SEARCH

SEARCH BY CITATION

Keywords:

  • Human embryonic stem cells;
  • Dopaminergic;
  • PA6 cells;
  • Stromal-derived inducing activity;
  • Differentiation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Producing dopaminergic (DA) neurons is a major goal of human embryonic stem cell (hESC) research. DA neurons can be differentiated from hESC by coculture with the mouse PA6 stromal cell line; this differentiation-inducing effect is termed stromal-derived inducing activity (SDIA). The molecular and biochemical nature of SDIA is, however, unknown. Various studies have suggested that SDIA involves either a fixation-resistant component located on the PA6 cell surface or factors secreted into the medium by PA6 cells. To address this question, hESC were cocultured with PA6 cells for 12 days and then further differentiated with sonic hedgehog homolog, fibroblast growth factor-8, and glial cell line-derived neurotrophic factor. After 18 days, 34% of cells were tyrosine hydroxylase (TH)+. When PA6 cells were fixed or irradiated, the number of TH+ cells was decreased by threefold, whereas mitomycin-c treatment of feeder cells decreased the number of TH+ cells by 32%. The neural-inducing effect of PA6 cells, as monitored by β-III-tubulin expression, was minimally affected by mitomycin-c treatment or fixation but was decreased 50% by irradiation. Medium conditioned by PA6 cells was ineffective in differentiating TH+ cells when used alone. Conditioned medium combined with heparin and/or fixed PA6 cells produced TH+ cell differentiation, although less effectively than PA6 cell coculture. Thus, PA6 cell surface activity is required for neural differentiation of hESC, but secreted factors are required for the specific DA neuron-inducing effect.

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


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Embryonic stem cells (ESC), because of their capacity for both multipotency and self-renewal, have a great potential for generating specialized cells for use as in vitro model systems and perhaps eventually for cell replacement therapy. Cells that can be generated from human embryonic stem cells (hESC) include multiple subtypes of central nervous system (CNS) cells, thus providing valuable research tools for understanding molecular mechanisms and biochemical pathways underlying cell development and neurodegenerative processes. The major challenge in differentiation of ESC is to control and direct the ESC toward a specific functionally distinct phenotype of interest. Midbrain dopamine (DA) neurons are of particular interest because of their importance for several disorders, especially Parkinson's disease (PD), schizophrenia, and drug abuse [1, 2]. Several laboratories have described methods to differentiate mouse and primate ESC into midbrain dopaminergic neurons in response to extrinsic cues [3, [4], [5], [6]7].

Methods for differentiating ESC into midbrain DA neurons fall into two distinct categories. The first involves stepwise differentiation, including generation of embryoid bodies (EBs) followed by selection and expansion of nestin-positive neural precursors and differentiation into neural subtypes [5]. This approach consists of multiple stages and requires a number of media and culture conditions. A second strategy for generation of midbrain dopaminergic neurons is a coculture method based on what has been termed stromal-derived inducing activity (SDIA). The stromal cell line PA6 or similar cell lines, such as MS5, derived from mouse skull bone marrow, serve as a feeder layer and exhibit inducing activity. This method has the advantage of simplicity and rapidity, requiring only a single step, but suffers from the limitation of requiring the presence of the feeder cell line [3, 8].

Our laboratory and others have adapted the SDIA method for generation of human dopaminergic neurons [3, 6, 9, [10], [11]12]. In previous experiments, tyrosine hydroxylase (TH)+ neurons generated by the coculture method expressed markers for mature dopaminergic neurons, including DAT, Nurr1, Lmx1b, and Ptx3 [9]. On the basis of markers that are expressed, a number of studies also suggest that the majority of TH+ neurons generated from hESC by SDIA have a midbrain dopaminergic identity [13, [14]15].

The present study had two main objectives. The first was to determine whether SDIA is required only for the initial stages of differentiation. A previous study by Perrier et al. [6] was able to obtain dopaminergic differentiation of hESC by 28 days of coculture with SDIA, followed by further differentiation for several weeks in the presence of growth factors, including sonic hedgehog homolog (SHH) and fibroblast growth factor 8 (FGF8). However, 28 days is generally a sufficient period of time to allow for nearly complete differentiation of dopaminergic neurons in the presence of SDIA alone [9, 12]. We therefore examined the possibility that differentiation could be obtained by coculturing cells with SDIA for only 10–12 days, until DA neurons first begin to appear, followed by completion of differentiation in the presence of the CNS patterning factors SHH and FGF8. SHH and FGF8, secreted by the notochord and the isthmus, respectively, are key molecules involved in the development of midbrain dopaminergic neurons [16, 17].

The second aim of this study was to gain a better understanding of the nature of SDIA. Initial identification of SDIA suggested that this activity is accumulated on the surface of PA6 cells, since it was reported that PA6 cells retain neural-inducing activity after being fixed with paraformaldehyde [3]. On the other hand, several studies suggest that soluble factors secreted by PA6 cells can induce, at least to some degree, neural differentiation and generation of DA neurons from hESC [18, 19]. To resolve this controversy, we attempted to separate the contributions of cell surface and soluble factors to the overall dopaminergic and neural differentiation-producing effect of SDIA.

