial (GFAP [A–C]), neuronal (MAP2, red [D–F]), and oligodendroglial (GalC⫹O4, red [G–I]) markers in progenitors (3 days in vitro ⫼) and differentiated IhNSCs (17 and 24 div). IhNSCs were differentiated as described in Materials and Methods. 4,6-Diamidino-2-phenylindole nuclear staining (blue) is also shown to detect total cells. Magnification, ⫻20. Abbreviations: GFAP, glial fibrillary acidic protein; MAP2, microtubule-associated protein 2.

www.StemCells.com

2316


Figure 4. Quantitative analysis of neuronal and glial marker expression during differentiation. (A): Western blot analysis of GFAP and MAP2 expression in immortalized human neural stem cell (IhNSC) NS, D, progenitors (third day), and differentiated progeny at 10, 17, and 24 days. (B–D): Quantitative analysis of the differentiation capacity of IhNSCs and parental human neural stem cells (hNSCs). The remainders of the cells that add up to 100% of the total cell number were mostly GFAP-immunoreactive cells (Fig. 3B, 3C). Data are expressed as percentages of cells positive for ␤-tubulin III (early neuronal marker) (B), MAP2 (late neuronal marker) (C), and GalCO4 (oligodendroglial markers) (D) over total number of cells at different times of differentiation. Abbreviations: D, doublets; GFAP, glial fibrillary acidic protein; MAP2, microtubule-associated protein 2; NS, neurospheres; WT, wild-type.

Figure 5. Analysis of proliferation decay during differentiation. (A): Immunofluorescence analysis with KI67 antibody of differentiating human neural stem cells (hNSC) and immortalized human neural stem cells (IhNSCs). Data are expressed as percentages of KI67⫹ cells over total number of cells. (B): Proliferation of neuronal and oligodendroglial progenitors during differentiation. IhNSCs differentiated at 17 days were coimmunostained with KI67 (a, c) and ␤-tubulin III (b) or GalC and O4 (d). Colabeled cells are indicated by white arrows. Magnification, ⫻20. (C): Segregation of neuronal (␤-tubulin III, in red) and astroglial (GFAP, in green) markers at 10 days in vitro of differentiation. 4,6-Diamidino-2-phenylindole nuclear staining (blue) is also shown. Magnification, ⫻40. Abbreviations: GFAP, glial fibrillary acidic protein; wt, hNSC.

generate oligodendrocytes. As much as 19.11% ⫾ 7.97% of total cells were GalC- or O4-positive oligodendrocytes (Fig. 4D) [25, 26], which evolved from typical immature bipolar precursors (Fig. 3G) to star-shaped cells, with numerous branched processes (Fig. 3H, 3I).

These findings describe a novel— heretofore unavailable— and truly multipotent IhNSC line, endowed with stable functional characteristics, with a broad capacity to generate human neurons and the ability to generate large numbers of human oligodendrocytes.