In this study, several strategies involving chemical inactivation by paraformaldehyde and ethanol and inhibition of mitosis by γ-irradiation or mitomycin-c treatment were tested in an attempt to kill or inactivate the PA6 cells. We investigated the effect of these treatments on neural-inducing activity of PA6 cells by quantifying expression of the neural precursor and neuronal markers neural cell adhesion molecule (NCAM), Sox-1, β-III-tubulin, and TH. In addition, we used the signaling molecules SHH, FGF8, and glial cell line-derived neurotrophic factor (GDNF), suggested to promote survival and differentiation of the dopaminergic phenotype [20, [21]22], and examined the efficacy of medium conditioned by PA6 cells alone and in combination with fixed PA6 cells.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Cell and Culture Methods

The hESC line BG01 was obtained from BresaGen (Athens, GA, http://www.bresagen.com.au). Cells were maintained in an undifferentiated state on a feeder layer of mitomycin-c-inactivated mouse embryonic fibroblast (MEF) feeder cells at a density of 1.25 × 106 cells per 60-mm dish coated with 0.1% gelatin (Millipore, Billerica, MA, http://www.millipore.com). The culture medium contained Dulbecco's modified Eagle's medium/nutrient mixture (1:1), supplemented with 10% knockout serum replacement, 2 mM l-glutamine, 1 mM nonessential amino acids, 4 ng/ml basic fibroblast growth factor, 50 U/ml Pen-Strep (Gibco, Gaithersburg, MD, http://www.invitrogen.com), and 0.1 mM β-mercaptoethanol (Millipore).

The culture medium was changed daily, with routine passage of hESC on fresh MEF layers carried out twice a week. Dissociation of hESC colonies from the MEF feeder layers was achieved by treating them with 1 mg/ml collagenase type IV for approximately 2 hours.

The PA6 mouse stromal cell line was purchased from Riken BioResource Center Cell Bank (Tsukuba, Japan, http://www.brc.riken.jp/inf/en) and maintained in PA6 culture medium consisting of α-minimum essential medium (Gibco) supplemented with 10% fetal bovine serum (Atlanta Biologicals, Lawrenceville, GA, http://www.atlantabio.com) and 50 U/ml Penn-Strep. Routine analysis of the ESC karyotype was carried out. The karyotype of BG01 cell line was initially normal but converted to trisomy 17 between passages 60 and 70. Critical experiments, including neural and dopaminergic induction of hESC by the SDIA strategy and isolation and differentiation of midbrain-specific neural progenitor cells, were confirmed with karyotypically normal hESC lines BG01 and BG03 in our previous studies [9, 10].

Coculture Experiments

For differentiation, PA6 cells were seeded at 1.5 × 106 cells per well in collagen type I-coated six-well plates. Twenty-four hours later, the medium was replaced with differentiation Glasgow minimum essential medium supplemented with 10% knockout serum replacement, 0.1 mM nonessential amino acids (Gibco), 1 mM sodium pyruvate (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), and 0.1 mM β-mercaptoethanol (Millipore). To differentiate hESC in vitro, ESC were dissociated from the MEF feeder layer by enzymatic treatment and added to the PA6 layer at a density of 50–100 small colonies, corresponding to approximately 50,000–100,000 cells, per well of a six-well plate (∼5,000–10,000 cells per cm2). The culture medium was changed on day 4 and every other day thereafter. The hESC were allowed to differentiate in the coculture system for 10–12 days.

Isolation of Neural Progenitor Cells and Inactivation of PA6 Cells

To isolate the generated neural progenitor cells, the coculture was subjected to papain digestion according to the manufacturer's instructions (16.5 U/ml; Worthington Biochemical, Lakewood, NJ, http://www.worthington-biochem.com), in Dulbecco's phosphate-buffered saline. The coculture was incubated in the papain solution at room temperature for 5–10 minutes. The PA6 cell layer sheet was carefully held by a sterile pipette tip while the colonies were gently washed and removed.

For some experiments, the PA6 cells were inactivated by irradiation, mitomycin-c treatment, or fixation prior to addition of hESC. For irradiation, the PA6 cells were grown to confluence and irradiated with a dose of 0.98 Gy/minute ionizing radiation for 30 minutes from a cesium-137 source (Gammacell 40; MDS Nordion, Ottawa, http://www.mds.nordion.com). For mitomycin-c inactivation, confluent layers of PA6 cells were incubated with 10 μg/ml mitomycin-c (Sigma-Aldrich) for 1 hour. Cells were washed three times with fresh medium and permitted to recover overnight. Fixation of cells was carried out by incubation with 100% ethanol or 4% paraformaldehyde for 15 minutes followed by five washes with phosphate-buffered saline (PBS). For all conditions, after removal from the feeder cell layer, the neural progenitor cells were seeded at a density of 10,000–20,000 cells per cm2 in small clusters in 6- or 12-well plates coated with poly-l-lysine (10 mg/ml; Sigma-Aldrich) and then with laminin (20 μg/ml; Sigma-Aldrich). The isolated progenitor cells were cultured in the presence of 40 ng/ml SHH, 40 ng/ml FGF8, and 20 ng/ml GDNF (all from R&D Systems Inc., Minneapolis, http://www.rndsystems.com). Experiments on fixation and mitotic inactivation of PA6 cells were performed at the same time as experiments on normal PA6 cells but are described separately in Results for the sake of clarity.

Immunocytochemistry

Cells were fixed with 4% paraformaldehyde for 15 minutes, washed with PBS, and then incubated with blocking buffer (PBS, 10% goat serum, and 0.2% Triton X-100) for 1 hour. Cells were incubated for 2 hours at room temperature with primary antibodies diluted in PBS containing 5% goat serum. The following primary antibodies were used: mouse anti-Oct 3/4 (1:50; Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com), goat anti-Oct3/4 (1:100; Santa Cruz Biotechnology), rabbit anti-TH (1:1,000; Pel-Freez, Rogers, AK, http://www.invitrogen.com), rabbit anti-β-III-tubulin (1:2,000; Promega, Madison, WI, http://www.promega.com), rabbit anti-γ-aminobutyric acid (GABA) (1:2,000; Sigma-Aldrich), mouse anti-NCAM (clone NCAM-0B11, 1:50; Sigma-Aldrich), mouse anti-nestin (1:50; R&D Systems), rabbit anti-glial fibrillary acidic protein (anti-GFAP) (1:2,000; Dako, Carpinteria, CA, http://www.dakousa.com), rabbit anti-microtubule-associated protein (MAP-2) (1:1,000), rabbit anti-dopamine β-hydroxylase (DBH) (1:200), mouse anti-stage-specific embryonic antigen 4 (SSEA-4) (1:200), mouse anti-polysialyinated-NCAM (PSA-NCAM) (clone 2-2B; 1:200), and rabbit anti-Sox-1 (1:500; all from Millipore).

The cultures were washed with PBS and then incubated with fluorescent-labeled secondary antibodies (Alexa Fluor 488-labeled [green] or Alexa Fluor 568-labeled [red] goat IgG; 1:500; Gibco) in PBS with 2% goat serum and 1% bovine serum albumin for 1 hour at room temperature. The cells were rinsed five times for 5 minutes each time in PBS. The major antibody used for quantification of TH+ cells was the Pel-Freez polyclonal TH antibody, which is produced using purified TH protein, similar to the original description of a TH antibody by [23]. TH antibody produced by immunization with purified TH protein cross-reacts with TH from multiple mammalian species and tissues but does not detect the related enzymes phenylalanine hydroxylase, tryptophan hydroxylase, dopamine β-hydroxylase, or tyrosine [23]. The Pel-Freez polyclonal TH antibody has been extensively studied and is known to be reliable and specific for immunocytochemical and immunohistochemical detection of TH. In addition, we used a monoclonal anti-TH antibody recognizing an epitope in the N terminus (1:100; Sigma-Aldrich) and rat brain sections to verify the specificity of the polyclonal TH antibody. Expression of the markers β-III-tubulin, MAP2, GFAP, and nestin was always examined with careful consideration of the expected morphology of cells associated with these markers and particularly the expected intracellular localization of these proteins. Negative controls included substituting the primary antibodies with nonimmune mouse and rabbit IgG (1:100; Santa Cruz Biotechnology) and preabsorption of the Oct3/4 primary antibody with its antigenic peptide (0.2 mg/ml N-terminal Oct3/4 of human origin; Santa Cruz Biotechnology).

Images were acquired on a Carl Zeiss Axiovert 200M (Carl Zeiss, Jena, Germany, http://www.zeiss.com) microscope, using a ×40 objective for quantification. Counting and quantification of cells positive for various markers was carried out by randomly choosing 20 colonies in each 35-mm dish in coculture conditions. For calculation of the number of TH+ cells and cells expressing other neuronal markers in feeder-free conditions, cells were costained with 4′,6-diamidino-2-phenylindole and quantified. For each experiment, five fields chosen at random, containing an average of 150 cells per field, were analyzed. The experiments were performed in sextuplicate wells and repeated at least three times. Fields where no cells were found were disregarded, and only TH-expressing cells with definitive cell bodies and processes were regarded as TH+.

Preparation of Conditioned Medium

Conditioned medium was prepared by the culture of confluent PA6 cells in differentiation medium. Pooled conditioned media were filtered to remove cell debris and stored at −20°C for a maximum of 2 days before use. Conditioned medium was changed every day because of possible degradation of secreted factors. To examine the effect of heparin, a solution containing 50 mg/ml heparin (Sigma-Aldrich) was added to the cultures to reach a final concentration of 100 μg/ml.

Statistics

Differences in numbers of colonies or cells were tested by analysis of variance followed by Tukey-Kramer multiple comparisons or by t tests for experiments with two conditions. Results of the analyses of variance are indicated in the legends of Figures 4, Figure 5., Figure 6.7, and results of the multiple comparison tests are indicated by asterisks on the graphs. Differences were considered significant at p < .05.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Induction of Neural Differentiation by PA6 Cells

After 3 days of growth on PA6 cells, hESC colonies started to show signs of differentiation, including loss of their compact appearance and tight borders (Fig. 1A). By day 8, the hESC developed into rosette-like structures (Fig. 1B). These rosette-like structures are believed to represent neural precursor cells in an arrangement resembling the neural tube. The radial arrangement and the circular morphology are reminiscent of the pattern that is present at the time of neural tube closure. The rosettes expressed signature markers of precursor cells in the neural plate, such as Sox1, nestin, NCAM, and PSA-NCAM (Fig. 1E– 1H). Nestin is expressed in a variety of precursor cells of neuroectodermal and mesodermal origin, and it distinguished differentiating hESC from PA6 cells. The majority of colonies contained cells that continued to express Oct3/4 and SSEA-4 at this stage (Fig. 1I, 1J). β-III-tubulin-positive neurons expressing TH could be detected in approximately 40% of colonies (Fig. 1K, 1L).

thumbnail image

Figure Figure 1.. Time course of differentiation of human embryonic stem cells (hESC) in coculture with PA6 stromal feeder cells. Phase-contrast images of hESC in coculture with PA6 cells for 3, 8, 10, and 12 days (A–D). (A): Two colonies after 3 days of coculture, illustrating the development of undefined borders and more elongated cells in the periphery of the colonies. (B): Demonstration of rosette-like structures at day 8 in the center of a colony (arrows). (C, D): Colonies after 10 and 12 days of differentiation containing cells with process-bearing neuronal morphology. (E–P): Immunocytochemical marker expression analysis of hESC after 8 days (E–L) and (M–P) 12 days of coculture. (E): Sox-1. (F): Nestin. (G): NCAM. (H): PSA-NCAM. (I): Oct3/4. (J): SSEA-4. (K): β-III-tubulin. (L): TH. (M): TH and Oct3/4. (N): SSEA-4. (O): β-III-tubulin. (P): TH. Scale bars= 200 μm. Abbreviation: PSA-NCAM, polysialyinated neural cell adhesion molecule; SSEA-4, stage-specific embryonic antigen 4; TH, tyrosine hydroxylase.