2317

active for glutamate (4.3% ⫾ 1.24% of total IhNSCs) (Fig. 6C). We could maintain GABA- and glutamate-positive neurons in vitro for up to 40 days, during which maturation occurred, leading to accumulation of neurotransmitters in vesicular structures along the neuronal processes (described below). This neuronal differentiation pattern was strikingly similar to that of parental hNSCs, in which 21.3% ⫾ 2.47% of the total ␤-tubulin III (␤-tubIII)-IR population was GABAergic (corresponding to 6.07% ⫾ 1.22% of total cells), whereas 28.4% ⫾ 2.58% of total ␤-tubIII-IR cells were glutamatergic (representing 9.2% ⫾ 2.24% of total cell numbers). In addition, 1.87% ⫾ 1.03% of total cells expressed the cholinergic marker ChAT (Fig. 6D), whereas immunoreactivity for somatostatin was observed only occasionally. In agreement with our previous report on hNSCs [18], we observed no serotoninergic immunoreactivity. However, the catecholaminergic marker TH, whose expression in parental hNSCs is both transient and sporadic, was detected in 0.536% ⫾ 0.105% of total IhNSC numbers (Fig. 6E) and corresponded to 2.86% ⫾ 1.11% of total neuronal progeny (as ␤-tubulin III-IR cells) (Fig. 6G, 6H). Intriguingly, we could significantly increase the number of TH-IR neurons upon exposure to medium conditioned by B49 cells, similar to parental cells [13]. Indeed, 6.87% ⫾ 0.929% of the total differentiated progeny labeled with the anti-TH antibody (Fig. 6F), corresponding to 49.24% ⫾ 1.53% of the ␤-tubIII-IR progeny (Fig. 6I– 6L). Previous studies showed that a catecholaminergic neurotrophic component in B49 medium is glial-derived neurotrophic factor (GDNF) [30]. We did test the effect of GDNF alone as well as in combination with other neuronal surviving factors, such as brain derived neurotrophic factor and ciliary neurotrophic factor, on IhNSCs. Although no TH induction could be detected, induction of the glutamatergic phenotype was observed. We would argue that additional, as yet unidentified factors may be required for TH induction in our system. Nonetheless, the IhNSCs described here do display a heretofore undocumented capacity to generate a large number of human catecholaminergic neurons. This, together with their ability to generate other major CNS neurotransmitter phenotypes, makes them a rather useful, renewable source of human neurons. Figure 6. Analysis of neuronal subtypes in immortalized human neural stem cell (IhNSC)-differentiated cells and induction of dopaminergic phenotype in IhNSCs-derived neurons. (A–D): Cultures were differentiated as in Figure 5; fixed; immunostained with antibodies against GABA, ␤-tubulin III, GLU, and ChAT, detected with a fluorescein-conjugated secondary antibody. ␤-Tubulin III was detected with rhodamine-conjugated antibody. (E): Quantitative analysis of neurotransmitter phenotypes distribution at 17 days of differentiation. (F): Time-course study of TH-immunoreactive cells in IhNSCs induced with B49C.M⫹LIF. Data are expressed as relative percentage of TH-immunoreactive cells over total number of cells. (G–L): Immunostaining of untreated (G, H) and B49C.M.⫹LIFinduced (I, L) neurons with anti-TH antibody (in red [G, I]) and anti-␤ tubulin III (in green [H, L]). 4,6-Diamidino-2-phenylindole nuclear staining (blue) is also shown in combination with TH labeling to indicate total number of cells. Magnification, ⫻20. Abbreviations: ChAT, choline acetyltransferase; CONTR, control; GLU, glutamate; LIF, leukemia inhibitory factor; TH, tyrosine hydroxylase.

Neurotransmitter Phenotypes We studied the neurotransmitter phenotypes generated by the neuronal IhNSC progeny. Twenty-four days after growth factor removal, 39.17% ⫾ 7.5% of total ␤-tubulin III-immunoreactive (IR) cells (i.e., 7.11% ⫾ 3.5% of total IhNSC progeny) were colabeled with anti-GABA antibodies (Fig. 6A, 6B), whereas 24.6% ⫾ 2.17% of total ␤-tubulin III-IR cells were immunorewww.StemCells.com

Morphological and Functional Analysis of Neuronal Cells By definition, mature neurons are polarized cells that segregate axonal from dendritic markers [31]. From this perspective, we investigated IhNSC neuronal progeny by multiple immunolabeling using axonal, dendritic, and synaptic markers. Changes in the morphology of IhNSC progeny, occurring between 3 and 24 div, are shown in Figure 2 (phase) and Figure 3D–3F. Marker segregation became apparent starting at day 17 with developmental progression and remained incomplete up to day 38. Yet neurites clearly showed numerous, well-defined spines (Fig. 7C) and many presynaptic varicosities within putative axons (described below). Also, IhNSCs neuronal progeny did possess the properties of functionally active neurons. These cells were analyzed at stages between 17 and 38 div, during which we monitored membrane voltage and currents by patch-clamp recordings in both cell-attached and whole-cell configurations. Notably, most cells exhibiting neuron-like morphology (n ⫽ 26 of 38) produced clear and spontaneous action potentials (Fig. 7D), which could be also elicited by somatic current injections (n ⫽ 4 of 7) (Fig. 7E). Voltage clamp analysis revealed the presence of voltage-activated sodium and potassium conductance on the order of a few hundred pA, although