Download figure to PowerPoint

After 12 days of coculture, expression of Oct3/4 and SSEA-4 was downregulated, whereas expression of β-III-tubulin and TH was elevated (Fig. 1M– 1P). The hESC were differentiated on PA6 cells for 10–12 days, depending on the stage of differentiation as measured by loss of expression of the undifferentiated ESC markers Oct3/4 and SSEA-4 and appearance of rosette-like structures.

Differentiation After Removal from PA6 Cells

At the time of isolation, Oct3/4 expression was substantially downregulated in all colonies and completely absent in approximately 60% of colonies. Neural rosette structures were readily visible in the center of the majority of colonies.

Upon removal from PA6 cell feeder layer, the differentiating hESC formed small clusters that were suspended in the medium (Fig. 2A). Within 3 days, the cell aggregates attached to the substrate and formed colonies in an organized radial or columnar pattern (Fig. 2B). Process-bearing cells with a neuron-like morphology were observed after approximately 3–5 days (Fig. 2C). Over the next several days, the colonies continued to expand, and cells with elongated morphology suggestive of neurogenic radial glial cells transitioned to more neuronal cell types, as indicated by radial migration and process formation in the periphery of colonies (Fig. 2D, 2E) [24, 25]. More undifferentiated areas in the colonies, as indicated by neural rosettes, appeared to be located in the centers (Fig. 2F), whereas organized radial migration of neurons was seen in the periphery (Fig. 2G).

thumbnail image

Figure Figure 2.. Differentiation of neural progenitor cells following enzymatic isolation. (A): Small clusters of neural progenitor cells immediately after isolation from the PA6 cell layer. (B): Cells attached to the laminin substrate after 3 days, showing cells in a typical columnar arrangement. (C): The first sign of neurons, based on morphological criteria, appearing after 5 days. (D): Further differentiation and expansion of cells in a colony at day 7. (E): Migration and process formation at the edge of a colony at day 7. (F): Formation of rosette-like structures in the center of a colony at day 10. (G): An area more distant from the center of the colony shown in (F). (H, I): Center and periphery of a colony at day 19. Scale bars= 200 μm.

Download figure to PowerPoint

By days 14–19 after dissociation from PA6 cells (Fig. 2H, 2I), most cells had neural characteristics, and no expression of Oct3/4 or SSEA-4 could be detected. After 18 days 46% ± 8% of all cells present were MAP2+, and of these 80% ± 11% were TH+ (Fig. 3A, 3B).

thumbnail image

Figure Figure 3.. Characterization of cellular phenotypes present in differentiated cultures 18 days after removal of cells from the PA6 cell feeder layer. Immunocytochemistry for neuron subtype-specific markers after 18 additional days of differentiation in the presence of SHH, fibroblast growth factor 8, and glial cell line-derived neurotrophic factor. (A, B): Coexpression of MAP2 and TH. Of MAP2+ cells, 80% ± 11% were positive for TH. (C): Expression of GABA was seen in a small number of cells. (D): Illustration of the large number of TH+ cells in a culture. (E): Bipolar morphology of TH+ neurons. (F): Expression of TH was seen in 34% ± 6% of the total cells. This frame shows an area in a typical colony. (G–I): Expression of S-100β and GFAP in astrocytes and glial cells in cultures costained for the neuronal marker β-III-tubulin. Scale bars= 200 μm. Abbreviations: DAPI, 4,6-diamidino-2-phenylindole; GABA, γ-aminobutryic acid; GFAP, glial fibrillary acidic protein; MAP2, microtubule-associated protein; TH, tyrosine hydroxylase.

Download figure to PowerPoint

Expression of MAP2, mainly associated with dendritic extensions of mature neurons, was absent in a minority of immature TH+ cells with short processes and dendritic branches. However, the relative intensity of the red color is higher than green in Figure 3A and 3B (the intensities were adjusted separately for each color to maximize visualization of all markers), which prevents the green-labeled MAP2 marker from being seen in a subset of TH+ neurons with relatively weak MAP2+ staining. Expression of serotonin, glutamate, and DBH was not detected.

Other studies involving neural differentiation from hESC using different techniques (e.g., formation of EBs followed by retinoic acid treatment) have resulted in the formation of primarily GABAergic and glutaminergic neurons, with a small minority of neurons differentiating toward a midbrain DA fate [26]. The GABAergic phenotype has been suggested as being a default differentiation pathway for primary rodent and human neural precursor cells isolated from the CNS [27]. In the present study, only a few GABA-expressing neurons (less than 1% of cells) were detected (Fig. 3C), compared with the 34% ± 6% of the total cells that were TH+ (Fig. 3D– 3F). It has been reported that DA neurons in the forebrain, specifically in the olfactory bulb, often coexpress GABA [28]. No coexpression of GABA and TH was observed, compatible with a midbrain dopaminergic phenotype.

By phase-contrast microscopy, we were able to determine that the population of cells in differentiated cultures exhibited two principal morphologies. The neuronal cells were relatively uniform, with compact cell perikarya and a generally bipolar morphology. Cells with flat polygonal shapes and structural characteristics of astrocytes and glial cells were also present. Immunocytochemical analysis of neuronal processes using anti-β-III-tubulin and the glial and the astrocyte markers GFAP and S100β confirmed that astrocytes comprised a large fraction of the MAP2-negative cells that were present (Fig. 3G– 3I). GFAP staining was limited to cells with clear astrocyte morphology and was not expressed in β-III-tubulin+ neuronal cells. The orange-yellow appearance in Figure 3H and 3I is caused by superimposition of distinct neuronal and glial cells and does not reflect coexpression of GFAP and β-III-tubulin within individual cells.