2318


Figure 7. Morphological and functional differentiation of neurons: action potentials, dendritic clusters of glutamate receptors, and axonal vesicle turnover. (A–C): An exemplar neuron obtained by neural stem cell differentiation labeled with two antibodies for the axonal marker Tau (A) and the dendritic marker MAP2 (B). (C): Merged image of (A, B). The inset in (C) represents the magnified view of the dendritic shaft. (D): Representative extracellular recordings of spontaneous action potentials from the soma of a neuronal cell. The bidirectional transients on the current trace are action potentials. (E): Representative whole-cell current clamp recordings of evoked action potentials. (F–G): Double immunofluorescence staining to reveal the presence of synaptic varicosities: axons are in red (Tau-immunoreactive), and synaptic varicosities are in green (synaptic vesicle antigen P38-immunoreactive). (H): Spontaneous vesicular turnover occurring in a putative axon. The green puncta indicate the presence of active synaptic varicosities with cycling vesicles. All results presented in Figure 7 were obtained from neural stem cells differentiated in vitro for 24 days. (I–N): In vivo labeling for active presynaptic varicosities (I) and surface GluR1 glutamate receptors (L). MAP2 antibody was used as counterstain (M). (N): Merged image of (I, L, M). Abbreviations: GLUR1, glutamate receptor 1; MAP2, microtubule-associated protein 2; ms, millisecond(s); sec, second(s).

no clear spontaneous synaptic events could be detected (n ⫽ 38 of 38). These results led us to ask whether these excitable cells established synaptic terminals. We examined the developmentinduced morphological appearance of punctuate immunoreactivity, using well-known presynaptic markers (e.g., synaptic vesicle proteins synaptotagmin-I, P38, and synapsin-I). The localization of synaptic vesicles is one of the earliest detectable events of synaptogenesis in vitro. We observed that as early as 17 days in vitro (Fig. 7F, 7G), most cells demonstrated thin putative axonal branches, showing clear punctuated staining for synaptic vesicles markers, reminiscent of the synaptic varicosities found in brain tissue. To understand whether these terminals might represent functional synaptic contacts, we analyzed both the presence and extent of synaptic vesicle cycling using anti-synaptotagmin-1 antibodies, tagged to the intralumenal domain of this synaptic vesicle protein [28]. Furthermore, to determine whether these terminals were actual synapses, endowed with both pre- and postsynaptic specializations, we colabeled postsynaptic structures with antibodies directed against the dendritic marker MAP2 and monitored the surface glutamate GluR1 receptors by antibodies raised against extracellular domain. Incubating cultured neurons with synaptotagmin-1 antibodies for periods ranging from 2 to 12 hours led to efficient internalization of antibodies into cells (Fig. 7H, 7I, 7N). The resultant punctuate staining patterns were consistent with the distribution of synaptic varicosities, as revealed by the colocalization of anti-synaptotagmin-1 with other presynaptic terminal markers. Exposing these cells to anti-GluR1 antibodies led to distinct and specific dendritic membrane staining patterns

(Fig. 7L, 7N), which were also punctuate, suggesting the expression of small glutamate receptors clusters on dendritic surfaces. In most cases, postsynaptic labeling for GluR1 receptors and presynaptic sites displaying clear vesicle turnover did not colocalize, suggesting that although synaptogenesis was clearly in place, formation of mature synapse was incomplete. These results are in agreement with the absence of spontaneous synaptic currents in electrophysiological recordings (described above). This observation notwithstanding, our results describe the inherent ability of the IhNSC progeny to carry out neuronal maturation to an extent never described before.