Fixation and Mitotic Inactivation of PA6 Cells

We tested several methods of inactivating the PA6 cell layer by mitotic inactivation or fixation. Expression of Oct3/4, β-III-tubulin, and TH was examined after 12 days of coculture with mitomycin-c-treated, irradiated, or fixed PA6 cells (Fig. 4). Expression of β-III-tubulin, representing the neural-inducing activity of PA6 cells, was most markedly affected by the irradiation and mitomycin-c treatment conditions (Fig. 4A, 4C). The percentages of β-III-tubulin+ colonies were 37% ± 14% and 66% ± 8%, respectively, compared with untreated PA6 cells, for which β-III-tubulin expression was present in 79% ± 12% of colonies (Fig. 4E). In contrast, high levels of β-III-tubulin expression were seen in cells differentiated on fixed PA6 cells, in which case the number of colonies expressing was 77% ± 11% (Fig. 4B, 4E). There was no significant difference in the percentage of β-III-tubulin+ colonies between cocultures with fixed or untreated PA6 cells, implicating a role of PA6 cell membrane factors in overall neuronal differentiation.

thumbnail image

Figure Figure 4.. Comparison of expression of Oct3/4, β-III-tubulin, and TH by immunocytochemistry among cells differentiated on Ir., Fix., Mit-c-treated, or untreated PA6 cells after 12 days of coculture. (A): Few cells expressing β-III-tubulin or TH were observed following differentiation on Ir. PA6 cells. (B): Expression of β-III-tubulin was unchanged in cells differentiated on Fix. PA6 cells, compared with untreated cells, whereas TH expression was greatly reduced. (C): Mit-c treatment of PA6 cells had a relatively small effect on differentiation, as measured by either β-III-tubulin or TH expression. (D): β-III-tubulin and TH expression in cells differentiated on untreated PA6 cells. (E): Quantitative analysis of expression of Oct3/4, β-III-tubulin, and TH in human embryonic stem cells differentiated on untreated, Mit-c treated, Fix., and Ir. PA6 cells. Data represent 60 colonies from three independent experiments. The overall effect of treatment was statistically significant for Oct3/4-expressing(p < .0005), β-III-tubulin-expressing (p < .0001), and TH-expressing (p < .0001) colonies. *, p < .05; **, p < .01; ***, p < .001, compared with the untreated condition. In addition, the Mit-c and the irradiation conditions were significantly different from each other for Oct3/4 and β-III-tubulin (p < .0001) and TH (p < .01). Scale bars= 200 μm. Abbreviations: Fix., fixed; Ir., irradiated; Mit-c, mitomycin-c; TH, tyrosine hydroxylase.

Download figure to PowerPoint

Expression of TH was observed in 70% ± 15% of colonies grown on untreated PA6 cells (Fig. 4D, 4E). For cells maintained on irradiated or fixed PA6 cells, the number of TH+ colonies was reduced to 18% ± 8% and 21% ± 10%, respectively, a decrease of more than two-thirds (Fig. 4A, 4B, 4E). The number of TH+ colonies in cells grown on mitomycin-c-treated PA6 cells was less markedly reduced, to 46% ± 13% (Fig. 4C, 4E). The number of TH+ cells within the TH-expressing colonies was also very low in the irradiated and fixed conditions, compared with cocultures using untreated PA6 cells.

Inactivated PA6 cells, in spite of their weak ability to induce neural or dopaminergic differentiation, were effective in promoting a loss of Oct3/4 expression. Mitomycin-c-treated or fixed PA6 cells behaved similarly to normal PA6 cells in inducing hESC to adopt the Oct3/4-negative phenotype (Fig. 4E).

After 12 days of coculture of hESC with normal, mitomycin-c-treated, irradiated, or fixed PA6 cells, colonies were isolated as previously described and further differentiated in the presence of SHH, FGF8, and GDNF. The overall viability of cells differentiated on inactivated PA6 cells was reduced upon subculture, compared with cells cocultured with normal PA6 cells. As shown in Figure 5, the neuronal and dopaminergic differentiation potential of progenitor cells isolated from inactivated or fixed cells was significantly reduced. We investigated further whether the differentiation was more directed toward glial lineage by examining GFAP immunoreactivity. Results indicated a similar reduction in differentiation to a glial lineage. Percentages of β-III-tubulin, MAP2, TH, GFAP-expressing cells, and MAP+ cells expressing TH, after 18 days of differentiation subsequent to removal from the feeder layer (30 days total), are shown in Figure 5.

thumbnail image

Figure Figure 5.. Differentiation of progenitor cells isolated from coculture with inactivated PA6 cells. (A–D): Expression of MAP2 and TH 18 days after isolation from Ir. (A), Fix. (B), Mit-c (C), and untreated (D) PA6 cells. (A): Cells that were grown on Ir. PA6 cells gave rise to the lowest number of MAP2+ and TH+ cells. (B, C): Cells isolated from Fix. PA6 or Mit-c-treated cells behaved similarly, resulting in a significantly decreased number of MAP2+ and TH+ cells. (D): Untreated PA6 cells were most efficient in generating MAP2+ and TH+ cells. (E): Comparison of expression of β-III-tubulin, GFAP, MAP2, and TH. The overall effect of treatment was statistically significant for β-III-tubulin-expressing (p < .0001), GFAP-expressing (p < .0005), MAP2-expressing (p < .0001), and TH-expressing (p < .0001) cells. (F): MAP2+ cells expressing TH following differentiation for 18 days after isolation from untreated, Mit-c-treated, Fix., and Ir. PA6 cells. The overall effect of treatment was statistically significant for MAP2+ cells expressing TH (p= .041). *, p < .05; **, p < .01; ***, p < .001, compared with the untreated condition. In addition, the Mit-c and the irradiation conditions were significantly different from each other for MAP2 (p < .05). Scale bars= 200 μm. Abbreviations: DAPI, 4,6-diamidino-2-phenylindole; Fix., fixed; GFAP, glial fibrillary acidic protein; Ir., irradiated; Mit-c, mitomycin-c; MAP2, microtubule-associated protein; TH, tyrosine hydroxylase.