DISCUSSION In this work, we describe the establishment of a novel immortalized human neural cell line (IhNSCs), by using wild-type human neural stem cells as a target of v-myc-mediated immortalization. Given its characteristics, this cell line will permit biotechnological and biomedical applications that use human rather than rodent brain cells. An additional advantage lies in the similarity of IhNSCs to their parental, normal hNSCs. In fact, this should enable a two-step approach, by which preliminary, broad tests can be rapidly performed on IhNSCs, to be eventually refined using normal hNSCs. We also show that IhNSCs are multipotent and, for the first time, show their ability to generate large quantities of oligodendrocytes and mature human neurons,

De Filippis, Lamorte, Snyder et al. the latter being capable of eliciting action potentials and of organizing the synaptic machinery. Our growth kinetics analysis shows that immortalization potentiates the self-renewal ability observed in normal parental hNSCs. Indeed, similar to other somatic stem cells, wild-type human neural stem cells normally enter into “crisis” at approximately 40 –50 passages [19], although some groups have reported a much earlier time point [32]. By v-myc transfection, we have currently propagated the same cells as IhNSC for up to 120 passages without any sign of crisis, senescence or proliferative decay. These results underline the fact that by v-myc transduction, we have generated what appears to be an immortal hNSC line. For immortalized, long-term-expanded cells to be considered a potential source of human cells for clinical applications, extensive validation of lack of transformation becomes a vital endeavor to prevent tumorigenicity upon transplantation. Because of their extensive proliferative capacity, their retention of an intrinsic growth factor dependence, and the stable karyotype observed in their parental, normal counterpart, immortalized cells are well-poised as the system of choice to expand safety studies in this area.

IhNSCs as an Unlimited Source of True Human Neuronal Cells A careful review of the data available on immortalized human neural cell lines reveals that they generate different neurons and astroglial cells [4], with differentiation often being regulated by cues other than those required by normal human neural cells [18]. It has also been demonstrated that whereas astroglial differentiation may require a signal that can only be provided by primary neural tissue, neurons can only be generated by exposure to serum [25]. In this view, unlike normal human neural cells, differentiation in immortalized cells largely depends on external cues rather than being a spontaneous phenomenon. This was not the case for our IhNSCs, whose differentiation occurred upon simple mitogen withdrawal, in a manner similar to their parental hNSC, in agreement with findings on analogous cell lines (i.e., HNSC.100 cells) [26]. Nonetheless, significant, advantageous differences do distinguish our IhNSCs from HNSC.100. The neurons and astroglia generated by HNSC.100 were 11% and 85% of total cell numbers, whereas the production of oligodendrocytes was anecdotal. Conversely, not only did the differentiated progeny of IhNSCs embody numerous oligodendrocytes (described below) and nearly twice as many neuronal cells (Fig. 4B, 4C), but these could be maintained as viable cells for up to 40 days from the onset of differentiation. This is in sharp contrast with the maximum lifespan of HNSC.100-derived neurons: they are limited to a maximum of 12 days in vitro, which does not allow for completion of neuronal differentiation and maturation [19, 26]. Furthermore, the spectrum of neurotransmitter phenotypes generated by our IhNSCs was also much broader than reported so far. Whereas HNSC.100 [26] generate exclusively GABAergic cells, IhNSCs produce a wide array of neuronal subtypes: 39.17% of the neurons were GABAergic, 24.6% were glutamatergic, and 10% were cholinergic. In addition, to the best of our knowledge, we report the first evidence of the ability of immortalized hNSCs to efficiently generate TH-immunoreactive human neurons (up to 49% of total neuronal cells) under appropriate conditions (Fig. 6F, 6I– 6L). This finding may be of primary relevance for studies in the area of therapeutics for Parkinson disease. The ability of IhNSC to generate mature and functional neurons without the support of cocultured primary cells [33] is an important factor for the in vitro modeling of human neurowww.StemCells.com

2319

genesis using these cells. Here we show that IhNSCs generate mature neurons endowed with ion channel machinery able to support the elicitation of true action potentials, both spontaneous and evoked, while expressing the dendritic and axonal markers MAP2 and Tau 1. A branched and mature morphology with presynaptic varicosities and postsynaptic spines was clearly evident (Fig. 7C). In neurons showing vesicle accumulation at putative synaptic terminals, both trafficking and active recycling of synaptic vesicles was demonstrated by in vivo labeling with anti-lumenal synaptotagmin-I antibodies (Fig. 7H, 7I). Although we did not detect spontaneous synaptic currents, the morphological and functional synaptic features evident in IhNSC-derived neurons are suggestive of advanced stages of maturation. The observed expression of late synaptic markers whose sublocalization is comparable to that of active synapses in vivo, as well as elicitation of action potentials, suggests IhNSCs as an abundant stable source of mature, human neuronal cells for neurochemical, biochemical, molecular, and electrophysiological studies and drug modeling.

Oligodendroglia as Generated by IhNSCs The ability of IhNSCs to give rise to bona fide oligodendrocytes is pivotal in cell lines to be used for modeling human neurogenesis ex vivo and, particularly, in drug discovery for demyelinating disorders. Oligodendrocytes derive from progenitor cells (OPCs), which express both the NG2 proteoglycan and platelet-derived growth factor ␣ receptor (PDGF-␣R) [34, 35]. Fluorescence-activated cell sorting analysis of NG2 and PDGF-␣R expression in IhNSCs identifies numerous human OPCs (supplemental online Fig. 4). Apart from the previously described v-myc-immortalized human neural stem cells [25, 26], which were unable to originate oligodendrocytes, we could also detect branching oligodendrocytes (more than 19% of total cell counts). These oligodendroglial cells, whose presence was detected at 17 days after the start of differentiation following mitogen removal, were GalC⫹ and O4⫹ (vs. 1%– 4% in parental cells) and were also shown to be positive for the expression of 2⬘,3⬘-cyclic nucleotide 3⬘-phosphodiesterase and ␤-tubulin IV (supplemental online Fig. 5), a subtype of tubulin protein specifically expressed in premyelinating oligodendrocytes and abundant in myelinating oligodendrocytes [36]. To our knowledge, this report is the first to describe a renewable source of multipotent human neural cells having the capacity to generate large numbers of bona fide human oligodendrocytes.

Immortalization but Not Transformation: A Critical IhNSC Feature Our studies show the generation of a new IhNSCs line through retroviral-mediated gene transfer of the v-myc oncogene [37, 38]. Although this results in a faster proliferation rate and enhanced levels of self-renewal, IhNSCs preserve many of the characteristics of their normal counterparts, without signs of transformation (Figs. 1B, 5A–5C), which does not come as a surprise. It is generally accepted that myc is not oncogenic by itself, likely because of its dual action as a promoter of cell division and as an apoptotic mediator in the absence of growth factors [39]. Furthermore, MLV-derived retroviral vectors, such as the one used here, are spontaneously downregulated in mammalian CNS in vivo [40]. In addition, Flax et al. [25] have reported that in genetically manipulated human NSC clones, the propagating gene v-myc was undetectable in human cells beyond 24 – 48 hours following orthotopic engraftment, whereas only a residual v-myc expression was found in HNSC.100 cells grafted into the brain [26]. Here, we show that a spontaneous and significant downregulation of v-myc expression occurs in IhNSC in concomi-