Download figure to PowerPoint

Effect of PA6-Conditioned Medium

We investigated whether dopaminergic differentiation of hESC could be induced by factors secreted from the PA6 cells by medium conditioned by PA6 cells to hESC cultures and monitoring their differentiation. When hESC were grown in the presence of conditioned medium (CM), the majority of colonies attached to the gelatin-coated culture dishes, spread, and started to differentiate (Fig. 6A). In contrast, hESC colonies in cultures containing unconditioned medium (UCM) survived poorly (Fig. 6B).

thumbnail image

Figure Figure 6.. Influence of PA6 cell-secreted factors on survival and neural differentiation of human embryonic stem cells (hESC). Colonies are shown 48 hours after culturing hESC in the presence (A) and absence (B) of PA6 CM. The majority of hESC in UCM did not attach to the gelatin-coated cell culture dishes, whereas CM promoted cell-to-substrate attachment and enhanced subsequent survival of hESC. (C, D): β-III-tubulin expression after 12 days of differentiation in the presence (C) and absence (D) of CM. Shown is the differentiation of hESC 48 hours after subculturing in CM (E) and UCM (F). (G, H): Expression of MAP2 and TH (G) and GFAP and TH (H) after 18 additional days of differentiation in CM. No surviving cells were found in cultures containing UCM at this stage. (I): Quantitative analysis of expression of Oct3/4, β-III-tubulin, and TH in colonies differentiated in CM or UCM. The effect of treatment was statistically significant for Oct3/4 (p= .0039) but not for β-III-tubulin or TH. Scale bars= 200 μm. Abbreviations: CM, conditioned medium; DAPI, 4,6-diamidino-2-phenylindole; GFAP, glial fibrillary acidic protein; MAP2, microtubule-associated protein; TH, tyrosine hydroxylase; UCM, unconditioned medium.

Download figure to PowerPoint

In spite of the poor survival of cells differentiated in UCM, the percentage of colonies expressing β-III-tubulin was similar to that of cells differentiated in CM (Fig. 6C, 6D). The percentage of β-III-tubulin-positive colonies was 52% ± 8% in cultures containing UCM and 60% ± 13% in cultures containing CM. However, the number of β-III-tubulin+ cells was significantly lower within colonies grown in UCM compared with CM. TH expression was not found in cultures containing UCM, whereas CM gave rise to TH expression in 12% ± 3% of colonies (Fig. 6I).

Partially differentiated hESC were subcultured after 12 days by enzymatic treatment in the same way as hESC cocultured with PA6 cells. CM continued to enhance survival and proliferation of cells after enzymatic passage, whereas cells cultured in UCM showed substantial cell death upon passaging (Fig. 6E, 6F). In the presence of CM, cells attached to the poly-l-lysine/laminin-coated plates and elaborated processes after 2 days (Fig. 6E).

Expression of Oct3/4 in cells cultured in UCM for 12 days was approximately 3% ± 3%, compared with 35% ± 9% of cells maintained in CM. This sustained Oct3/4 expression may indicate that PA6 cell CM is able to maintain the pluripotency of hESC while producing a limited degree of differentiation. The number of cells expressing MAP2 and TH was elevated after an additional 18 days. GFAP-expressing cells were dominant in the cultures (Fig. 6G, 6H).

Combination of PA6 Cell Surface Activity and Secreted Molecules

Insofar as both CM and inactivated PA6 cells had effects on differentiation of hESC when used individually, it seems reasonable to expect that the complete inducing activity, including neural and dopaminergic induction by the PA6 cells, might be mediated by secreted molecules from the PA6 cells in combination with PA6 cell surface activity. We therefore sought to determine whether dopaminergic differentiation could be produced by culturing hESC on fixed PA6 cells in combination with CM collected from live cells. Cultures were examined after 18 days of differentiation. This time point was chosen because of our previous observation that under coculture conditions, differentiation of hESC to TH-expressing neurons was nearly complete at this time. With regard to neural and dopaminergic differentiation, assessed by the number of colonies expressing β-III-tubulin or TH, no significant difference was observed between cultures exposed to CM or UCM when combined with fixed PA6 cells (Fig. 7A, 7B).

thumbnail image

Figure Figure 7.. Combinatorial effect of PA6 cell surface activity and secreted factors. Expression of β-III-tubulin and TH was examined after 18 days of coculture of human embryonic stem cells (hESC) with fixed PA6 cells under the influence of CM (A), UCM (B), CM plus Hep (C), and UCM plus Hep (D). Hep dependence of secreted factors involved in dopaminergic differentiation is indicated by the high number of TH-expressing cells seen in the CM plus Hep condition, shown in (C). (E, F): Phase-contrast images of differentiated hESC colonies cultured in CM and UCM, respectively, in the presence of Hep, showing neural rosette-like structures in the core of colonies. Images shown in (G) and (H) illustrate random differentiation of hESC cultured in CM and UCM, respectively, in the absence of Hep. (I): Quantitative analysis of expression of β-III-tubulin and TH and appearance of neural rosette-like structures in colonies cultured in CM or UCM in the presence or absence of Hep. The overall effect of treatment was statistically significant for β-III-tubulin-expressing and TH-expressing (p < .0001) colonies and for rosette-containing colonies (p < .0003). *, p < .05; **, p < .01; ***, p < .001, compared with the UCM condition. In addition, the UCM+Hep and CM+Hep conditions were significantly different from each other for β-III-tubulin and TH (p < .0001) and for rosette-containing colonies (p < .01). Scale bars= 200 μm. Abbreviations: CM, conditioned medium; Hep, heparin; TH, tyrosine hydroxylase; UCM, unconditioned medium.