2320

tance with mitogen removal and the onset of cell differentiation. Although inactivation of the Mo-MLV promoter-enhancer and of other endogenous proviruses has been previously documented, both ex vivo and in vivo [41– 49], the cellular and molecular mechanisms that underlie this phenomenon in our system remain to be elucidated. In view of the fact that we are analyzing a mixed, bulk population, an effect related to the integration site [44, 46, 50, 51] on v-myc downregulation is most unlikely. Intriguingly, previous studies [52] showed that mitogens such as EGF and prolactin can act on a specific enhancer at the 5⬘ end of the MMLV-LTR sequence. Thus, it may be possible that v-myc downregulation in our cultures, which follows EGF and FGF2 removal, may occur through as yet unidentified neural responsive elements in the MoLV-LTR sequence. Additional studies are required to unravel these issues. Nonetheless, these findings support the nontransformed nature of IhNSCs, which also remains strictly dependent on growth factor to undergo proliferation and self-renewal (Fig. 1B). We confirmed this by assessing the potential tumorigenic ability of IhNSCs through intraparenchymal injection in the adult mouse brain (n ⫽ 6): no tumors could be detected by magnetic resonance imaging (MRI) analysis at time points well beyond 6 months following transplantation, similar to mice transplanted with parental, nonimmortalized cells (n ⫽ 2). As positive controls, we used animals receiving human glioblastoma stem cells under identical conditions. As expected, invasive tumors developed as early as 1 month after injection [13]. The candidate metastatic capacity of IhNSCs was further tested by i.v. tail injection of IhNSCs: no tumor formation could be detected by whole-body MRI analysis (resolving up to 100 ␮m) or histological analysis (data not shown).

CONCLUSION The findings of this work describe a novel human neural stem cell line, whose functional properties permit its application in the areas of applied biotechnology and human developmental neurobiology. The readiness by which high-throughput drug discovery assays can be established on human neurons, astroglia, and oligodendroglia and by which models of human neurogenesis ex vivo can be developed and the possibility of confirming the results obtained by using the normal parental hNSCs (if necessary) pave the way for the development of new and meaningful endeavors in the area of drug modeling, development, and validation.

ACKNOWLEDGMENTS We thank Luigi Carlessi for Western blot analysis; Tatiana Lavazza for confocal microscopy; Letterio Politi for MRI analysis; and Fraser McBlane, Daniela Ferrari, and Laura Rota Nodari for editing suggestions. Valuable support was provided by Manuel Fiorillo and Vincenzo Fiorillo. This work was supported by generous funding from Neurothon Onlus Association, Fondazione Cariplo, BMW Italia S.p.A, Instituto Superiore di Sanita and Framework VI funding from EEC.

DISCLOSURE

16

2 3 4

5 6 7 8 9 10 11 12 13 14

15

Anderson DJ. Stem cells and pattern formation in the nervous system: the possible versus the actual. Neuron 2001;30:19 –35. Gage FH. Stem cells of the central nervous system. Curr Opin Neurobiol 1998;8:671– 676. Gage FH. Mammalian neural stem cells. Science 2000;287:. Martinez-Serrano A, Rubio FJ, Navarro B et al. Human neural stem and progenitor cells: in vitro and in vivo properties, and potential for gene therapy and cell replacement in the CNS. Curr Gene Ther 2001;1:279 –299. McKay R. Stem cells in the central nervous system. Science 1997;276: 66 –71. Morrison SJ, Shah NM, Anderson DJ. Regulatory mechanisms in stem cell biology. Cell 1997;88:287–298. Ming GL, Song H. Adult neurogenesis in the mammalian central nervous system. Annu Rev Neurosci 2005;28:223–250. Weiss S, Reynolds BA, Vescovi AL et al. Is there a neural stem cell in the mammalian forebrain? Trends Neurosci 1996;19:387–393. Gage FH, Coates PW, Palmer TD et al. Survival and differentiation of adult neuronal progenitor cells transplanted to the adult brain. Proc Natl Acad Sci U S A 1995;92:11879 –11883. Kilpatrick TJ, Bartlett PF. Cloning and growth of multipotential neural precursors: Requirements for proliferation and differentiation. Neuron 1993;10:255–265. Gritti A, Cova L, Parati EA et al. Basic fibroblast growth factor supports the proliferation of epidermal growth factor-generated neuronal precursor cells of the adult mouse CNS. Neurosci Lett 1995;185:151–154. Gritti A, Parati EA, Cova L et al. Multipotential stem cells from the adult mouse brain proliferate and self-renew in response to basic fibroblast growth factor. J Neurosci 1996;16:1091–1100. Galli R, Gritti A, Bonfanti L et al. Neural stem cells: An overview. Circ Res 2003;92:598 – 608. Gritti A, Fro¨lichsthal-Schoeller P, Galli R et al. Epidermal and fibroblast growth factors behave as mitogenic regulators for a single multipotent stem cell-like population from the subventricular region of the adult mouse forebrain. J Neurosci 1999;19:3287–3297. Levison SW, Goldman JE. Both oligodendrocytes and astrocytes develop

CONFLICTS

The authors indicate no potential conflicts of interest.

REFERENCES 1

OF POTENTIAL OF INTEREST

17 18 19

20

21 22 23 24 25 26 27 28

from progenitors in the subventricular zone of postnatal rat forebrain. Neuron 1993;10:201–212. Doetsch F, Garcı´a-Verdugo JM, Alvarez-Buylla A. Regeneration of a germinal layer in the adult mammalian brain. Proc Natl Acad Sci U S A 1999;96:11619 –11624. Johansson CB, Svensson M, Wallstedt L et al. Neural stem cells in the adult human brain. Exp Cell Res 1999;253:733–736. Vescovi AL, Gritti A, Galli R et al. Isolation and intracerebral grafting of nontransformed multipotential embryonic human CNS stem cells. J Neurotrauma 1999;16:689 – 693. Vescovi AL, Parati EA, Gritti A et al. Isolation and cloning of multipotential stem cells from the embryonic human CNS and establishment of transplantable human neural stem cell lines by epigenetic stimulation. Exp Neurol 1999;156:71– 83. Ostenfeld T, Svendsen CN. Requirement for neurogenesis to proceed through the division of neuronal progenitors following differentiation of epidermal growth factor and fibroblast growth factor-2-responsive human neural stem cells. STEM CELLS 2004;22:798 – 811. Carpenter MK, Cui X, Hu ZY et al. In vitro expansion of a multipotent population of human neural progenitor cells. Exp Neurol 1999; 158:265–278. Nunes MC, Roy NS, Keyoung HM et al. Identification and isolation of multipotential neural progenitor cells from the subcortical white matter of the adult human brain. Nat Med 2003;9:439 – 447. Ryder EF, Snyder EY, Cepko CL. Establishment and characterization of multipotent neural cell lines using retrovirus vector-mediated oncogene transfer. J Neurobiol 1990;21:356 –375. Sah DW, Ray J, Gage FH. Bipotent progenitor cell lines from the human CNS. Nat Biotechnol 1997;15:574 –580. Flax JD, Aurora S, Yang C et al. Engraftable human neural stem cells respond to developmental cues, replace neurons, and express foreign genes. Nat Biotechnol 1998;16:1033–1039. Villa A, Snyder EY, Vescovi A et al. Establishment and properties of a growth factor-dependent, perpetual neural stem cell line from the human CNS. Exp Neurol 2000;161:67– 84. Bull ND, Bartlett PF. The adult mouse hippocampal progenitor is neurogenic but not a stem cell. J Neurosci 2005;25:10815–10821. Malgaroli A, Ting AE, Wendland B et al. Presynaptic component of long-term potentiation visualized at individual hippocampal synapses. Science 1995;268:1624 –1628.