Download figure to PowerPoint

In an attempt to increase the biological activity of the secreted molecules, the same experiments were repeated in the presence of heparin. In the presence of heparin, substantial differences were observed in dopaminergic differentiation of the hESC in CM compared with UCM. In contrast to previous results from cultures of hESC differentiated in CM and UCM in the absence of fixed PA6 cells (Fig. 6C, 6D), the percentage of β-III-tubulin+ colonies was significantly increased in cultures differentiated on fixed PA6 cells in CM compared with UCM in the presence of heparin. The average number of β-III-tubulin+ colonies was 36% ± 17% in cultures containing CM, in contrast to 17% ± 1% in cultures containing UCM. Moreover, 89% ± 10% of β-III-tubulin+ neurons were TH+ in cultures differentiated in CM, compared with 25% ± 14% in UCM. Percentages of colonies containing TH+ cells were 32% ± 10% and 9% ± 3% in CM and UCM, respectively (Fig. 7C, 7D).

An interesting observation in cultures containing heparin was the appearance of neural rosette-like structures composed of columnar cells in the core of colonies. The columnar arrangements of cells were more frequent in colonies cultured in CM plus heparin, but they could also be detected in cultures containing UCM and heparin (Fig. 7E, 7F, 7K). The latter condition (UCM plus heparin) gave rise to a combination of rosette-like structures and cells with epithelial morphology (Fig. 7F). The percentages of colonies that exhibited rosette-like structures were 37% ± 4% and 12% ± 2% in cultures containing CM or UCM, respectively (Fig. 7I). In the absence of heparin, on the basis of morphology, the population of cells appeared to undergo spontaneous multilayer differentiation, and rosette-like structures could not be detected (Fig. 7G, 7H). The increase in the number of TH+ neurons in cultures containing heparin suggests a link between heparin-binding growth factors and dopaminergic differentiation of hESC. To further evaluate the direct interaction of heparin with molecules secreted from PA6 cells, hESC were allowed to differentiate in CM containing heparin in the absence of PA6 cells. This condition yielded 54% ± 6% TH+ colonies after 18 days of differentiation, compared with 81% ± 1% TH+ colonies in cocultures with PA6 cells. The number of TH+ neurons present within colonies, however, was substantially decreased in the CM plus heparin condition, compared with the PA6 cell coculture condition (supplemental online Fig. 1).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Dopaminergic neurons have been generated from mouse, primate, and human hESC by coculture with the PA6 stromal cell line [3, 4, 7, 9, [10], [11]12]. PA6 cells exert a differentiation-inducing effect on ESC that has been termed SDIA. The differentiation of midbrain dopamine neurons from hESC using this strategy as described here is highly efficient, with up to 80% of all neurons and 34% of all cells in the colonies expressing TH. The protocol is technically simple and rapid (supplemental online Fig. 2). In contrast to previous protocols, this method does not require the formation of EBs, selection of neural precursor cells, retinoic acid, or multiple treatment phases. Thus, this method is well suited as an in vitro model to study development of dopaminergic neurons and for examination of the mechanism of dopaminergic neurodegeneration in PD.

Although the procedure described here may be a step toward improving the utility of embryonic stem (ES) cells for replacement therapy in PD, we would not suggest that this procedure or anything similar should be used to prepare cells for transplantation in human subjects. Several unresolved issues remain, including the possible presence of rare undifferentiated hESC with the potential to form teratomas, cells with undetected mutations that could predispose them to forming other types of tumors [29, 30], and the possible presence of other undesirable cell types. In addition, it is unclear to what degree transplanted hESC could substitute for fetal dopaminergic neurons in transplantation [31, 32]. Additional issues, such as the use of animal cells in generation of DA cells, would need to be addressed before this approach could be applied in clinical settings. The PA6 cells can be partially eliminated by isolation of precursors as shown in this study, but a feeder-free system would presumably be required for use in cell-based therapy.

Several recent studies have reported that ESC-derived TH+ cells can innervate the host striatum, form functional synaptic connections, and reverse motor impairments in rat models of PD [33, 34]. Nonetheless, relatively little is known about long-term survival of these transplanted dopamine neurons and their functional characteristics or the ability to generate new connections.

Many factors have previously been implicated in the regulation of dopaminergic differentiation, including FGF8, SHH, interleukin (IL) 1, IL11, hepatocyte growth factor, insulin-like growth factor 2, pleiotrophin (PTN), brain-derived neurotrophic factor, GDNF, and others [16, 17, 20, [21]22, 35, [36], [37], [38], [39], [40], [41], [42]43]. The present study identifies several apparently independent facets of SDIA as a first step toward elucidating the mechanisms underlying dopaminergic neuron development from hESC.

PA6 CM was not able to induce differentiation of dopaminergic neurons as efficiently as live PA6 cells after 12 days in coculture with hESC, but it did promote survival of hESC and sustained Oct3/4 expression. Maintenance of Oct3/4 expression has been correlated with SDIA-induced neurogenesis, neuroectoderm formation, and neural differentiation from ESC, suggesting that continued Oct3/4 expression may be involved in the initial phases of neuronal differentiation [44, 45].