29 Galli R, Binda E, Orfanelli U et al. Isolation and characterization of tumorigenic, stem-like neural precursors from human glioblastoma. Cancer Res 2004;64:7011–7021. 30 Lin LF, Zhang TJ, Collins F et al. Purification and initial characterization of rat B49 glial cell line-derived neurotrophic factor. J Neurochem 1994;63:758 –768. 31 Horton AC, Ehlers MD. Neuronal polarity and trafficking. Neuron 2003; 40:277–295. 32 Villa A, Rubio FJ, Navarro B et al. Human neural stem cells in vitro. A focus on their isolation and perpetuation. Biomed Pharmacother 2001; 55:91–95. 33 Craig AM, Graf ER, Linhoff MW. How to build a central synapse: clues from cell culture. Trends Neurosci 2006;29:8 –20. 34 Nishiyama A, Lin XH, Giese N et al. Interaction between NG2 proteoglycan and PDGF alpha-receptor on O2A progenitor cells is required for optimal response to PDGF. J Neurosci Res 1996;43:315–330. 35 Trapp BD, Nishiyama A, Cheng D et al. Differentiation and death of premyelinating oligodendrocytes in developing rodent brain. J Cell Biol 1997;137:459 – 468. 36 Terada N, Kidd GJ, Kinter M et al. Beta IV tubulin is selectively expressed by oligodendrocytes in the central nervous system. Glia 2005; 50:212–222. 37 Lee CM, Reddy EP. The v-myc oncogene. Oncogene 1999;18:2997–3003. 38 Yeh E, Cunningham M, Arnold H et al. A signalling pathway controlling c-Myc degradation that impacts oncogenic transformation of human cells. Nat Cell Biol 2004;6:308 –318. 39 Chang DW, Claassen GF, Hann SR et al. The c-Myc transactivation domain is a direct modulator of apoptotic versus proliferative signals. Mol Cell Biol 2000;20:4309 – 4319. 40 Martı´nez-Serrano A, Bjo¨rklund A. Gene transfer to the mammalian brain using neural stem cells: a focus on trophic factors, neuroregeneration, and cholinergic neuron systems. Clin Neurosci 1995–1996;3:301–309. 41 Barklis E, Mulligan RC, Jaenisch R. Chromosomal position or virus mutation permits retrovirus expression in embryonal carcinoma cells. Cell 1986;47:391–399.

2321

42 Conklin KF, Groudine M. Varied interactions between proviruses and adjacent host chromatin. Mol Cell Biol 1986;6:3999 – 4007. 43 Dorner AJ, Stoye JP, Coffin JM. Molecular basis of host range variation in avian retroviruses. J Virol 1985;53:32–39. 44 Fincham VJ, Wyke JA. Differences between cellular integration sites of transcribed and nontranscribed Rous sarcoma proviruses. J Virol 1991; 65:461– 463. 45 Fincham VJ, Wyke JA. On the structure, genesis and significance of DNA duplications at the Rous sarcoma proviral insertion sites in Rat-1 cells. J Gen Virol 1992;73:3247–3250. 46 Hoeben RC, Migchielsen AA, van der Jagt RC et al. Inactivation of the Moloney murine leukemia virus long terminal repeat in murine fibroblast cell lines is associated with methylation and dependent on its chromosomal position. J Virol 1991;65:904 –912. 47 Ja¨hner D, Jaenisch R. Chromosomal position and specific demethylation in enhancer sequences of germ line-transmitted retroviral genomes during mouse development. Mol Cell Biol 1985;5:2212–2220. 48 Palmer TD, Rosman GJ, Osborne WR et al. Genetically modified skin fibroblasts persist long after transplantation but gradually inactivate introduced genes. Proc Natl Acad Sci U S A 1991;88:1330 –1334. 49 Xu L, Yee JK, Wolff JA et al. Factors affecting long-term stability of Moloney murine leukemia virus-based vectors. Virology 1989;171: 331–341. 50 Gu¨nzburg WH, Groner B. The chromosomal integration site determines the tissue-specific methylation of mouse mammary tumour virus proviral genes. EMBO J 1984;3:1129 –1135. 51 Ucker DS, Firestone GL, Yamamoto KR. Glucocorticoids and chromosomal position modulate murine mammary tumor virus transcription by affecting efficiency of promoter utilization. Mol Cell Biol 1983;3:551–561. 52 Haraguchi S, Good RA, Engelman RW et al. Prolactin, epidermal growth factor or transforming growth factor-alpha activate a mammary cellspecific enhancer in mouse mammary tumor virus-long terminal repeat. Mol Cell Endocrinol 1997;129:145–155.

See www.StemCells.com for supplemental material available online.

www.StemCells.com