We also examined whether cell surface activity of PA6 cells was sufficient for the induction of hESC toward the dopaminergic lineage by using fixed PA6 cells. In contrast to other studies [3], the present results indicated that fixed PA6 cells were able to induce significant neural differentiation of ES cells, but fixed PA6 cells were relatively ineffective in producing dopaminergic differentiation. The ability of PA6 cells to induce dopaminergic differentiation was decreased somewhat by mitomycin-c treatment but markedly decreased by irradiation. Thus, PA6 cells have a dopaminergic-inducing effect that is drastically decreased by both fixation and irradiation but less affected by mitomycin-c treatment; conversely, the neural-inducing effect is not altered by fixation but is decreased by mitotic inhibition. Subsequent to isolation from mitotically inhibited PA6 cells, the partially differentiated hESC showed limited survival, and dopaminergic differentiation was significantly reduced.

Synergistic effects of PA6 cell surface activity and secreted molecules were confirmed by increased dopaminergic induction of hESC cultured on fixed PA6 cells in the presence of PA6-conditioned medium and heparin compared with dopaminergic induction by the two elements separately. However, the addition of heparin to cultures differentiated in unconditioned medium in the presence of fixed PA6 cells decreased the number of neuronal cell-containing colonies. An explanation of this inhibitory effect of heparin on general neural induction will require further investigation.

It is likely that heparin increases the stability of secreted factors in the extracellular matrix, thus increasing ligand-receptor interactions. A number of investigations suggest that glycosaminoglycans serve as biological regulators of cellular signaling [46, 47]. Factors secreted by stromal cells have been shown to have high affinity for heparin, and the heparin interaction appeared to have an important role in their biological activity [48]. Nevertheless, the combinatorial effects of cell surface activity and secreted factors were still not as effective as the SDIA mediated by untreated PA6 cells. This discrepancy could be due to proteolytic degradation of secreted factors that were not completely stabilized by the concentration of soluble heparin used and alteration of cytoskeletal arrangements and extracellular matrix molecules induced by chemical fixation.

In an attempt to identify the molecular nature of the dopaminergic-inducing effect of PA6 cells, a cytokine growth factor-focused array and reverse transcription-polymerase chain reaction were used in a previous study to compare PA6 cells with MEFs [9]. Higher levels of mRNA expression for several growth factors, including hepatocyte growth factor, vascular endothelial growth factor, and FGF7, were found in PA6 cells. Addition of these growth factors to hESC cultures was not sufficient to induce TH-positive colonies. The conditions under which these factors were tested, however, may not have been optimal.

The current data suggest that several factors cooperatively promote the maturation of hESC cultures to the midbrain dopaminergic phenotype. Prior studies have proposed that PA6 cell surface activity is responsible for the SDIA effect and dopaminergic induction in particular. The nature of SDIA has remained very controversial. Our data taken together show that the PA6 cell surface primarily triggers signaling, which promotes cell survival and cell adhesion, and is responsible for enhanced neurogenesis of hESC, rather than providing lineage-specific instructions. The specific dopaminergic-inducing effect appears to reside in soluble factors that are secreted by PA6 cells. Nevertheless, the presence of cell surface material enhances the survival of the hESC while differentiating, thus markedly increasing the overall yield of dopaminergic cells. Whether this effect occurs through a direct effect of PA6 cell surface and extracellular matrix molecules or by stabilization of secreted factors has not yet been resolved.

Using more comprehensive array studies (unpublished data), we have now identified a large number of candidate molecules potentially responsible for the SDIA effect. IGF2 and PTN are among these factors. High levels of IGF2 were found in PA6 cells (unpublished data), as well as mesencephalic-restricted neural progenitor cells [10]. In addition, IGF2 has been reported to assist in neural induction by inhibiting the activity of antineuralizing transcription factor Smad1 [49]. High levels of PTN, a heparin-binding growth-associated molecule, were also found in PA6 cells compared with other cell lines, which lacked the dopaminergic-inducing effect. Several recent studies have demonstrated a role of PTN in dopaminergic differentiation and survival [39, [40]41]. In view of the present data illustrating the importance of heparin in dopaminergic induction, PTN is a possible candidate for contribution to SDIA. To determine whether these molecules that we have identified or other specific proteins can mimic SDIA, appropriate conditions isolating hESC from feeders and feeder cell-conditioned medium, eliminating unknown interactions, are required. Studies in progress suggest that the issue of duplicating the effect of PA6 cells is likely to be complex, because of obstacles such as limited cell survival in the absence of feeder cell layers and possible instability of the molecules of interest.

Summary

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Thus, PA6 cells appear to produce a number of effects on hESC, suggesting that their overall effect is complex and possibly the result of several factors. The effects include (a) a neural-inducing effect, (b) a dopaminergic-inducing effect, (c) a survival-promoting effect, and (d) an Oct3/4 expression-maintaining effect. These effects are, to some degree, separable by treatments, including fixation, mitotic inactivation, and the use of CM both alone and in combination with heparin. Overall, this suggests that the effects of PA6 cells on hESC are complex and may be the result of several different proteins present on the cell surface and secreted into the medium. To mimic the effects of SDIA, the nature of the secreted molecules that are required to produce each of these four effects and whether they involve instructive or inhibitory signaling cues would have to be determined.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Research was supported by the Intramural Research Program of National Institute on Drug Abuse, NIH, Department of Health and Human Services.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information
FilenameFormatSizeDescription
SC-08-0039_Supplemental_Figure_1.tif2215KSupplemental Figure 1
SC-08-0039_Supplemental_Figure_2.tif2217KSupplemental Figure 2

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.